ASTM (1978) - Significance of Tests and Properties of Concrete and Concrete-Making Materials

Copyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:48:22 EDT 2014
Downloaded/printed by
Universidade do Gois pursuant to License Agreement. No further reproductions authorized.
SIGNIFICANCpER OF
TESTS AND
OPERTIES
OF CONCRETE AND
CONCRETE-MAKING
MATERIALS
Sponsored by
ASTM Committee C-9 on
Concrete and Concrete Aggregates
AMERICAN SOCIETY FOR
TESTING AND MATERIALS
ASTM SPECIAL TECHNICAL PUBLICATION 169B
04-169020-07
~~I~
AMERICAN SOCIETY FOR TESTING AND MATERIALS

1619 Race Street, Philadelphia, Pa. 19103
Copyright 9 by American Society for Testing and Materials 1978

Library of Congress Catalog Card Number: 78-51628
NOTE
The Society is not responsible, as a body,
for the statements and opinions
advanced in this publication.
Printed in Baltimore, Md.

Dec. 1978
Second Printing, Ann Arbor, MI
December 1984
Foreword
This publication is a revision and expansion of the report on Significance
of Tests and Properties of Concrete and Concrete-Making Materials published in 1966. That publication in turn replaced editions of Report on
Significance of Tests of Concrete and Concrete Aggregates published in
1935, 1943, and 1956. The present report covers several types of materials
not referred to in the earlier reports but whose importance has increased
greatly since preparation of the 1966 report.
As was true for the previous publications in this series, the separate
chapters have been prepared by individuals selected because of their
knowledge of the respective subjects and because of their participation in
development ofpertinent methods of testing and specifications. While independent expression by the authors has been encouraged, individual
chapters have been reviewed and the entire report has been coordinated by a
special committee appointed for that purpose by ASTM Committee C-9 on
Concrete and Concrete Aggregates. The special committee in charge consisted of: C. H. Best, Kansas State University, Manhattan, Kansas, chairman; B. Mather, U. S. Army Engineers, Waterways Experiment Station,
Vicksburg, Miss.; J. F. McLaughlin, Purdue University, West Lafayette,
Ind.; R. C. Mielenz, The Master Builders Co., Division of Martin Marietta,
Cleveland, Ohio; J. T. Molnar, The Standard Slag Co., Youngstown, Ohio;
M. Polivka, University of California, Berkeley, Calif.; and R. D. Walker,
Virginia Polytechnic Institute and State University, Blacksburg, Va.
Related
ASTM Publications
Concrete Pipe and the Soil Structure System, STP 630 (1977),
04-630000-07
Living with Marginal Aggregates, STP 597 (1976), 04-597000-07
Contents
Chapter 1--Introduction--R. E. PHILLEO
PART I
GENERAL
Chapter 2--Techniques, Procedures, and Practiees of Sampling of

Concrete and Concrete-Making Materials--E. A. ABDUN-NUR
Chapter 3--Statistical Considerations in Sampling and T e s t i n g - - n .
T. ARNI
Chapter 4--Quality in Concrete Testing--J. R. DISE

Chapter 5--Needed Researeh--R. D. WALKER
5
24
44
49
PART II
TESTS AND PROPERTIES OF CONCRETE

Chapter 6 - - T h e Nature of Concrete--T. c. POWERS
Freshly Mixed Concrete
Chapter 7 - - U n i f o r m i ~ and Workability--o. T. SMITH
Chapter 8 - - M a k i n g and Curing Concrete Specimens--R. F. ADAMS
Chapter 9--Setting T i m e - - J . n. SPROUSE AND R. B. PEPPLER
Chapter 10--Air Content and Unit Weight--F. F. BARTEL
Hardened Concrete
Chapter 11--Petrographlc Examination--KATHERINE MATHER
Chapter 1 2 - - S t r e n g t h - - g . N. DERUCHER
Chapter 13--Aceelerated Strength Tests--M. a. WILLS
Chapter 14--Elastic Properties and Creep--R. E. PHILLEO
Chapter 15--Nondestructive Tests--E. A. WmTEHURST AND V. M.
MALHOTRA
Chapter 16--Volume Change--J. L. SAWYER
Chapter 1 7 - - T h e r m a l Properfies--j. A. RHODES
Chapter 18--Pore Structure--GEORGE VERBECK
Chapter 19--Corrosion of Reinforcing Steel--P. D. CADY
Chapter 20--Corrosion of Embedded Materials Other than Reinforcing Steel--BERNARD ERLIN AND HUBERT WOODS
Chapter 21--Bond with Reinforcing Steel--L. A. LUTZ
Chapter 22--Abrasion Resistanee--R. o. LANE
Chapter 23--Resistance to Weathering--HOWARD NEWLON, JR.
59
74
102
105
122
132
146
162
180
200
226
242
262
275
300
320
332
351
Chapter 24--Resistance to Chemical Attack--L. H. TUTHILL

Chapter 25--Resistance to High Temperatures--PETER SMITH
Chapter 26--Radiation Effects and Shielding--MILOS POLIVKAAND
H. S. DAVIS
Chapter 27--Air Content and Unit Weight--s. a. HELMS
Chapter 28--Cement Content--w. G. HIME
Special Categories
Chapter 29--Ready-Mixed Concrete--R. D. GAYNOR
Chapter 30--Lightweight Concrete and Aggregates--v. w. LEWIS
Chapter 31--Packaged, Dry, Combined Materials for Mortar and
Concrete--A. C. CARTER
Chapter 32--Preplaced Aggregate Concrete--a. g. LAMBERTON
369
388
420
435
462
471
503
525
528
PART I l l
TESTS AND PROPERTIES OF CONCRETE AGGREGATES
Chapter 33--Petrographic Examination--R. c. MIELENZ

Chapter 34--Grading--w. Ia. PRICE
Chapter 3S--Shape, Surface Texture, Surface Area, and Coatings-M. A. OZOL
Chapter 36--Weight, Density, Absorption, and Surface Moisture-W. G. MULLEN
Chapter 37--Porosity--w. L. DOLCH

Chapter 38--Abrasion Resistance, Strength, Toughness, and Related Properties--R. c. MEININGER
Chapter 39--Thermal Properties--8. K. COOK
Chapter 40--Chemical Reactions Other than Carbonate Reactions--
539
573
584
629
646
657
695
708
SIDNEY DIAMOND
Chapter 41--Chemical Reactions of Carbonate Aggregates in Cement Paste--H. N. WALKER

Chapter 42--Soundness and Deleterious Substances--L. DOLAR-
722
744
MANTUANI
PARTIV
TESTS AND PROPERTIES OF OTHER MATERIALS
Chapter
Chapter
Chapter
Chapter
Chapter
43--Mixing and Curing Water for Concrete--w. J. MCCOY

44--Curing Materials--R. E. CARRIER
45--Air-Entraining Admixtures--PAUL KLIEGER
46--Mineral Admixtures--L. H. TUTHILL
47--Chemical AdmixtureS--BRYANT MATHER
765
774
787
804
823
Chapter 48--Cellular Concrete--L M. LEGATSKI

Chapter 49--Organic Materials for Bonding, Patching, and Sealing
of Concrete--R. j. SCHUTZ
Chapter 50--Pumpability Aids for Conerete---a. c. VALORE, JR.
836
852
860
STP169B-EB/Dec. 1978
R. E. Philleo ~
Chapter I--Introduction
When an ASTM Committee undertakes the preparation of a statement on

the significance of tests and properties of materials under its jurisdiction, it
embarks on a job that will never be finished. New varieties of materials are
developed, new uses for old materials surface, and testing techniques are
refined. Also, for materials such as concrete with time-dependent properties,
the age at which users demand test results diminishes.
ASTM Committee C-9 on Concrete and Concrete Aggregates articulated
its first statement in 1935 with Report on Significance of Tests of Concrete
and Concrete Aggregates, ASTM STP 22. Since then, this volume has been
updated every ten years. This volume presents the point of view of the 1970s.
If we are curious to discover how the outlook of the 1970s differs from that of
the 1960s, we need only compare the contents of the current volume with the
contents of its predecessor, Significance of Tests and Properties of Concrete
and Concrete-Making Materials, ASTM STP 169-A, published in 1966.
In Part II on Tests and Properties of Concrete, there are new chapters on
accelerated strength testing and nondestructive testing of hardened concrete.
Both reflect the growing impatience of users with the 28-day strength
cylinder, which, although providing an excellent evaluation of the material
leaving a concrete mixer, yields data which are neither timely nor directly applicable to the structure under consideration. There is also a new chapter on
preplaced aggregate concrete, an old technique which only recently has been
standardized by ASTM because of its applicability to structures for radiation
shielding.
In Part III on Tests and Properties of Concrete Aggregates, there is now an
entire chapter devoted to alkali-carbonate reactions in recognition of the increasing number and complexity of such reactions which recently have been
identified. The ability of technologists to tailor concrete to highly specific and
unusual applications is highlighted in Part IV on Tests and Properties of
Other Materials, where the number of chapters on concrete admixtures has
been nearly doubled by the introduction of new chapters on admixtures for
IChief, Structures Branch, Engineering Division, Civil Works, Office of Chief Engineers,
Department of the Army, Washington, D.C. 20314
Copyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:38:37 EDT 2014 1
Universidade
do Gois pursuant
Agreement.
No further reproductions
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
SIGNIFICANCE OF THE PROPERTIES OF CONCRETE
production of cellular concrete and pumping aids for concrete. The latter
bears witness to a minor revolution in the placing of ready-mixed concrete
made possible by the development of portable concrete pumps.
The volume is the handiwork of a special subcommittee of ASTM Committee C-9, who organized the subject matter, selected the authors, and reviewed
the papers. Most of the papers were authored by present or recent chairmen
of the technical subcommittees having jurisdiction over the subjects of the
papers. T h e subcommittee was chaired by Professor C. H. Best, and the
members included Bryant Mather, J. F. McLaughlin, R. C. Mielenz, J. T.
Molnar, Milos Polivka, and R. D. Walker. ASTM Committee C-9, as well as
all those who produce and use concrete, are indebted to them for their timely
statement.
E. A . A b d u n - N u r 1
Chapter 2nTechniques, Procedures,

and Practices of Sampling of
Concrete and Concrete-Making
Materials
Introduction
Millions upon millions of dollars change hands daily in various segments

of the concrete industry, based on evidence obtained from samples. Yet,
in most instances, the sample is obtained in an indifferent manner by
someone who is ignorant of the basic principles of sampling (frequently a
laborer or warehouseman), the use to which the sample is to be put, the
tests to be performed on the sample and the final decisions to be based
on or derived from the test results--all very important factors that should
be considered in the sampling process. There appears to be no adequate
appreciation of the importance of sampling by those who should be concerned with the problems resulting from poor sampling--those who have
to make important decisions and set policy based on test results from
samples--despite the cost of the large volume of samples and testing in
the concrete field and the larger economic significance of conclusions which
are derived from these.
On the other hand, other industries confronted with similar problems
have gone to great lengths to develop reliable and efficient methods and
procedures. The coal, fertilizer, ore, and abrasives industries [1-3] 2 that
have sampling problems similar to those found in aggregates, for example,
have developed criteria and methods that increase the probabilities of proper
sampling. Basic principles and approaches that can be helpful in providing
sampling details for the various segments of the concrete industry have
been developed by the American Society for Testing and Materials (ASTM)
Committee E-11 on Statistical Methods and others [4-15]. Other sampling
1Consulting engineer, Denver, Colo. 80222.
2The italic numbers in brackets refer to the list of references appended to this chapter.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
GENERAL
procedures for concrete and concrete ingredients, that range from indifferent to very good, may be found in various ASTM standards (see Appendix)
and other handbooks and manuals [16-18].
What Is Sampling
The Sample
A sample is a small portion of a larger volume of group of materials
(such as a lot, shipment, stockpile, batch, carload, or truckload) about
which information is desired. The characteristics of the sample are presented as evidence of the properties of the larger unit from which it is
taken. A series of such samples provides a pattern of the variations in
properties of the total universe. Universe, in this instance, means the aggregate of lots, stockpiles, carloads, and so on, which comprise any specific
material used in a project or produced under regular production for the
izrescribed time interval under consideration. The measurements on many
samples from the universe constitute a statistical population.
Sampling
Sampling is the process of obtaining samples from the larger universe.
When the universe is perfectly homogeneous, sampling becomes the simple
physical act of taking a sample from the unit about which information is
desired. In this case, any sample can truly represent the larger homogeneous whole.
Unfortunately, nature, in general, and the concrete field, in particular,
rarely if ever, present us with a perfectly homogeneous universe of any
material or process. If the rare occasion of perfect homogeneity should
occur, one would not know it ahead of time and would have to treat it as
a variable. Each lot, truckload, batch, or stockpile, not only varies in some
measure from similar units intended to be identically the same, but frequently varies within each such unit. The more heterogeneous the universe,
the more work it takes to develop a reliable estimate of its characteristics.
Therefore, it pays to save sampling and testing funds where little variability
occurs and expend them where more heterogeneity is prevalent.
Sampling is thus much more than the physical act of taking a sample of
the larger unit as evidence of the properties of the latter. First of all, a
sampling plan must be formulated to reflect the variability characteristics
of the universe. After that, individual samples taken in accordance with
such a plan must be obtained in such a manner that each sample is truly
representative of the unit from which it comes.
This broadened concept of sampling is in effect an acceptance of the
fact that sampling is a complex business and the needed skill lies as much
or more in the one exercising overall control of the work as in the one
ABDUN-NUR ON SAMPLING OF CONCRETE
actually taking the sample. A major reorientation of ideas on the subject

is needed.
Sampling Plan
The development of a sampling plan requires the mastery of the sophisticated fundamentals of probability sampling, an intimate knowledge of the
product being sampled, and a high degree of skill, experience, background,
and creativeness. In some cases, the help of a statistician will prove advantageous in developing the theoretical aspects and procedures for the
practical man to follow. Models for probability sampling, significance,
and interpretation have been developed by ASTM Committee E-11: Recommended Practice for Probability Sampling of Materials (E 105), Recommended Practice for Choice of Sample Size to Estimate the Average
Quality of a Lot or Process (E 122), and Recommended Practice for Acceptance of Evidence Based on the Results of Probability Sampling (E 141).
ASTM Committee D-4 on Road and Paving Materials recently proposed
a standard on sampling: Recommended Practice for Random Sampling
of Paving Materials (D 3665). Setting up sampling plans without the use of
probability techniques will in most instances introduce subconscious bias.
From the standpoint of economics, sampling plans either provide a
cheaper method of achieving a given reliability or result in a better reliability
for a given cost. The degree of reliability should be set to fit the economic
need or justification. The end result for which samples are taken affects
to a large extent the details of the sampling plan.
Once a sampling plan is developed for any given situation or set of conditions, it can be followed easily by the average field man under proper
supervision. Details of the many approaches that can be applied to the
development of sampling plans for concrete and its various ingredients are
beyond the scope of this short discussion. The references in this chapter
will provide the reader with an operational starting point.
Taking of Sample
The actual physical manipulation needed to provide a sample representing
a larger unit is simpler than formulating the sampling plan, and the Annual
Book of ASTM Standards provides guidelines under the various designations as shown in the Appendix to this chapter. The procedures can be
mastered easily, and if the instructions are followed carefully, one can expect a reasonably reliable sample to result. Other guidelines are found in
various handbooks and manuals [16-18]. In general these procedures were
developed many years ago, and they are becoming outmoded with the new
concepts of sampling. They can and are being improved materially in the
continuing process of reexamination by the issuing body.
GENERAL
The relationship of a sampling plan to the actual taking of a sample

may best be illustrated by an example in which the universe is the total
concrete of any given class on a project. The concrete reaches the project or
is made at the site in a series of batches. Although all batches are intended
to be exact duplicates, this is not so in practice. Therefore, for proper
evidence of the variations of the concrete properties in the universe, a plan
is needed that designates which of the batches to sample in order to develop
the same pattern of variation found in the universe; this is known as random
or probability sampling. The second step is to sample each batch that has
been tagged for sampling, in conformance with the sampling plan, in such
a manner as to reflect the variation pattern in the concrete as it comes out
of the mixer. This is best done by taking several subsamples at different
stages of discharge from each batch and compositing them for a total
sample. In case, on the other hand, it is desired to study the efficiency of
mixing of the mixer, then the subsamples are tested separately to develop
a variation pattern of the mixing operation to determine if such a pattern
falls within the tolerances applicable to the situation.
This example shows strikingly that sampling in the concrete field is not
a simple matter and must not be delegated to the untrained or careless, or
relegated to a laborer for convenience.
Significance of Various Facets of Sampling
Economic Significance
Adequate sampling plans and procedures have more far reaching effects
than appear on the surface. In the first place, sampling involves the costs
of sampling, packaging, and shipping, which represents an investment.
But samples are usually taken for the preparation of specimens for testing,
which normally costs far more than the sampling and involves time, sophisticated equipment, and skilled personnel, all of which add to the sample
cost. However, the greatest economic significance based on the test results
from the samples is manifested in the decisions and conclusions which run
into millions of dollars daily in the concrete industry alone. In many cases,
such decisions are not only economic factors, but also involve safety and
life.
Thus it is seen that sampling is the most important link in the whole
chain of construction events leading to decisions, exchange of funds between
parties, contractual relations, and even safety. Under such circumstances,
can one sufficiently stress the importance of sampling plans, methods,
and procedures?
Random or probability sampling permits one to attach measures of confidence to the estimates of population parameters. Biased or subjective
sampling precludes the use of such confidence limits. This affects the
reliability of decisions based on test results.
Adams (Chapter 8) shows the importance of sampling and of the making

of specimens in strict accordance with the required standards. Disc (Chapter 4) stresses the importance of the quality of testing. The question may
well be raised whether the quality of testing or the proper making of specimens has any significance if the latter were prepared from samples that
have been improperly selected, taken, and handled prior to testing. These
are some of the factors that seem to have been frequently neglected if one
were to judge by observations of actual sampling in the field.
Too often, written or verbal instructions require a "representative"
sample from a "typical" batch or unit. A sample, for certain purposes as
previously described, must be representative of the batch or unit; but as
far as the universe is concerned, there is no such thing as a single representative or typical batch or unit. A sample really represents only itself as a
point under the probability curve of the universe. Therefore, such instructions are misleading, and result in improper samples. Yet, results of tests
made on samples taken with such unrealistic instructions are used to make
most important decisions even when the data appear most improbable.
Why, then, has sampling been apparently neglected in the concrete field?
Many factors have caused this situation to develop.
1. In the concrete field, samples are heavy, bulky, and dirty, and therefore, in many cases, the engineer passes the buck by sending someone at
the lower echelon to take care of sampling, feeling that he would be wasting his expensive time to do so himself.
2. In the past, when testing was not used to reach conclusions by statistical or probability means, probability sampling was not nearly as important
as it is now, when statistics has come into the picture. Yet, sampling procedures have not been revised substantially.
3. Automation, which makes "quality control" possible and places a
premium on homogeniety has only recently come into being in the concrete field. "Perfect" homogeneity was not a practical aim or achievement
as long as most operations were conducted by hand, and the influence of
the personal differences between individuals was a paramount factor.
4. Decisions have been made to a large extent by whether the engineer
felt the job to be satisfactory or not and not so much by test results. The
latter were something to stick in the file for legal protection, and when an
occasional test turned out to be low, everyone got into a huff and blamed
the sample or the contractor and made the latter the goat. Since the engineer in those days was very close to the job and observed everything that
was taking place, his judgment was very effective.
No one ever dreamed that occasional low test results are part of everyday
patterns of variation and are to be expected. If they fall within the proper
variation pattern, they should not be a source of concern. Fortunately,
these low test results showed up only occasionally since the natural tendency
for bias in sampling is towards the better side rather than the poorer side.
This problem has been discussed in detail by Abdun-Nur [19].
10
GENERAL
Sampling and Inspection

In most instances, sampling (the actual taking of the sample), which
becomes one of the duties of the inspector, is frequently and unfortunately
passed on to a laborer furnished by the contractor. In many instances, it
has been observed that the inspector handles sampling as a matter of convenience, neglecting the fact that sampling is a most important duty to be
performed in exact accord with specified techniques.
As an example, on one project, written instructions, the specifications,
and the formal inspector indoctrination meetings all stressed the obtaining
of three sets of cylinders from three distinct batches of concrete to represent
the variations in a given volume. Field observations indicated that several
inspectors were making the three sets of cylinders, all at one time, from
one single batch. This of course yielded erroneous records, inasmuch as
the test results were being analyzed as though the tests had come from
three distinct batches rather than from one batch. The record thus indicated
a higher degree of homogeneity than actually existed on the project. Inasmuch as variability was tied to strength requirements, this permitted the
contractor to get by with lower strenghts than desired and required for the
integrity of the design by the engineer. When an inspector was asked why
he was not following instructions, the answer was that "It saves making
three messes, and gets the job done all in one mess." In general, when the
effect of such disregard of instructions is explained, inspectors are more
careful.
Thus, it is not only important that the inspector be adequately trained
and qualified to do his job properly and to pay special attention to his
sampling plan and procedures, but it is also just as important to have
surveillance to assure that such procedures are being carried out properly.
Better trained and higher caliber personnel are badly needed as sampling
carries a burden of responsibility that has usually been ignored.
Inspection can, of course, be either internal--by the producer or contractor, to control his product--or external--by the owner, his engineer,
or his representative to ascertain compliance with specifications. At the
present time, there is need for both internal and external inspections to
assure proper construction. A detailed discussion of this subject is outside
the scope of this paper and has been treated by Abdun-Nur [20].
Protection of Samples
Sampling plans, methods, and procedures are not enough to assure
proper samples reaching the laboratory for the preparation of specimens
for testing. The proper protection, packaging, and handling of the samples
is just as important since carelessness in these factors can invalidate the
usefulness of the best sample. Wills [21] has shown the undesirable effects
11
on concrete properties due to the use of bags contaminated with sugar or

flour, or bags treated with chemical preservatives.
Every engineer has seen concrete cylinder specimens unprotected, permitted to dry out in the hot sun or to freeze on the job. Occasionally,
carelessness has resulted in heavy equipment running over specimens which
were damaged beyond usefulness; yet, they were sent to the laboratory for
testing with the expectation that usable results would be obtained. Proper
protection, packaging, and care of samples in the field and in shipping
are essential if test results obtained from specimens prepared from these
samples are to mean much and if reliable conclusions are to be drawn from
such results. These are serious hazards; yet data obtained in such a manner
are being used every day as a guide for the uninitiated to follow.
Compositing Samples
The validity of composited samples depends on the purpose and end use
of the test results from the specimens prepared therefrom. If a representative sample of the properties of a unit of concrete or aggregate is to be
obtained, then subsamples taken from various portions of the unit and
composited are in order. But if the sampling is for the study of the variations
within a unit, then the subsamples should be kept separate and individual
specimens should be made and tested separately. This question of cornpositing should be decided upon in the development of the sampling plan
and passed down to the sampler in the form of instructions to be followed
explicitly; otherwise, the analyzed results may be unknowingly very misleading.
Sampling and Quality Control

The control of any product is only possible through proper sampling.
This is true whether the product be aggregates, various packaged ingredients
that enter into the concrete as constituents at various times, or the concrete itself in the plastic stage. Sampling permits the determination of
whether the process is producing what is needed and when adjustments
and their extent may be required to bring the process within tolerances
and up to the desired quality. To attempt to control a product by occassional sampling of the finished material is unrealistic, ineffective, and too
late--by then it has already been produced and in most cases cannot be
modified. Yet in the concrete field, occasional sampling of this latter type
is almost standard routine. It is naive to think that this is quality control.
To obtain adequate quality control of concrete, every step of the process
or procedure must be controlled as well as every ingredient that enters into
it. Such control has to be developed by the producer or Contractor (who
is a producer in the broad sense of the word) and be built into his proCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:38:50 EDT 2014
12
GENERAL
cesses or operations. It cannot be imposed by the purchaser or owner from

outside.
The most the owner or purchaser can do is to sample and test to satisfy
himself that the product complies with the specification requirements. Incentives built into the specifications by the owner or purchaser are of tremendous help in getting the producer to develop proper quality control
in his operations.
Modern automated computer-controlled batch and mixing plants are
now available that will record, control, and correct weigh~.s and slumps-even reject a batch that is outside of tolerances [22].
Sampfing Concrete and Its Constituents
General
The Appendix gives a list of the various ASTM standards under the
jurisdiction of ASTM Committee C-9, or useful in their work, but originating in other committees. The list of references at the end of this chapter gives papers, manuals, handbooks, and bibliographies that treat approaches to sampling plans, methods, and procedures in various fields,
which can be adapted to the needs of the concrete field. In these references,
the reader can also find references to other items in case he desires to go
deeper into the subject. Arni (Chapter 3) treats the statistical facets of
sampling concrete and its ingredients.
The purpose of sampling will determine to a large extent the desirable
sampling plan and procedures. Is the purpose of sampling to investigate
the quality of a deposit or a product, to determine the homogeneity for a
quality control operation, or to determine whether a material meets specifications requirements?
In general, where an operation is functioning under proper and effective
quality control procedures, a variable sampling plan can be most effective
and economical [23]. In such a plan, a basic full sampling plan and inspection requirement is first developed. When results indicate control and
acceptable homogeneity for a long period of time, the plan can be reduced
to less frequent sampling, but the basic plan is reinstated immediately upon
the appearance of any signs that control has been relaxed.
Aggregates
The workhorse of sampling methods used in aggregate work has been
issued by ASTM Committee D-4 under the title, Sampling Aggregates
(D 75). It treats various procedures for sampling aggregates at various
13
locations such as quarries, pits, railroad cars, bins, and stockpiles. It is

quite definite in its instructions and cautions about homogeneity or variability of the material in the locations being sampled, but not about the
tools. It requires the taking of a sufficient number of samples to reflect the
variability of the material being sampled but does not indicate how to develop
a sampling plan for such a series of samples or under what conditions tompositing is valid. ASTM Recommended Practices E 105, E 122, E 141, and
D 3665 cited previously provide guidelines for developing such sampling
plans.
Various methods for sampling, whether natural deposits, finished aggregates, other concrete-making materials, or concrete itself, can be found
elsewhere [16-18]. The references on coal sampling, bulk sampling, and
sampling in general [1-15, 23-25] provide additional information that is
applicable, even though not standardized specifically for the aggregate
field.
Admixtures, Curing, Patching, Bonding, and Sealing Materials

These materials reach the job or the concrete products plant in packaged
form (cans, buckets, barrels, bags, and so on) or are delivered in some
instances in bulk solution form. Others like fly ash (ASTM Sampling and
Testing Fly Ash or Natural Pozzolans for Use as a Mineral Admixture in
Portland Cement Concrete (C 311)) involve the same problems that are
found in cement sampling, but obviously are different from those for aggregates. All of such materials have had some factory processing, and invariably the processor claims very close quality control.
In practice, most large users test the source for approval and after that
rely on the processor's certification that the material is the same as that
submitted for original approval. This is, in many ways, a similar procedure
to that used for cement for practically all ordinary jobs except the large
government contracts.
The acceptance of such a procedure is based on the false assumption
that because it is a processed material there are no significant variations
in the final product. This assumption is made for convenience rather than
because it represents a true picture--it eliminates the need for continuous
testing that is costly, takes time, and can delay the work.
This works fine as long as everything turns out smoothly, but if some
problem develops in the concrete, it becomes impossible to determine with
any certainty what the significant factors might be. There are not only the
normal variations in the materials that are not on record, but it is always
possible that some human error somewhere along the line of processing,
warehousing, and delivery may have been made and gone undetected.
Whether this calculated risk is worth taking is up to the responsible en-
14
GENERAL
gineer to determine in each case. But if proper original or continuous

sampling is desired, the procedures and plans are much simpler than in
the ease of bulk aggregates.
The proper sampling for these materials becomes a problem similar to
that of sampling any manufactured packaged article or item. Many plans
have been worked out and proven satisfactory and usually can be adapted
to the concrete field [1-16, 23-25]. Usually this adaptation consists of a
two-step procedure. The first step is the random sampling of a lot or shipment which results in the selection of a number of the packages out of the
whole to be used as evidence of the variation pattern of the lot. The second
step is to take individual samples from each package that has been set
aside for sampling.
The first step is based on probabilities and statistical principles, while
the second step requires manipulative care to ensure that the sample is
representative of the container. When the package is relatively small and
the product is in the form of a liquid of some type, the second step is easy,
as proper shaking or stirring will permit the obtaining of a representative
sample without much trouble. But when one gets into the larger packages,
such as a 50-gal drum, a tank car, or tank truck, there is always the possibility that there is stratification, and the sample has to be taken in such a
manner that it is a composite of the various strata.
Compositing assumes that in the unloading, transferring, and handling
the various strata will be intermixed. If this is not the case, then the sampling has to be related to the handling and dispensing procedures in order
to develop the variation patterns. This is likely to get rather complicated,
and can be avoided if the storage container from which the material is
dispensed is kept continuously stirred to maintain the homogeneity of its
contents (ASTM Specification for Liquid Membrane-Forming Compounds
for Curing Concrete (C 309)).
In the case of a powdered material that is to be dissolved prior to use, a
composite sample from various portions of each of the shipping containers
set aside from the lot or shipment for sampling should be taken for the
second step of sampling (ASTM Sampling and Testing Calcium Chloride
for Roads and Structural Applications (D 345) and Specification for Chemical Admixtures for Concrete (C 494)). For products such as form insulating
mats, curing mats, single use cylinder molds, and so on, procedures have
to be adapted for each material. To date, ASTM Specification for Molds
for Forming Concrete Test Cylinders Vertically (C 470) is the only standard
available in this area, to this writer's knowledge.
Detailed guidelines for sampling these classes of materials, which have
proved satisfactory in general industrial applications and which in many
cases can be adapted to the concrete field, may be found in the industrial
product control literature [1,4,10,23,25].
15
Fresh Concrete
ASTM Sampling Fresh Concrete (C 172) outlines in detail the manipulations in taking a sample from a batch but does not describe sampling plans
that permit the selection of the batches to be sampled. ASTM Recommended Practice D 3665 provides guidance for developing such sampling
plans. In general, the same problems that are encountered in bulk aggregate sampling from truckloads are found also in sampling batches of concrete; it is fortunate that the most difficult problem of sampling aggregates
from a stockpile does not exist for concrete. Sampling procedures, cornpositing, details of manipulations, and sampling plans should depend on
the purpose for which the sampling is being done.
Hardened Concrete
ASTM Obtaining and Testing Drilled Cores and Sawed Beams of Concrete (C 42) governs the taking of cores or the sawing of beams from hardened concrete, but no mention is made of how to decide at what point or
points such specimens should be secured. Some suggestions are given by
Abdun-Nur [26]. The American Concrete Institute (ACI) Committee 214
on Evaluation of Results of Tests Used to Determine the Strength of Concrete is trying to develop a recommended practice which it is hoped will
cover this detail. What is as important is that the recommended practice
will cover the significance of the tests made on such cores or beams. Standardization of this item is more difficult because it involves a lot of judgment which cannot be standardized. These are important as safety decisions
involving potential life hazard, and economic decisions involving large sums
are frequently made on the basis of testing of specimens from hardened
concrete.
A recent example of such problems is a case where a large number of
cores were taken and tested by several laboratories. These tests indicated
a certain level of strength and a variation pattern--all with the exception
of one laboratory, where three cores indicated consistently a 50 percent
higher strength than those tested by other laboratories--a most improbable
situation if sampling had been properly conducted.
Needed Improvements and Their Benefits

A S T M Standards
In many cases, the existing ASTM standards contain the manipulative
instructions for taking samples of the various concrete ingredients and
other materials used in the curing, protection, and repair of concrete, and
16
GENERAL
also for the sampling of hardened concrete. In most cases, modifications or

revisions and broadening of scope of these procedures are desirable.
Recognition of Variability
Essentially, the big problem in sampling in the field of concrete technology is the dissemination of the ideas regarding variability. Since it must
become commonly accepted that there is no such thing as "representative
batch" or "typical batch," a sampling plan that reflects the variations in
any given situation or universe is needed. Such a plan would provide the
proper information for control of concrete and its ingredients as a basis for
determining specification conformance and permit making realistic decisions when problems arise.
Probability Sampling
Probability sampling is our greatest need at the present time. To bring
it about requires training of many people, from the engineer down to the
inspector-sampler. The importance of proper sampling and its effect on
testing and making the final most important decisions needs to be stressed
repeatedly. The basic fundamentals for probability sampling are available,
and industries other than the concrete industry with similar problems,
such as the coal and ore industries, have already done a lot of development
work adaptable to the concrete field, particularly in the aggregate area
[1-15 and 23-25]. Just as the continued stirring of a liquid or solution
keeps it so homogeneous that one can take a sample at any time and at
any place in the container, probability sampling achieves the equivalent of
homogenization mathematically and permits a calculation of its reliability.
It has been said that the probability testing of aggregates is of no importance since the testing of the plastic concrete is the final criterion. This
may be true for determining compliance with specifications in the few cases
where the latter are based on final performance only. But, for the concrete
producer who is trying to control his product, it is essential that he use the
best probability sampling techniques to determine the pattern of variation
of the aggregates or any other ingredients that go into his concrete. This
enables him to make the proper allowances in the production of his concrete to meet a given specification requirement.
Inasmuch as the probability approach to sampling provides the required
degree of reliability from the smallest number of samples of any other
approach, in most cases the reduced quantity of sampling will result in an
overall decrease of the total sampling cost. Added to this is the fact that
the cost of testing specimens prepared from unreliable samples will be
eliminated, and the total saving becomes very appreciable. Of course, the
main advantage is that the resulting data are more reliable, and therefore
the decisions based on such information become more realistic. This perCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:38:50 EDT 2014
17
mits designing to closer tolerances or lower factors of safety, and in the long
run would result in safer and more economical engineering structures.
Notwithstanding all that has been stressed about the need and usefulness
of probability or random sampling, it must be stressed that because of the
widespread availability of computers and because engineers and statisticians
are so comfortable with large amounts of data, one must keep in mind that
a balance must be kept between the practical and theoretical; and frequently judgment must be used in lieu of complicated theoretical considerations.
Trained Personnel
Such an approach requires sophisticated and properly trained persons
at various levels, but once plans and procedures have been set, the actual
sampling becomes a question of following instructions and being sure that
there is adequate supervision by the more sophisticatedly trained personnel
of the lower grade personnel who actually take samples. Thus, the cost of
supervision and higher grade personnel might increase, but this will be
compensated for by the lower cost of the less skilled samplers.
Summary
Sampling is probably the most important step in the testing sequence.
On this information, millions of dollars change hands daily--yet, it is the
most neglected activity in this area of engineering.
The one thing needed above all else is the realization that variations in
composition and properties of concrete and concrete making materials are
basic and can only be reduced but never completely eliminated. The degree
of homogeneity that can be attained in practice depends on the cost of reducing variability as compared with the benefits derived therefrom [19].
Probability sampling and sampling plans are essential to the determination of the true pattern of the existing variations, and fundamental to the
problem of controlling any of the products going into concrete and of the
concrete itself.
Control must be made an integral part of every step in production. Control cannot be imposed from outside by the job owner or the purchaser
of over-the-counter products; all the owner or purchaser can do is determine whether the material as produced does or does not meet specifications.
Incentives built into the specifications can motivate the producer to develop
proper control [27-30].
It is necessary to sample not only at all steps of production to guide production procedures, but sampling should also be done on the finished
product as close as possible to the point of use to ascertain compliance
with specifications. Handling of the materials after they have been sampled
18
GENERAL
needs much attention; otherwise, the samples will not reflect the materials
as they are incorporated into the structure.
More detailed instructions and standards can provide better guidance.
Better qualified and trained personnel, under proper supervision, are a
necessity to achieve maximum reliability with minimum sampling as a basis
for final important economic and safety decisions. This would improve the
present sampling situation immensely.
References
[1]
[2]
[3]
[4]
Bulk Sampling, A S T M STP 114, American Sociaty for Testing and Materials, 1951.
Coal Sampling, A S T M STP 162, American Society for Testing and Materials, 1954.
Bulk Sampling, A S T M STP 242, American Society for Testing and Materials, 1958.
"Usefulness and Limitations of Samples," Proceedings, American Society for Testing
and Materials, Vol. 48, 1948, pp. 857-895.
[5] Tanner, Louis and Deming, Edwards, "Somg Problems in the Sampling of Bulk Materials," Proceedings, American Society for Testing and Materials, Vol. 49, 1949, pp.
1181-1188.
[6] Slonim, M. J., Sampling in a Nutshell, Simon and Schuster, New York, 1960.
[7] Tanner, Louis, "Probability Sampling Methods for Wool," A S T M Materials Research
and Standards, March 1961, pp. 172-175.
[8] Cochran, W. G., Sampling Techniques, 2nd ed., Wiley, New York, 1963.
[9] Bicking, C. A., "Bibliography on Sampling of Raw Materials and Products in Bulk,"
Technical Association of the Pulp and Paper Industry, Vol. 47, No. 5, May 1964.
[10] Symposium on Some New Techniques in Sampling and Quality Control, presented at
67th Annual Meeting, American Society for Testing and Materials, June 1964 (unpublished).
[11] Bicking, C. A., "The Sampling of Bulk Materials," A S T M Materials Research and
Standards, March 1967, pp. 95-116.
[12] Duncan, A. J., "Contributions of ASTM to the Statistical Aspects of the Sampling of
Bulk Materials," A S T M Materials Research and Standards, Nov. 1967, pp. 477-485.
[13] Visman, J., "A General Sampling Theory," A S T M Materials Reseach and Standards,
Nov. 1969, pp. 8-13, 51-56, 62, 64, and 66.
[14] Sampling, Standards, and Homogeneity, A S T M STP 540, American Society for Testing
and Materials, 1973.
[15] "Acceptance Sampling," A S T M Standardization News, Sept. 1975.
[16] Handbook for Concrete and Cement, Waterways Experiment Station, U.S. Army, Corps
of Engineers, Vicksburg, Miss., 1949 and subsequent revisions.
[17] ACI Manual of Concrete Inspection, 6th ed., American Concrete Institute, 1975.
[18] Concrete Manual, 8th ed., U.S. Bureau of Reclamation, 1975.
[19] Abdun-Nur, E. A., "How Good is Good Enough," Proceedings, American Concrete
Institute, Vot. 59, No. 1, Jan. 1962, pp. 31-46 and 1219-1244.
[20] Abdun-Nur, E. A., "Inspection and Quality Control," 6th Annual Concrete Conference,
Utah State University, Logan, Utah, March 1964.
[21] Wills, M. H., Jr., "Contamination of Aggregate Samples," National Sand and Gravel
Association, July 1964.
[22] Abdun-Nur, E. A., "Accelerated, Early, and Immediate Evaluation of Concrete
Quality," International Symposium on Accelerated Strength Testing of Concrete, American
Concrete Institute Convention, Mexico City, Mexico, 24-29 Oct. 1976.
[23] Sampling Procedures and Tables for Inspection by Variables for Percent Defective,
MIL-STD-414, Superintendent of Documents, Washington, D.C., 11 June 1957.
[24] Singlg-Level Continuous Sampling Procedures and Tables for Inspection by Attributes,
Handbook H-I07, Superintendent of Documents, Washington, D. C., 30 April 1959.
[25] Sampling Procedures and Tables for Inspection by Attributes, MIL-STD-IOSD, Superintendent of Documents, Washington, D. C., 29 April 1963.
19
[26] Abdun-Nur, E. A., "Sampling of Concrete in Service," Highway Research Board Special
Report 106, 1970, pp. 13-17.
[27] Abdun-Nur, E. A., "Designing Specifications--A Challenge," Proceedings, Journal of
the Construction Division, American Society of Civil Engineers, Separate No. 4315,
May 1965.
[28] Abdun-Nur, E. A., "Product Control and Incentives," Proceedings, Journal of the Construction Division, American Society of Civil Engineers, Separate No. 4900, Sept. 1966.
[29] Abdun-Nur, E. A., "Adapting Statistical Methods to Concrete Production," Proceedings,
National Conference on Statistical Quality Control Methodology in Highway and Airfield
Construction, Nov. 1966.
[30] Adbun-Nur, E. A., "What is the Quality Assurance System?" Transportation Research
Board, 55th Annual Meeting, Washington, D. C., 19-23 Jan. 1976.
APPENDIX
ASTM DESIGNATIONS UNDER JURISDICTION OR RELATED TO WORK
OF ASTM COMMITTEE C-9
ASTM
Designation
Title
Sample Origin or
Requirement"
SAMPLING AND TEST METHODS

C 29
C 30
C 31
C 39
C 40
C 42
C 70
c 78
C 85
C 87
C 88
Unit Weight of Aggregate

Voids in Aggregate for Concrete
Making and Curing Concrete Test Specimens in the
Field
Compressive Strength of Cylindrical Concrete
Specimens
Organic Impurities in Sands for Concrete
Obtaining and Testing Drilled Cores and Sawed
Beams of Concrete
Surface Moisture in Fine Aggregate
Flexural Strength of Concrete (Using Simple Beam
With Third-Point Loading)
Cement Content of Hardened Portland Cement
Concrete
Effect of Organic Impurities in Fine Aggregate on
Strength of Mortar
Soundness of Aggregates by Use of Sodium Sulfate
or Magnesium Sulfate
D 75
D 75
C 172
C 31, C 192
D 75
yes
D 75
C 31, C 192
yes
D 75
D 75
"This column indicates the best source or method of sampling for the particular designation
available in existing ASTM standards.
1. "yes" indicates that the standard includes sampling instructions.
2. "no" indicates that no ASTM standard for sampling the material in question is available.
3. "ina" indicates that specific sampling procedures are inapplicable.
4. When more than one reference is shown, each reference reflects one of the ingredients
needed for the standard.
20
GENERAL
ASTM
Designation
C 116
C 117
C
C
C
C
123
127
128
131
C 136
C 138
C
C
C
C
142
143
156
157
C 172
C 173
C 174
C 183
C 192
C 215
C 227
C 231
C 232
C 233
C 234
C 235
C 289
C 293
C 311
C 341
C342
C 360
C 403
C 418
Title
Sample Origin or
Requirement a
Compressive Strength of Concrete Using Portions

of Beams Broken in Flexure
Materials Finer Than No. 200 (75-~m) Sieve by
Mineral Aggregates by Washing
Lightweight Pieces in Aggregate
Specific Gravity and Absorption of Coarse Aggregate
Specific Gravity and Absorption of Fine Aggregate
Resistance to Abrasion of Small Size Coarse Aggregate
by Use of the Los Angeles Machine
Sieve or Screen Analysis of Fine and Coarse Aggregates
Unit Weight, Yield, and Air Content (Gravimetric)
of Concrete
Clay Lumps and Friable Particles in Aggregates
Slump of Portland Cement Concrete
Water Retention by Concrete Curing Materials
Length Change of Hardened Cement Mortar and
Concrete
Sampling Fresh Concrete
Air Content of Freshly Mixed Concrete by the Voiumetric Method
Measuring Length of Drilled Concrete Cores
Sampling Hydraulic Cement
Making and Curing Concrete Test Specimens in the
Laboratory
Fundmental Transverse, Longitudinal, and Torsional
Frequencies of Concrete Specimens
Potential Alkali Reactivity of Cement-Aggregate
Combinations (Mortar-Bar Method)
Air Content of Freshly Mixed Concrete by the Pressure
Method
Bleeding of Concrete
Air Entraining Admixtures for Concrete
Comparing Concretes on the Basis of Bond Developed
With Reinforcing Steel
Scratch Hardness of Coarse Aggregate Particles
Potential Reactivity of Aggregates (Chemical Method)
Flexural Strength of Concrete (Using Simple Beam
With Center-Point Loading)
Sampling and Testing Fly Ash for Use as an Admixture in Portland Cement Concrete
Length Change of Drilled or Sawed Specimens of
Cement Mortar and Concrete
Potential Volume Change of Cement-Aggregate
Combinations
Ball Penetraton in Fresh Portland Cement Concrete
Time of Setting of Concrete Mixtures by Penetration
Resistance
Abrasion Resistance of Concrete by Sandblasting
C 78, C 293
D
D
D
D
75
75
75
75
D 75
D 75
C 172
D 75
C 172
no
yes
yes
C 172
C 42
D 75
D 75, yes
C 192
D 75, yes
C 172
C 172, C 192
D 75, C 192
C 192, no
D 75
D 75
C 31, C t92
yes
C 42
D 75
C 172
C 172, C 192
no
ASTM
Designation
C 441
C 469
C495
C 496
C 512
C 513
C 535
C 566
C 567
C 586
C 597
C 617
C 641
C 642
C 666
C 671
C 672
C 683
C 684
C 702
C 779
C 796
C 803
C 805
C 827
D 75
D 345
D 545
D 1191
Title
Sample Origin or
Requirement a
Effectiveness of Mineral Admixtures in Preventing

Excessive Expansion of Concrete Due to the
Alkali-Aggregate Reaction
Static Modulus of Elasticity and Poisson's Ratio of
Concrete in Compression
Compressive Strength of Lightweight Insulating
Concrete
Splitting Tensile Strength of Cylindrical Concrete
Specimens
Creep of Concrete in Compression
Securing, Preparing, and Testing Specimens from
Hardened Lightweight Insulating Concrete for
Compressive Strength
Resistance to Abrasion of Large Size Coarse Aggregate by Use of the Los Angeles Machine
Total Moisture Content of Aggregate by Drying
Test for Unit Weight of Structural Lightweight
Concrete
Potential Alkali Reactivity of Carbonate Rocks for
Concrete Aggregates (Rock Cylinder Method)
Pulse Velocity Through Concrete
Capping Cylindrical Concrete Specimens
Staining Materials in Lightweight Concrete Aggregates
Specific Gravity, Absorption, and Voids in Hardened
Concrete
Resistance of Concrete to Rapid Freezing and Thawing
Critical Dilation of Concrete Specimens Subjected to
Freezing
Scaling Resistance of Concrete Surfaces Exposed to
Deicing Chemicals
Compressive and Flexural Strength of Concrete
Under Field Conditions
Making, Accelerated Curing, and Testing of Concrete
Compression Test Specimens
Reducing Field Samples of Aggregate to Testing Size
Abrasion Resistance of Horizontal Concrete Surfaces
Foaming Agents for Use in Producing Cellular Concrete Using Preformed Foam
Penetration Resistance of Hardened Concrete
Rebound Number of Hardened Concrete
Early Volume Change of Cementitious Mixtures
Sampling Aggregates
Sampling and Testing Calcium Chloride for Roads
and Structural Applications
Preformed Expansion Joint Fillers for Concrete
(Nonextruding and Resilient Types)
Concrete Joint Sealers
21
yes
C 31, C 192
C 172
C 31, C 192
yes
yes
D 75
D 75
C 192
D 75, yes
no
no
D 75, no
no
C 192, yes
yes
yes
C 172
C 172
D 75
no
C 192
no
yes
C 192, yes
yes
yes
no
no
22
GENERAL
ASTM
Designation
D 1851
D 1855
D 3042
D 3407
D 3408
Title
Sample Origin or
Requirement ~
Concrete Joint Sealers, Cold-Application Type

Jet-Fuel-Resistant Concrete Joint Sealer, Hot-Poured
Elastic Type
Insoluble Residue in Carbonate Aggregates
Joint Sealants, Hot-Poured for Concrete and Asphalt
Pavements
Joint Sealants, Hot-Poured, Elastomeric-Type, for
Portland Cement Concrete Pavements
yes
no
yes
yes
yes
SPECIFICATIONS
C 10
C 33
C 94
C 150
C 171
C 260
C 309
C 330
C 331
C 332
C 387
C 470
C 494
C 595
C 618
C 637
C 685
D 98
D 448
D632
D 994
D l190
D1751
D1752
D 1850
Natural Cement
Concrete Aggregates
Ready-Mixed Concrete
Portland Cement
Sheet Materials for Curing Concrete
Air-Entraining Admixtures for Concrete
Liquid Membrane-Forming Compounds for Curing
Concrete
Lightweight Aggregates for Structural Concrete
Lightweight Aggregates for Concrete Masonry Units
Lightweight Aggregates for Insulating Concrete
Packaged, Dry, Combined Materials for Mortar and
Concrete
Molds for Forming Concrete Test Cylinders Vertically
Chemical Admixtures for Concrete
Blended Hydraulic Cements
Fly Ash and Raw or Calcined Natural Pozzolan for
Use in Portland Cement Concrete
Aggregates for Radiation-Shielding Concrete
Concrete Made by Volumetric Batching and Continuous Mixing
Calcium Chloride
Standard Sizes of Coarse Aggregate for Highway
Construction
Sodium Chloride
Preformed Expansion Joint Filler for Concrete
(Bituminous Type)
Concrete Joint Sealer, Hot-Poured Elastic Type
Preformed Expansion Joint Fillers for Concrete
Paving and Structural Construction (Nonextruding
and Resilient Bituminous Types)
Preformed Sponge Rubber and Cork Expansion Joint
Fillers for Concrete Paving and Structural Construction
Concrete Joint Sealer, Cold-Application Type
C 183
D 75
C 172
C 183
yes
yes
yes
D 75
D 75
D 75
C 192, yes
yes
yes
C 183
yes
D 75, yes
D 75, C 172,
C 192
D 345
D 75
yes
yes
yes
yes
yes
D 140
ASTM
Designation
D 1854
D 2628
D 3405
D 3406
Title
Sample Origin or
Requirement"
Jet-Fuel-Resistant Concrete Joint Sealer, Hot-Poured

Elastic Type
Preformed Polychloroprene Elastomeric Joint Seals
for Concrete Pavements
Joint Sealants, Hot-Poured, for Concrete and
Asphalt Pavements
Joint Sealants, Hot-Poured, Elastomeric-Type, for
Portland Cement Concrete Pavements
D 140
yes
yes
yes
RECOMMENDED PRACTICES
C 295
C 457
C 682
C 801
C 802
C 823
D 3665
E 105
E 122
E 141
Petrographic Examinations of Aggregates for Concrete

Microscopical Determination of Air-Void Content and
Parameters of the Air-Void System in Hardened
Concrete
Evaluation of Frost Resistance of Coarse Aggregates
in Air-Entrained Concrete by Critical Dilation
Procedures
Determining Mechanical Properties of Hardened
Concrete Under Triaxial Loads
Conducting an Interlaboratory Test Program to
Determine the Precision of Test Methods for
Construction Materials
Examination and Sampling of Hardened Concrete in
Constructions
Random Sampling of Paving Materials
Probability Sampling of Materials
Choice of Sample Size to Estimate the Average
Quality of a Lot or Process
Acceptance of Evidence Based on the Results of
Probability Sampling
23
D 75, yes
yes
C 192, yes
C 192
yes
yes
ina
ina
ina
ina
H. T. A r n i t
Chapter 3--Statistical
Considerations in Sampling and
Testing
Introduction
It is not the purpose of this chapter to provide a short text on general

statistical methods and procedures. Many sources, some of which are listed
in the references, are already available that provide such information in
great detail. Rather, the purpose is to identify situations that are peculiar
to the testing and evaluation of cement and concrete in which statistical
techniques can be of value, to suggest techniques that are appropriate to
use in these situations, to provide, by references to other sources as much
as possible, information about how to use these techniques in the given
situations, and, last but not least, to point up some of the pitfalls and
misconceptions that often cause problems in the application of statistical
techniques.
Sampling 2
The first consideration in any process of evaluation of characteristics

of a material is, How shall the material be sampled so that the results of
whatever measurement process is to be applied will provide the needed
information and the evaluation can be made with the desired degree of
confidence?
Variations in samples of concrete and concrete-making materials make
it difficult to obtain samples with the assurance that they are completely
representative of the source or production of these materials. Test results
l Retired; formerly materials research engineer, Federal Highway Administration, Department of Transportation, Washington, D.C. 20590.
2Adapted from Cordon, W. A., "Size and Number of Samples and Statistical Considerations in Sampling," Significance of Tests and Properties of Concrete and Concrete-Making
Materials, A S T M STP 169-,4, Americal Society for Testing and Materials, 1960, p. 21.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
ARNI ON STATISTICAL CONSIDERATIONS
25
are more reliable as the number and size of samples increase, but additional samples or samples of greater size increase the cost of testing.
It is necessary, therefore, to establish the accuracy desired in each case
commensurate with funds and facilities available and with how much information it is deemed necessary to obtain.
It may be better to make no tests than to make tests with poor samples
which do not portray the actual properties of the materials. An engineer or
architect who must rely on samples or tests that do not accurately represent
materials or structures could probably make more objective decisions if
there were no samples or tests available. He could at least make conservative assumptions rather than rely on fallacious information.
Tests for concrete and concrete materials may conveniently be separated
into the following general categories.
1. Acceptance Tests--Tests which demonstrate that the materials in
question will meet specific requirements of the work as specified. Examples
are grading of aggregate, soundness of aggregate, slump, and air content
of concrete.
Performance tests may also be used as acceptance tests. These tests
demonstrate how a given material will perform in concrete under field or
simulated field conditions. Examples are freezing and thawing of concrete
prisms, expansion of concrete prisms due to alkali reactivity, strength of
concrete, drying shrinkage, and rate of hardening.
2. Construction Control Tests--Construction control tests are made
at intervals throughout construction of a project. These tests not only provide a check on the performance of materials and may be used for acceptance or rejection of the work, but they also measure the uniformity of
concrete produced. The strength of 6 by 12-in. (or 150 by 300 mm) cylinders
is universally accepted as a standard of control in the United States, since
most properties of concrete and concrete materials, as well as the influence of testing and construction practices, are reflected in the strengths
of the cylinders.
3. Research Procedures--Tests which are conducted in the laboratory
or field to establish relationships among the variables that influence the
properties of concrete. Examples include relationship between strength
and water-cement ratio, influence of admixtures on the properties of
concrete, and rate of hardening at various temperatures.
Concrete Aggregates
Acceptance tests for concrete aggregates establish the suitability of a
deposit, source, or quantity of finished production. During preliminary
reconnaisance of available deposits, aggregate quality evaluation may be
based on test results of a composite sample. This is possible since the
general characteristics of aggregate quality can be estimated from a relaCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:38:51 EDT 2014
26
GENERAL
tively small n u m b e r of tests. After preliminary analysis, the properties of

and variations in the most promising source can be established with
samples from locations t h r o u g h o u t the area and depth o f the deposit.
As a general rule, aggregate samples need only be large enough to include a representative portion of all materials present in the source and
provide ample material for all tests contemplated as specified by standard
test procedures. The n u m b e r of samples will vary with the size of the deposit, and the size of samples will vary with the m a x i m u m size of aggregate
occurring in the deposit. For gradation analysis, for example, samples
should be large enough to assure occurrence of particles of the largest
dimension in sufficient n u m b e r so that the inclusion or exclusion of one of
these large particles will not affect the grading significantly. Variations in
the size of samples required for petrographic examination are shown in
Table 1 [1]. 3
TABLE 1--Size of sample requiredfor petrographic examination.
Size Fraction
Weight of 300
Particles
3 to 11/2in.
11/2 to 3/4 in.
3/4 to 3/a in.
3/8 to 3/[6 in.
No. 4 to 8
No. 8 to 16
No. 16 to 30
No. 30 to 50
No. 50 to 100
57 lb (26 kg)
19.1 lb (8.7 kg)
2.6 lb (1.2 kg)
0.75 lb (340 g)
15 g
2.1 g
0.28 g
0.033 g
0.0066 g
Conversionfactors-1 in. = 25.4 mm, and

1 lb = 453.6 g.
Other Materials
Materials used in making concrete (other than portland cement and
aggregates) include water and admixtures. Other materials, used to treat
hardened concrete surfaces, are also discussed.
W a t e r - - I t is generally taken for granted that available water will be
suitable for making concrete. Water may contain harmful salts, however,
and where there is a question, a chemical analysis of a water sample is
made to establish its suitability.
U n d e r this procedure it is assumed that the chemical composition of a
water source will not vary during the construction of the project.
27
Mineral Admixtures--Pozzolans and other mineral admixtures may be

proprietary materials or may come from natural deposits.' A natural deposit will require sufficient samples to establish the extent and depth of the
deposit. ASTM Sampling and Testing Fly Ash or Natural Pozzolans for
Use as an Admixture in Portland Cement Concrete (C 311) requires an
3.6-kg (8-1b) sample for each 91 Mg (100 tons) of material.
Chemical, physical, and performance tests for compressive strength and
drying shrinkage are required.
Statistical considerations in testing mineral admixtures require establishing a reliable average for the deposit, lot, or shipment under test
and uniformity from sample to sample.
Chemical Admixtures--Nearly all chemical admixtures are proprietary
materials sold under a particular brand name. With minor exceptions, the
suitability of chemical admixtures is established by their performance when
used in concrete. The ASTM Specification for Air-Entraining Admixtures
for Concrete (C 260) and Specification for Chemical Admixtures for Concrete (C 494) give specific instructions for size and number of samples to
be taken.
Statistical considerations are concerned with the sample being representative of the lot or production to be used on the specific work.
Curing Materials--Membrane curing compounds are used to retain the
mixing water in concrete until curing has taken place. These materials
are composed of various combinations of chemicals and are sold under
various trade names. The performance of these materials is the responsibility of the manufacturer. The ASTM Test for Water Retention by Concrete Curing Materials (C 156) determines the performance of curing
materials in preventing evaporation. This test establishes the acceptance
of the material and is assumed to be representative of the production or lot
under consideration. The ASTM Specification for Liquid Membrane Forming Compounds for Curing Concrete (C 309) specifies that samples be
taken for each lot, batch, or other unit of production.
Bonding and Patching Materials--The use of bonding and patching
materials, particularly epoxy resins, has increased in importance in the last
several years. The composition of these materials is complex and is varied
to produce specific desired results. It is necessary to rely on the manufacturer's recommendations when using these materials.
Sealing Materials--The application of sealing materials to hardened concrete reduces the absorption of moisture and is a promising method of
reducing freezing and thawing deterioration. Proprietary materials for this
purpose are on the market and, in addition, combinations of 50 percent
linseed oil and 50 percent thinner are recommended in the literature. The
ASTM Test for Scaling Resistance of Concrete Surfaces Exposed to Deicing Chemicals (C 672) is used to measure the effectiveness of sealing
materials in reducing absorption.
28
GENERAL
Concrete
Acceptance and Performance Tests--Acceptance and performance of
concrete is established by tests of fresh and hardened concrete. Tests of
fresh concrete provide a means of checking the proportions and properties
of the concrete mixture. The slump test and other measures of consistency
indicate the water content and, indirectly, water-cement ratio of a given
mixture and may be used to reject concretes having excessive water. The
measurement of the air content (ASTM Recommended Practice for Microscopical Determinations of Air-Void Content and Parameters of the AirVoid System in Hardened Concrete (C 457)) and distribution and size of air
voids enables the engineer to control properly the use of air-entraining
agents. The density or unit weight of the concrete is used to measure the
volume of a given batch of concrete and the proportions of ingredients of
fresh concrete based on the volume of concrete produced [2]. This provides
an accurate means of determining the cement content per unit volume of
concrete.
The required number of tests of fresh concrete will depend upon the uniformity of production. The concrete mix will generally be accepted, corrected, or rejected on the basis of individual tests. As in the case of aggregate acceptance tests, tests of fresh concrete may have limited statistical
considerations since individual tests serve as a basis for action. This does
not preclude analysis of the uniformity and control of these properties.
In the final analysis, the strength of concrete measured by the strength
of 6 by 12-in. (or 150 by 300-mm) cylinders broken in compression after 28
days of moist curing is generally accepted as one measure of concrete
quality. Flexural strength of beams is also used extensively for pavements,
and entrained air content is also required for concrete that is to be exposed
to freezing and thawing.
It is assumed that samples taken each day represent the concrete placed
during that day, and like other performance tests the acceptance of concrete is based on these tests. Evaluation of strength tests is discussed
further in the next section.
Evaluation of Strength Tests
One of the earliest and most widely used applications of statistics in the
concrete field has been in the area of evaluation of strength tests both of
mortar cubes for the testing of cement strength and, more extensively, for
the analysis of strengths of concrete specimens, usually, in the United
States, in the form of 6 by 12-in. cylinders. The chief pioneer in this effort
was Walker who published his study in 1944 [3]. In 1946, largely at the
instigation of Walker, the American Concrete Institute (ACI) began
work on statistical evaluation of compression tests that eventually resulted
in the publication of ACI Standard Recommended Practice for Evaluation
29
of Strength Test Results of Concrete (ACI 214-77). First published as a

standard in 1957, this document has undergone a number of revisions, the
most recent of which was adopted in 1977.
In 1971 a symposium was conducted at the ACI Fall Convention on the
subject "Realism in the Application of ACI Standard 214-65." This symposium presented valuable information on the meaning and use of ACI 214
and resulted in a symposium volume that included seven papers presented
at the symposium, a reprint of ACI 214-65, and reprints of two earlier
papers dealing with evaluation of concrete strengths [4].
Evaluating Standard Cylinders

The methods used for evaluating strength measurements in ACI 214
involve several important statistical concepts that are useful in other
situations. These concepts include the following.
1. Results obtained in the testing of carefully made concrete specimens
from a concrete production that is as uniform as is possible to make it
exhibit a wide scatter.
2. The aggregate of such results exhibits a distribution pattern of a
characteristic form that has a characteristic central tendency, measured by
the mean or arithmetic average, and a characteristic spread on either side
of the high point of the curve that describes the distribution, measured by
the standard deviation or the coefficient of variation. The meaning of these
terms and directions for calculating them are given in ACI 214 as well as in
other statistical references [5-8].
3. Mean and average are synonymous terms, but standard deviation and
coefficient of variation are not. These terms are not merely alternative ways
of measuring the same thing and are not equally applicable in all situations. This point is discussed further.
4. It is a practical impossibility to design concrete so that no standard
test will ever fall below any realistic so-called minimum design strength.
5. The frequency with which routine tests will fall below a so-called
specified minimum is determined by the average strength of the production and the spread of the distribution, whether the latter is measured by
the standard deviation or the coefficient of variation.
Most of these topics are adequately covered in ACI 214 and in other
statistical references, and the coverage will not be duplicated here. A word
is appropriate, however, about the difference in use between standard
deviation and coefficient of variation.
Coefficient of Variation versus Standard Deviation

It is a common misconception that the standard deviation and the coefficient of variation are merely alternative ways of expressing the spread
30
GENERAL
of measurements, and that which one should be used in a given situation

is merely a matter of convenience or preference. This is not true, and over
the years the problem of which measure of variability is appropriate to use
in evaluation of concrete strengths has been the subject of much discussion
and controversy.
The standard deviation is the most generally accepted and widely used
measure of variability. The range, which is often used as a simpler statistic
is actually most often used, either explicitly or implicitly, as a means of
estimating the standard deviation, and its use is based on the same underlying assumptions about the data. So far as compressive strength is concerned, the assumption is that, although the average strength level of a
group or population of measurements from a given production may change,
the standard deviation remains the same. The use of the coefficient of
variation is appropriate whenever the standard deviation varies with the
level of strength, specifically when the standard deviation is proportional to
the average.
In the 1957 and 1965 versions of ACI 214 it was assumed that the coefficient of variation was the appropriate statistic for concrete strength
measurements. The ACI Building Code Requirements for Reinforced Concrete (ACI 318-71) adopted the standard deviation, although previous
editions of this standard merely referred to ACI 214-65. More recent research has indicated that neither standard deviation nor coefficient of
variation is appropriate over the whole range of concrete strengths encountered in practice. Thus, the most recent revision of ACI 214 recognizes
both measures and recommends that the standard deviation be used for
strengths greater than 211 k g f / c m 2 (3000 psi). (See the references listed in
ACI 214 and for a more detailed technical discussion of the differences
between standard deviation and coefficient of variation, see Ref 9.)
Accelerated and In-Place Tests for Evaluating Concrete

The traditional method of evaluating strength or quality of concrete in
general in the United States has been by the use of compressive strength
tests of concrete cylinders, usually 6 by 12-in. (or 150 by 300 mm) cylinders
tested at 28 days. Other countries have used cubes, but the method of
analysis is similar. This method of evaluation has always been subject
to two main criticisms. (a) The test measurements do not adequately
represent the strength of the concrete in place, since, at the very least,
they are not subject to the same curing conditions as concrete in the structure and are tested in a very different environment from that in which the
concrete in the structure exists. (b) Even to the extent that the specimens
represent the strength of concrete in the structure, they represent past
history because of the fact that results are usually not obtained until 28
days after the concrete is placed. This too makes it difficult to do anyCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:38:51 EDT 2014
31
thing about remedying the situation if serious difficulties are found,

especially with the more rapid construction techniques that now prevail.
In an attempt to remedy or minimize the effects of one or both of
these difficulties, a number of other methods of measurement have been
developed. These involve two main categories of measurement systems:
(a) accelerated methods that attempt to assess the strength of standard
specimens at earlier ages, such as 24 to 48 h, and (b) various so-called inplace methods that attempt to determine the strength of concrete in place
in the structure without relying on separately cast and cured specimens.
The first category of methods is still subject to the limitation that the
strength measurements are still not actually representative of strengths
existing in the structure, especially since they rely on various special curing
conditions that cause an acceleration of strength development. Three of
these methods have been standardized in ASTM Standard Method of
Making, Accelerated Curing, and Testing of Concrete Compression Test
Specimens (C 684).
The second category includes a number of systems that measure different aspects of concrete quality. These methods include Swiss hammer or
rebound tests, pulse velocity measurements, penetration tests, and pullout tests. The first three of these are now standardized ASTM tests (ASTM
Test for Rebound Number of Hardened Concrete (C 805), ASTM Test for
Pulse Velocity Through Concrete (C 597), and ASTM Test for Penetration
Resistance of Hardened Concrete (C 803)) and a standard for the fourth is
now being developed.
All of these methods in both categories have their advantages and disadvantages for the purpose of assessing concrete quality, and any one or
any combination of them could be used for the purpose of evaluation of
concrete quality. It is not the purpose of this chapter to discuss the relative
merits of the methods for the purpose of such evaluation, but rather to
examine the statistical methods involved in analyzing the results obtained
by use of the methods, and, in particular, to discuss what is involved in the
investigation of interrelationships between the methods.
Relationships between Accelerated or In-Place Tests and Standard Cylinders
Testing of concrete specimens for compressive strength occupies a

unique position in the field of testing of construction materials because of
the fact that concrete structures have traditionally been designed on the
theory that the concrete in the members of the structure achieved, in a
reasonable length of time after placement of the concrete, a certain specified degree of strength, designated in design formulas as f c ' . In the past,
this specified strength has often been treated as a minimum strength that
the concrete should have. More recently it has been recognized that the
specified strength as used was really not a minimum and it was cornCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:38:51 EDT 2014
32
GENERAL
pletely impractical for it to be so. Thus later specifications and other

documents have tended more and more to define f c ' as a strength level
below which no more that a certain percentage of strength tests would be
allowed to fall (ACI 214 and 318). The problems connected with how
closely in time and character the results of strength tests approximate
those in the structure remain, however, and many attempts have been
made to develop methods for using accelerated or in-place tests to "predict" what the 28-day cylinder strengths would be.
In published reports on these attempts, a statement often occurs to the
effect that a "good correlation" was found between the results obtained
by the use of two alternative test methods when applied to supposedly
identical concretes. More specifically, this applies to the relationship between standard 28-day strength tests and those obtained by one of the
alternative methods.
In these reports, what is meant by a good correlation is often not specified. In other cases, correlation coefficients calculated from paired data obtained by the two methods are calculated and reported. In either case there
is often a lack of understanding of what is really meant by a good correlation or a "significant correlation coefficient."
W h a t is meant by a good correlation is that as one of the variables being
correlated increases, the other tends to increase also, or alternatively, the
second variable tends to decrease as the first increases, as in the case of
a negative correlation. All that is indicated by a good correlation is that the
trend exhibited by the data when one variable is plotted against the other
is sufficiently well defined to permit rejection of the hypothesis that there is
no relationship between the two variables whatever, in other words that the
relationship is completely random.
In the case of strength evaluation, all of the in-place tests mentioned
tend to show increases in the measurements as concrete strength increases
due to continued hydration of the cement or to whatever cause. There will
always be a good correlation between the results of any test that exhibits
such behavior and the results of standard strength tests. Thus a good correlation or a high correlation coefficient dees not in itself constitute evidence
that the relationship is sufficiently close to permit the use of one type of
test as a means of prediction of what will happen in another type. In fact,
the correlation coefficient has very limited use in the field of analyzing engineering data. For further discussion of this point, see Refs 10 and 11.
Reference 11 contains quotations from a number of statistical authorities on
the use of the correlation coefficient.
Regression Lines
The standard method of using paired data as a source of developing
means for predicting one variable from another is by calculating a linear
33
regression line using the method of least squares. This method is described
elsewhere and will not be detailed here [5,8,12-14]. However, some problems with and misconceptions about such so-called prediction lines need to
be discussed. This method is very useful in situations that involve actual
calibration of a measuring device and in which the plotted points approximate the calculated regression line very closely with small scatter. An
example is the calibration of proving rings that are used for calibrating
testing machines. This use is discussed by Hockersmith and Ku [15].
In the situation where two measurement situations bear a statistical
relationship to each other and one of the test methods is used to obtain
measurements that are used to predict measurements of the other type,
there is almost always a large amount of scatter in the data. This is due to
the fact that, although both sets of measurements are affected by changes
in the property of interest, they are actually measuring two different
quantities and are thus affected by different sets of influences that are
extraneous to the property of interest. The fact that this is so is usually
taken account of by the recognition that a large quantity of data are
needed to calculate a reliable line, and, in fact, the more data, the better.
It is curious, however, that recognition of the scatter of the data often
does not extend beyond calculation of the regression line to its use. The
user of the line often acts as though the fact that he has calculated the line
from a lot of data that exhibited a certain scatter ensures that future data
will now realize that this is the true line and conform to it exactly. He
makes a measurement of his independent variable, usually the easier of the
two measurements to make, refers to the line, and reads off the corresponding value of the dependent variable without making any recognition
of the fact that, if the regression line really represents the behavior of the
current data in the same way that it represented that of the data from
which it was calculated, then the future data are subject to the same scatter
as the original data, and the calculated Y value has a wide uncertainty associated with it. Very often, even when the existence of this uncertainty is
recognized, there is lack of recognition of how wide this uncertainty is and
of its effect on the calculated data. This point is illustrated in the case of
prediction of compressive strength measurements from penetration and rebound tests in Refs 1 6 and 17.
If regression lines showing the relationship between data from two
measurement systems are to be calculated and used for any purpose other
than to illustrate what kind of relationship existed in the data, several
things have to be considered.
1. The calculation of a linear regression line involves the calculation
of two parameters: the slope of the line, b~, and the Y intercept of the
line, b 0. Due to the departures of the data points from the line, each of
these parameters has an estimated variance and corresponding standard
error (the square root of the variance): S b l 2 for the slope and Sbo 2 for the
34
GENERAL
intercept. In addition, there is an estimated variance, Sy 2, and corresponding standard deviation, Sy, calculated from the sum of the squares of the
deviations of the measured Ys above and below the fitted line. The three
measures of variation (variances or their corresponding square roots)
should always be given whenever a regression line is reported. The standard
errors of the slope and intercept are frequently omitted. These quantities,
however, give an indication of how significant the relationship is. In
general, ~if the slope is not at least twice as large as its standard error,
there is insufficient evidence to conclude that the true slope is other than
zero, and that the relationship is other than a purely random one. In addition, the number of pairs of data from which the regression was calculated and the range of spread of both X and Y values should be reported
along with the other parameters.
2. In addition to the quantities mentioned in 1., a confidence interval
for the fitted line should be shown on the graph of the line (Fig. 1). How to
COMPRESSIVE
15
I
30
2jO
r,
29 t
STRENGTH, MPa
25
~
30
i
35
4p
28
27
26
--~t . . . . . . . . ~....
Ru
1;
25
~ 24
R I__
2 3 - - - -
I
22
i!
I
2 1 - -
:1
i
0 I
2000
3000
4000
si
sl
~S u 5000
~'u
6000
COMPRESSIVE STRENGTH" PSI

FIG.
1--Regression curve and confidence limits f o r compressive strength versus rebound
numbers,
35
calculate the quantities needed to plot the confidence interval is described

elsewhere [5]. It should be noted that the upper and 'lower confidence
limits are represented by two branches of a hyperbola which are closest
together at the point where X = ~(, the average of all the X values used in
calculating the relationship. Thus uncertainty of the calculated Y increases
as the X value departs from )( in either direction. Often a confidence interval is plotted by multiplying the Sr by the t value for the number of points
used and drawing parallel straight lines above and below the regression
line. This is not correct and gives a too optimistic picture of the uncertainty
of estimated Y values.
3. There are three kinds of confidence intervals that can be calculated
for a fitted regression line: the line as a whole, a point on the line, or a
future value of Y corresponding to a given value of X. Reference 5 describes how to calculate all three of these confidence intervals. It is important to remember that the first case, a confidence interval for the location of the line as a whole, is the only one that is appropriate to use if the
calculated line is to be used repeatedly as a line for predicting future values
of Y from future observed values of X.
4. It does not infrequently happen that, even when all the parameters
and the estimates of their uncertainty and the proper confidence limits are
correctly calculated and reported, thetrue meaning ofthe various quantities
is not understood. The meaning of the confidence interval given with a
predicted Y value involves an understanding of confidence intervals in
general.
Figure 1 taken from Ref 16 illustrates the point. The regression line
shown is based on 16 plotted points relating strengths of 28-day standard
cylinders to averages of 20 Swiss hammer rebound numbers obtained on
slabs made from the same batches of concrete as the cylinders. The figure
also shows the hyperbolic curves representing the upper and lower 95
percent confidence limits for location of the line referred to above.
For a hypothetical rebound number of 25, this figure indicates a calculated average compression strength of 31.2 MPa (4530 psi), with the
95 percent confidence interval extending from 30.0 to 40.5 MPa (4350 to
4710 psi). Since the rebound numbers themselves have a distribution
with a gharacteristic scatter, illustrated in this case by a standard deviation of 0.50 for averages of 20 rebound numbers, the 95 percent confidence
interval for the average rebound measurement is from 24 to 26. These
figures combined with the confidence interval for the line give an approximately 90 percent confidence interval for the predicted compressive
strength of 27.6 to 35.1 MPa (4010 to 5070 psi).
Unfortunately, this is not the limit of the final uncertainty of the predicted result. A 95 percent confidence limit is often interpreted as meaning
that 95 percent of future results will be within the limits given. What it
actually means, however, is that if the experiment is repeated a large
36
GENERAL
number of times, each with the same materials and conditions and with the
same number of determinations, and each time the 95 percent confidence
interval is calculated, then 95 percent of the intervals so calculated will
include the true average. This does not mean that any particular one of the
intervals will actually contain the true average in its exact center. For any
given determination of the line and its confidence interval, the circumstance that the calculated line coincides exactly with the true line is highly
unlikely.
Frequently situations are reported in which a previously calculated regression line is used to make predictions as though the calculated line
exactly represented all future relationships. The fact that the data exhibit
scatter is obvious when they are plotted and is recognized by the circumstance that as large a block of data as possible is used in order to make the
calculated line as reliable as possible. The line is then used as though all
future data conformed to it exactly, whereas all future data are subject to
the same degree of scatter that characterized the original data. There have
been proposals that regression lines relating 28-day cylinder strengths to
results of accelerated tests be used, and then the principles of ACI 214, in
which concrete is fabricated on the assumption that a certain percentage of
tests will be permitted to fall belowfc ', would be applied. Actually, due to
the uncertainties involved in transferring from accelerated to standard
strengths, there is no way that such a process will provide complete assurance that any selected percentage of below-strength tests will be achieved.
In general, then, if any system of measurements is a valid system for
evaluating quality of concrete, it must have such validity within itself. The
use of a regression equation to convert measurements obtained to measurements that might have been obtained by some other system involves multiplying the measurements obtained by a constant and addition of another
constant. This process in no way augments or enhances whatever information about quality resided in the original measurements. Thus, if accelerated or inplace measurements are to be used to evaluate concrete, the
target numbers appropriate to those measurements should be developed
and used as criteria of quality, rather than using a conversion that introduces a large and unnecessary degree of uncertainty into the overall process.
The Maturity Concept

A method of using accelerated tests to get numbers appropriate for use
in design formulas that call for standard 28-day tests that appears to be
somewhat more reliable than the straight linear equations just discussed is
one that involves application of what is known as the maturity concept.
This concept goes back a long way, but recently more detailed work has
been done on its application to evaluation of concrete strengths [18-21].
37
Since this method has not been published as an ASTM method, it will
not be described in detail here. However, Subcommittee C09.02.09 of
ASTM Committee C-9 on Concrete and Concrete Aggregates is working on
a draft of a test method, and some comments on statistical features related
to the method are in order.
Briefly, the method involves establishing a relationship between the
logarithm of strength and the logarithm of a quantity called maturity of
the concrete that is defined as the product of temperature at which the
curing is taking place and the time in days. The most recent reference of
this work produced by Hudson et al in 1976 shows a number of calculated
log-log relationships for which linear confidence intervals have been
plotted. How these confidence intervals were derived is not given, but they
should be hyperbolic like any other confidence intervals for a plotted line.
Also a method of fitting a parabolic curve of the form Y = a x b is given
that involves only four pairs of points. Such a curve would be of doubtful
value for use in future predictions. In general, this method shows some
promise of being more useful than the methods already described.
Monitoring Production 4
One of the most effective means of maintaining the quality of a manufactured product is by continuously monitoring the quality by means of
regularly performed tests throughout the process of production. The best
tool for doing this is by means of control charts.
The control chart became a well-established technique in production
quality control during the World War II era. The control chart is a combination of both graphical and analytical procedures. The basis of the
theory arises from the fact that the variation of a process may be divided
into two general categories. One portion of the variation can be described
as random or chance variation of the process and the other as the variation due to assignable causes. A process that is operating with only chance
variation should result in some distribution of the measured characteristic,
and one should be able to predict a range within which a certain percentage of the data should fall. If some assignable cause (such as an
increased water-cement ratio) results in a change in the distribution, then
the values of the measured characteristic could fall outside the predicted
range.
Control charts are particularly useful in controlling the quality of a
product manufactured with tools or dies that wear with time. As these tools
or dies wear, the variation in size of the product will increase. A control
chart will graphically show the trend, and when the tolerance level is
4Adapted from McLaughlin, J. F. and Hanna, S. J., "Evaluation of Data," Significance
of Tests and Properties of Concrete and Concrete-Making Materials, A S T M STP 169-A,
American Society for Testing and Materials, 1966, p. 36.
38
GENERAL
reached, the machine may be retooled. Once the limits f o r a process have
been established, it is only necessary to plot the data on the control chart
and be ready to take corrective action when the data exceed the control
limits.
Three types of control charts that are frequently used are control charts
for averages, control charts for standard deviations, and control charts for
ranges. Also, control charts for other measures, such as percent defective,
may be useful. Detailed treatment of this subject and tables of control
chart constants for determining upper and lower control limits are presented in tests on statistical quality control [22-24]. Table 27 in Ref 7 contains the control chart constants for averages, standard deviations, and
ranges.
Precision Statements and Interlaboratory Tests

One of the most important concepts associated with test methods, but
one that unfortunately has often been ignored in the past, is that a test
method is a measuring tool that has a characteristic precision. One of the
most important items of information that should accompany the tool,
therefore, is information about the degree of precision that should be
expected to characterize measurements obtained by the use of that tool or
test method. Often in the past it has been assumed that writers of a test
method have done their job when they have written down the procedures to
be followed in conducting a test. There may be situations in which this is
all that is required, and the person conducting the test does not care how
closely his results should be expected to agree among themselves or with
someone else's. However, in most cases in which people go to the trouble
of obtaining measurements by means of a test method, and especially in
cases where the acceptance or rejection of materials depends on the
results, information about precision is a vital part of the test method.
As a result of concern about problems connected with precision statements and how to develop and use them, a joint task group of ASTM
Committees C-1 on Cement, C-9, D-4 on Road and Paving Materials,
and D-18 on Soil and Rock for Engineering Purposes developed two
recommended practices: Recommended Practice for Preparing Precision
Statements for Test Methods for Construction Materials (C 670) and
Recommended Practice for Conducting an Interlaboratory Test Program to
Determine the Precision of Test Methods for Construction Materials
(C 802). ASTM Recommended Practice C 670 gives directions and a
recommended form for writing precision statements when the necessary
estimates (usually standard deviations) of precision are in hand. The
second (ASTM Recommended Practice C 802) describes a recommended
method for conducting an interlaboratory study and analyzing the results
in order to obtain the necessary estimates. Both of these recommended
39
practices appear in Parts 13, 14, and 15 of the 1976 Annual Book of
A S T M Standards and should be studied and followed closely by any task
group of ASTM Committee C-9 that is charged with writing precision a
statement. Material given in the standards will not be duplicated here, but
certain aspects of precision statements and their derivation that are not
explicit in the standards are given.
One of the most important uses of precision information given in test
methods is in connection with specification writing. It is beyond the scope of
this chapter to discuss the writing of specifications in detail. PhiUeo has written on this topic elsewhere [25]. However, it is important to understand that
any numerical limits placed on properties in a specification must be developed in the light of knowledge about the precision of the test method
that is used to produce the results on which the decision to pass or fail the
material are to be based. Thus any test method for which there is no information on precision, whether in the form of a precision statement in the
test method or of some other body of information, is useless as an instrument for making decisions on acceptance or rejection of materials.
One of the problems that regularly bothers people in connection with a
precision statement is what interpretation should be placed on a particular
failure to meet the criterion of precision given in a test method. The
criterion selected for use as an index of precision in ASTM Recommended
Practices C 670 and C 802 is the (D2S) limit or the (D2S%) limit. The first
is derived from the standard deviation of test results and the second,
from the coefficient of variation. In each case the limit is derived by
multiplying the appropriate standard deviation or coefficient of variation by
the factor 2x/2. In the normal and proper operation of the test method in a
laboratory of average competence, this is the difference between two results obtained on the same material that is expected to be exceeded 5 percent of the time. s In the case of single-operator or within-laboratory
precision, it is the criterion to be met by two results obtained by the same
operator in the same laboratory, and in the case of multilaboratory precision it is the criterion to be met by two results obtained in two different
laboratories.
I f two results differ by more than the (D2S) 6 limit, a number of interpretations are possible. Which interpretation is most appropriate depends on
various circumstances connected with the situation, and in most cases a
degree of judgment is involved.
In the first place the reason for supplying any limit in a precision statesOther criteria with different numbers of tests or different percentages by which they will
be exceeded are possible and allowed by the documents of ASTM Committee E-11 on Statistical Methods. The (D2S) limit is the one deemed most appropriate and recommended for use
by ASTM CommitteesC-1, C-9, D-4, and D-18.
6Anything said about the (D2S) limit also applied to the (D2S%) limit, unless otherwise
stated.
40
GENERAL
ment is to provide a criterion for judging when something is wrong with the
results. Thus the failure of a pair of results to meet the (D2S) criterion
causes suspicion that the conditions surrounding the two tests are not
really the same as those surrounding the original tests from which the precision statement was derived or that the samples used in the two tests did
not really represent the same materials. What action is appropriate to take
depends on how serious the consequences of failure are. In most cases, a
single isolated failure to meet the criterion is not cause for alarm, but an
indication that the process under consideration should be watched to see if
the failure persists. I f it is feasible to do so, usually the tests should be
repeated, and almost always the procedures of the laboratory or laboratories
involved should be examined to make sure that the test method is being
applied in the manner envisioned by the written instructions and used
when the precision statement was developed.
Usually failure to meet a multilaboratory precision limit entails more
serious consequences than those connected with failure to meet a singleoperator criterion. The latter is sometimes used to check the results and
procedures of a single operator in a laboratory, and failure to meet the
criterion leads to reexamination of the materials and procedures. I f the test
is being used to determine compliance with a specification, the singleoperator (D2S) limit should be used to check whether or not the results
obtained are a valid test for the purpose.
Multilaboratory tests are most often used in situations where there is
a dispute about acceptance of materials. In these cases, both laboratories
should obtain two results by the same operator who is to be used in the
between laboratory tests, and use the single-operator difference as a check
on proper application of the test method within the laboratories.
One aspect of comparisons of results with an index of precision that is
often overlooked is that a (D2S) limit applies only to the difference between a randomly selected pair of results. For instance, one cannot obtain
three results and apply the (D2S) limit to the three differences thus available. The probabilities involved in the process have changed. ASTM Recommended Practice C 670 gives methods for deriving indexes of precision
appropriate for the total range of any number of test results up to ten.
These must be used if more than two results are to be evaluated.
Another feature of many test methods that is related to the question of
how many results to use is the practice of stating that " a t least" a certain
number or a certain number "or more" tests should be made. Precision
limits of whatever form must be tied to a definite number of results, and
the group responsible for the method has the responsibility of deciding
how many tests should be made for their particular purposes and of
requiring that number.
Another frequently overlooked aspect of precision statements is that the
41
indexes of precision given in a test method are not immutable. Over the
years conditions, materials, apparatus, operators, etc. change, aometimes
a large number of failures to meet precision criteria within or between
laboratories is an indication that the indexes of precision in the test method
are no longer valid and need to be revised. In many cases, it is not necessary to conduct another large interlaboratory study to revise the precision
statement if the subcommittee responsible for the test method is aware of
this situation and institutes a systematic program of accumulating new
data from laboratories that are routinely using the method. From these, an
appropriate revision of the precision statement often can be made. Results obtained in a reference sample program such as that conducted by
the Cement and Concrete Reference Laboratory can often be helpful in
this regard [26].
Test methods are constantly being revised. Very often the purpose of a
revision is to provide more precise results but the precision statement is not
revised, the excuse being that the revision has not changed the precision.
If this is true, then there would seem to be no reason for making the revision. Whenever a test method is revised other than editorially, its precision should be reexamined to see if a change in the precision statement
is needed.
Another practice that sometimes introduces vagueness into a test
method and into any assessment of its precision is that of introducing
indecision as the conditions under which estimates of precision apply.
Sometimes a statement of the conditions and materials that obtained when
the precision statement was developed is given, and then the statement is
made that applications to other conditions and materials should be made
or interpreted "with caution." Of course, how one applies the results with
caution is impossible to determine and means nothing in practice. One
either uses the criterion in a given situation or one does not. A definite
statement about conditions and materials under which an index of precision
was obtained must be given, and in the absence of hard evidence that the
limit is applicable to other situations, its use should be limited to the known
conditions.
Nonparametric Tests
There are some test methods that do not provide numbers for which the
customary processes of calculating means, standard deviations, (D2S)
limits, and other so-called parametric statistics are applicable. These
tests include those involving measurement of what is called a nominal
or classificatory scale and those involving measurement on an ordinal or
ranking scale [271. Test methods of the latter type sometimes cause problems because of the fact that numbers are assigned to the different levels of
42
GENERAL
quality of performance in the method, and then the numbers are treated as
though they represented measurement on an interval scale, which is the
type of measurement scale appropriate to most of our test methods.
An example of measurement on an ordinal scale is provided by ASTM
Test for Scaling Resistance of Concrete Surfaces Exposed to Deicing
Chemicals (C 672). In this test, the surfaces of specimens are observed at
various times during treatment and are rated on a scale of 0 to 5, 0 representing no scaling and 5 representing severe scaling. Because of the numbers there is a temptation to average results of several specimens and
even to calculate standard deviations to indicate scatter. Such calculations
are inappropriate, because the magnitudes of the numbers indicate only
order or rank and are not measurements of quantities. When lengths are
measured, for instance, the difference between an object that measures
5 cm and one that measures 6 cm is a length of 1 cm. The same difference
applies to two objects that measure 9 and 10 cm, respectively. The increment of one between scaling ratings of one and two, however, is not
necessarily the same increment as that between three and four. Ranking
numbers of this type are not amenable to the ordinary processes of arithmetic, and adding them up and dividing by the number of measurements
has little significance. Central tendency and scatter can be indicated by
giving the median and the highest and lowest ranking. See Ref 9 for other
statistical procedures applicable to this type of data.
A test that provides measurement on a nominal or classificatory scale is
one in which results merely fall into different categories without any
judgment being made that one category is higher or lower than another.
This type of test is not very important in the concrete field. An example is
ASTM Test for Inorganic Impurities in Sands for Concrete (C 40). In one
procedure a solution from the test sample is compared to a reference
solution and judged to be lighter, darker, or the same. In another procedure, five color standards may be used. This may be treated as an ordinal
scale if one end of the scale is judged to be better than the other end and
the stages in between represent progression from one level to another. Reference 9 gives procedures that are appropriate to measurement on a
nominal scale.
References
[1] Mielenz, R. C., "Petrographic Examination," in Signi~cance of Tests and Properties
of Concrete and Concrete-Making Materials, ASTM STP 169, American Society for
Testing and Materials, 1956, 13. 253.
[2] Concrete Manual, 7th ed. U.S. Bureau of Reclamation, 1963.
[3] Walker, Stanton, "Application of Theory of Probability to Design of Concrete for
Strength," Concrete, Vol. 52, No. 5, Part 1, May 1944, pp. 3-5.
[4] "Realism in the Application of ACI Standard 214-65," AC! Publication SP-37, American Concrete Institute, 1973.
43
[5] Natrella, N. G., Experimental Statistics, National Bureau of Standards Handbook 91,
1963.
[6] MeLaughlin, J. F. and S. J. Hanna, "Evaluation of Data," in Significance of Tests and
Properties of Concrete and Concrete-Making Materials, A S T M STP 169-A, American
Society for Testing and Materials, 1966.
[7] A S T M Manual on Presentation of Data and Control Chart Analysis, A S T M STP 15D,
American Society for Testing and Materials, 1976.
[8] Dixon, W. J. and Massey, Frank, Jr., Introduction to Statistical Analysis, McGrawHill, N.Y., 1957.
[9] Siegel, Sidney in Nonparametric Statistics for the BehavioralSciences, McGraw-Hill, N.Y.,
1956, pp. 26-30.
[10] Arni. H. T., "The Correlation Coefficient in Analysis of Engineering Data--Its Significance of Limitations," Public Roads, Vol. 36, No. 8, June 1971, pp. 167-174.
[11] Arni, H. T., "The Significance of the Correlation Coefficient for Analyzing Engineering
Data," Materials Research and Standards, Vol. 11, No. 5, May 1971, pp. 16-19.
[12] Daniel, C. and F. S. Wood, Fitting Equations to Data, Wiley, N.Y., 1971.
[13] Acton, F. S., The Analysis of Straigth Line Data, Wiley, 1959.
[14] Mandell, J., The Statistical Analysis of Experimental Data, Wiley, 1964.
[15] Hockersmith, T. E. and Ku, H. H., "Uncertainties Associated with Proving Ring
Calibration," Instrument Society of America, Preprint Number 12.3-2-64, 1964.
(Reprinted in Precision Measurement and Calibration, National Bureau of Standards
Special Technical Publication 300, Vol. 1, 1969.)
[16] Arni, H. T., "Impact and Penetration Tests of Portland Cement Concrete," Report
No. FHWA-RD-73-5, Federal Highway Administration, Feb. 1972.
[17] Arni, H. T., "Impact and Penetration Tests of Portland Cement Concrete," Highway
Research Record No. 378, Highway Research Board, 1972, pp. 55-67.
[18] Hudson, S. B. and Steele, G. W., "Prediction of Potential Strength of Concrete from the
Results of Early Tests," Highway Research Record No. 370, Highway Research Board,
1971, pp. 25-28.
[19] Hudson, S. B. and Steele, G. W., "Developments in the Prediction of Potential
Strength of Concrete from Results of Early Tests," Transportation Research Record,
No. 558, Transportation Research Board, 1975, pp. 1-12.
[20] Lew, H. S. and Reichard, T. W., "Prediction of Strength of Concrete from Maturity," to
be published in ACI Symposium Volume on Accelerated Strength Testing.
[21] Hudson, S. B., Bowery, F. J., and Higgins, F. T., "Research Study to Refine Methods
and Procedures for Implementing the Method of Early Prediction of Potential Strength
of Portland Cement Concrete," West Virginia Department of Highways Research
Project 47, Final Report, Woodward-Clyde Consultants, Rockville, Md., 1976.
[22] Burr, I. W., Engineering Statistics and Quality Control, McGraw-Hill, N.Y., 1953.
[23] Duncan, A. J., Quality Control and Industrial Statist&s, Revised Edition, Richard
D. Irwin, Inc., Homewood, I11., 1959.
[24] Bennett, C. A. and N. L. Franklin, Statistical Analysis in Chemistry and the Chemical
Industry, Wiley, N.Y., 1954.
[25] Philleo, R. E., "Establishing Specification Limits for Materials," presented at ASTM
Symposium on Preparation, Use, and Interpretation of Precision Statements, Chicago,
June 1976, to be published.
[26] Arni, H. T., "Precision Statements Without an Interlaboratory Test Program," presented at ASTM Symposium on Preparation, Use, and Interpretation of Precision Statements, Chicago, June 1976, to be published.
[27] Siegel, Sidney in Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill,
N. Y., 1956, pp. 21-28.
J. R . D i s e 1
Chapter 4--Quality in
Concrete Testing
Introduction
The purpose of inspection and testing of concrete and concrete-making

materials is to determine whether their characteristics and qualities as
used in construction comply with contract documents. Quality of testing
has a direct bearing on the reliability of such determinations, and on
quality control, and is therefore a key element in the construction process.
Concerns in Testing
Testing of concrete and its component materials may be said to consist

of sampling, performing specified tests, and reporting.
Procurement of specimens is often regarded as the most important step
in the testing process, and the literature repeatedly advises that poor specimens and faulty sampling techniques will defeat the purposes for which
tests are made. Adequate instruction of sampling personnel is essential,
and supervisors at all levels must be well acquainted with sampling plans
and procedures, and fully prepared to ensure that the plans and procedures are faithfully followed. No laboratory can produce satisfactory information if the samples it receives are carelessly taken or are altered by
mistreatment in shipment and do not represent the material under consideration.
The properties of concrete enumerated in construction contracts are
often based on the results of tests developed by the American Society for
Testing and Materials (ASTM). For example, it is noted that the Standard Building Code of the American Concrete Institute (ACI 318) requires
that all tests of material and concrete be made in accordance with the
ASTM standards. But no matter who the originator of the chosen standards may be, it is imperative that every effort be made to avoid the use
1Manager, Cement and Concrete Reference Laboratory, National Bureau of Standards,
Washington, D. C. 20234.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
DISE ON QUALITY IN CONCRETE TESTING
45
of obsolete, rejected, or modified testing techniques. Employment of an

unsatisfactory procedure is a waste of time and effort, and is potentially
dangerous because it can lead to completely erroneous conclusions about
the characteristics of the concrete in a structure.
In studying quality assurance for highway construction materials, it has
been found that variance in quality can be divided into material or process
variance, sampling variance, and testing variance; and in some instances,
50 percent or more of the overall variance can be attributed to the sampiing and testing phases [1]. 2 These observations indicate that these two
processes must be constantly monitored if optimum results are to be obtained.
On completion of a piece of test work, a laboratory customarily submits
a written report to its client. Reports should be complete and factual,
citing the methods used in obtaining specimens, the tests performed, the
specified values for the measured characteristics, the values obtained, the
parts of the structure involved, and similar pertinent data [2]. Pertinent
data is generally considered to include comments about the appearance
or behavior of specimens that might in any way have affected the results
obtained.
When a final report is available, it can be determined if either a material complies with specification requirements and can be used, falls into
the category of substantial compliance and can be used with reservations,
or fails to comply with requirements and cannot be used [2]. If there is
any significant uncertainty about the reliability of one or more features
of the report, concern immediately develops as to whether the testing was
done in a satisfactory manner.
Qualifications of the Testing Laboratory

An evaluation of the capabilities of a laboratory to do good work helps
to alleviate concern about the quality of testing and the reliability of the
test results.
The principal areas to be covered in an evaluation are (a) qualifications
of personnel, (b) adequacy of the plant and equipment, and (c) reliability
of measurements. Overall proficiency is necessary since adequacy in one
area can be offset by inadequacies in another, resulting in a failure of
testing to conform to the desired standard of excellence.
The managers of the laboratory should be professional men of good
standing in their field. Their academic training is relevant as well as the
length and stature of their professional and technical experience. A supervising laboratory technician should have several years of experience in the
2The italic numbers in brackets refer to the list of referencesappended to this paper.
46
GENERAL
testing of construction materials and should be able to demonstrate his

ability to perform the conventional tests for concrete and its component
materials in the manner stipulated under the governing procedures. Advanced academic training is most beneficial, however, the principal
technician and the workers he supervises need not be graduates of engineering or scientific institutions. It is in order to expect that the laboratory management will have provided training in laboratory techniques
and related topics in all instances where such instruction is needed.
Every laboratory should be able to furnish a description of its physical
resources on request. As a minimum, this description should include a
plan of its plant and an inventory of its equipment and reference literature.
Review of this information, combined with a visit to observe the condition
of the equipment and the working environment, will provide general assurance that the capability to perform the specified tests does indeed exist.
A detailed review of the laboratory's approach to quality assurance is
necessary in order to assess the reliability of its measurements. In such a
review, particular attention should be given to equipment maintenance
and calibration practices. The principal elements of this phase of quality
assurance have been identified as: (a) inventory of equipment requiring
calibration or standardization, (b) documentation of required frequency of
calibration or standardization, (c) inventory of standards, (d) external
sources for needed calibration or standardization services, (e) documentation of procedures and environmental conditions, (f) periodic checks for
prevention of errors and inaccuracies, (g) record keeping, and (h) labeling
to show calibration status [3].
As part of its quality control program, each laboratory should participate periodically in available standardization and interlaboratory test
programs. In this connection, it is noted that several hundred concrete
laboratories are now taking advantage of the consulting services of the
Cement and Concrete Reference Laboratory (CCRL) at the NationalBureau
of Standards (NBS). The CCRL, a joint project of NBS and ASTM Committees C-1 on Cement and C-9 on Concrete and Concrete Aggregates,
was established in 1929 to promote uniformity in the testing of cement
and concrete. If the laboratory has been inspected previously, the evaluation should include a check to ensure that all deviations from specification
requirements reported by the CCRL have been corrected. If the laboratory
is not participating in the program, it should consider doing so on the
premise that it would be difficult to find a more convenient and effective
way in which to obtain comprehensive and impartial evaluations of its
methods and equipment at regular intervals. In addition, the laboratory
should enroll in the comparative testing programs conducted by the CCRL,
and make use of any appropriate NBS Standard Reference Materials which
might be available, such as the standard chemical samples of portland
cement and glass spheres for sieve calibration.
DISE ON QUALITY IN CONCRETE TESTING
47
Advancements in the Quality of Testing

Over the years, impressive progress has been made in upgrading the
quality of concrete testing. Education and training efforts by such organizations as the American Concrete Institute, the Portland Cement Association,
and the National Ready Mixed Concrete Association and the implementation of certification programs and requirements for technicians have made
great contributions. The requirements of the Nuclear Regulatory Commission concerning quality assurance programs for laboratories testing concrete
and other materials for use in nuclear power plants are having a strong
impact. A new manual for establishment and maintenance of a qualtiy
control system prepared by the American Council of Independent Laboratories provides basic information about the qualification of laboratory services [4].
The recently developed ASTM Recommended Practice for Generic Criteria for Use in the Evaluation of Testing a n d / o r Inspection Agencies
(E 548), provides much needed general criteria for the evaluation of testing
laboratories. The more familiar ASTM Recommend Practice for Inspection
and Testing Agencies for Concrete, Steel, and Bituminous Materials as
Used in Construction (E 329), contains specific information for use in
evaluating concrete testing laboratories. Jointly, these two documents provide important guidance for improving the quality of testing in the concrete field.
Another important recent event was the publication of the National
Voluntary Laboratory Accreditation Program promulgated by the Department of Commerce [5]. The goal of this program is to provide a national
voluntary system to examine the technical competence of private and public
testing laboratories that serve regulatory and nonregulatory product evaluation and certification needs, and to accredit those laboratories that meet
the qualifications pursuant to these procedures. The program will also
require those laboratories that are accredited to maintain an acceptable
level of competence and thus assure national interests that such capability
is available.
It is now the practice to provide a precision statement for each ASTM
method of test, and ASTM Committee C-9 is strongly supporting this endeavor on behalf of the concrete field. Customarily, these statements provide estimates of the difference that may be expected between duplicate
measurements made on the same material in the same laboratory by the
same operator using the same apparatus within a time span of a few days.
These statements also provide estimates of the difference that may be expected between measurements made on the same material in two different
laboratories. This information is probabilistic in nature. Accordingly,
when a difference larger than that indicated as being acceptable occurs,
the requirements of the method should be reviewed, calculations should
be checked, and the working condition of instruments should be evaluated.
48
GENERAL
When properly developed and used, precision statements will do much to

advance the quality of testing.
Conclusions
High quality in testing is needed to support technological advancements
in the utilization of concrete. Continuing recognition of this fact, with due
regard for the influencing factors, helps to ensure that specification requirements faithfully reflect the properties of this important building material.
References
[1] McMahon, T. F. and Halstead, W. J., "Quality Assurance in Highway Construction,
Part I--Introduction and Concepts," Public Roads, Vol. 35, No. 6, Feb. 1969, p. 1929.
[2] Waddell, J. J., "Quality of Testing," in Significance of Tests and Properties of Concrete
and Concrete-Making Materials, A S T M STP 169-A, American Society for Testing and
Materials, May 1975, p. 32.
[3] Wening, R. J., "Quality Assurance In The Laboratory," Standardization News, Vol. 4,
No. 3, March 1976, p. 13.
[4] Quality Control System Requirements for a Testing and Inspection Laboratory, American
Council of Independent Laboratories, Washington, D.C., 1976.
[5] "Procedures for a National Voluntary Laboratory Accreditation Program," Federal
Register, Title 1S, Subtitle A, Part 7, 25 Feb. 1976.
R. D. Walker'
Chapter 5--Needed Research
Introduction
Portland cement concrete, because of its versatility and relatively low
cost, continues to be one of the most widely used construction materials
throughout the world. As developments occur in the use of concrete as a
construction and architectural material, new demands are placed on the
quality and quantity of the materials used in concrete. These demands are
magnified by increasing shortages, environmental requirements, and
zoning regulations.
Inasmuch as the title of this book refers directly to concrete and concrete-making materials, which necessarily constitute a broad subject, it
seems appropriate to emphasize in this chapter concrete and its largest
constituent, aggregate. Cement, admixtures, steel, and other concretemaking materials also will be covered.
Need for Continuing Research

In a period of limited economic and natural resources, research which
contributes to improved service life of concrete is important. Improved
service life can be anticipated if efforts are made to discover ways to: reduce
the permeability of concrete to water and dissolved salts, improve resistance to attack by aggressive media (such as the sulfates), improve the
engineering properties (such as tensile strength, volume change stability,
etc.), accomplish quicker and more permanent methods of repairing existing concrete, and increase our knowledge of measuring effects of the environment.
In all aspects of producing concrete and concrete-making materials, it is
desirable to use less energy. How are we to do this? Substituting waste
pozzolanic materials for cement might help, or perhaps more heat-efficient
cement manufacturing procedures are possible.
Additional research in the areas of concrete mixture design, test methods
for concrete and materials, and specifications is necessary to obtain optiI Head, Department of Civil Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Va. 24061.
49
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
50
GENERAL
mum utilization of our available materials. Special attention should be

given to specification limits as a function of economic considerations and
failure probability.
Many continuing problems exist, particularly regarding effects of aggregates on concrete, which, if solved, will result in economic savings. Such
problems as pavement D-line cracking, alkali-silica, alkali-carbonate, and
other cement-aggregate reactions must not be forgotten. Learning how to
upgrade marginal aggregates as well as to develop synthetic aggregates is
essential.
Problems concerning portland cement concrete require intensified research. Detailed descriptions of specific problems are given in other portions of this chapter.
The Role of ASTM Specifications and Tests

As economic and environmental factors increase in importance, the
American Society for Testing and Materials (ASTM) and, in particular, its
Committee C-9 on Concrete and Concrete Aggregates must seek to develop
specifications and tests that are more responsive to evaluating the suitability of new aggregate sources and the performance of concrete. Clearly,
better test methods and specifications are needed. The latest revision of
ASTM Specification for Concrete Aggregates (C 33) represents an attempt
to establish modular requirements for coarse aggregates based upon the
use and exposure to weathering of the concrete in which the aggregate is
to be incorporated. New research is needed to provide a more reliable basis
for the requirements included in this revised specification.
A standard developed by ASTM Committee C-9 in 1977 is used to detect
the damaging effects of the corrosion of reinforcing steel in bridge decks
(Standard Test Method for Determining Half Cell Potential of Reinforcing
Steel in Concrete (C 876)). This method will be useful in evaluating the
extent of damage caused by salt corrosion.
The ultimate objective of future research as it affects the development of
ASTM standards concerning concrete and the materials used in concrete is
to provide a framework for evaluating the materials in relation to different
elements such as:
1. Environmental conditions
2. Aggregate description
3. Condition of enclosure
4. Response in service
S. Failure criteria
6. Exposure control
7. Aggregate beneficiation
These elements must be understood completely before this objective can be
achieved.
WALKER ON NEEDED RESEARCH
51
Elements Affecting Evaluation of Concrete

Given the current state of knowledge, some discussion of the elements
listed in previous paragraphs is justified.
Environmental Conditions
Environmental factors, of course, are less important where the concrete
is incorporated into structures that are protected from the weather as
compared to the relatively severe weathering exposure received by concrete in pavement, bridge decks, and marine structures. In addition to the
need for research in predicting long-range weather trends (beyond the
scope of this publication), much is yet to be learned concerning how to
draw geographic lines of demarcation related to particular uses of concrete. This is a difficult question at best, because many environmental
factors enter into the problem such as maximum and minimum temperatures, moisture environment, freezing index, and amount of rainfall. For
example, the location of the existing water table and subgrade conditions
can have a strong effect, especially on highway or airfield pavement.
In ASTM Specification C 33 (the specification for concrete aggregates),
there is a printed map of the contiguous 48 states of the United States
where lines are drawn delineating areas of negligible, moderate, and
severe weathering. The regions so designated come from a map developed
many years ago for the building brick industry and are used because of the
lack of better data. Research that might improve the usefulness of the
map, even slightly, will provide better means of interpreting the significance of environmental conditions for a given situation. Continued research in the general area of concrete durability with respect to weathering is essential.
Aggregate Description
Aggregates may be described on the basis of the assemblage of particles
or on the basis of individual particles. Both descriptions are important and
both contain areas of limited understanding. Adequate description is a
cornerstone to intelligent testing and specification writing.
With regard to the assemblage of particles, the practice has been to
ascribe a greater degree of homogeneity to aggregate than actually exists.
Essentially all sources of gravels and many ledge rock quarries produce
heterogeneous aggregates. Thus, data on specific gravity, absorption, frost
susceptibility, etc., of the assemblage may not relate to performance in a
meaningful way.
Persons responsible for producing quality concrete should be more aware
of the science of petrography. More research in the application of petrographic methods to the aggregate evaluation process is needed. ASTM has
52
GENERAL
two standards concerning the application of these methods: Recommended

Practice for Examination and Sampling of Hardened Concrete in Constructions (C 823), and Recommended Practice for Petrographic Examination of Hardened Concrete (C 856).
Condition of Enclosure
Both the conditions of enclosure of the concrete and the conditions of
enclosure of the'aggregates in the concrete need to be considered. Whether
or not the concrete is enclosed in such a way as to encourage the increase
or retention of moisture has an effect on its performance. Further, the
performance of aggregates is affected by the nature and condition of the
paste enclosure. The physical and chemical nature of paste, its structure,
and the type of discontinuities that exist are reasonably well documented.
The relationship of these factors or characteristics to aggregate performance has not been well documented. For example, it is not known how,
why, or when water moves preferentially from the aggregate to the paste
phase, or vice-versa.
Response in Service
How does concrete respond as it exists in the working elements of highway pavement, bridge decks, or other service conditions? We should have
the ability to predict how long a concrete pavement will remain structurally sound without significant D-line cracking and without detrimental
aggregate polishing. How does one relate the response in service to such
durability tests as ASTM Test for Resistance of Concrete to Rapid Freezing and Thawing (C 666)? Although we may expect a concrete having a
100-cycle durability factor (DF) of 85 to perform longer in a given freezethaw environment than a concrete having a DF of 40, we cannot adequately
predict how long either concrete will perform under field conditions.
Failure Criteria
When looking at the problem of "needed research," perhaps a beginning
step is to determine what constitutes failure. If a concrete floor slab in a
building falls, it may not be too difficult to determine that failure has
occurred. However, with large quantities of concrete and concrete aggregates going into highway or similar construction, the questions, "What
constitutes pavement failure?" and "When is cracking, deflection, etc.,
excessive in highway pavement?" become most difficult to answer.
It is possible that for each service environment and each level of service,
failure criteria must be established. Such criteria might consider factors
such as micro- and macrocracking, deflection, disintegration, abrasion,
53
volume change, and aesthetic qualities. The American Association of State

Highway Officials (AASHO) road test in the late fifties and early sixties
attempted to establish a failure criteria based upon rider response as it
affected comfort. This was later correlated to "slope variance"--a form of
roughness measurement. Further efforts such as this are required to make
the use of ASTM test methods and specifications more significant.
Exposure Control
Is it possible to control concrete field response through exposure control? A better question may be, "Is it conceptually feasible to exercise
exposure control?", since if it is, the answer to the first question is probably
yes. Quantitative information is needed to answer the latter question.
It is obvious, for example, that bridge decks are exposed more severely
than are bridge piers. However, no scale is available to measure these
differences. Attempts are made to lessen the severity of the deck exposure
by surface treatments with linseed oil, etc. If this seems reasonable, then
in other situations alternate design approaches or a different construction
method might serve to modify the exposure the concrete must withstand.
Aggregate Beneficiation
Another approach involved modifying the concrete aggregate on the
assumption that the exposure is known and unalterable. Modification is
performed so that no significant portion of the aggregate is likely to be
stressed beyond its capability. The weak fractions are simply removed.
A procedure suggested by Lin et alZ would have the water in the aggregate replaced by 100 percent ethylene glycol. In the laboratory a unifor~nly
poor aggregate in concrete exposed to freezing and thawing had the 100
cycle DF of the concrete changed from 3.6 to 80 percent.
In many cases, removing the weak fractions does not leave enough
strong aggregate, and the use of ethylene glycol is obviously expensive.
However, perhaps both attempts help illustrate another area that needs
research.
Some Specific Examples

For a number of years, the Transportation Research Board, through its
committees, has been publishing lists of Research Problem Statements of
immediate urgency. Typical are those found in Transportation Research
2Lin, C. H., Walker, R. D., and Payne, W. W., "Chert-Aggregate Concrete Durability
After AntifreezeTreatment," in Living with MarginalAggregates, ASTM STP 597, American
Societyfor Testing and Materials, 1976, pp. 76-83.
54
GENERAL
Circular Number 179. 3 Other statements were developed under a National

Cooperative Highway Research Program (NCHRP) project and published
in NCHRP Report 100. 4 The following paragraphs abstract a few of the
statements from these publications for illustrative purposes.
Bridge Decks
One such statement describes the problem of the performance of bridge
decks with deep reinforcement. It indicates that corrosion of steel reinforcement has become the nation's most troublesome bridge maintenance
problem. The objectives of proposed research are to develop information
on the time relationship among chloride concentration at the level of reinforcing steel, the onset and progression of corrosion, and the development
of spalling at the deck surface or delamination at the level of the reinforcement.
Alkali-Aggregate Reactivity
A need exists for exact determination of rocks containing alkalireactive silica or silicate. Several instances have been noted of alkali-silica
reaction leading to distress in concrete structures in which the reactive
aggregates have been rocks of types not expected to be reactive with alkalies
in cement using current criteria. The specific objectives suggested are to:
1. Develop criteria that will permit demonstration of reactivity in slowly
reactive rocks within reasonable laboratory testing times
2. Advance means other than the specification of low alkali cement that
will effectively control slowly developing alkali-silica reaction
3. Establish more perfectly the mechanisms of the reaction, including
those mechanisms that cause low quartz to become reactive
Concrete Strength
Alternatives to testing of concrete cylinders and cores are needed for
acceptance purposes. There are currently procedures such as ultrasonic
resonance, radioactive and electrical procedures, and others such as the
pullout, probe, or penetration tests. Controversy still exists as to what inplace strength is and whether it is determinable. One of the more specific
objectives recommended for needed research is to compare the currently
3"Research Problem Statements," Transportation Research Circular Number 179, Transportation Research Board, 1976.
4Walker, R. D., Larson, T. D., and Cady, P. D., "Research Needs Relating to Performance of Aggregates in Highway Construction," National Cooperative Highway Research
Program Report 100, HighwayResearch Board, 1970.
55
used methods and recommend appropriate methods and develop standard

procedures for these methods.
Improving Concrete with Chemicals or Other Means

Weaknesses of portland cement concrete include low tensile strength and
the resulting tendency for cracking, and limited resistance against chemical
aggression. Perhaps these properties can be improved by modifying the
concrete by combining it with polymers, prepolymers, polymerizable
monomer systems, sulfur, or other suitable chemicals. Little research has
been done, and the technology is very complex, especially when dealing
with fresh concrete. The addition of fibers to concrete with the purpose of
improving the tensile strength and cracking resistance needs additional
research.
Role of Aggregates in D-Line Cracking

At least three theories have been advanced to explain the cause of D-line
cracking. None of these theories indicate how aggregate quality relates to
this deterioration. Objectives of the research would include:
1. Determination of how D-line cracking develops in pavement
2. Relationship of aggregate quality to D-line cracking
3. Selection or development of tests that would discriminate aggregates
causing this type of failure
Performance of Concrete When Pop-Outs Occur

Pop-outs are essentially superficial failures of concrete caused by near
surface pressure development. Research has provided a reasonable understanding of why pop-outs occur. However, other specific problems need to
be answered:
1. What is the aesthetic value of highway concrete? Can it be quantified
and, if so, by whom?
2. Can pop-out occurrence be quantitatively related to aesthetic value?
3. How many pop-outs can be tolerated from the several performance
viewpoints? How does this vary with different structural applications of
concrete?
Continuing Responsibility
Page after page of publications (such as those in footnote 3) could be
abstracted, but that would defeat the purpose of this chapter and unnecessarily duplicate information already available. The committees of the
56
GENERAL
Transportation Research Board are charged with the continuing responsibility of bringing to light needed research.
Energy and Energy Reducing Alternatives

Energy is required to produce, transport, modify, beneficiate, and
manufacture aggregates. Research is needed to study the energy consumption trade-offs when choosing an aggregate source or making a decision to
manufacture a synthetic aggregate. Can gradation specifications be modified to permit less energy use in their production?
Cities, in addition to being generators of construction, are also the chief
generators of solid waste. It makes little sense to haul solid waste out of
town while hauling aggregates into town if there is a possibility of processing some of the waste into usable aggregates.S Can research overcome the
barriers in this area?
This section deserves more attention, but we as a people are only beginning to become conscious of the nation's energy needs.
Conclusions
It should now be obvious that the directions of needed research in the
area of concrete and concrete-making materials are multitudinous. Such
research in some cases may be narrow, but to a large degree must be sufficiently broad and applied to include a concern for reaching a better
understanding of the relationship between service and environment. The
research must include such things as a continued search for basic mechanisms. Such research must persist in time and expenditure.
Acknowledgments
It should be noted that the work of the author, along with T. D. Larson
and P. D. Cady, of the Pennsylvania State University, in preparing NCHRP
Report 100 (see footnote 4) is freely drawn upon in writing this chapter.
Further, both Larson and Cady contributed their thoughts and ideas which
were of great help to the author.
sPhilleo, R. E., Introduction to Living with MarginalAggregates. ASTM STP 597, American Societyfor Testing and Materials, 1976, p. 1.
C Powe~ ~
Chapter 6--The Nature of Concrete
Introduction
The term cancrete can be construed to include a considerable variety

of products made from portland cement or other cementing media, but in
this publication the term concrete usually refers to a material which was at
first a plastic mixture (or mixture that became plastic as a result of manipulation, especially vibration) of portland cement, water, air, and mineral
aggregate. Therefore, this discussion of the nature of concrete will have the
scope indicated by that description.
A writer's concept of the nature of concrete can hardly be revealed in a
few words, but his treatment of certain topics and his definitions are indicative. Here are some examples: in 1878 Trautwine, in the l l t h edition
of his Pocket Book for Civil Engineers, said "Cement concrete, or beton,
i s . . . cement mortar mixed with gravel or broken stone, brick, oyster
shells, etc., or with all together." He described mortar as sand containing
a volume of cement equal to the volume of voids in the sand. 2 In 1907, L.
C. Sabin in a book on concrete said, "Concrete is simply a class of masonry
in which the stones are small and of irregular shape. The strength of concrete depends largely on the strength of the mortar; in fact, this dependence
will be much closer than in the case of other classes of masonry, since it
may be stated as a general rule, the larger and more perfectly cut are the
stone, the less will the strength of the masonry depend on the strength of
the mortar." Feret in 1896 considered water and air to be definite components of mortar (and presumably also of concrete), but it is not clear
that he thought of cement paste as an entity. Zielenski, once head of the
Hungarian Association for Testing Materials, in 1910 called concrete a
conglomerate body; he considered the conglomerate to be composed of
1Retired; formerly research counselor, Research Department, Portland Cement Association,
Skokie, Ill.
2 He said also, "Nearly all the scientific principles which constitute the foundation of civil
engineering are susceptible of complete and satisfactory explanation to any person who really
possesses only so much elementary knowledge of arithmetic and natural philosophy as is
supposed to be taught to boys of 12 to 14 in our public schools."
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
60
mortar and coarse aggregate, and the mortar to be coLnposed of paste and
sand, with or without air voids. Taylor and Thompson, authors of perhaps
the best of early books on concrete, in the 1912 edition said, "Concrete is
an artificial stone made by mixing cement, or some similar material which
after mixing with water will set or harden so as to adhere to inert material,
and an aggregate composed of hard, inert, materials of varying size, such
as a combination of sand or broken stone screenings, with gravel, broken
stone, cinders, broken brick, or other coarse material." D. A. Abrams, in
the first bulletin from the Structural Materials Research Laboratory, Lewis
Inst., Chicago, Ill. (1918), emphasized the significance of the ratio of water
to cement in concrete, and he abandoned the notion that concrete is a
mixture of mortar and coarse aggregate, pointing out that the whole aggregate, fine and coarse combined, should be considered as one, even though
fine and coarse aggregates are proportioned separately. F. R. McMillan in
his book, Basic Principles of Concrete Making (1929), said, "Expressed in
the simplest terms, concrete is a mass of aggregates held together by a
hardened paste of portland cement and w a t e r . . , the paste is the
active element." In their textbook on concrete, Troxell and Davis (1956)
wrote, "Concrete is a composite material which consists essentially of a
binding medium within which are embedded particles or fragments of a
relatively inert mineral filler. In portland-cement concrete the binder or
matrix, either in the plastic or in the hardened state, is a combination of
portland cement and water." This definition was adopted by the American
Concrete Institute (ACI) Committee 116 on Nomenclature in 1964. The
Encyclopedia Britannica, 1963 edition, says, "Concrete is a building material consisting of a mixture in which a paste of portland cement and
water binds inert aggregates into a rock-like mass as the paste hardens
through chemical reaction of cement with water."
Although it is possible to discern an evolution of concepts in the above
definition and descriptions, the early concept of concrete as a mixture of
mortar and coarse aggregate tends to persist, despite Abrams' contention
that the total aggregate functions as a unit. The idea lingers with us, perhaps, because it has simplicity and plausibility and partly because it is not
altogether unrealistic, especially when there is a gap between the largest
size in the sand and the smallest size in the coarse aggregate. However, gap
gradings are not common, and in any case the correctness of Abrams' conclusion can hardly be questioned.
Even the latest definitions provide but a superficial idea of the nature of
concrete. They give no hint as to how and why the originally plastic mass
becomes hard and strong and, indeed, say nothing as to how and why
the mixture had plasticity in the first place. Moreover, they give no adequate basis for understanding such aspects of concrete as volume change
characteristics and stress-strain-time phenomena. We must, therefore, go
far beyond a superficial definition or description to attain today's underCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:39:08 EDT 2014
POWERS ON NATURE OF CONCRETE
61
standing of the nature of concrete. To do the subject full justice would

be a book-length project; I cannot do more than touch on a few fundamental topics.
Gross Structure of Concrete

When we inquire into the nature of concrete we find it necessary to
regard concrete not as an entity--a s u b s t a n c e J b u t as a structure having
component parts, as has already been indicated. The predominant component of concrete is an aggregation of mineral particles, called the aggregate, and this aggregation requires a certain minimum of space per unit
weight of material. The volume of space occupied by a properly compacted
fresh concrete mixture is slightly greater than would be the compacted
volume of the aggregate it contains. The difference is significant; it shows
that the individual rock 3 particles in concrete need not be in contact with
each other. There is clear evidence that the rock particles in concrete
are, in fact, not in contact with each other: freshly mixed concrete could not
be plastic if the solid particles were not dispersed to some degree; inspection of broken sections of hardened concrete show that not only are the rock
particles in a dispersed state while the mixture is fresh, but also they remain dispersed, although generally not exactly to the same degree that
prevails immediately after mixing, owing to settlement under the force of
gravity before setting occurs.
Rock particles in plastic concrete are dispersed in a matrix composed
of paste and air bubbles; 4 the paste is composed of portland cement and
water. The degree of dispersion actually depends upon the consistency of
the paste and the volume of air; the stiffer the consistency and the higher
the air content, the greater the mean clear distance between aggregate
particles. In practical terms, this means that the degree of dispersion of
rock particles is greater the lower the water-cement ratio of the paste, and
the higher the air content; in any case, the volume of concrete seldom if
ever exceeds the compacted volume of the aggregate by more than 10 percent, and usually it does not exceed it by more than 3 percent when no
air-entraining agent is used.
Without an air-entraining agent, concrete placed by a standardized
procedure contains a characteristic amount of air in the form of bubbles,
the amount being a function of certain variable factors. The main factors
are the consistency of the paste and the gradation of the aggregate. At a
given paste consistency, and with a standardized mixing procedure, air
content depends mainly on those features of aggregate grading that control
3Throughout this discussion, the term rock refers to the particulate mineral matter that
makes up the aggregate, regardless of the size or shape of the particles.
4Voids due to incomplete filling of mold or form are not considered here, only the voids
that are normal componentsof the mixture.
62
the mean size of the voids in the aggregate, voids being here defined as the
space occupied by paste and air. The mean size and the size range of the
air bubbles in concrete are also significant structural features, particularly
with respect to the ability of concrete to withstand the effects of freezing.
These characteristics also are subject to systematic variation as will be
discussed later.
The foregoing statement as to the effect of aggregate grading on air content is applicable principally to the leaner types of mixtures. In richer mixtures, water-cement ratio less than about 0.5 by weight, aggregate characteristics have little effect. Throughout the range of the most frequently
used mixtures, air content and void characteristics are strongly influenced
by water-cement ratio, the air content and average air void size increasing
with an increase of water-cement ratio under given conditions.
When a suitable air-entraining agent is used, the air content can be
raised to almost any desired level. At an air content higher than that normally present in a given mixture, the mean size of air bubbles is controlled
by the characteristics of the air-entraining agent used and, in the leaner
range of mixtures, by the same factors that control the normal air voids.
Specifically, if the normal air content of a given mixture is relatively high
and the mean size of the voids relatively large (the two usually go together),
raising the air content by means of a given approved air-entraining agent
will result in a system of voids having a relatively large mean size, as compared with the mean size when the amount of air normally present is
smaller, and the original mean size also smaller. This means that with a
given air-entraining agent used in different mixtures a wide range in average void size may be observed.
In short, the gross structure of concrete appears to be that of an aggregation of rock particles slightly dispersed in a matrix of paste and air bubbles, the proportion and size characteristics of air bubbles being subject
to systematic variation just as is the proportion of aggregate.
A conclusion arising directly from consideration of the gross structure of
concrete is that the firmness or mechanical stability of concrete cannot be
attributed to mechanical stability of the aggregation of rock particles; it
is clearly due to the mechanical stability of the matrix and to the mechanical stability of individual particles of rock. Also, it is clear that the stability
of the matrix is due to that of hardened cement paste.
Interparticle Forces in Freshly Mixed Concrete

We have already seen that plasticity of freshly mixed concrete is possible
because the rock particles of the aggregate are slightly separated from one
another by matrix material; otherwise, any deformation would necessarily
be dilatant rather than plastic. The matrix, itself, is plastic because the
cement particles and air bubbles are dispersed in water and especially
63
because the interparticle forces tend to hold particles together while at the
same time preventing actual point to point contact. Such a state is due to
the coexistence of forces of attraction and repulsion between cement particles. Attraction is due to relatively long range intermolecular forces
known as van tier Waals forces; about which more later. The forces of repulsion are, in this case, due to electrostatic repulsion and to a "disjoining
pressure" (Deryagin) s maintained by adsorbed water molecules covering the
surfaces of the grains. Electrostatic repulsion is due to what is called a
Gouy diffuse layer of ions, in this case negative ions, the negative ions
being held near each cement particle by positive ions selectively adsorbed
from the surrounding aqueous solution.
Owing to the existence of opposing interparticle forces, a pair of cement
grains has a minimum of potential energy with respect to those forces when
the particles are separated from each other by a certain small distance,
the distance amounting to perhaps ten water molecule diameters, more or
less; hence, cement particles tend to assume positions with respect to
each other corresponding to minimum potential energy with respect to
balance of internal forces, and this is an essential condition for the plastic
state.
When particles are in positions of minimum potential energy with respect to the forces acting between them, they are said to be in "potential
troughs," and any mechanical displacement of particles with respect to
each other requires a certain amount of work to "lift" the particles out of
their potential troughs. Thus, we see that interparticle forces give freshly
mixed paste in the quiescent state, a structure having a low degree of
firmness or shearing strength. When shearing stress exceeding shearing
strength is applied and maintained, a paste is caused to flow continuously
if its solid content is considerably smaller than that at normal consistency,
which is the case for pastes used in concrete. The shearing strength (yield
value) and the resistance to continuous shearing stress (the mobility or
structural viscosity) is often used as a measure of paste consistency, but
such a measure pertains only to paste in the fluid state maintained by a
sufficiently high shearing stress.
Consistency of freshly mixed paste can be made softer by diminishing
the depth of the potential troughs, and this can be done by using an appropriate surface-active material able to increase interparticle repulsion;
indeed, repulsion can be raised to such a degree as to destroy plasticity,
changing paste from a plastic to a fluid material. In some concrete mix5In previous publications I have used terms such as film pressure, swelling pressure, or
spreading pressure, to indicate the force developed by adsorbed films of water in spaces too
narrow to accommodate the normal thickness of the films. Several years ago Deryagin, a
surface physicist of the Soviet Union, introduced the term disjoining pressure and another,
disjoining action; when thinking of the tendency of water to disperse a coherent system of particles, Deryagin's term seems most apt.
64
tures, the water content of the paste is so high that the paste has very little plasticity to begin with; even though the interparticle forces discussed
above exist, the particle concentration is so low that interparticle forces
are relatively ineffective. The paste in a properly constituted freshly mixed
concrete has an optimum consistency, neither too soft nor too firm (stiff);
under no circumstances should it be completely fluid.
Chemical Nature of Hydrated Cement

The chemical compounds found in hydrated cement are complex; most
of them are impure in the sense that they contain elements not ordinarily
given in their formulas, and they do not have exactly the same composition when formed under different conditions, especially with respect to
temperature and original cement concentration. For our present purpose
it will suffice to mention a few outstanding characteristics.
All the components of hydrated cement are basic; the hydroxyl-ion concentration is always at least as high as that of a saturated solution of calcium hydroxide, and it is usually considerably higher because of the presence of alkali hydroxides. Any of the compounds can be decomposed by
carbonic acid and, therefore, by ordinary rain water. That concrete is not
generally destroyed this way is explainable mostly in terms of physical
factors: under ordinary conditions of exposure the quantity of acid in contact with concrete during a given time is small relative to the quantities of
basic material available, and the permeability of concrete to water is so low
that the action of weak acids is only superficial. Even contact with softwater streams usually causes decomposition at a negligible rate. In cities
where rain falls through industry-polluted air and becomes distinctly acid,
acid action is evidenced by surface etching. But, whenever a continuous
supply of strong acid is encountered, concrete must be protected or it will
be destroyed.
Hydrated cement is able to react also with carbon dioxide gas in the
presence of water vapor, but the effect is not destructive; actually such reaction may increase chemical stability.
The reactions between the anhydrous components of portland cement
and water are remarkable in that they involve a doubling of the volume
of space required by solid material while the apparent volume of the system remains constant. There is one exception: the formation of calcium
sulfo-aluminate by reactions involving gypsum and tricalcium aluminate
tends to cause volume expansion, and when the sulfate ion concentration
is too high, expansion can be destructive. Normally, the amount of gypsum
needed to control the setting of cement gives only tolerable expansion,
but when an ufilimited supply of sulfate ion is present in the environment,
concrete may be destroyed by it. Practically, such destruction is avoidable
by using a cement of low tricalcium aluminate content.
65
Chemical Aspects of Mineral Aggregates

The petrographers and mineralogists contributing to this publication
have much to say on this subject; I shall mention only some chemical
characteristics that are of special interest because of the structural and
chemical characteristics of the paste component of concrete.
Rocks used for concrete aggregates are generally materials that have
survived geologic ages and are thus those that have demonstrated some degree of chemical and physical stability. Some are chemically basic; some
are acidic. Limestone is, of course, basic and vulnerable to acid attack,
but when used as concrete aggregate it is less vulnerable than the hydrated
cement paste which envelopes it. There are at least two other kinds of
undesirable chemical attack on certain minerals that seem to be the result
of conditions peculiar to the interior of concrete; one kind involves siliceous
rocks; the other, certain dolomites.
Some kinds of siliceous rocks, opal being an outstanding example, may
be decomposed by the caustic solutions found in concrete. Under some
conditions this reaction is accompanied by destructive expansion, and under
other conditions, no expansion. The conditions mentioned involve chemicophysical factors too complex to be described here, as may be ascertained by
referring to the literature on the alkali-silica reactions in concrete. A principal factor determining the physical effect of such reactions is the special
and selective hindrance to the diffusion of various ions through the structure of cement paste; another is the quantity and specific surface area of
the reactive form of silica.
Also, some dolomites react with the caustic aqueous solution in concrete and expand destructively; the reaction involves decomposition of the
dolomite and the formation of magnesium hydroxide (brucite); this reaction is called dedolomitization.
Structure of Hardened Cement Paste

We have already seen that freshly mixed cement paste is a dispersion of
cement particles in water and that is has a certain structure owing to the
forces of attraction and repulsion among the cement particles. This structure is the starting pattern of the structure that subsequently develops from
the materials produced by reactions between the components of cement
and water. These reaction products are collectively that which we have
already called hydrated cement; now we shall stress the physical aspects
of hydrated cement.
Although hardened cement paste looks the way we might expect an
amorphous continuum to look, we know that it actually comprises a hierarchy of aggregations of matter. Moreover, we know that cement paste
contains submicroscopic voids, its void content usually being upwards of
66
40 percent, although a lower void content is possible. Since, in general,

we have learned to think of matter as intrinsically granular, we are inclined to regard the pore space as having the character of interstices in a
granular aggregation. By dispersing the structure and examining the fragments by electron microscopy, we have seen particles and have been impressed by their smallness and irregularity of shape. These particles, not
all of the same kind chemically, may be regarded as the primary particles of paste structure, even though atoms and molecules are the primary
and secondary aggregations of matter in general.
The term gel particle refers to particles having dimensions in the submicroscopic range of sizes called colloidal; in this particular case colloidal
bodies can be defined as molecular or ionic aggregations having a very
high specific surface area, such as is possible only in the submicroscopic
range. The colloids observed in cement paste are mostly quasicrystalline,
lack of normal crystallinity being due to the extremely small size and
imperfect atomic or molecular organization of the solid material. The high
specific surface area is due mostly to the thinness of the particles, one of
the three dimensions, and perhaps two, being greater than the limit usually
stipulated for the colloidal state; these particles are usually only three or
four molecules thick.
Along with the colloidal material in hardened cement paste is crystalline
calcium hydroxide having relatively low specific surface area. The amount
of calcium hydroxide is different for cements having different chemical
compositions, but it is usually between 20 and 30 percent of the weight
of the dry hydrated cement. Calcium hydroxide crystals are usually surrounded by and intergrown with colloidal material, and thus they constitute an integral part of the solid structure.
The most abundant colloidal constituent of hydrated cement is an impure calcium silicate hydrate of somewhat indefinite stoichiometry. It
has characteristics resembling those of a natural mineral called tobermorite
and has thus come to be called tobermorite gel. (The term gel designates
a rigid aggregation of colloidal material.) There are also amounts of calcium aluminate hydrate and calcium alumino ferrite hydrate; the physical
states of these materials are not known exactly, but it appears that they are
colloidal but with a lower order of specific surface area than those of tobermorite gel.
An outstanding characteristic of the colloidal matter in hydrated cement
paste is that its specific surface area is virtually the same in all pastes
made of the same cement regardless of differences in paste density, and
among pastes made with different portland cements it is not much influenced by differences in chemical composition. This observation is one of
the cornerstones of our concept of paste structure.
The colloidal matter together with calcium hydroxide appear as a continuous solid structure apparently occupying the whole volume of any specCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:39:08 EDT 2014
67
imen of hardened cement paste. As already indicated the structure is

porous; normally the solid matter occupies 45 to 60 percent of the apparent volume, and the highest possible solid content (exclusive of unhydrated cement if any) has been found to be about 72 percent or perhaps a
little more of the total volume occupied by hydrated material. In other
words, a specimen of hydrated cement paste may have a porosity of not
less than about 28 percent, and it usually has a porosity between 40 and 55
percent; the porosity will have a higher range if the paste is not fully mature, which means if complete curing has been deliberately or inadvertently
omitted. The apparent volume of a specimen of paste, expressed as a
volume per unit quantity of cement, is determined by the net volume of
mixing water per unit volume of cement. The net volume of mixing water
is that which remains within the specimen at the end of the period of
settlement (bleeding) which is normally from 1 to 2 h'after mixing.
The solid content of any freshly mixed paste is a little over 62 percent of
what the solid content will be after chemical reactions have converted all
the cement to the hydration products described above. As already mentioned, the increase of solid volume takes place without appreciable change
of overall volume,6 regardless of how high the cement content may be or
how little water per unit of cement. One cubic centimeter of cement, solid
volume, produces about 1.6 cm a of hydrated cement. On this basis we
might expect that the hydration products will require 0.6 cm 3 of space
in addition to the space originally occupied by 1 cm 3 of cement, or 0.19
cm 3 of space per gram of cement; this amounts to saying that if the watercement ratio is 0.19 by weight, there will be ample room within the specimen for all hydration products that can be derived from 1 g of cement, and
the material would become a voidless soild. But experimentally it was
found that void space cannot be eliminated. This experimental observation
is another cornerstone on which our concept of paste structure rests.
As just indicated, it was found that under no circumstance can completely mature paste be made entirely solid; the highest possible solid content is about 72 percent, and the rest of the unit volume remains full of
water or is void if all the water in such space is caused to evaporate. The
densest possible hydrated cement paste contains a continuous system of
pores, as evidenced by its permeability to water. Because of the intrinsic
porosity of the structure, which limits the solid content to 72 percent of the
apparent volume, it follows that instead of 1.6 cm3/cm 3 of cement, the
volume of paste must be at least 1.6/0.72 = 2.2 c m a / c m 3 of cement to
provide enough space for all the hydration products that can be derived
from 1 cm 3 of cement, and this means that the water-cement ratio by
weight must be at least 0.38. In any paste containing less than this amount
6The microscopic changes that may occur later as the result of drying and wetting are
properly ignored here.
68
of water-filled space, some of the cement remains anhydrous regardless

the duration of curing, and a residue of the original cement remains a permanent feature of the structure of the paste, the residue appearing in
cross sections as scattered remnants of the largest grains.
Of course, when the volume of cement paste is greater than 2.2 cm3/cm 3
of cement, as it usually is, there is more than enough space to accommodate the hydration products and thus less than 72 percent of the space can
become filled with solid matter. When there is more space available than
the minimum required by the hydration products, the extra space is a
feature of the structure of paste.
Various observations lead to the concept that in every paste the hydration products tend to become locally concentrated to the m a x i m u m degree
possible, even when excess space is available, but at the same time they
form a continuous structure having an overall volume equal to the apparent volume of the paste. One line of evidence supporting this view develops from consideration of the structure of freshly mixed paste. Although
cement particles in freshly mixed paste are individually dispersed throughout the volume of mixing water, they cannot be uniformly spaced because
of the interparticle forces, already discussed, that hold the particles practically (but not exactly) in point to point contact. This being true, it seems
that if we could subdivide a freshly mixed specimen of paste having a
certain water-cement ratio into a large number of cubical cells each having
an edge length of say 100 #m (about the same as the mean diameter of
the largest cement particles but about 100 times as large as the spherical
equivalent of a "particle" of tobermorite gel), we would not find the same
volumes of cement and water in each cell; we would find some cells almost filled with a single grain, some would contain many small grains, and
some might contain few if any grains. In other words, we would find that
the overall water-cement ratio is an average of many different local watercement ratios, some higher and some lower than the average. Since some
of the cells must be nearly full of solid material to start with, considering
the size of many of the cement grains, and since interparticle attractior
tends to hold particles close together, it seems almost certain that the 72
percent limit of the content of solid hydrated material will be achieved in
many cells after all the cement has become hydrated, even though some of
the cells cannot become filled to this extent. In other words, it is reasoned
that if the hydrated material in a specimen of paste can reach an average
density of 72 percent but no higher, at any sufficiently low water-cement
ratio, that same degree of density can be and is produced locally at various
places throughout any paste, however high the water-cement ratio. If any
of the imagined cells contain excess cement, which presumably is the case
wherever the local water-cement ratio is lower than about 0.38 by weight,
such cells can get rid of their excess material by diffusion of material into
69
a d j a c e n t cells l a c k i n g cement, so t h a t eventually all the c e m e n t can b e c o m e

h y d r a t e d if the average w a t e r - c e m e n t ratio is high enough. 7
I have used the t e r m cement gel to designate h y d r a t e d c e m e n t paste
in its densest form. It should be noted t h a t by this definition c e m e n t gel
is not synonymous with gel as used above, for e x a m p l e , t o b e r m o r i t e gel.
The t e r m c e m e n t gel is convenient for d e s i g n a t i n g the p r e d o m i n a n t l y
colloidal m a t e r i a l f o u n d in h y d r a t e d c e m e n t paste, b u t it m u s t be k e p t
in m i n d t h a t the t e r m includes noncolloidal c a l c i u m h y d r o x i d e a n d other
noncolloidal m a t e r i a l , if any, a n d therefore it does not c o n f o r m exactly to
the t e r m gel, p r o p e r l y defined as a solid c o m p o s e d of colloidal m a t e r i a l .
T h e c o n c e p t of p a s t e s t r u c t u r e d e s c r i b e d above, which r e q u i r e s us to
visualize an uneven d i s t r i b u t i o n o f c e m e n t gel, entails a c o r r e s p o n d i n g uneven d i s t r i b u t i o n of t h e sizes of interstitial spaces; surface to surface distances range f r o m zero at chemically b o n d e d points to a m a x i m u m distance t h a t is p r o b a b l y g r e a t e r the g r e a t e r the capillary porosity of the
paste, b u t which in any case is not known exactly. The o r d e r o f m e a n pore
size is i n d i c a t e d by the q u o t i e n t o f the volume of pore space by the b o u n d ary a r e a of t h a t space, which q u o t i e n t in h y d r a u l i c e n g i n e e r i n g is called
the h y d r a u l i c radius. F o r a porosity of 28 p e r c e n t a n d for solid m a t t e r
having a specific surface a r e a of 5.2 106 c m 2 / c m 3 of solid m a t t e r , which
is a b o u t the.specific surface a r e a of c e m e n t gel, the q u o t i e n t is 7.5 10 -a
c m or 7.5 A . 8 T h e c o r r e s p o n d i n g average d i s t a n c e f r o m solid surface to
solid surface is between two a n d four times the h y d r a u l i c r a d i u s , d e p e n d ing u p o n the s h a p e of the interstitial spaces; in the p r e s e n t case where
m a n y of the spaces are believed to be slit-like, a b o u t 1.8 n m (18 /~) seem
to be a r e a s o n a b l e estimate. W h e n the porosity exceeds 28 percent, which
is to say when the paste c o n t a i n s spaces other t h a n gel pores, the m e a n
pore size is, of course, g r e a t e r t h a n the m e a n size of gel pores.
W e can now describe h a r d e n e d paste as a solid c o m p o s e d o f cement
gel, r e m n a n t s o f c e m e n t grains, if any, a n d space not filled with cement
gel, if any. Any space not filled with c e m e n t gel or g r a i n r e m n a n t s is reg a r d e d as interstitial spaces a m o n g masses of c e m e n t gel a n d is called
capillary space, capillaries, or capillary cavities. T h e l a t t e r t e r m is a p p l i e d
in pastes so dense t h a t the capillaries are discontinuous, the original inter7Theoretically, if 50~mm or (2-in.) cube, for example, of fresh paste containing excess
cement is joined to one with a deficiency of cement, and if after hardening both are kept
saturated with water, the excess cement in one cube could, after the density of hydrated material had reached 72 percent, diffuse into the excess space in the other; but this conclusion
would be hard to prove because the process would probably require geological ages for completion. However, diffusion for a distance of a few micrometres can occur in a normal curing
period, as has been verified by microscopic observations.
8To acquire an idea of the magnitude of an angstrom unit, try the following: if one could
place two marks 1 ,~ or 0.1 nm apart on a piece of superrubber about 25 mm (1-in.) long,
and then if one could stretch the rubber until it could encompass the earth, the two marks
would have become a little over about 150 mm (6-in.) apart.
70
stitial space among the cement grains having become segmented into isolated
cavities by the growth of cement gel. The pores within cement gel are called
gel pores.
Rock Structure
The structure of concrete is characterized not only by the structure of
cement paste but also by the structure of individual pieces of rock making
up the aggregate. Some pieces, particularly the smallest, may consist of
practically voidless crystals or fragments of crystals; but in most concrete
aggregates most of the particles have granular structure and are porous,
and most are permeable to fluids. The pores in permeable rock are usually
larger than those in cement paste; also rocks are less porous than paste, as
a rule. A rock having a porosity of say 1 percent, may have a coefficient of
permeability to water equal to that of a specimen of hardened paste having
a porosity of 50 percent; thus, we know that the pores in rocks are usually
larger than those in hydrated cement paste; in other words, rocks have a
relatively coarse texture. A few kinds of rock do have fine texture, and
unless they are also practically nonporous they do not make satisfactory
concrete aggregate; certain argillaceous limestones are examples of unsuitable fine textured rocks, certain cherts also.
Some rocks contain pores that reduce their apparent specific gravity but
do not affect their permeability. Among these are vesicular rocks and artificially expanded shales in which isolated voids were formed by expansion
of trapped gas. Such rocks are useful when concrete of low unit weight is
desired.
States of Water in Concrete

The solid matter in mature concrete may contain water molecules or ions
derived from water, as indicated by the formulas for the compounds in
hydrated cement, and by those for some minerals found in concrete aggregate. Of principal interest, however, is the water that remains chemically
free, and which is found in gel pores, or in capillary spaces in paste or
rock. The discussion shall be confined for the most part to chemically free
water found in cement paste since it influences the properties of concrete
to a very important degree.
It is commonly known that water in a vessel open to the atmosphere
will eventually evaporate unless the atmosphere is saturated with water
vapor. Cement paste can be regarded as a vessel open to the atmosphere
so far as contained water is concerned, but not all of the water it is capable
of holding can evaporate unless the surrounding atmosphere is practically
empty of water vapor; if the cement paste was originally saturated, some of
71
the water can evaporate, but a definite fraction of it will be retained, the
amount retained being a larger fraction of the total evaporable water the
higher the degree of saturation of the ambient atmosphere, that is, the
higher the humidity. At low ambient humidities, water molecules are restrained from evaporating by the van der Waals forces of attraction between them and the surfaces of the gel particles, which attraction holds
a condensed film of water molecules. In other words, the restrained molecules are held by van der Waals forces.
Because of cement paste the internal surface area is very extensive, a
large fraction of the total evaporable water can exist as a thin film spread
over solid surface. As has already been pointed out, opposite surfaces are
necessarily quite close together. In spaces up to about 1.8 nm (18 ,~) wide
(see above) adsorption of two molecular layers on each surface is sufficient
to fill most of such space, and such will be the condition when the ambient
humidity is about S0 percent. At higher ambient humidities, the pore
space may be almost completely full, adsorption being aided by hydrostatic
tension maintained by curved meniscuses of the water in capillary spaces.
A state of mechanical tensile stress between water molecules in the condensed state restrains evaporation in about the same way that van der
Waals attraction restrains evaporation from a thin layer on a solid surface.
Of the total capacity of mature cement paste for evaporable water, from
about one third to two thirds of it will be full when the ambient humidity
is only 50 percent. In rocks having porosities upwards of 1 percent the relatively large pores can retain very little water at that humidity, but at humidities upwards of 90 percent, a considerable fraction of the total capacity
of rock for evaporable water may be retained. Of course, different rocks
differ considerably in this respect; the fine textured rocks usually retain
a comparatively large amount of water at an intermediate humidity.
The failure of water in concrete to evaporate as it normally does shows
that it is altered to some degree by the material with which it is in contact.
Some of the alteration may be due to dissolved material, particularly the
alkalies, but removal of solutes only modifies the situation. The principal
forces acting on water and preventing it from evaporating into an unsaturated atmosphere have already been identified as van der Waals forces
of attraction and capillary-induced tension. Naturally, such forces are
balanced by counterforces, the counterforces corresponding to elastic
strains of one kind or another in the solid structure within which the water
is held. Such strains correspond to the reversible part of volume changes
due to drying and wetting of concrete.
There is good evidence that some of the evaporable water normally held
in concrete contributes to its strength, that is, van der Waals attraction between water molecules and between solid material and water molecules
seems to contribute to the total cohesive force. However, there is a somewhat greater weakening effect due to swelling.
72
Strength
In the above discussion dealing with cohesion and adhesion we necessarily
spoke of strength. But there are other aspects of the strength of concrete
depending on the nature of its structure, and some of these will be discussed
in the following paragraphs.
It would be expected that with a given aggregate the more cement gel
per unit volume of paste and the more paste per unit volume of matrix,
which means the lower the air content of the matrix, the greater the strength
of the concrete. Strength tests of many kinds of concrete show that such is
indeed the case; it was the fact underlying the cement-space ratio law for
strength given by Feret (1897), and the less general water-cement ratio law
given by Abrams (1918), and earlier by Zielinski (1908). For aggregates
composed of strong particles, the upper limit of concrete strength tends to
be established by the upper limit of the density of the matrix, that limit
usually being established by the means at hand for compacting the mixture.
On the other hand, with aggregates composed of relatively weak particles,
the upper limit tends to be established by the strengths of the aggregate
particles. This does not mean, however, that concrete cannot be stronger
than the strength of individual aggregate particles, as is evident when one
considers air bubbles as aggregate.
The strength of a given kind of concrete is not a single valued property
of the material. Compressive strength, in particular, is a function of the
rate at which stress is applied, the function being such that strength appears higher the higher the rate of loading. If a stress less than the strength
indicated by an ordinary test procedure, and yet greater than about two
thirds of that stress, is maintained without increase, the specimen will
eventually fail. There are good reasons to believe that failure is essentially
a random (stochastic) process, and therefore the stress existing at time of
failure is intrinsically a variable number. In other words, two absolutely
identical specimens would not be likely to fail at exactly the same stress,
or, under a sustained high stress less than the mean "instantaneous"
strength, they are not likely to fail at the same elapsed time after loading.
Thus, it seems that even with a perfect testing machine and with specimens
exactly alike, a variation of test results about the mean value from a large
number of identical specimens would still be found.
When a concrete cylinder is subjected to a steadily increasing axial
stress, as in the ordinary compression test, the observed increase in diameter at low stress is a certain constant fraction (Poisson's ratio) of the longitudinal shortening (compressive strain). There is no tensile stress associated with the lateral strain just mentioned, but after the compressive
stress becomes about two thirds of the stress at failure, further increase in
compressive stress causes lateral dilation to increase more than can be
accounted for by Poisson's ratio. This extra strain does denote the developCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:39:08 EDT 2014
73
ment of tensile stress. When such stress begins to exceed tensile strength,
vibrations due to internal splitting can be detected wifh suitable instruments. At failure, fractures appear as uneven surfaces having an inclination to the axis somewhat less than 45 deg, or fractures appear as cracks
parallel to the axis, or both kinds of fractures appear. Such experimental
evidence indicates that the failure of concrete under compression, or when
subjected to shear stress, is essentially failure in tension. If failure under
compressive stress were limited only by strength in pure shear, the principal strains preceding failure would involve only the sliding of one smooth
surface over another, and thus no increase in volume other than that
accounted for by Poisson's ratio would be necessary; hence, the observed
extra dilation with simultaneous internal cracking shows that tensile stress
develops across the incipient fracture surfaces and causes separations at
such surfaces.
Current theory of fracture indicates that during a compression test tensile stress necessarily develops around holes, cracks, or flaws in the material, and thus tensile stress would develop even in a continuum, structureless except for flaws. Without questioning this deduction one can at
the same time suggest that tensile stress would arise anyway, simply because of the granular nature of the material. Since the individual particles
of which paste or rock are composed are much stronger than the structure,
it is impossible for the surface of a fracture in paste to be smooth. Specimens of neat cement show evidence of tension failure in a compression
test; they have a strong tendency to split. In concrete made of strong aggregate material and weak paste, the nominal shearing surfaces are likely
to follow the contours of rock particles, in which case the cause of tensile
stress is obvious. In concrete made with strong pastes, fracture surfaces
usually pass through rock particles, but even so the fracture surface is not
smooth, and again it is not difficult to account for the development of tensile stress as one rough surface is forced to move away from the other as it
tends to slide. These considerations amount to classifying cement paste or
concrete as an intrinsically dilatant system, just as compact, uncemented,
granular systems are intrinsically dilatant.
Conclusion
While discussing the structure of concrete and the internal forces that
give it stability, various properties were discussed. It would be possible to
elaborate these brief discussions and to add other topics; however, other
papers in this publication are concerned with such reviews, and no doubt
interpretations in terms of structure will be found in them. Therefore, this
paper is concluded without further development of the subject.
D. T. S m i t h 1
Chapter 7--Uniformity and

Workability
Introduction
Concrete is a manufactured product. Any successful manufacturing
operation requires that the products produced from day to day are uniform
and similar in appearance and quality. Concrete manufacturing differs
from other types of manufacturing in many important respects which enhance the versatility of concrete for building construction, but increase
the problems of quality control and uniformity.
Uniformity of Concrete
Concrete is a heterogenous mixture of natural or artificial aggregates,
cement, water, and quite often entrained air, pozzolans, and other admixtures. The materials used in the manufacture of concrete have wide ranges
in physical and chemical composition, and these variations are reflected in
the workability, appearance, uniformity, consistency, and quality of the
concrete.
Nonuniformity may be evident in freshly mixed or hardened concrete.
There are two distinct classes of nonuniformity in freshly mixed concrete,
that is, within-batch variations and batch-to-batch variations.
Concrete initially discharged may vary in many important respects from
the last concrete in the mixer, or variations may occur throughout the
batch. These differences, which are classed as within-batch variations, may
be attributed to blade wear on the mixer, inadequate mixing, improper
loading sequence, or overloading of the mixer.
Concrete produced by equipment employing volumetric batching and
continuous mixing by an auger is sampled and tested, and has to meet
the same uniformity requirements (ASTM Specification for Concrete made
I Manager, Technical Services, Marquette Cement Manufacturing Company, Nashville,
Tenn. 37238.
74
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
SMITH ON UNIFORMITY AND WORKABILITY
75
by Volumetric Batching and Continuous Mixing (C 685)) as concrete

produced in central and truck mixers (ASTM Specification for Ready-Mixed
Concrete (C 94)). Concrete from this style of equipment may have variations
throughout the batch due to batching and mixing equipment as well as
possible segregation of aggregates in the storage hoppers.
Differences in appearance, workability, and quality of concrete placed
on different days or from different mixers on the same day are designated
batch-to-batch variations. These variations may occur as the result of
changes in one or more of the following:
1. Mixture proportions of aggregate, cement, water, and admixtures,
including batching errors
2. Moisture content and absorption of the aggregates
3. Aggregate grading
4. Type, surface texture, and shape of aggregate
5. Quantities of deleterious materials
6. Specific gravities of materials
7. Type and brand of cement
8. Different lots of cement from the same source
9. Concrete temperature
10. Air temperature
11. Humidity
12. Wind velocity
13. Type and condition of mixing equipment
14. Testing procedures
After mixing, every operation involved in the transportation, depositing,
and consolidation of concrete in the forms can contribute to segregation and
nonuniformity. Inadequate preparation of grades and forms may also influence nonuniformity by permitting sand, cement, and water to escape,
with a resulting accumulation of coarse aggregate, in certain areas.
The properties of the concrete in the finished structure are of primary
importance. Properties and tests of the individual components and of the
fresh concrete are valuable aids in predicting and controlling the uniformity,
homogeneity, and quality of the finished product. Experience has shown
that variations affecting the properties of plastic concrete will also affect
the properties of hardened concrete.
Control of Concrete Production
Production of uniform concrete can be obtained only through systematic
control of all operations from selection and production of materials through
batching, mixing, transporting, conveying, placing, consolidation, finishing,
and curing. The materials supplied batch after batch to the mixer must
have uniform properties. Control must be maintained at the batching and
76
mixing plants. Aggregates should be tested for gradation, specific gravities,

and moisture content, and mixture adjustments must be made to correct
for changes in these properties. Mixed concrete should be tested for consistency, air content, temperature, and unit weight. Concrete cylinders
should be fabricated for compressive strength tests.
Routine instructions for measuring, mixing, and placing concrete are
given by the American Concrete Institute (ACI) [I]. 2 Additional information
on practices that lead to better uniformity are found in the ACI Standard
Recommended Practice for Measuring, Mixing, Transporting, and Placing
Concrete (ACI 304). Some variation must be accepted, but consistent concrete of satisfactory quality can be obtained if proper control is maintained.
Concrete uniformity generally has been measured in terms of compressive
strength, slump, unit weight, air content, and content of coarse aggregate
or cement. Uniformity tests have been used to establish required mixing
time, mixing speed, mixer capacity, and efficient batching procedures.
Tests by a Number of investigators have been considered in the preparation of ASTM Specification C 94.
Methods of Measuring Uniformity

Dunagan Test--Dunagan [2] proposed a method for measuring the proportions of cement, water, sand, and coarse aggregate in fresh concrete by
a series of wash separations and weighing in air and water. This method
has been used by Slater [3], Hollister [4], Cook [5], and others to study the
effects of different rates of rotation of truck mixers, effect of time of haul,
and effect of mixing time on uniformity of concrete. The Dunagan test
has limited usefulness because of sampling errors and difficulties in distinguishing between cement and very fine sand.
U.S. Bureau of Reclamation Test of Mixer Performance--The U.S.
Bureau of Reclamation's mixer performance tests [6] are used to evaluate
the ability of a mixer to mix concrete that will be within prescribed limits
of uniformity. The uniformity of fresh concrete is evaluated by comparing
variations in quantity of coarse aggregate and unit weight of air-free mortar
of two samples, one taken from the first and last portions of the batch.
Large variations in this test may indicate that batching procedure is
inefficient, or mixer blades are worn. Additional mixing time may be required if the unit weight varies more than 24.0 kg/m 3 (1.5 lb/ft 3).
Bureau data for one series of mixing time tests on 3.06 m 3 (4 yd 3) mixers
showed that variation in air-free unit weight was reduced from 1.36 kg
(3.0 lb) to approximately 0.50 kg (1.1 lb) when the mixing time was increased from 1 to 2 min, but increased mixing beyond 2 min had little
additional influence on this variable. Variation in water-cement ratio was
77
reduced from 29 to S percent when mixing time was increased from 1 to 2

min in this same series of tests.
Air-Free Unit Weight Test--A study designed to establish test methods
and limits for variations in truck mixed concrete was reported by Bloem
et al [7]. Variations in slump, air content, percent of coarse aggregate,
air-free unit weight of mortar, water content by oven drying, and compressive strength of concrete obtained after approximately 1/6, 1/2, and s/6 of
discharge from a truck mixer were determined. They concluded that the
air-free unit weight test was an improvement over the unit weight test because the number of variables was reduced and excessive changes in this
property reflected changes in water or in proportions of cement and sand.
According to their data, a difference in air-free unit weight of mortar of
17.6 k g / m 3 (1.1 lb/ft. 3) correspond to a change in water of about 9.90
litres/m 3 (2 gal/yd 3) when the proportions of sand to cement were maintained constant and the water alone was varied. They suggested that a
variation of more than 16.0 k g / m 3 (1 lb/ft 3) in this test indicates real
differences in the proportions of the mortar ingredients, and differences of
more than 32.0 kg (2 lb/ft 3) should be considered evidence of unsatisfactory uniformity.
U.S. Army Corp of Engineers Method of Test for Concrete Mixer Performance (CRD-C55-73)--This method of evaluation tests 3 samples of
concrete for water content, cement content, unit weight of air-free mortar,
and coarse aggregate content. Tests are usually performed on samples of
concrete from the first and last 10 percent of the load, but not the very
first and very last portions of the batch to be discharged. The tests are
performed also to determine the feasibility of altering the mixing time. A
variability index for each test can be calculated by dividing the smallest
value by the largest value of three tests and divide by 100.
ASTM Specification C 94, Table XI, contains tolerances of test results
which are requirements of uniformity of a single batch of concrete.
Centrifuge Test--This test, also known as the Willis-Hime method, is
described in detail elsewhere [8]. It provides a basis for within-batch comparisons of cement content of concrete and employs a liquid with a density
greater than sand but less than cement. This liquid is used to separate the
components of a carefully prepared mortar sample extracted from the
concrete. This test is one of several mixer performance tests specified in
ASTM Specification C 94. Variability of 7 percent or more in the cement
content of two samples from the mixer is evidence of incomplete mixing.
The centrifuge test was used in the study by Bloem et al [7]. They concluded that the test is quite involved, and in most cases the information
gained is not commensurate with the time and labor required. Data cited
from previous tests indicated that the test was highly reproducible provided
extreme care was used and corrections were made for sampling errors in
coarse aggregate content of the small test portions.
78
Unit Weight--Unit weight of the concrete is influenced by the specific

gravity and amount of coarse aggregate, air content, proportions of sand
to cement, and water content. Consequently, variations in unit weight are
difficult to evaluate as to cause or significance. The test is more definitive
when the weight and solid volume of coarse aggregate and volume of air
are eliminated as in the air-free unit weight of mortar test.
The unit weight test is recommended as a job control measure for lightweight aggregate concrete in conjunction with air determinations and slump.
If the slump and air content are maintained constant, a change in unit
weight indicates a change in weight of aggregate. If the weight of aggregate per cubic metre of lightweight concrete changes, it may be the result
of change in moisture content, gradation, or density of the aggregate.
Additional tests, including unit weight, moisture content, and gradation
of the aggregate are required in order to determine the cause of the change.
Air Content--Air content has an important influence on workability,
compressive strength, and durability of concrete. The strength of concrete
decreases uniformly with increase in air content of the fresh concrete provided the water-cement ratio is held constant. Air entrainment also increases
slump, each 1 percent of additional air being approximately equivalent to
2.5 cm (1 in.) in slump. When the slump is maintained constant by reduction in water content, strength reduction due to air entrainment is not so
great and the strength of lean mixes may be slightly increased. Strength
reductions of 16 to 20 percent at 28 days have been reported for concretes
containing 5.5 to 6.5 sacks of cement per cubic metre when 5 percent entrained air was added to the concrete.
Too little air will not provide workability and durability. Sudden loss of
workability may indicate a reduction in air content. Resistance of concrete
to freezing and thawing is increased several hundred percent by incorporating the correct amount of air in the concrete. Consequently, it is important that the concrete contain a uniform quantity of air. Within-batch
variations should not exceed 1 percent.
Variation in air content obtained from a given dosage of air-entraining
agent may result from changes in concentration of the agent, brand or type
of cement, sand grading, pozzolan, or admixture, temperature of the mix,
slump, or length of mixing. Air content is generally maintained at the correct level by increasing or decreasing the dosage of air-entraining agent
because it may not be practical or desirable to remove the cause of the
variation. When air-entraining cement is used, control of air content is
more difficult because any change in the dosage of air-entraining agent
must be accompanied by a corresponding change in cement content. For
this reason, many engineers prefer to use regular cement and add the airentraining agent at the mixer.
Air content of normal weight concrete may be computed by comparing
the actual unit weight of concrete with the theoretical weight based on
79
the specific gravity of the materials used, as outlined in ASTM Test for Unit
Weight, Yield, and Air Content (Gravimetric) of Concrete (C 138). Results
obtained by this method are influenced by variations in mix proportions,
specific gravity of ingredients, and changes in moisture content of aggregates. Furthermore, any change in batch weight requires a new computation.
Several meters have been developed to determine directly the air content
of fresh concrete. Determination of air content by this method is generally
preferred to the gravimetric or unit weight method. ASTM has standardized
two methods for determining air content by means of meters.
The pressure meter (ASTM Test for Air Content of Freshly Mixed Concrete by the Pressure Method (C 231)) consists of a special pressure-tight
container and accessories designed to hold an accurate volume of concrete.
The container is filled with concrete, the lid is securely fastened, and pressure is applied by a hand pump to a predetermined amount. The apparatus
is so calibrated that the percentage of air is read directly when the pressure
is released. This method is used more than any other and is considered
satisfactory for all types of concrete and mortar except those made with
highly porous lightweight aggregate. This apparatus must be calibrated
periodically to guard against changes caused by rough usage, and, if the
elevation of the place at which the apparatus is used changes by more than
183 m (600 ft), it should be recalibrated. An aggregate correction factor
should be determined with the materials used and subtracted from the
apparent reading to determine the actual air content. The aggregate correction factor varies only slightly for the same type of aggregate and need
only be checked when there is a definite change in materials.
The volumetric method (ASTM Test for Air Content of Freshly Mixed
Concrete by the Volumetric Method (C 173)) consists of removing air from
a concrete sample by mixing it with water in a special container. The volume of air is determined from the difference in volume of the sample containing entrained air and the volume of the sample after it has been agitated
to permit the air to escape. This method is recommended particularly for
lightweight concrete, but it may be applied to other types of concrete as
well.
The air indicator is a miniature device that uses the volumetric principle.
A small sample of carefully selected mortar is obtained from the concrete
and placed in a brass cup measuring 1.9 cm (3/4 in.) in diameter by 1.3 cm
(1/2 in.) high and compacted with a wire or knife blade. The glass tube that
comes with the device is filled to the top line with isopropyl alcohol, the
brass tube is inserted in the tube, and the liquid level is adjusted to the
top line. The finger is placed over the stem to prevent alcohol from escaping, and the indicator is rolled several times until all mortar has been removed from the cup. With the indicator in a vertical position, the finger
is carefully removed from the stem, and the number of graduations from
80
the top to the new liquid level gives an indication of the air content in the
mortar sample. A correction factor, based upon the design mix, must be
applied to convert to percent air in concrete. Meticulous care must be used
in selection of the mortar sample, method of inserting the stopper, agitation
of the sample, and removal of the finger from the tube. The test can only
provide an indication of relative air content and cannot be considered as
reliable as the pressure meter or the volumetric method.
Slump--The slump test (ASTM Test for Slump of Portland Cement
Concrete (C 143)) is essentially an indication of the wetness of the concrete.
Within-batch variations in slump indicate incomplete mixing and nonuniform distribution of water or other ingredients throughout the batch.
Batch-to-batch variations may result from hatching errors, uncorrected
changes in moisture content or grading of the aggregate, or variations in
temperature. In reasonably uniform concrete, slump should not vary more
than about 2.5 cm (1 in.) within a batch.
Cement Content of Fresh Concrete--Constant neutralization by 3 N
hydrochloric acid (HCI) for a fixed period of time (1 h in this procedure)
has been employed by the California Department of Transportation [9].
Accuracy of this procedure was reported to be "within about ~h sack per
cubic yard." Field tests indicate the procedure is most useful for evaluating
the performance of a concrete mixer. Relative cement contents of various
portions of a batch can be determined in about 1 h. The report states that
"adjustments to aggregate a n d / o r cement feed can be made as needed to
improve mixing uniformity."
This test procedure was apparently used on concrete that did not contain
calcareous aggregates; at least, there was no discussion of the reaction of
HC1 on the aggregates, and this is a phenomemon that would have a profound effect on the end results of the test.
The authors of the report state in their conclusions that the test was not
proven to be of sufficient accuracy for routine control of cement content
during normal concrete production, and at present, the most applicable
use of the procedure is testing mixer efficiency. For the latter purpose they
found the titration method to be more effective than the "wash-out-method"
and the "heavy media method" that were also tried in this program.
The procedure requires the development and use of calibration curves
when determining the relative variation of cement content within a batch.
"In the event of any change in source of aggregate, cement, or working
solution of acid, a new calibration curve must be established" according
to the procedure outlined in the report.
Rapid Analysis Method--This method of test is for the determination of
cement and water content in fresh concrete. It can be used in the laboratory
or in the field. The proposed method is essentially the same as that originally proposed by Kelly and Vail with the exception that this system [10]
81
uses a titration method rather than flame photometer for calcium determination.
The equipment and method are described in the report [10]. Statistical
analysis of the Kelly-Vail test data shows the development of a predictive
relationship between 28-day strength (S), the Kelly-Vail water/cement (w/c)
ratio (KVC), and air content (A) of the concrete; also presented is an
evaluation of the repeatability of KVC measurements and 28-day strengths
within and among concrete batches.
The method relies on chloride ion titration for determining water content
in the fresh concrete, and calcium titration for cement content. Also, the
test can be completed in the field in 15 min. The system has been proven,
by extensive field testing, to be accurate enough to estimate the strength
potential of fresh concrete. The reagents required for water content determination are: 0.5 N sodium chloride (NaC1) solution, a 0.05 N potassium
thiocyanate (KSCN) solution, nitrobenzene (C6H~NO2), saturated ferric
alum solution, and a 50 percent nitric acid (HNOa) solution. The reagents
needed for the calcium titration (cement content determination) are: a 5
percent HNO3 solution, tap water, 0.01 M solution of disodium ethylenediaminetetra-acetate (EDTA) in distilled water, eriochrome black T indicator, and ammonia-ammonium chloride NH3/NH2 CI buffer solution. EDTA
should be stored in polythene bottles.
The limitation of the method is that the 1-kg ( - 2.S-Ib) sample being
used limits the maximum size coarse aggregate in the concrete to 3.8 cm
(11/2 in.). This limitation can be removed by modifying the procedures and
using larger samples by applying the restriction that the largest aggregate
must weigh less than 10 percent of the specimen weight. The cement test
has a second limitation caused by the technique measuring all calcium
passing through the smallest sieve rather than cement only. Tests have been
made on both fine and coarse calcareous aggregates, and they indicate that
only the fine aggregates cause a significant interference. At the present
time, the operational accuracies of the CERL/K-V water and cement content tests as determined by extensive laboratory and field tests indicate an
average error of 4 to 6 percent for water contents and 6 to 8 percent for
cement contents.
This method seems to be a very useful tool for measuring and controlling
the uniformity of the w/c ratio of fresh concrete in the field--much better
than the other methods in use for measuring uniformity as it is related to
cement and water contents. It also appears to have excellent potential for
concrete strength values at a very early age, that is, in 15 min by using w/c
versus strength curves.
Compressive Strengths--Measured concrete strength is used widely as a
criterion of concrete quality. Other factors such as durability, abrasion
resistance, thermal properties, dimensional stability, placeability, and corn-
82
pactability may be more critical, but strength tests are easily made and
variations in strength are assumed to be indicative of variations in other
properties.
Compressive strength is a control test used to determine the degree of
uniformity of concrete. ACI Committee 214 has developed a Standard
Recommended Practice for Evaluation of Compressive Test Results of Field
Concrete (ACI 214-65). This evaluation is statistical in nature and is a
method for control of strength based on the coefficient of variation. Table
XI of ASTM Specification C 94 lists requirements for uniformity of concrete, and these are expressed as maximum permissible difference in results
of samples taken from two locations in a concrete batch. The permissible
difference for the average compressive strength at 7 days for each sample
(not less than 3 cylinders), based on the average strength of all comparative
specimens is 7.5 percent. In a well controlled laboratory, compressive
strength of cylinders fabricated from the same concrete sample may vary
3 to 5 percent. Variations in excess of this amount must be attributed to
variations within the batch of concrete.
Workability
Workability is a term that is used every day in concrete construction,
and it is a factor easily appreciated in practice. Workability means different
things to different people and for different placing conditions. Various
methods have been developed for its measurement. None of these tests
evaluate all characteristics that are involved in this property.
Glanville [11] defined workability as: "that property of the concrete which
determines the amount of useful internal work necessary to produce full
compaction." Powers [12] defined it as: "that property of the plastic concrete mixture which determines the ease with which it can be placed and
the degree to which it resists segregation." Both of these definitions are
expressed in terms related to the physical characteristics of the concrete
alone, being independent of the methods of placing and compacting.
Workability of concrete is defined in ASTM Definition of Terms Relating
to Concrete and Concrete Aggregates (C 125) as: "that property determining the effort required to manipulate a freshly mixed quantity of concrete with minimum loss of homogeneity."
In actual practice, workability is related directly to the type of construction and methods of placing, mixing, and working. Concrete that can be
placed readily without segregation in a mass dam would be entirely unworkable in a thin structural member. Workable concrete compacted by
means of high-frequency vibrators would be unworkable if vibrators could
not be used and hand tamping and spading were required. Concrete having
suitable workability for a pavement might be unsuitable for use in a thin,,
heavily reinforced section.
83
Properties involved in workability include finishing characteristics, consistency or fluidity, pumpability, mobility, segregation, and bleeding. None
of the test methods proposed or in use today simultaneously measure all of
these properties. Consequently, measurement of workability is determined
to a large extent by judgement, based on experience.
Workability is dependent upon the physical and chemical properties of
the individual components and the proportions of each in the concrete.
The degree of workability required for proper placement and consolidation
of concrete is governed by the type of mixing equipment, size, and type of
placing equipment, method of compaction, and type of concrete.
Factors Affecting Workability

Some of the factors that affect the workability [11] of concrete are quantity of cement, characteristics of cement, consistency, grading of fine aggregate, shape of sand grains, grading and shape of coarse aggregate, proportion of fine to coarse aggregate, percentage of air entrained, type and
quantity of pozzolan, quantity of water, and amount and characteristics of
admixture used.
Cement--Very lean mixes tend to produce harsh concrete having poor
workability. Rich mixes are more workable than lean mixes, but concrete
containing a very high proportion of cement may be sticky and difficult to
finish. An increase in the fineness of cement increases the cohesiveness
of the concrete mix as well as the rate at which cement hydrates and the
early strength development. Differences in bleeding tendency, accumulation of laitance, and other properties of concrete made with cements having
the same fineness and chemical analysis have been observed.
Consistency--Consistency and plasticity are terms often used to indicate
workability. Consistency generally denotes the wetness of the concrete which
is commonly measured by the slump test. It must not be assumed that the
wetter the mix the more workable the concrete. If a mix is too wet, segregation may occur with resulting honeycomb or sand streaking on the exposed
surface; finishing properties will be impaired because of the accumulation
of laitance on the surface. If a mix is too dry, it may be difficult to place
and compact, and segregation may occur because of the tendency for larger
particles to roll towards the outer edge of the heap, formed when it is deposited. It is agreed generally that concrete should have the driest consistency that is practicable for placement with proper vibration. The consistency necessary for full compaction varies with the type of structure,
type and size aggregate, and type of compaction equipment available.
Sand--Concrete containing fine sand requires more water for the same
consistency, as measured by the slump test, than an equivalent amount of
coarse sand. Very coarse sand can have an undesirable effect on finishing
quality. Neither very fine or very coarse sand is desirable. Rounded river
84
sand gives greater workability than crushed sand composed of sharply

angular pieces with rough surfaces. Angular sand particles have an interlocking effect and less freedom of movement in the freshly mixed concrete
than smooth rounded particles. Natural sand may give satisfactory results
with a coarser grading than would be permitted with crushed sand. In
addition, concrete must contain 2 to 3 percent more sand by absolute volume of total aggregate and 3.70 to 5.6 kg (10 to 15 lb) more water per
0.765 m (1 yd 3) when crushed sand is used.
Coarse Aggregate--The particle size distribution of coarse aggregate influences water requirements and workability of concrete. Coarse aggregates
meeting standard grading requirements, such as ASTM Specification for
Concrete Aggregates (C 33), should be used. After the grading is established,
it should be maintained within rather close tolerances to avoid sudden
changes in workability and other concrete properties. Segregation is reduced and uniformity improved by separating the aggregate into several size
fractions and recombining these fractions when concrete is manufactured.
Breakage, segregation, and contamination of aggregate can occur during
handling and stockpiling. Introduction into the mixer of a large quantity
of undersize material which may have accumulated will result in a sudden
change in workability resulting in a demand for additional water. ACI
Committee 304 has suggested rescreening at the batch plant as the method
most likely to eliminate undesirable undersize and promote uniformity.
Production of workable concrete with sharp, angular, crushed aggregates
generally requires more sand than similar concrete made with rounded
aggregates. The water content may be increased 8.9 to 14.8 kg/m 3 (15 to
25 lb/yd 3). If the w/c ratio is held constant, more cement is required. Flat
or elongated particles which are defined as particles having a ratio of width
to thickness or length to width respectively, greater than 3:1, are detrimental to concrete workability. More sand, cement, and water are required
when the coarse aggregate contains flat and elongated particles.
The maximum size of aggregate that can be used to produce workable
concrete is limited by practical considerations including type and size of
structure, amount and spacing or reinforcing bars, method of placing,
and availability of materials. Generally, aggregate should not be larger
than three fourths of the maximum clear spacing between reinforcing bars
nor larger than one fifth of the wall thickness or narrowest dimension
between sides of forms. ACI Recommended Practice for Selecting Proportions for Concrete (ACI 211.1) provides recommendations on the maximum
sizes of aggregate for various types of construction. U.S. Bureau of Reclamation [13] experience in pumping concrete indicates that concrete
containing 6.4 cm (21/2 in.) maximum size aggregate can readily be pumped
through an 20.3 cm (8 in.) pipe, but aggregate larger than 6.4 cm (21/2 in.)
may cause difficulty. Concrete containing aggregate graded to 2.5 cm (1 in.)
maximum size can be placed by pneumatic equipment.
85
Air Entrainment--Entrained air increases the paste volume, acts as a

lubricant, and improves the workability of concrete. It reduces bleeeding
and segregation during handling and placing of concrete and increases
cohesiveness or "fattiness" of the concrete. Improvement in workability
resulting from air entrainment is more pronounced in lean mixes that are
harsh and unworkable because of poor aggregate grading or type of aggregate used,
Finely Divided Materials--Addition of finely divided material, including
inert or cementitious materials or pozzolans, generally improves the workability of the concrete. Improvement is more noticeable in lean mixes than
in rich mixes. These materials have been used to improve the grading of
sands deficient in fines. Cementitious and pozzolanic materials are usually
substituted for 10 to 25 percent of the cement. Workability will be improved if these materials are added as replacement for part of the sand, instead of substituted for part of the cement.
Chemical Admixtures--Water-reducing admixtures, when added to concrete, permit a reduction in mixing water with no loss in slump, or, if the
water content is held constant, produce an increase in slump. Set retarding
admixtures reduce the early rate of hardening and permit concrete to be
handled and vibrated a longer period of time.
It has been reported that there is a decrease in the frequency of plugged
pump lines when water-reducing retarders are used in the concrete. In
addition, less power may be required to pump the concrete.
Mixture Proportions--Workability can be controlled by proper proportioning of the constituent materials. As the proportion of mortar, including
sand, cement, water, and air, is increased, the grading and angularity of
the coarse aggregate become less important. There should be sufficient
mortar to fill the voids in the coarse aggregate plus a sufficient amount to
permit the concrete to be placed readily in forms and vibrated around
reinforcement. An excess of mortar increases workability, but excess workability is inefficient. It should not be more than is required for consolidation by modern equipment. The quantity of mortar required to produce the
desired workability with a given coarse aggregate can be determined more
effectively by laboratory tests. ACI Recommended Practice for Selecting
Proportions for Concrete provides a basis for estimating the proportions
of coarse aggregate to be used in trial mixes.
Methods of Measuring Workability
Slump Test--The slump test (see Fig. 1) is the most commonly used
method of measuring consistency or wetness of concrete. It is not suitable
for very wet or very dry concrete. It does not measure all factors contributing to workability, nor is it always representative of the placeability of the
concrete. However, it is used conveniently as a control test and gives an
86
FIG. 1--Slump test.
indication of the uniformity of concrete from batch to batch. Repeated

batches of the same mix, brought to the same slump, will have the same
water content and w/c ratio provided weights of aggregate, cement, and
admixtures are uniform and aggregate grading is within acceptable limits.
Additional information on workability of the concrete can be obtained if,
after removing the slump cone, the concrete is tapped on the side with the
tamping rod. Two concretes with the same slump may behave differently,
that is, one may fall apart after tapping and be harsh with a minimum
of fines, and the other may be very cohesive with surplus workability. The
first concrete may have sufficient workability for placement in pavements
or mass concrete, but the other concrete may be required for more difficult
placement conditions.
The slump test should be performed ~n strict accordance with the requirements of ASTM Test C 143. Tests are usually made at the point of placement and should be made whenever cylinders are molded for compressive
strength testing. Slump tests may be made at the mixing plant in order
to check the uniformity of batching operations.
Popovics [14] has presented data indicating that the relationship between
consistency values, as measured by the slump test and the water content of
concrete, is parabolic, that is, the percentage change in water content required to increase the slump 2.5 cm (1 in.) may vary from 2.0 percent
when the initial slump is 12.7 cm (5 in.) to approximately 4.5 percent when
the initial slump is 5.1 cm (2 in.). An average change in water content of 3
percent generally is considered necessary for a 2.5 cm (1 in.) change in
slump.
As the temperature of the concrete increases, the slump decreases. Concrete placed at a slump of 10.2 cm (4 in.) at 21~ (70~ may only have a
7.6 cm (3 in.) slump when placed at 32~ (90~
or the same concrete
may have a slump of 14 cm (5.5 in.) when placed 10~ (50~
Air entrainment and water-reducing admixtures will increase slump of
concrete if all other conditions remain the same. Each 1 percent increase
or decrease in air content will produce approximately the same influence as
a change in water content of 3 percent.
87
The slump generally is reported to the nearest 0.6-cm (1/4-in.). Slump

reported by different operators on the same batch of concrete may vary
by as much as 1.3 cm (1/2 in.). The most unsatisfactory form of slump is
the shear slump, that is, a falling away or shearing off of a portion of the
concrete from the mass. It is impossible to lay down rules on the correct
point to measure the slump. If this condition exists, the concrete probably
lacks the necessary plasticity for the slump test. It may be difficult to consolidate in the forms, but if consolidation of the concrete is practical, the
slump test is not a valid control test for this type of concrete.
K-Slump Tester--The K-slump tester (see Fig. 2) is reported to measure
slump directly in 1 min after the tester is inserted in the concrete [15]. The
apparatus is comprised of the following four principal parts.
1. A chrome-plated steel tube with external and internal diameters of 1.9
and 1.6 cm (0.63 and 0.75 in). The tube is 25 cm (9.84 in.) long, and its
lower part is used to make the test. The length of this part is 15.5 cm
(6.10 in.) which includes the solid cone that facilitates inserting the tube
into the concrete. Two types of openings are provided in this part: 4 rectangular slots 5.1 cm (2.01 in.) long and 0.8 cm (0.31 in.) wide, and 22
round openings are distributed uniformly in the lower part.
2. A disk floater 6 cm (2.36 in.) in diameter and 0.24 cm (0.09 in.) ir.
thickness which divides the tube into two parts; the upper part serves as a
handle and the lower one is for testing. The disk also serves to prevent the
tester from sinking into the concrete beyond the preselected level.
3. A hollow plastic rod 1.3 cm (0.51 in.) in diamter and 25 cm (9.84 in.)
long which contains a graduated scale in centimetres. This rod can move
freely inside the tube and can be used to measure the height of mortar that
flows into the tube and remains. The rod is plugged at each end with a
plastic cap to prevent concrete or any other material from entering.
4. An aluminum cap 3 cm (1.18 in.) in diameter and 2.25 cm (0.89 in.)
long which has a little hole and a screw that can be used to set and adjust
the reference zero of the apparatus. In the upper part of the tube there is
also a small pin which is used to support the measuring rod at the beginning
of the test.
The K-slump tester is reported to measure an index which is related to
workability after the device is removed from the concrete. The first reading
is taken after the tester has been in place (the disk resting on the concrete
surface) for 60 s. This reading in centimetres is referred to as the K-slump.
The device is removed from the concrete and the measuring rod is again
lowered to rest on the surface of the concrete remaining in the tube; this
reading in centimetres represents the workability of the mix.
Studies have been made and reported on 420 concrete batches by five
laboratories. Statistical determinations and equations are reported else
where [15].
Placeability--An article by Angles [16] described a placeability apparatus
88
TESTS A N D P R O P E R T I E S OF C O N C R E T E
63,smA.
--I
;~__~
(2,501
-0
"
2
5
COLLAR FLOATER
MAT.,IS GAUGE STEEL
sO,O ,I
I2.SS(
"6
HOLE O(A., AND SLOT

WIDTH 6 . S m m
(255)
~o
(.o61
THROUGH HOLE FOR 1.5 DIA.
RIVET PROTRUDING 2.5
INSIDE TUBE
(.Jol
_ ~ r ~ 28.0
. - - ( . = O_~_).
t
- -
-"" ~z--5
[3351
~T
(.591
(.591
(.59J (2551
li!
Jill
TUBEPATTERN
MAT., 18 GAUGE STEEL
TOTAL WEIGHT OF PARTS NO 4,5.7.= 22S GMS
(71111 OZ.)
ASSEMBLED
PARTIAL
AND
TESTER
SIDE
VIEW
SECTION
NOTE
: - CHROME PLATE
TUBE, COLLAR AND P O I N T
- A L L DIMENSIONS IN M t L L I M E T R E S . (INCHES).
- T O T A L WEIGHT 3 8 2 GRAMS. ( 1 3 4 8 4 OUNCES)
FIG. 2--2-K slump tester.
SMITH ON U N I F O R M I T Y A N D W O R K A B I L I T Y
~16.50.D.
(.491
14.50.D.
11rl
AS SUPPLIED
WT : I0 BMS (0:55OZ)
MODIFIED
WT : 07 GMS { 024 OZ)
PLASTIC CAPS
6
.
.D ~
2 .9
(''~' 0
l93
:~.,I'D' ,
IE SO.D.
(50)
15.9 1.0.
20.5 OIA.
(.545~
'
t.,~;
DRILL AND TAP FOR
NO IO-Z4 UNC 4ram LONG (16)
SOCKET SET SCREW
-~T-Ii
i ~ol
.....
WT : 21.8 GMS.(?,t oz)

3 SCALE DATUM
MAC, ALUMINUM
b-~-I
20.50.D.
(.el}
17.5 I.O.
(6,)
~0~
(.81)
z~176
T-=
-..L
7 TUBE POINT
2 MEASURING ROD
MAT, CLEAR ACRYLIC TUBE
MAT., STEEL
WT; 19.5 GMS ( 6e40Z I

FIG. 2--Continued.
89
90
that would simulate the operation of placing concrete under typical conditions. The equipment is comprised of three units:
1. A rectangular, heavy gage container of suitable size with two vertical
channels at the midpoints on the inside of the longer sides
2. A screen, secured by removable clamps, made of round horizontal bars
with openings between them of a size appropriate to the maximum size
of aggregate in the mix under test
3. A plate to act temporarily as a shutter during the placing of the concrete in Compartment B of the container.
The test is performed by placing a known weight, for example 25 kg
(55 lb) of concrete into Compartment B with a standardized scoop. The
shutter is removed and the vibrating table, on which the test apparatus
rests, is switched on. The time required for the concrete to pass through
the screen and attain the same level in both compartments is measured
with a stop watch. Placeability is the time expressed in seconds required
from turning on the switch to leveling of the sample along the full length
of the box.
The author also suggests that internal vibration could be used by fixing
diameter, frequency, amplitude, and location of the vibrator. Other data
and observations that could be obtained, according to Angles, are bleeding,
segregation, and bulk density.
Remolding Test--The remolding test apparatus (see Fig. 3) was developed
by Powers [12] to measure "the relative effort required to change a mass
of concrete from one definite shape to another by means of jigging." The
FIG. 3--Remolding apparatus.

91
equipment consists of a metal cylinder mounted inside a larger cylinder

and a suspended plate which fits inside the smaller cylinder. A slump cone
is placed inside the smaller cylinder so that the bottom rests on the base.
It is filled with concrete, the slump cone is removed, and the plate is placed
on top of the concrete. The flow table on which the apparatus is mounted
is then operated. The number of 0.6-cm (l/4-in.) drops, required to mold
the concrete to a cylindrical form, is a measure of the workability of the
concrete. This method has not found widespread use and no ASTM standard has been written about it.
The remolding test and slump test were used by Cordon [17] in an extensive series of tests on air-entrained concrete. He found that workability,
as measured by the remolding test, was increased when air content was
increased, and slump and percent sand were maintained constant. When
percent sand was reduced as the air content was increased and slump was
constant at 10.2 cm (4 in.), the remolding effort was also constant at
approximately 42 jigs. It might be concluded that the remolding test is
more sensitive to changes in air content than the slump test.
Ball Penetration Test--The Kelly ball test (see Fig. 4) [18] was developed
principally as a convenient method of measuring and controlling consistency
in the field. The ball test can be performed on concrete in the forms, and
it is claimed that tests can be performed faster and precision is greater
than with the slump test. One disadvantage of this test is that it requires
a large sample of concrete.
The apparatus weighs 13.6 kg (30 lb) and consists of a 15.2-cm (6-in.)
diameter ball and stem which can slide through the center of a stirrup,
the legs of which rest on the concrete to be tested. The depth of concrete
must be at least 20.3 cm (8 in.), and the minimum distance from the center
of the ball to the nearest edge of the concrete is 22.9 cm (9 in.).
The surface of the concrete is struck off level, avoiding excess working.
FIG. 4--Ball penetration apparatus,

92
The ball is lowered gradually onto the surface of the concrete, released,
and the depth of penetration read immediately on the stem to the nearest
0.6 cm (l/4-in.). The ratio of slump to the penetration of the ball is between
1.5 and 2 and is fairly constant for a given mix but varies according to the
mix. The method has been adopted as the ASTM Test for Ball Penetration
in Fresh Portland Cement Concrete (C 360).
The ball penetration test was compared with the slump test by Howard
and Leavitt [19]. Twenty tests were made using each method. The ball
penetration averaged 1.32 with a standard deviation of 0.45 and the slump
average 2.5 with a standard deviation of 0.81. They concluded that the
slump test required approximately 10 min, and the ball penetration only
required 10 s when used on paving concrete.
The Thaulow concrete tester (see Fig. 5) [20] consists of a 10-1itre container with a gradation mark at 5 litres, equipped with a slump cone, and
fitted with a stainless steel handle, and a drop table actuated by a crank
which drops the table through 1 cm (0.394 in.) four times for each revolution.
The slump cone is fastened in the container, filled with concrete in the
usual manner, and lifted off. The handle is allowed to fall freely from the
vertical position alternately on each side of the container until the concrete
is remolded and the entire periphery is at the 5-1itre mark. The number of
blows required is an indication of the workability.
If the concrete is very dry, the container is fastened to the drop table.
The slump cone is fastened, filled in the usual manner, and further compacted by 15 drops of the table. The cone is removed, and the number of
revolutions of the crank handle necessary to bring the concrete to the 5-1itre
mark is a measure of the consistency.
ACI has prepared a Recommended Practice for Selecting Proportions for
FIG. 5 - - T h a u l o w
c o n c r e t e tester.
93
No-Slump Concrete (ACI 211). The concrete has a slump less than 0 cm
(0 in.). Differences in consistency of very dry mixes cannot be measured
with the slump cone, but the Thaulow equipment is considered to have
merit for this application. Concrete with a slump of 2.5 to 5.1 cm (1 to 2 in.)
requires 14 to 28 drops, and concrete with a slump of 7.6 to 10.2 cm (3 to
4 in.) requires less than 7 drops.
Vebe Apparatus--The Vebe consistometer (see Fig. 6) [21] consists of a
vibrating table, a sheet metal pan, slump cone, and plastic plate attached
to a graduated, free-moving rod which serves as a reference end point. The
cone is placed in the pan, filled with concrete, and removed. The plastic
disk is brought into position on top of the concrete, and the vibrating table
is set in motion. The number of seconds required to remold the cone of
truncated concrete to the shape of the cylinder is the measure of consistency
and is reported as the number of Vebe seconds or degrees. This method is
very suitable for very dry concrete, but the variation is too vigorous for concrete with a slump greater than about 5.1 cm (2 in.). For example, 0 to 3 s
is required for concrete with a slump of 7.6 to 10.2 cm (3 to 4 in.), and 10
to 52 s may be required for concrete with less than 0-cm (0-in.) slump.
Compacting Factor--The compacting factor test (see Fig. 7) has been
developed in Great Britain [11]. The apparatus consists of two conical
hoppers fitted with strong doors at their base and a 15.2 by 30.4-cm (6 by
12-in.) cylinder. The top hopper is filled with the concrete to be tested and
struck off without compacting it. The door at the bottom of the hopper is
opened, and the concrete drops by gravity into the somewhat smaller hopper below. The door of the second hopper is opened, and the concrete is
allowed to fall into the cylinder which is struck off and weighed. The ratio
of the weight of concrete in the cylinder mold to the weight of concrete
I'
/-1
',1
FIG. 6-- Vebe apparatus.

94
FIG. 7--Compacting factor apparatus.
from the same batch fully compacted in the mold is the compacting factor. The sensitivity of this test is considered good for medium consistency concrete, but less than some other tests for very dry concrete.
An average compacting factor of 0.75 corresponds to a slump to 0 to
2.5 cm (0 to 1 in.), and 0.90 is approximately 7.6 to 10.2-cm (3 to 4-in.)
slump.
Wigmore Consistometer--The Wigmore consistometer (see Fig. 8) is
described by Orchard [22].
This apparatus consists of a galvanized container and handoperated
compaction table. A 5.1-cm (2-in.) diameter ball which is fastened to a
sliding stem is mounted in the lid of the container. The container is filled
with concrete which is compacted on the table by 8 drops. The container
is again filled with concrete and leveled off; then the lid and the ball are
placed in position with the ball resting on the surface of the concrete. The
apparatus is placed on the table and the concrete is compacted by turning
95
FIG. 8 - - Wigmore consistometer.
the handle attached to the cam at the rate of about 1 rps. The table drops
0.6 cm (%2 in.) four times per revolution of the cam and the number of
drops required to lower the ball and stem 19.7 cm (73/4 in.) into the concrete is considered a measure of the consistence of the concrete.
The number of drops-required may vary from 20 for very wet concrete
15.2-cm (6-in.) slump to 200 to very stiff concrete.
It is claimed that the Wigmore consistometer is an improvement over
the slump test because work is actually done on the concrete in a way which
resembles field conditions. Variations in results may be expected if the
ball comes in contact with large aggregate.
Flow Cone--Contraction joints in dams, cleavage planes in rock fbundations, cavities behind tunnel linings, voids in preplaced aggregate, and
openings around post-tensioned cables may be filled with grout pumped
under pressure. Grouts consist of cement and water or combinations of
sand, cement, water, finely divided filler, and admixtures. Grouts must be
very fluid to penetrate small cavities. The slump test and other described
methods of measuring concrete consistency are unsuitable.
The U.S. Corps of Engineers has prepared a standard test procedure
for measuring the flow of grout mixtures of means of the flow cone (Method
of Test for Flow of Grout Mixtures (CRD-C79-58)) (see Fig. 9). This method
96
FIG. 9 - - F l o w cone.
outlines the procedure to be used in the laboratory and in the field for
determining the consistency of grout mixtures by measuring the time of
efflux of a specified volume of grout from a standardized flow cone or
funnel. The flow cone is mounted firmly with the top surface level, the
discharge tube is closed by placing the finger over the end, and 1725 ml
of mixed grout poured into the cone. The finger is removed, and the number of seconds until the first break in the continuous flow of grout is the
efflux time. When comparing grouts, the speed of mixing and the mixing
time have an influence on efflux time and should be maintained constant.
Grout Consistency Meter--A meter for measuring consistency of grout
has been developed at the University of California and is described elsewhere [23]. The grout consistency meter (see Fig. 10) is essentially a torque
meter. The sample of grout is placed in a metal pan mounted on a platform that can be rotated at a constant speed of 60 rpm. Suspended from a
music wire is a 7.3-kg (16-1b) paddle assembly to which a torque is applied
as the sample of grout is rotated. The angle of twist or consistency factor
is read by an index pointer attached to a cross strut.
Otto Graf Viscosimeter--The Otto G r a f viscosimeter (see Fig. 11) has
been used in Europe to measure fluidity of grouts. It consists of a brass
tube, approximately 90.2 c m (351/2 in.) long and 6.2 c m (27/16 in.) inside
diameter mounted on a base in an upright position, and an immersion body
weighing 5 kg (11.1 lb) that is 28.2 cm (113/32 in.) long. The immersion
body has protuding cams on the sides so that grout can pass between them
when the immersion body is placed in the tube containing the grout. In
practice, the tube is filled to a prescribed level with grout. The immersion
body is then placed in the tube and released. The number of seconds reCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:39:10 EDT 2014
97
1"3
~7 r--"
I
I
FIG. lO--Grout consistency meter.
quired for the body to assume its final position is a measure of the fluidity
of the grout.
This equipment was specified as a control measure for grouting operations on a post-tensioned bridge. Consistency was also measured with the
flow cone. Tests were made immediately after mixing and again 30 min
after mixing. The tests indicate that the flow cone provided the same information as the viscosimeter and was as effective for control purposes.
The ratio of viscosimeter reading to flow cone reading was approximately
1:5.
Conclusions
Concrete knowledge and technology have increased during the years, but
additional knowledge will be required if concrete is to maintain the position
it has established as the universal building material. New products are
being developed in all phases of industry at a rapid rate, and improvements
in concrete production, control, delivery, and placing techniques must also
be developed. At the same time, more widespread use of available knowlCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:39:10 EDT 2014
98
FIG. 11--Otto Graf viscosimeter.
edge for controlling uniformity, quality, and workability will improve the
competitive position of concrete.
New methods of mixing, placing, consolidating, and finishing concrete
may permit the use of much less water and improve concrete quality. Eventually it may be possible to place and consolidate 0 to 5.1-cm (0 to 2-in.)
slump concrete in walls, floors, and so on, as readily as 7.6 to 10.2-cm
(3 to 4-in.) slump concrete is placed today. New tests and methods for
quality control and measurement of workability must be developed concurrently with methods of mixing and placing concrete. The slump test
is an old friend and has served its purpose well. We do not like to see old
friends pass away, but change is inevitable. The slump test and other tests
99
used to measure uniformity may be replaced by more efficient test methods

in the future.
Meters have been developed to measure moisture content of sand and
coarse aggregate in the bins at concrete plants and promote better control
of mixing water. Van Alstine [24] attributed the uniformity obtained at
Denver Reservoir No. 22 Dam, where sand bins were filled 17 to 20 times
each day with sand of widely varying moisture content, to the use of an
electrical resistance moisture meter. More general use of devices of this
type will improve quality control of concrete throughout the industry.
Methods of measuring consistency of concrete while it is being mixed in
the mixer drum should be developed so that corrections can be made
immediately. The slump test and other tests used to indicate consistency
can be made only after the concrete is discharged, and corrections can be
applied only to subsequent batches. Polatty [25] reported that such a device,
called the "plastograph," gave a better indication of workability than the
slump test and indicated a moisture change in time to make a correction
in the same batch. The equipment was used at Allatoona Dam with mixes
containing 15.2-cm (6-in.) maximum aggregate and three bags of cement
per cubic metre. When used with 7.6 or 3.8-cm (3 or llA-in.) aggregate,
the equipment was less efficient.
Two-Point Workability Test

Tattersall [26] discusses the principles of measurement of the workability
of fresh concrete and a proposed simple two-point test. In this paper the
author points up that an understanding of workability of fresh concrete is
important: (a) to make possible the design of mixes for particular purposes;
(b) to provide a method of control in the manufacture of the mix; and (c)
to contribute to the efficient use of manufacturing processes such as vibration, pumping extrusion, and the like.
Tattersal contends that in spite of all the efforts (papers that have been
written and the many proposed tests) over the past 30 years, there is no
test that is fully satisfactory and the property of workability cannot even be
defined except in the most general terms. Each of the test methods is
capable of classifying, as identical, concretes that can be shown to be dissimilar.
In this paper the author suggests the following summary of proposed
terminology as an effort toward standardization.
I. Qualitative
A. Workability
1. Flowability
2. Compactability
3. Stability
100
4. Finishability
5. Pumpability
6. Extrudability
II. Quantitative Empirical
A. Slump
B. Compacting Factor
C. V-B Time
III. Quantitative Fundamental
A. Viscosity
B. Yield Value
C. Mobility
The objection to the several workability tests are that they were almost
without exception single-point tests, whereby only one measurement was
made at one specific rate of shear or set of shearing conditions. Tattersall
states that such a procedure is valid only for a simple Newtonian liquid
whose flow properties are completely defined by the constant ratio of stress
to shear rate, and that ordinary observation shows that fresh concrete is
not a Newtonian liquid and it consequently follows that any test based
explicitly or implicitly on the assumption that it is, will be inadequate.
The author states there is evidence to indicate that in practice it may be
sufficient to treat the material as conforming with the Bingham model and
that it approximately obeys the following relationship: the stress at a rate
of shear--the yield value is equal to the product of the plastic viscosity and
the rate of shear. The yield value and plastic viscosity are constants, and it
follows that measurements at two shear rates are required to determine
them.
The balance of this paper [26] discusses test procedures, results, and
conclusions, and further modifications to allow application of vibration.
Tattersall [27,28] further explains the rationale of a two-point workability test and the relationships between slump, compacting factor, Vebe
time, and the two-point test.
The test procedure constitutes the measurement of power required at
three separate speeds to operate a 18.9-1itre (20-qt.) food mixer when
empty; then repeat the power measurement at all 3 speeds when the bowl
contains 21 kg (46.5 lb) of concrete. Values for yield value and plastic
viscosity are obtained by plotting (P--PE)/w against w, where w is speed, P
is power under load, and PE is power when the bowl is empty.
The two-point test appears to yield more information concerning the
performance of a concrete mixture than any of the present test procedures
being used to determine workability. The rapid analysis method of determining water and cement contents of freshly mixed concrete appears to
have good potential for determining and consequently controlling concrete
uniformity. The development of improved testing procedures for freshly
mixed concrete is still a requirement of the concrete industry.
101
References
[1] AC1 Committee 311, ACI Manual of Concrete Inspection, 5th ed., American Concrete
Institute, 1970.
[2] Dunagan, W. M., "A Method of Determining the Constituents of Fresh Concrete,"
Journal, American Concrete Institute, Proceedings, Vol. 26, 1930.
[3] Slater, W. A., "Tests of Concrete Conveyed From a Central Mixing Plant," Proceedings,
American Society for Testing and Materials, Vol. 31, Part II, 1931.
[4] Hollister, S. C., "Tests of Concrete From a Transit Mixer," Proceedings, American Concrete Institute, Vol. 28, 1952, pp. 405-417.
[5] Cook, G. C., "Effect of Time of Haul on Strength and Consistency of Ready-Mixed
Concrete," Journal, American Concrete Institute, Proceedings, Vol. 39, 1952, pp. 413-428.
[6] "Variability of Constituents in Concrete (A Test of Mixer Performance), (Test 26),"
Concrete Manual, 7th ed., U. S. Bureau of Reclamation, 1963.
[7] Bloem, D. L., Gaynor, R. D., and Wilson, I. R., "Testing Uniformity of Large Batches
of Concrete," Proceedings, American Society for Testing and Materials, Vol. 61, 1961.
[8] "Proposed Tentative Method of Test for Cement Content of Freshly Mixed Concrete,"
Bulletin, American Society for Testing and Materials, No. 239, July, 1959, p. 48.
[9] California Department of Transportation, Report No. CA-DOT-TL-5149-1-76-65,
Sept. 1976.
[10] CERL Technical Report M-212, U. S. Army Construction Engineering Research Laboratory, May, 1977.
[11] Glanville, W. R., Collins, A. R., and Mathews, D. D., "The Grading of Aggregate and
Workability of Concrete," Road Research Technical Paper No. 5, H. M. Stationary
Office, London, 1947.
[12] Powers, T. C., "Studies of Workability of Concrete," Journal, American Concrete
Institute, Vol. 28, Feb. 1932, p. 419
[13] Concrete Manual, 7th ed., U. S. Bureau of Reclamation.
[14] Popovics, Sandor, "Relations Between the Change of Water Content and the Consistence of Fresh Concrete," Magazine of Concrete Research, London, July 1962.
[15] Nasser, K. W., Journal, American Concrete Institute, Oct. 1976.
[16] Angles, John, Concrete, Dec. 1974.
[17] Cordon, W. A., "Entrained Air--A Factor in the Design of Concrete Mixes," Journal,
American Concrete Institute, Vol. 42, 1946.
[18] Kelly, J. W. and Milos, Polivka, "Ball Test for Field Control of Concrete Consistency,"
Journal, American Concrete Institute, Proceedings, Vol. 51, May 1955, p. 881.
[19] Howard, E. L. and Leavitt, G., "Kelly Ball versus Slump Cone," Journal, American
Concrete Institute, Proceedings, Vol. 48, 1952, pp. 353-354.
[20] Thaulow, Sven, "Field Testing of Concrete," Horak Cementforening, Oslo, 1952.
[21] Bahrner, V., "Report on Consistency Tests on Concrete Made by Means of the Vebe
Consistometer," Report No. 1, Joint Research Group on Vibration of Concrete, Svenska
Cemenfforeningen, March 1940.
[22] Orchard, D. F., Concrete Technology, Vol. 2, Wiley New York, 1962.
[23] Davis, R. E., Jansen, E. C., and Neelands, W. T., "Restoration of Barker Dam," Journal,
American Concrete Institute, Proceedings, Vol. 44, April 1948.
[24] Van Alstine, C. B., "Water Control by Use of a Moisture Meter," Journal, American
Concrete Institute, Proceedings, Nov. 1955, pp. 341-347.
[25] Polatty, J. M., "New Type of Consistency Meter Tested at Allatoona Dam," Journal,
American Concrete Institute, Proceedings, Oct. 1949, pp. 129-136.
[26] Tattersall, G. H., "Principles of Measurement of the Workability of Fresh Concrete and
a Proposed Simple Two-Point Test," Fresh Concrete, Rilem Seminar Proceedings, 1973,
Vol. 1.
[27] Tattersall, G. H., "The Rationale of a Two-Point Workability Test," Magazine of Concrete Research, Vol. 25, No. 84, Sept. 1973.
[28] Tattersall, G. H., "The Relationships Between the British Standard Tests for Workability and the Two-Point Test," Magazine of Concrete Research, Vol. 28, No. 96,
Sept. 1976.
R . F. A d a m s ~
Chapter 8--Making and Curing

Concrete Specimens
Introduction
This chapter covers in general the importance of properly making and

curing concrete specimens. In this discussion, the word "making" means
paying attention to all the details of mixing the concrete (in the case of
laboratory-mixed concrete), specimen molds, placing, consolidation, and
finishing the specimens of any kind for either field or laboratory-mixed
concrete. "Curing" pertains to temperature, moisture, and time, which are
the factors of the curing process of concrete governing the rate of strength
gain and many other properties of concrete for which tests are made.
By far the number of specimens are made for strength tests. However,
specimens are made for many other purposes. ASTM Making and Curing
Concrete Test Specimens in the Field (C 31) provides detailed procedures
for making and curing concrete specimens in the field, that is, on construction jobs. ASTM Making and Curing Concrete Test Specimens in the
Laboratory (C 192) provides detailed procedures for making and curing
concrete specimens in the laboratory. These are reference procedures and
other ASTM standards refer to C 31 and C 192 either for a portion or all
of the procedures for making and curing the required specimens. Other
test procedures may supply additional details or make exceptions to ASTM
Methods C 31 and C 192 to accommodate them to the specific specimens
or tests to be made.
It is most important that the tests and specimens be made by trained
personnel and that the details of the various methods be followed precisely.
Only in this way can meaningful and reproducible test results which are
not open to question be obtained. The significant differences resulting
from having strength tests made by trained and untrained personnel (contractor's personnel in this case) has been pointed up by Wagner.2 Wagner's
study showed that strength tests of the same concrete made by trained
1Consulting concrete engineer, Sacramento, Calif. 95822.
2Wagner, W. K., Materials Research and Standards, Vol. 3, No. 8, Aug. 1963, p. 629.
102
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
lntcrnational
www.astm.org
ADAMS ON MAKING AND CURING CONCRETE SPECIMENS
103
personnel were higher and more uniform than those made by untrained
personnel.
Engineers and technicians responsible for supervising and making these
tests must be thoroughly familiar with the test procedures and should be
trained in making the tests strictly in accordance with the detailed procedures. Also, they should be supplied with the required tools and equipment. Personnel should periodically review the test procedures in order to
refresh their memory of the details. Also supervisors should periodically
review the work of those making the tests to determine whether it is being
done satisfactorily. When job strength tests fail to meet specification requirements and the rejection of concrete is considered, the tests are almost
always questioned, particularly if it can be shown that the person making
the test had not complied with the test procedures. The competency of the
testing and personnel becomes the subject of controversy, particularly if
litigation results. Training has been provided to job personnel not responsible for making tests so that they can observe the tests to determine
whether proper procedures are being followed. Deviations are noted so that
a defense is available in case of controversy.
ASTM Method 31 provides strength specimens for determining: (a) the
adequacy of designated mix proportions for producing the required job
strength, (b) the basis for acceptance of job concrete, (c) quality control
purposes, and (d) the attainment of sufficient strength of the concrete in
the structure so that forms or shoring may be removed or the structure may
be put in service. This method provides for two curing procedures: (a) job
curing where the specimens are stored at the job under temperature and
moisture conditions as near like those of the structure as possible or (b)
standard conditions where the specimens are taken to the laboratory and
moist cured at 21 + 1.7~ (73.4 + 3~ after the first day in the field.
One or both of these options for curing must be selected, depending upon
the purpose for making the tests. The importance of making the proper
choice between field and standard curing can be appreciated when the
considerable strength differences which result from curing cylinders under
standard conditions or on the job when the ambient temperature is near
freezing are recognized. Likewise, the importance of storing cylinders to be
job cured so they will be as near like the structure with respect to temperature and moisture conditions must be appreciated, and a conscientious
effort must be made to achieve this condition when the curing option has
been selected.
ASTM Method C 192 provides detailed procedures for mixing concrete
and making specimens for a wide variety of purposes such as: (a) job trial
mixtures, (b) evaluation of different mixtures or materials, or (c) providing
specimens for a variety of research purposes.
Two recent additions to ASTM standards should be mentioned and discussed briefly.
104
ASTM Method of Test for Compressive Strength of Concrete Cylinders

Cast in Place in Cylindrical Molds. (C 873) provides for making and testing
concrete cylinders which are molded in place in the concrete member in
special plastic molds which can be easily removed when the desired age for
strength tests is reached. These are more commonly referred to as cast on
place pushout cylinders (CIPPOC). Such specimens are more readily used in
slab placements, not in columns or walls. These cylinders are surrounded by
the concrete in the structure until they are removed for testing and are
subjected to nearly the same temperature and moisture environment as the
structural member.
ASTM Method of Making, Accelerated Curing, and Testing of Concrete
Compression Test Specimens (C 684) provides for accelerated curing of
strength specimens by three different procedures so that strength test results
are obtained at either one or two days age. All three procedures utilize curing
at elevated temperatures. These strength results are not the same as 28-day
strength results. However, they do produce results which are reproducible
and correlate well with the 28-day strength. 3-5 Therefore these accelerated
tests can be reliably used to judge the quality of the concrete or to predict the
28-day strength, provided correlation of the two procedures have been made
using the same materials and mixtures. The value of having meaningful
strength results at one or two days is obvious. It is much easier to correct
conditions causing low strength concrete when this is known at an early age.
3Malhotra, V. M. and Berwanger, C., Proceedings, American Concrete Institute, Vol. 68,
1971, p. 963.
4Malhoti:a, V. M. and Zoldners, N. G., Proceedings, American Concrete Institute, Vol.
66, 1969, p. 894.
SSmith, P. and Chojnacki, B., Proceedings, American Societyfor Testing and Materials,
Vol. 63, 1963, p. 1079.
J. H. Sprouse and R. B. Peppler ~
Chapter 9--Setting Time
Introduction
Scripture's paper on Setting Time [1] 2 published in 1955, summarized
the problems involved and the progress made, as of that time, in developing
methods for measuring the setting time of concrete. He pointed out that
the setting time of cement does not adequately define the setting time of
concrete and warned that very misleading conclusions may result from assuming such to be the case. Many criteria for defining setting time of
concrete had been suggested, including loss of workability before placing,
impairment of bond of freshly placed concrete to concrete in place, time of
finishing for floors, ability to strip forms, and early strength gain. Scripture
was of the opinion that none of these criteria represents the setting time
of concrete, stating that as soon as concrete has any appreciable strength it
has certainly set, and conversely, as long as concrete maintains any significant degree of workability, it is not set. He concluded that one or more
points between these two extremes represent the setting time of concrete
and, like the setting time of cement, will necessarily have to be defined by
a particular test method and apparatus. Subsequent investigations of the
rate of hardening of concrete have tended to substantiate these conclusions.
Methods of Test
The various methods of test for determining the setting time of concrete
which had been proposed or investigated prior to 1955 were described by
Scripture, and their merits were discussed as summarized below.
Electrical Measurements
The electrical measurement method [2,3] consisted of measuring the
change in electrical resistance of concrete mixtures with time. For a cont Director, Engineering, and director, TechnicalDevelopment,respectively,Master Builders,
Cleveland, Ohio 44118. Originallywritten for STP 169A by T. M. Kelly.
2The italic numbers in brackets refer to the list of references appended to this paper.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
106
crete mixture without any additive, the electrical resistance was found to
increase fairly steadily until it reached a point at which the curve leveled
off. A similar concrete mixture containing calcium chloride, however, had
an almost constant and very low electrical resistance throughout the period
of test. Therefore, it appeared that such a test method would be very limited
in its application and would be of little use where it is desired to show the
effects of chemical admixtures on the setting time of concrete.
Consistency Measurements
Apparatus suggested for use in relating change in consistencywith setting
time were the Vebe apparatus, modified Gillmore and Vicat needles,
penetration rods, and the Kelly ball. The penetration methods, in general,
were considered to be unsatisfactory because of interference of large pieces
of aggregate. Penetration tests, if applied to mortar sieved through a No. 8
sieve, gave more reproducible results than those obtained on neat cement
paste, but doubt was expressed that the results really represented the time
of setting of concrete as a whole in view of the discrepancy between the
setting times of neat cement and mortar pats.
Troweling Method
A method based on determining the point at which it is just possible to
finish the surface of the concrete by steel troweling was described as giving
reasonably reproducible results, though somewhat subject to personal
factors. It was considered questionable whether this method measures the
setting time of concrete for those interested in anything except floor finishing.
Velocity and Frequency Measurements

The soniscope method [4], based on measurement of wave velocity,
was considered to have application as a research tool but was cited as
being too high in cost and requiring too delicate a technique for consideration as a standard test method. Results obtained using conventional sonic
apparatus to measure fundamental frequencies and employing a mix confined in ,a metal tube were considered sufficiently promising to merit
further investigation.
Bleeding Characteristics
This method had been suggested but no data were available on its application to the determination of setting times. Because of a wide variation
SPROUSE AND PEPPLER ON SETTING TIME
107
in bleeding characteristics of different concrete mixtures, the proposed

method was not considered promising.
Heat of Hydration
Though curves obtained by this method showed sharp changes where the
setting time of cement would be expected, its applicability in determining
the setting time of concrete was doubted on the basis that the method
measures a chemical property of cement and that the shape of the curve
could be completely changed by modifying the gypsum content of the
cement without actually changing the setting time of the cement or concrete.
Changes in Volume
Determination of the setting time of concrete from measurement of
changes in volume of the setting concrete in a dilatometer was considered
to be open to the same objections as those stated in connection with heat of
hydration determinations.
Strength Determinations
One method based on strength determinations consisted of casting
specimens in flexible containers and determining deflection. Some work
had also been done on both compressive and flexural strength determinations, but sufficient reliable data were not available to demonstrate the
suitability of any of these methods for measuring setting time. From limited
results obtained by a tension test method it appeared that this method
showed some promise, since a fairly sharp break in the curve appeared to
occur at a definite time.
Deformability Changes
The pressure-type air meter had been employed to determine changes in
deformability of air-entrained concrete. Curves in which percentage of air
was plotted against time after mixing showed very definite points of inflection when the concrete became sufficiently rigid to resist deformation
by the air pressure commonly used in this type of meter. Such a method
would, of course, be applicable only to air-entrained concrete.
Scripture tentatively concluded that probably more than one definition
of setting time would be needed, since setting time of concrete means
something different to each individual, depending on the specific application and that setting time would probably be defined in terms of a maximum
108
change in some physical rather than chemical property of the concrete,

determined by a particular test method. Therefore, any definition would
necessarily await the establishment of a suitable test method.
Penetration Resistance Method
Substantial progress has been made in measuring the rate of hardening

of concrete subsequent to publication of the 1955 paper on Setting Time.
Tuthill and Cordon [5] developed a method based on the use of Proctor
penetration needles to determine the hardening characteristics of concrete
mortar sieved from concrete. Their work and application of their method
by others indicates that correlation can be established between the hardening of concrete and of the mortar sieved from it. The authors state, "Results of tests indicate that this equipment provides an accurate, rapid,
and economical method of determining hardening characteristics of concrete mortar which, though not equal to those of corresponding concrete,
are of similar character and reliably indicative of what may be expected of
the concrete." In their test procedure the mortar was sieved from the concrete and placed in a container at least 15.2 cm (6 in.) deep, the container being large enough to permit a minimum of ten undisturbed readings of penetration resistance. After the mortar had been vibrated in the
container, it was covered and placed in a room maintained at the desired
temperature of test. Bleeding water was poured off before making a penetration test. Penetration resistance readings were taken at such intervals as
necessary to define the hardening characteristics by means of penetration
needles of appropriate end areas. The first reading was taken with a needle
having a 6.45-cm 2 (1-in. 2) bearing area, and subsequent readings were
taken with this needle until penetration resistance reached or approached
the maximum capacity of the spring. Smaller needles were used in subsequent readings, the size being dictated by the penetration resistance of
the mortar and the maximum capacity of the spring. In making penetration resistance readings the pressure on the needle was applied steadily and
gradually until the needle had penetrated the mortar to a depth of 2.54 cm
(1 in.). In conjunction with penetration resistance tests, a 14 dm 3 (Ih-ft3)
container was filled with concrete for observation of its hardening progress
and its response to vibration during the period of penetration tests of its
mortar, and a concrete test cylinder was made from the same concrete
for determination of compressive strength of the concrete when the penetration resistance of its mortar reached 27.6 MPa (4000 psi). A second
concrete cylinder was made for test at 48 h. The authors stated that these
tests were made to provide further information on the early hardening
characteristics and to give confidence that the retarded concrete would
proceed to develop proper strength.
Tuthill and Cordon defined the vibration limit of concrete as that point
109
during hardening of concrete when it no longer can again be made plastic

by revibration. That is, it is a point beyond which a delayed second layer
of concrete will not become monolithic with its preceding layer. They
stated that a running vibrator will not sink of its own weight into concrete that has passed the vibration limit and that, as measured by resistance
of the concrete mortar to the penetration test, it has been demonstrated
that the vibration limit will be reached at a resistance of approximately
3.4 MPa (500 psi). They indicated that a mortar was considered completely
hardened when its penetration resistance reached 27.6 MPa (4000 psi), at
which time the same concrete from which the penetration resistance mortar
specimen had been made had developed a compressive strength, determined on a 15.2 by 30.5-cm (6 by 12-in.) cylinder of approximately 0.7
MPa (100 psi).
Subcommittee C09.03.14 on Methods of Testing of Setting Time of
Concrete, ASTM Committee C-9 on Concrete and Concrete Aggregates,
subsequently proposed a test method based on the procedure developed
by Tuthill and Cordon. This method has undergone many revisions since it
was first adopted and is presently entitled, ASTM Test for Time of Setting
of Concrete Mixtures by Penetration Resistance (C 403). In scope this
specification provides a procedure for determining the time of setting of
concrete with a slump greater than zero by testing mortar sieved from the
concrete mixture. The definitions of times of setting are admittedly arbitrary,
since the hardening of concrete is a gradual process. Under this method
times of initial and final setting of concrete are determined on the basis
of a rate of hardening test made by means of penetration resistance
needles on mortars sieved from the concrete mixture. The times of setting
are determined from rate of hardening curves obtained from a linear pioL
of rate of hardening data, with elapsed time as the abscissa and penetration
resistance as the ordinate. This specification defines setting times as follows: (a) time of initial setting--the elapsed time, after initial contact of
cement and water, required for the mortar sieved from the concrete to
reach a penetration resistance of 3.4 MPa (500 psi) and (b) time of final
setting--the elapsed time, after initial contact of cement and water, required for a mortar sieved from the concrete to reach a penetration resistance of 27.6 MPa (4000 psi).
It is apparent that the above definitions of times of setting are based on
Tuthill's and Cordon's vibration limit and the penetration resistance at
which they considered the mortar to be hardened. Users of the test method
are not in complete agreement that 3.4 MPa (500 psi) is, in fact, the vibration limit or that mortar should be considered to be completely hardened
at 27.6 MPa (4000 psi) penetration resistance. Such agreement, however, is
not essential. These arbitrarily defined points serve as convenient reference
points for determining the relative rates of hardening of mortars from
different concretes during both the early and later stages of hardening and
110
TESTS
AND
PROPERTIES
OF
CONCRETE
for determining the effects of variables such as temperature, type and

brand of cement, cement factor, mixture proportions, slump, additions,
and admixtures upon the time of setting and hardening characteristics of
the mortar. Where the test method is being used as a part of performance
specifications, these arbitrarily defined points also make it simpler to
specify time of setting requirements. It should be kept in mind, however,
that the entire body of data from a rate of hardening test provides much information of value which may not be apparent if only the initial and final
setting times are reported. Changes in the rate of hardening during the
hardening period can be of considerable interest and value in scheduling
concrete placing and finishing operations.
Typical rate of hardening curves obtained by the penetration method are
shown in Figs. 1 to 4. The rate of hardening tests were made in accordance
with ASTM Test C 403 except that each curve represents a single rate of
hardening test, rather than an average of three or more. Figure 1 demonstrates the effect of a low-storage temperature on rate of hardening of
mortars sieved from two concrete mixtures. The concrete mixtures were
both batched and mixed at approximately 23~ (73~ and were similar
in all respects, except that after mixing and sieving to provide mortar for
penetration tests, mortar specimens from the two mixtures were stored at
23 and 13~ (73 and 55~
respectively. Initial and final setting times of
the mortar were retarded approximately 4 and 7 h, respectively, by the
lower storage temperature. Retardation of initial set probably would have
been greater if the mortar specimens for low-temperature storage had been
at 10~ (50~ at the start of the test. Figure 2 demonstrates the effects of
a retarding admixture at two storage temperatures, namely, 23 and 32~
(73 and 90~
Temperature of all of the mixtures after mixing was ap-
34.5
0
n
:E
27.6
dr
I/
i~
20.7
Q;
n-
13.8
!
Cs
C:
,0
m
4,-
,o
6.9
eQ;
rl
---1--- F-- V- - -
0
0
I0
12
14
16
Time, h
FIG. 1 - - E f f e c t o f low t e m p e r a t u r e on setting time. 1 M P a = 145 psi, ~
= (~
-- 3 2 ) / 1 . 8 .
SPROUSE
AND
P E P P L E R ON S E T T I N G
TIME
111
proximately 24~ (75~

The concrete mixtures were similar in all respects except that concretes containing the admixture, which was added at
the rate of 0.266 weight percent of cement, were reproportioned to maintain a constant cement factor. After sieving mortar from the concrete
mixtures, companion mortar specimens from the two types of mixtures
were stored at 23 and 32~ (73 and 90~
respectively. With this particular cement and admixture, and mortar temperatures at 24~ (75~ at
the start of test, the retarding effect of the admixture was greater at the
higher storage temperature. In Fig. 3 the rate of hardening of a mortar
sieved from a concrete mixture is compared with the rate of hardening of a
separately mixed cement:sand mortar. Cement of the same brand and
o 34.5
/32oc/
3 ; 27.6
~
u
Cement
20.7
-(/)
13.8
6.e
FINAL SET.=
//;
"
E
tie"
23 ~
// /'
~ 3.5
........
J~k~A_ _SgZ~
e
Time,
o---wlthretarder
1o
~z
~4
~6
FIG.
~
2--Effect of temperature and retarder on rate of hardening. 1 MPa :

= ( ~ --32)/1.8.
145 psi,
/l
34.5
FINAL SET,~I
27.6
9~
Cement
t3.8
g
~
6.9
~.
~
0
10
12
t4
le
Time, h
F I G . 3--Rate of hardening of screened versus mixed mortar. 1 MPa = 145 psi.
112
type, and sand from the same lot were used in both the concrete and the
mortar mixtures, and the water cement (w/c) ratio and cement:sand ratio
were the same for both mixtures. Tests on the two mortars were conducted
simultaneously. Initial and final setting times were approximately 1/2 and 1 h
greater, respectively, for the separately mixed mortar. Figure 4 illustrates
differences in the rate of hardening of concretes made with two different
brands of Type I portland cement. The concretes contained sand and coarse
aggregates from the same sources, and ambient temperatures were the
same during the mixing and testing period. Though the concrete made
with Cement B had a higher cement factor than that made with Cement A,
it reached final set approximately 1 h later.
Bond Pullout Pin M e t h o d
In 1957 a paper by Kelly and Byrant [6] was published describing a
bond pullout pin method for measuring the rate of hardening of concrete.
In this test procedure the whole concrete was utilized, rather than the
mortar portion, and the method was based on measurement of the rate of
development of concrete bond strength. Stainless steel pins, 9.5 m m (3/8 in.)
in diameter, were embedded vertically in concrete to a depth of 12.7 cm
(5 in.) by vibrating the concrete around them in a beam mold immediately
after the concrete was mixed. The pins were held firmly in a vertical position
by a special jig during specimen fabrication. Individual pins were subsequently pulled out of the concrete at increasing time intervals by means
of a spring scale and the load at bond failure measured. Rate of hardening curves were obtained by plotting bond stress against time on a linear
scale.
0
tl.
,~
34.5
A,Type
C.F.= 280 Kg/m a
Cement
w 20,7
.w
f~
Q
n."
t- 13.8
0
.e-.
~
6.9
3.5
2r.6
U
c
0
~_ I
!t
Cement B, Type !
C.F.= 307 Kg/m3
Tests at 23"C
FIN!L SlIT !
8
10
Time, h
FIG.
|Z
14
16
4 - - E f f e c t o f b r a n d o f c e m e n t on rate o f hardening. 1 M P a -~ 145 psi, ~

32)/1.8, 1 k g / m a = 1.686 l b / y d . 3
= (~
--
113
The authors stated that in the hardening process no sharp breaking

point occurs which can be interpreted as time of setting. Instead, concrete hardens at a gradually increasing rate until it reaches a definitely
hard or set condition. On the basis of the observed relationship between
time and the degree of hardening of concrete, they were of the opinion that
a test method to measure setting time p e r s e is not feasible. They believed,
however, that the rate of hardening of concrete under any given set of conditions can be determined and that setting times can be arbitrarily defined
on the basis of such rate of hardening determinations. Subsequently, a
proposed tentative test method was drafted for consideration by ASTM
Subcommittee C09.03.14 in which the test procedure was essentially the
same as that described in the above paper, except that the depth of pin
embedment was increased to 15.2 cm (6 in.) to permit the use of a conventional 15.2 cm (6 in.) beam mold. This method is under consideration
by ASTM Subcommittee C09.03.t4 as an alternative method for determining time of setting of concrete.
A favorable feature of the bond pullout pin test method is the fact that
the rate of hardening determinations are made on the whole concrete,
rather than on the mortar sieved therefrom. Therefore, the results can be
interpreted without concern as to the possible effects of removal of coarse
aggregate or of manipulation and exposure of the sample during the process of sieving the concrete to obtain mortar for the penetration tests.
On the other hand, modification and refinement of the method are
desirable. For example, the relation between the embedded surface area of
a pin and combined weight of beam and specimen are such that the maxim u m load is limited unless the specimen is physically restrained from being lifted during the later stages of hardening. Preliminary studies indicate
that higher bond stresses may be reached by reducing pin diameter or
reducing depth of embedment or both.
Typical rate of hardening curves obtained by the bond pullout pin
method are shown in Figs. 5 to 7. All of the concrete mixtures for which
rate of hardening curves are shown were similar in all respects, expect for
differences in brand of cement used or the addition of a chemical admixture. The temperature of the mixing room was maintained at approximately 23~ (73~ and all concrete mixtures were at approximately
the same temperature upon completion of mixing. Immediately after their
fabrication, the rate of hardening specimens were stored for the duration
of the tests at the temperatures indicated on the charts. Figure 5 illustrates
the effect of different storage temperatures on rate of hardening. Figure 6
illustrates the differences in rate of hardening that can result from a
change in the brand of Type I cement being used in a given concrete
mixture. Figure 7 shows the rate of hardening at various storage temperatures of a reference mixture and of a mixture containing a water reducing
admixture. The mixtures differed only in that the cement factors of the
114
7.0
5.6
132~23"C
CementC,Type
i /
/
v"
jr
4.2
O*C
2.8
qo
o
rn
1,4
0
0
Time,h
10
12
14
16
FIG. S - - E f f e c t o f temperature on rate o f hardening. 1 k g f / m m 2 = 67. 6 psi, ~

32 )/1.8.
= (~
--
6 - - E f f e c t o f b r a n d o f c e m e n t on rate o f hardening. 1 k g f / m m 2 = 67.6 psi, ~

-- 32)/1.8.
7.0
E
i 4.2 CementC~
CementD
Sr~y/
,~.
L.
c
o
m
5.6
/4-~ementE
/
TiDeI Cements
Temperature
25~
Z.8
1.4
Time,h
FIG.
(~
10
12
14
16
reference and admixture-containing concretes were 307 and 262 k g / m 3

per yard, respectively. Rates of hardening of the two types of concrete were
very similar at 21 and 32~ (70 and 90~ and during the early stages of
hardening at 10~ (50~
In the later stages of hardening at this temperature the water-reducing admixture accelerated the hardening rate. The
acceleration started at about 10 h and appeared to increase at a uniform
rate, reaching a maximum of about 1.5 h when tests were discontinued.
7.0
NE
32"C ~
iv'
//
4.2
Cement F, Type 1
21"C
'/
IO*C , ' /
tD
"o
n'l
115
1.4
J
#
~
0
0
~
4
I
6
O---NO ADMIXTURE
O'--W1THIADMIXTUREI
10
12
14
16
Time, h
FIG. 7 - - E f f e c t o f water reducer on rate o f hardening. 1 k g f / m m 2 = 6 Z 6 psi. ~

-- 32)/1.8.
= (~
Effect of Variables
The results of setting time tests are influenced by several variables.

Variations in temperature have appreciable effects on setting times as
illustrated in Figs. 1 and 5. In general, the change in setting time per degree change in temperature increases with decreasing temperature. Setting
times of concrete vary with type and brand of cement. As shown in Figs.
4 and 6, differences in setting times resulting from a change in brand of
Type I cement can be as great as those resulting from a substantial variation in ambient temperature. Tuthill et al [7] concluded that the time
to the vibration limit and the time rate of setting may be influenced more
by the brand and possibly the type of cement than by differences in the
variety of water-reducing retarders investigated. Vollick [8], in Fig. 7 of
his paper on water-reducing, set-retarding admixtures, illustrates substantial differences in times of setting of concretes that are similar in all
respects except for type of cement. His work indicates that Type III and
Type II cements accelerate and retard, respectively, as compared with
Type I cement. Mixture proportions, cement factor, slump and, of course,
additions and admixtures all influence setting time to some degree. In
general, setting times are accelerated by an increase in cement factor and
retarded by an increase in slump. In regard to admixtures, recent studies
have indicated that time of addition of set-retarding admixtures may have
an effect on time of setting of the concrete, the retardation usually increasing with delay in introduction of the admixture [ 9 , 1 0 ] . However, the
effect varies with the nature of the cement and the admixture.
116
Reporting Results
In a previous publication Polivka and Klein [11] and Loughborough [12]
plotted rate of hardening test data on a semilogarithmic scale, demonstrating
that the resulting curves are approximately straight lines. The ASTM Subcommittee C09.03.14 on Setting Time of Concrete is presently considering
the desirability of specifying in ASTM Test C 403 that rate of hardening test
data be plotted on a semilogarithmic scale. In well-controlled tests, data points
so plotted generally do not deviate appreciably from a straight line, whereas a linear plot yields a curved line. Some, who prefer the semilogarithmic
scale to the linear scale, maintain that an average straight line can be drawn
more accurately than an average curve through the plotted points derived
from the required three separate rate of hardening tests. They also point
out that the slope of the curve obtained by a linear plot is undergoing the
most rapid rate of change in the critical area of initial set, and unless
an adequate number of determinations are made at this stage of hardening
the estimation of initial setting time may be inaccurate. Others, who prefer the linear plot, feel that it provides a more meaningful picture of the
manner in which the concrete is hardening. They also feel that, since
the data points on a semilogarithmic plot seldom form a perfectly straight
line, the position of the line is a matter of judgment on the part of the
operator, while with a linear plot a smooth curve can usually be fitted
fairly closely to all data points. From data available, it appears that the
setting times determined by either method of plotting are essentially the
same. Figure 8 is a semilogarithmic plot of the penetration data plotted
linearly in Fig. 1. Figure 9 is a semilogarithmie plot of the bond pullout
data plotted linearly in Fig. 5. Data from both test methods yield approximately straight lines when the penetration resistance and bond stress are
plotted on the logarithmic scale.
Repeatability
The repeatability of rate of hardening determinations made an accordance with ASTM Test C 403 has not been positively established. However, in 1973, ASTM Subcommittee C09.03.14 voted to approve a proposed revision to the precision statement of ASTM Test C 403. The following
statement is based on data received from five laboratories which had
participated in a round-robin test program:
1. The single-operator coefficient of variation for time of initial set has
been found to be 7.1 percent. 3 Therefore, the range of results obtained
3These numbers represent, respectively, the (1S percent) limit as described in ASTM
Recommended Practice for Preparing Precision Statements for Test Methods for Construction Materials (C 670), and the value for the range of a set of three values which should
be exeeeded not more than 5 percent of the time.
--t
o)
F
m
m
m
-o
,-Q
>
z
C~
"(3
3O
~1
.9
co
r-
ro
r
11>
~.
Penetration Resistance, MPa
r~
\
-'t
e:
"I
:r
co
Bond
Strength,
Kgf/mm 2
.~
e,
o~
L~
-I
=7'
co
o~
.p.
r~
oo
(I
118
on three separate batches by the same operator with the same apparatus,
using similar materials, on three different days should not exceed 23 percent of their average (see footnote 3).
2. The multilaboratory coefficient of variation for time of initial set,
when results are based on the average of three tests, has been found to
be 5.2 percent. 4 Therefore, results of two different laboratories on similar
materials should not differ by more than 15 percent of their average (see
footnote 4).
3. The single-operator coefficient of variation for time of final set has
been found to be 4.7 percent (see footnote 3). Therefore, the range of
results obtained on three separate batches by the same operator with the
same apparatus, using similar materials, on three different days should
not exceed 16 percent of their average (see footnote 3).
4. The multilaboratory coefficient of variation for time of final set,
when results are based on the average of three tests, has been found to
be 4.5 percent (see footnote 4). Therefore, results of two different laboratories
on similar materials should not differ by more than 13 percent of their
average (see footnote 4).
It is anticipated that the repeatability of ASTM Test C 403 will be improved by further refinement of testing procedures as additional experience
is gained in use of the method. For example, there are indications that
penetration resistance values may differ substantially depending upon the
size of needle being used, in which case it may be desirable to specify the
sizes of the needles to be used over various ranges of penetration resistance.
The influence of temperature variation from center to edge of specimen
and the effect of boundary restraint also need further elucidation.
Application of Setting Time Tests
In Acceptance Specifications
A suitable method for measuring the rate of hardening and determining
the setting time of concrete was essential to the development of ASTM
Specification for Chemical Admixtures for Concrete (C 494). This specification set tolerance limits for effect on setting time of five types of chemical
admixtures as determined by the penetration resistance method. The
tolerances were established on the basis that the temperature of each of
the ingredients of the concrete mixtures, just prior to mixing, and the
temperature at which the time of setting specimens are stored during the
test period shall be 23 + 1.7~ (73 + 3~
A retarding admixture must
4These numbers represent, respectively, the (1S percent) and (D2S percent) limits as
described in ASTM RecommendedPractice C 670.
119
reach initial set at least 1 h later but not more than 3 h later than the
reference concrete and must reach final set not more than 3 h later than
the reference concrete. An accelerating admixture must reach initial set
at least 1 h sooner but not more than 3 h sooner than the reference concrete and must reach final set at least 1 h sooner than the reference concrete. An admixture that is required neither to accelerate nor retard must
reach both initial and final set not more than 1 h sooner or 1 h later than
the reference concrete. The purpose of ASTM Specification C 494 is to
provide methods and criteria for evaluating chemical admixtures under
highly standardized testing conditions in the laboratory that are not intended to simulate actual job conditions.
In Field Concreting
Job conditions frequently require a greater degree of and closer control of retardation or acceleration than is required under ASTM Specification C 494. The more stringent requirements may result from unusual
temperature or exposure conditions at time of concrete placement; structural
considerations, such as avoiding cracking of the concrete deck when steel
bridge girders deflect during placement of the deck; avoidance of cold
joints in mass concrete; or special types of concreting operations, such as
slip form construction. By prior studies in the laboratory under simulated
job conditions, utilizing the concreting materials to be used in the work,
time of setting tests are invaluable in planning concreting operations and
determining whether special measures will be required for set control.
If control is required, the type and dosage of admixture can be tentatively
selected, subject to verification and adjustment based on time of setting
tests made on the job during actual construction.
Precise correlations between time of setting test results and slab finishing
operations have not been reported in any published work. It seems reasonable to believe, however, that setting time tests made on slab concrete at
the jobsite would provide guidance in scheduling finishing operations.
Both the penetration resistance method and the bond pullout pin method
have been used advantageously in the field. Because it is a tentative standard
test method and utilizes less cumbersome apparatus, the former method
has been used more extensively. However, an engineer for one construction
agency s utilizes the latter method stating: "Although the pin pullout apparatus is quite bulky and awkward for general field use, it is advantageous
in that the rate of hardening test is made on actual job concrete." In this
instance the method was being used in the field on three large highway
5Peter Smith, Sr., materials engineer, Ontario Department of Highways, private communication.
120
projects to determine the influence of different retarders on rate of hardening of pavement concrete under field conditions.
Determinations of setting times made on small specimens may not be
alrectly applicable to the same concrete in a structure. It has been noted
by one construction engineer (see footnote 5) that concrete in the structure
sets up at a slightly faster rate than concrete in beam specimens used for
the bond pin pullout test, probably due to the differences in the concrete
volume and resultant temperature rise.
Wallace and Ore [13] reported that tests performed on samples of concrete at the batch plant do not reflect the influence of the increasing temperature occurring within massive blocks in the dam. They conducted
tests in which mortar samples for penetration resistance tests, sieved from
the mass concrete at a temperature of about 11 ~ (52 ~
were immediately
taken into each of three rooms wherein the air temperature was maintained
at 10, 24, and 38~ (50, 75, and 100~ while rate of hardening tests were
performed. The influence of predictable temperature cycles for various job
conditions on the rate of hardening of mass concrete could be estimated
from the data thus obtained.
Difference in conditions of exposure between the time of setting specimens and the concrete in a structure can lead to misapplication of time
of setting test results. This is particularly true in the case of flat slab work
where the time of setting of the slab may be strongly influenced by wind
velocity, effect of the subgrade, and the radiant heat of the sun as pointed
out by Schutz [14].
Summary
The development of test methods for measuring the rate of hardening
of concrete, on the basis of which setting times can be defined and determined, has been an important contribution to concrete technology. Factors
that affect the setting time of concrete can now be evaluated with reasonable accuracy and concreting operations planned or adjusted accordingly.
By means of time of setting tests the need for set-controlling admixtures
can be established with certainty and type and dosage can be determined
on a quantitative basis. In interpreting results of setting time tests for
application to an actual concreting operation, however, one should give
due consideration to disparities between the mortar or concrete in the test
specimens and the concrete in the structure, such as relative surface to
volume ratios, exposure to wind and sun, or exchange of heat with the
subgrade.
121
References
[1] Scripture, E. W., Jr., "Setting Time," Significance of Tests and Properties of Coner#te
and Concrete Aggregates, A S T M STP 169, American Society for Testing and Materials,
1955, p. 53.
[2] Calleja, J., "New Techniques in the Study of Setting and Hardening of Hydraulic
Materials," Journal, American Concrete Institute, March 1952; Proceedings, Vol. aS,
p. 525.
[3] Calleja, J., "Effect of Current Frequency on Measurement of Electrical Resistance of
Cement Pastes," Journal, American Concrete Institute, Dec. 1952; Proceedings, V,ol.
49, p. 329.
[4] Whitehurst, E. A., "Use of the Soniscope for Measuring Setting Time of Concrete,"
Proceedings, American Society for Testing and Materials, Vol. 51, 1951, p. 1166; discussion, p. 1176.
[5] Tuthill, L. H. and Cordon, W. A., "Properties and Uses of Initially Retarded Concrete,"
Journal, American Concrete Institute, Nov. 1955; Proceedings, Vol. 52, p. 273; discussion, Journal, American Concrete Institute, Dec. 1956, Part 2; Proceedings, Vol. -~2,
p. 1187.
[6] Kelly, T. M. and Bryant, D. E., "Measuring the Rate of Hardening of Concrete by
Bond Pullout Pins," Proceedings, American Society for Testing and Materials, Vol. 67.
[7] Tuthill, L. H., Adams, R. F., and Heroine, J. M., Jr., "Observations in Testing and
Use of Water-Reducing Retarders," Effect of Water-Reducing and Set-Retarding ~ldmixtures on Properties of Concrete, A S T M STP 266, American Society for Testing ~nd
Materials, 1960.
[8] Vollick, C. A., "Effect of Water-Reducing Admixtures and Set-Retarding Admixtu~res
on the Properties of Plastic Concrete," Effect of Water-Reducing and Set-Retarding
Admixtures on Properties of Concrete, A S T M STP 266, American Society for Testing
[9] Bruere, G. M., "Importance of Mixing Sequence When Using Set-Retarding AgQnts
with Portland Cements," Nature, Vol. 199, 1963, p. 32.
[10] Farkas, E. and Dodson, V. H., "Delayed Addition of Set-Retarding Admixtures to
Portland Cement Concrete," Proceedings, American Society for Testing and Materials,
Vol. 64, 1964, p. 816.
[11] Polivka, Milos and Klein, Alexander, "Effect of Water-Reducing Admixtures ~nd
Set Retarding Admixtures as Influenced by Portland Cement Composition," Effect of
Water-Reducing and Set-Retarding Admixtures on Properties of Concrete, A S T M STP
266, American Society for Testing and Materials, 1960.
[12] Loughborough, M. T., "Early-Setting Behavior of Concrete--A Brief Discussion,"
Ontario Hydro Research Quarterly. Vol. 15, No. 4, 1963.
[13] Wallace, G. B. and Ore, E. L., "Structural and Lean Mass Concrete as Affected by
Water-Reducing, Set-Retarding Agents," Effect of Water-Reducing and Set-Retarding
Admixture on Properties of Concrete, A S T M STP 266, American Society for Testing ~nd
Materials, 1960.
[14] Schutz, R. J., "Setting Time of Concrete Controlled by the Use of Admixtures," Journal,
American Concrete Institute, Jan. 1959; Proceedings, Vol. 55, p. 769.
F. F. B a r t e l 1
Chapter 10--Air Content and Unit

Weight
Introduction
Tests for air content and unit weight of freshly mixed concrete are made
frequently both in the field and in the laboratory. Procedures are outlined
in various ASTM standards which are straightforward and well within the
capabilities for performance and understanding of the average concrete
technician. Further, a m i n i m u m of training is required before proficiency
in the testing technique is acquired. Nevertheless, an understanding of the
limitations of the various methods, together with an appreciation for carefully following details of the procedures, is essential to obtaining and interpreting results.
Tests for air content and unit weight are made of fresh concrete to provide a control on these properties in the hardened concrete and also to
determine volume of concrete being produced from a given batch and its
unit cement content. The significance of air content and unit weight in
hardened concrete is discussed in detail in this publication by Helms. 2 Air
content most commonly is determined to ensure the presence of small
amounts of air prescribed in concrete in order to obtain its beneficial effects
on resistance to freezing and thawing and effects of deicing agents applied
to concrete for snow and ice removal. Tests for unit weight are made to
control weight p e r s e of both lightweight and high-density concretes. For
all concrete, unit weight determinations are used to establish the volume of
batch produced and to establish actual cement contents.
The interrelation between air content and unit weight should be obvious;
an increase in volume of air results in a lower unit weight, whereas a reduction in air content increases unit weight. Therefore, testing procedures
for these properties are combined in ASTM methods and should be performed in sequence in the field.
l President, Tews Lime and Cement Co., Milwaukee, Wis. 53210.
2See p. 435.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
BARTEL ON AIR CONTENT AND UNIT WEIGHT
123
Test Methods
Three tests for air content of fresh concrete, using gravimetric, pressure,
or volumetric methods, currently are standards of ASTM. Each of the procedures has its advantages and disadvantages, and it is expected that the
method most appropriate for the conditions and materials will be used.
Tests for unit weight are part of the gravimetric method.
Gravimetric Method
In the gravimetric procedure the unit weight of concrete is determined in
accordance with ASTM Test for Unit Weight, Yield, and Air Content
(Gravimetric) of Concrete (C 138). The test is made of a sample of concrete
in a measure of specified volumetric capacity. Concrete containing aggregate
graded up to 50 mm (2 in.) may be tested in a 14 dm 3 (0.5 ft 3) container;
where aggregate u p to 75 mm (3 in.) in size is used, the 28 dm 3 (1.0 ft 3)
measure should be employed. Larger containers should be used for concrete
containing still larger aggregate as specified in ASTM Test C 138. The
measure is filled in three layers, and each layer of concrete is consolidated
by rodding and tapping the sides of the measure a specified number of
times. The strikeoff of the top surface must be done " . . . with a flat cover
plate using great care to leave the measure just level full." From the known
volume of the measure and the weight of its contents, the weight per cubic
metre (foot) of concrete can be computed. In ASTM Test C 138 details
of the calculations are outlined to determine: (a) weight per cubic metre
(foot), (b) volume of concrete produced per batch, (c) relative yield, or ratio
of actual volume produced to the designed volume of the batch, (d) actual
cement content as kg/m 3 (lb/yd 3) of concrete, and (e) air content of concrete.
When the gravimetric procedure for determining air content of fresh
concrete is used, the unit weight of concrete as determined is compared
with the theoretical unit weight of air-free concrete. The theoretical unit
weight is calculated from the weight and bulk specific gravity of each ingredient in the concrete mixture following the calculations outlined in ASTM
Test C 138. The gravimetric method for determining air content has serious
limitations as a field test. One disadvantage is that an accurate knowledge
of batch proportions, specific gravities, and moisture contents of aggregates
is essential. For example, an error of 2 percent in the moisture content of
sand in an average concrete mix, in which a constant amount of water is
added at the mixer, will result in an error of one percentile in computed air
content. In the field where, in many cases, average moisture contents of
aggregates are assumed and consistency is controlled by varying the amount
of water added at the mixer to maintain slump, the gravimetric method
should not be relied upon for accurate results. Also, an error of 0.02 in the
specific gravity of the aggregates will result in an error of about one-half
124
percentile in air content computed under this method. Thus, where lightweight aggregates are used, whose specific gravity is exceedingly difficult to
determine, it is not recommended that the gravimetric procedure be used.
In the laboratory, however, where specific gravities and moisture contents are accurately known, reliable results can be obtained for concrete
containing normal weight aggregates. Furthermore, the concrete used in
the test need not be discarded, as in other procedures; it can be used for
other tests or in the preparation of specimens.
Pressure Method
The pressure method for determining air content of fresh concrete is

based on the fact that, in most concrete mixes, the only compressible ingredient is the air entrained or entrapped in the mixture. Klein and
Walker [1] 3 applied Boyle's law, involving the relationship of pressure and
volume of gases, to the determination of air content of fresh concrete.
Menzel [2] refined the apparatus and proposed a standard test procedure.
Two types of air meters are described in ASTM Test for Air Content of
Freshly Mixed Concrete by the Pressure Method (C 231). In Meter Type A,
a predetermined pressure is applied to a column of water above a sample
of concrete in a container of known volume. When the proper pressure is
applied, the drop in water level in the neck of the calibrated apparatus
indicates the air content of the concrete directly. The instrument must be
calibrated for use at various localities if differences in altitude are considerable. Also, the results vary with changes of barometric pressure. A
complicating feature of the test is that the pressure may compress air within
the interstices of nonsaturated, porous aggregates, and a higher than true
air content of concrete may thus be indicated. This complication may be
compensated for by determining a correction factor for the air within the
aggregate as outlined in ASTM Test C 231. In the case of slag vesicular
volcanic aggregates or lightweight aggregates, however, this aggregate
correction factor may be quite large and difficult to determine accurately.
The ASTM test requires the measuring bowl to have a capacity of at least
0.006 m 3 (0.20 ft 3) for testing concrete containing coarse aggregate particles
up to 50 mm (2 in.). For concrete containing larger aggregate, wet sieving
is specified to remove particles larger than 37.5 mm (11/2 in.).
The major advantage of the pressure method is that no knowledge of
specific gravities, moisture contents, or batch quantities of the concrete
mixture need be available to determine its air content. Possible inaccuracies
when porous or nonsaturated aggregates are used have been referred to
previously. Samples of concrete tested in this apparatus must, of course, be
125
discarded and cannot be used in specimens or for further testing because

of the contact of concrete with water during the test.
In Meter Type B in ASTM Test C 231 a known volume of air at an established pressure is released to contact concrete in a sealed container; the
drop in pressure provides a measure of the amount of air within the concrete. The apparatus has the advantage of not having test results influenced
by changes in barometric pressure. Compressibility of air within the aggregate must be taken into account in test results as it must be with results
from the Klein-Walker apparatus.
Volumetric M e t h o d
Inaccuracies in determining air content of fresh concrete containing

porous aggregates by the gravimetric and pressure methods have been
described. The volumetric procedure, however, can be used with such
materials to obtain accurate measurements of air content. Further, as with
the pressure method, no knowledge of specific gravities or moisture contents of ingredients need be known.
With the apparatus described in ASTM Test for Air Content of Freshly
Mixed Concrete by the Volumetric Method (C 173), water is filled to mark
over a sample of concrete in a container of known volume. The apparatus
is sealed; then concrete and water are intermingled and agitated until the
air in the concrete is displaced by washing action. The drop in water level
from its original mark provides a direct measure of the air content of the
concrete. In the test, mixing of water and concrete is repeated until no
further drop in water level indicates the removal of all air from the mixture.
The method requires the use of a container for concrete of not less than
0.002 m 3 (0.075 ft 3) when aggregrates up to 37.5 mm (11/2 in.) are used.
For concrete containing larger aggregate, wet sieving is specified to remove
particles larger than 25 mm (1 in.).
The chief disadvantage of the volumetric method is in the physical effort
required to agitate water and concrete sufficiently to remove the air.
Other volumetric procedures, such as the rolling method developed by
Menzel [2], the hook-gage or "Indiana Method" described by Miesenhelder [5], and the "Ohio Method" [6] using a hook gage, test larger
samples of concrete. With Menzel's volumetric apparatus the quantity of
water added to compensate for air removed from a known volume of concrete by mixing with water (on a rolling apparatus) provides a measure of
the air content. In the "Indiana Method," air is removed from a known
weight of concrete by stirring in water; the air content is computed from
the unit weight of airfree concrete determined by displacement in the test,
and the unit weight measured prior to removal of the air. Miesenhelder [5]
reported the disadvantages of the latter method to be the difficulty in mainCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:39:19 EDT 2014
126
taining the necessary accuracy of scales under field conditions and the
possibility of incomplete removal of air because of the physical exertion
required in the stirring process.
A device known as the Chace Air Indicator has been used extensively to
measure air content of fresh concrete. It involves a volumetric procedure
utilizing a pocket-sized graduated glass tube into which is fitted a brass
cup 25 mm (1 in.) in diameter by 12.5 mm (1/2 in.) high attached to a
rubber stopper. A sample of mortar is secured from the concrete, consolidated in the brass cup, and struck off to make the cup level full. The
graduated glass tube is filled with isopropyl alcohol, and the brass cup
inserted into it. After adjustment of the alcohol level to mark, the tube is
inverted from horizontal to vertical several times until all mortar has been
displaced out of the cup into the alcohol. In the course of this intermingling
of mortar and alcohol, the air is washed from the mortar. The drop in
level of the alcohol from its original mark is a measure of the air content
of the mortar; calculations dependent on the volume of mortar in the concrete yield an indicated air content of concrete. Tests reported by Willetts
and Kennedy [7] under laboratory conditions showed the Chace Indicator
to be accurate within plus or minus 0.6 percentile two thirds of the time
for air-entrained concrete. Additional studies of the air indicator were reported by Grieb [8] and Newlon [9]. Its principal use would appear to be
as a speedy screening device to indicate when air contents apparently are
approaching or do not conform with specification limits. When this is indicated, the appropriate ASTM test method for air content should be used
for acceptance or rejection of the concrete.
Comparison of Test Results

The existence of three ASTM procedures for determining air content of
fresh concrete logically raises the question as to which is most accurate.
Many studies have been made on this subject. In the ASTM Symposium on
Measurement of" Entrained Air in Concrete [10] the results presented by
the various authors are in substantial agreement. Results of gravimetric
tests agreed with other procedures as long as specific gravities and moisture
contents of ingredients were known, pressure method test results were in
accord with other data when concrete was made with dense aggregates,
and volumetric procedures gave substantially the same air contents as did
the other methods as long as all of the air was removed from the sample
tested. Tests by Britton [11] using ASTM gravimetric, pressure, and volumetric procedures gave comparable results, aside from the limitations mentioned above, for tests carefully conducted.
Significance of Test Results

Tests for air content of fresh concrete are most commonly made in the
127
field on concrete that is assumed to contain entrained air. Air entrainment

is obtained by using an air-entraining cement or by adding .an air-entraining
admixture to concrete or by using an air-entrained admixture to supplement
the air entrainment achieved by use of an air-entrainment cement. Numerous tests have demonstrated that purposefully entrained air increases the
resistance of concrete to destructive agencies such as freezing and thawing,
seawater, and salts used for removal of snow and ice from pavements. Along
with these beneficial effects, unfortunately, there is usually a decrease in
the strength of concrete, especially in relatively rich mixes. Thus, the problem in the case of air-entrained concrete is to control the air content to
ensure the beneficial effects without unduly reducing its strength.
Studies by Wuerpel [12], Gonnerman [13], and others show the need for
a minimum of about 3 percent entrained air in concrete mixtures containing 50 mm (2 in.) maximum size aggregate to secure the full benefits of
increased resistance to freezing and thawing. On the other hand, tests by
the Portland Cement Association [13] and the National Sand and Gravel
Association [14] indicate that, for a given water-cement ratio, each percentile of entrained air reduces compressive strength about 5 percent. This
reduction in strength, however, can be counteracted by reproportioning the
mix by reducing water and sand contents to maintain the slump and volume
of mortar in the concrete. Thus, by reproportioning of the mix, it is possible that the strength of lean, air-entrained concrete mixtures (containing
up to about 275 kg/m a or 470 lb/yd 3) actually may be greater than that of
normal concrete. In the case of rich mixtures, the reduction in strength
may be only 10 to 15 percent instead of the 20 to 25 percent which could
be expected were the mix not reproportioned. It is from this background of
information, together with field experience, that most specifications require
4 to 7 percent air in freshly mixed air-entrained concrete.
There are exceptions, however, where the 4 to 7 percent limits on air
content should not be applied. In mass concrete, for example, in which
cobbles of aggregate 152 mm (6 in.) or more in diameter are used, these
limits may be applicable to that portion of the concrete which would pass a
50 mm (2 in.) or 75 mm (3 in.) sieve, but they would be too high for the
concrete as a whole. As a matter of fact, the Corps of Engineers [15] wet
screens its concrete containing a nominal maximum size aggregate of
75 mm (3 in.) to 152 mm (6 in.) through a 37.5 mm (11A in.) or 75 mm
(3 in.) sieve before testing for air content. The Bureau of Reclamation
generally requires removal of larger sizes of coarse aggregate by handpicking before a test for air content is made. In concrete mixtures used in
Kansas pavements in which the maximum size of aggregate is about 9.4 mm
(3/8 in.), the 4 to 7 percent limits on air content should probably be raised.
The above discussion of exceptions to the usual specification limitations
on air content indicates that the air content of the mortar or cement paste
portion of the concrete should be of principal concern. Wuerpel [16] has
pointed out that the optimum air content for the mortar phase of concrete
128
should be 8.5 to 9.5 percent. Powers [17] and Klieger [18] have shown that
the air content of fresh concrete should be governed by the spacing and
size of voids in the cement paste. In concrete of a given air content, numerous small air bubbles closely spaced throughout the mass are more
effective in securing the beneficial effects of air entrainment than are a
smaller number of relatively large voids.
The air content of concrete containing a fixed amount of air-entraining
admixture added at the mixer, or as an addition interground with the
cement, varies with many factors. It is important that the effects of these
factors be understood so that test results obtained may be evaluated and
adjustments made to secure compliance with specifications. Less air will be
entrained in rich mixtures than in lean ones under given conditions. Concrete of high slump will ordinarily entrain more air than that of low slump.
The temperature of the concrete is important in that more air is entrained
by the same dosage of air-entraining admixture at 20~ (70~
than at
38~ (100~
and more at 4~ (40~ than at 20~ (70~ [19,20]. The
quantity of air will increase with mixing up to a point and then will gradually decrease; in transit-mixed ready-mixed concrete, for example, tests
have shown air content to increase during about the first 12 min of mixing
and then slowly to decrease [21]. The grading of the aggregate, particularly
that of the sand, also has its effect on air content; Walker and Bloem [19]
found that the air content of concrete increased as the proportion of sand
in the Nos. 30 to 50 size increased.
So far discussion has been confined to air content of concrete as mixed.
However, handling and placing of concrete on the job can affect the air
content of concrete in place. And, of course, it is the air voids in the hardened concrete that secure the beneficial effects attributed to air-entrained
concrete. There are indications that pumping of concrete may remove some
air from certain mixtures. Repeated handling and agitation of concrete
prior to placing might cause loss of air content. Vibration in consolidating
concrete in the forms may also remove air from the mass [22]. I t should
not be overlooked that the air in the lower portion of a deep section will be
compressed by the weight of concrete above it; this will affect the volume of
air but probably not the spacing, number of air voids, or the durability of
the concrete.
Nevertheless, even with all these complications, tests of fresh concrete for
air content by the ASTM standards described here constitute the best
means for controlling the end results in the field. These test methods,
coupled with intelligently prepared specification limits, should ensure the
desirable properties of air-entrained concrete.
Air content determinations are sometimes made to ensure that nonairentrained concrete is " n o r m a l " - - t h a t it does not contain more than about
1 to 1.5 percent of air. This is particularly necessary for mixtures for heavyduty floors or for very high-strength concrete which might be used in preCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:39:19 EDT 2014
129
stressed designs or heavy-load-carrying columns. Many manufacturers of

metallic aggregates for heavy-duty floors, for example, recommend that
their products not be used with air-entrained concrete. It is difficult to produce high compressive strengths with air-entrained concrete, particularly in
hot summer temperatures. This makes apparent the desirability of making
an air content test, at the time high-strength concrete is placed, to ensure
that even a moderate amount of air is not being entrained.
Unit weight tests of fresh concrete, as previously stated, are used to control weight p e r se of both lightweight and high-density concretes and to
establish the size of batch (or yield) being produced. In the case of lightweight aggregates, it is extremely difficult to determine accurately specific
gravity and absorption. For a given aggregate, the density of individual
particles varies with size, with the larger particles being less dense because
they have been expanded more. Thus, as gradation of the lightweight aggregate changes even a relatively small amount, the specific gravity will change
significantly. Further, it is impractical to bring these aggregates to an exact
condition of saturated-surface dry, especially when particle shape is rough
and irregular.
This difficulty in accurately determining specific gravity and absorption
of lightweight aggregates makes conventional concrete mix design procedures inaccurate. The unit weight procedure for proportioning and controlling concrete mixtures made with lightweight aggregates, however,
circumvents these difficulties. Details of the procedure are outlined in the
Recommended Practice for Selecting Proportions for Structural Lightweight
Concrete of the American Concrete Institute [23]. The weight of all materials to be used to produce a cubic metre (foot) of concrete is estimated
on the basis of experience, from data supplied by the aggregate manufacturer, or from weights per cubic metre (foot) of dry loose aggregate. On
the basis of the estimate, a trial mix of concrete is mixed and tested. If the
actual unit weight ef the fresh concrete differs from the estimate appreciably, it will be necessary to increase or decrease quantities of materials
until the yield approaches a full cubic metre (yard).
The unit weight of high-density concrete has been of interest for many
years where it was used in such things as counterweights, dam gates, and
so on. Recently the need for radiation shielding in various nuclear devices
has increased the use of high-density concrete. In general, the greater the
density of a material, the greater the ability to shield against radiation.
Concrete, with its relatively high-weight to low-cost ratio, has been used
extensively for this purpose. Where space limitations prevail, special highdensity aggregates have been used to increase unit weight of concrete. The
subject of radiation effects and shielding is discussed in detail in this publication by Polivka and Davis. 4 Obviously, it is the weight of hardened (not
4See p. 420.
130
fresh) concrete which is of concern for radiation shielding; however, the

control must be exercised through unit weight determinations of freshly
mixed concrete.
The use of ASTM Test C 138 to determine yield of concrete has been
referred to previously. Yield as such is defined as the volume of concrete
produced per batch. Relative yield is defined as the ratio of actual volume
of concrete obtained to the volume as designed for the batch. Relative yield
is of special importance to the commercial transactions between a readymixed concrete producer and his customer. It indicates whether a full
cubic metre (yard) is being delivered or whether the batch is short of its
design volume.
Summary
Tests for air content and unit weight of fresh concrete, carefully made in
accordance with the appropriate ASTM test method, will yield an accurate
measurement of the amount of air, weight, and volume of concrete being
produced. Tests for air content, coupled with intelligently selected specification limits, can ensure the beneficial effects of entrained air in hardened
concrete. Unit weight tests are reliable for controlling yield of concrete,
and for evaluating unit weight when stated in specifications.
References
[lJ Klein, W. H. and Walker, S., "A Method for Direct Measurement of Entrained Air in
Concrete," Journal, American Concrete Institute, June 1946; Proceedings, Vol. 42, p.
657.
[2] Menzel, C. A., "Procedures for Determining the Air Content of Freshly-Mixed Concrete
by the Rolling and Pressure Methods," Proceedings, American Society for Testing and
Materials, Vol. 47, 1947, p. 833.
[3] Tremper, B. and Gooding, W. L., "Washington Method of Determining Air in Fresh
Concrete," Proceedings, Highway Research Board, Vol. 28, 1948, p. 210.
[4] Pearson, J. C. and Helms, S. B., "The Effect of Sampling Errors on Unit Weight and
Air Determinations in Concrete," Proceedings, American Society for Testing and Materials, Vol. 47, 1947, p. 914.
[5] Miesenhelder, P. D., "Indiana Method for Measuring Entrained Air in Fresh Concrete,"
Proceedings, American Society for Testing and Materials, Vol. 47, 1947, p. 865.
[6] Barbee, J. F., "The Ohio Method of Determining the Amount of Air Entrained in Portland Cement Concrete," Proceedings, American Society for Testing and Materials, Vol.
47, 1947, p. 901.
[7] Willetts, C. H. and Kennedy, T. B., "A Limited Investigation of the Chace Air Meter,"
Miscellaneous Paper No. 6-189, Waterways Experiment Station, Corps of Engineers,
U.S. Army, Nov. 1956.
[8] Grieb, W. E., "The AE-55 Indicator for Air in Concrete," Bulletin 176, Highway Research Board, 1958.
[9] Newlon, H. H., Jr., "A Field Investigation of the AE-S5 Air Indicator," Bulletin 305,
Highway Research Board, 1961.
[10] Symposium on Measurement of Entrained Air in Concrete, Proceedings, American
Society for Testing and Materials, Vol. 47, 1947, p. 832.
131
[1l] Britton, R. E., "Report of Investigation of Different Methods for Determining the
Amount of Air Entrained in Fresh Concrete," Pennsylvania Slag Association, April 11,
1949.
[12] Wuerpel, C. E., "Laboratory Studies of Concrete Containing Air-Entraining Admixtures,"
Journal American Concrete Institute, Feb. 1946; Proceedings, Vol. 42, p. 305.
[13] Gonnerman, H. F., "Tests of Concrete Containing Air-Entraining Portland Cements or
Air-Entraining Materials Added to Batch at Mixer," Journal, American Concrete Institute, June 1944; Proceedings, Vol. 40, p. 477.
[14] Walker, S. and Bloem, D. L., "Control of Quantity of Entrained Air in Concrete,"
National Ready Mixed Concrete Association, 1950.
[15] "Investigation of Field Method for Determining Air Content of Mass Concrete," Technichal Memorandum No. 6-352, Corps of Engineers, U.S. Army, Waterways Experiment
Station, Nov. 1952.
[16] Wuerpel, C. E., "Purposeful Entrainment of Air in Concrete," Marquette Cement
Manufacturing Co., 1953.
[17] Powers, T. C., "Void Spacing as a Basis for Producing Air-Entrained Concrete,"
Journal, American Concrete Institute, May 1954; Proceedings, Vol. 50, p. 741.
[18] Klieger, P., "Effect of Entrained Air on Concretes Made w~,th So-Called 'Sand Gravel'
Aggregates," Journal, American Concrete Institute, Oct. 1948; Proceedings, Vol. 45, p.
149.
[19] Walker, S. and Bloem, D. L., "Studies of Concrete Containing Entrained Air," Journal,
American Concrete Institute, June 1946; Proceedings, Vol. 42, p. 629.
[20] Symposium on Entrained Air in Concrete, Proceedings, American Concrete Institute,
Vol. 46, 1946, pp. 601-699.
[21] Mumford, R. P., "Effect of Time of Mixing," Technical Information Letter No. 28,
National Ready Mixed Concrete Association, Sept. 1946.
[22] Crawley, W. O., "Effect of Vibration on Air Content of Mass Concrete," Journal
American Concrete Institute, June, 1953; Proceedings, Vol. 49, p. 909.
[23] "Recommended Practice for Selecting Proportions for Structural Lightweight Concrete,"
Journal American Concrete Institute, Sept. 1958; Proceedings, Vol. 55, p. 305.
Katharine Mather 1
Chapter 11 Petrographic
Examination
Petrography is the scientific description of the composition and texture of

rock, including the systematic classification of rocks. Petrographic examination of hardened concrete--a man-made rock--is the examination of concrete by the techniques used in petrography to determine the formation,
composition, and internal structure of the concrete and to classify it as to its
type, condition, and serviceability.
This paper is intended to show that petrographic examination provides information useful in evaluating hardened concrete, to note the problems inherent in its use, to outline what it involves, to describe the kinds of information that it can produce, and to show how this information can be applied.
The questions that materials testing and evaluating tries to answer are: (a)
"How will this material behave in use?", (b) " W h y did this material behave
in use in the way it did?", and (c) What can be anticipated as to the future
service of this material?" The first question is never answered unequivocally,
except after-the-fact, by determining how the material did behave in use.
The most useful method for developing practical information upon which to
make decisions that depend on prediction of probable behavior of materials
is the study of why materials behaved in use as they did.
Testers of materials are unable to compress time or to anticipate ann
reproduce the environment that the material will experience. Generally, they
use standardized procedures not directly related to the specific environment
or that do not determine the particular properties relevant to performance in
the specific instance. Thus, testing construction materials amounts to obtaining certain kinds of information about certain samples in specified conditions and extrapolating to the conditions of intended use insofar as they can
be predicted.
Petrographic examination of hardened concrete is included among the
l Geologist;Chief, EngineeringSciencesDivision,StructuresLaboratory,U.S. ArmyEngineer
Waterways ExperimentStation, Vicksburg, Miss. 39180.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
lntcrnational
www.astm.org
K. MATHER ON PETROGRAPHIC EXAMINATION
133
subjects in this volume because it helps to improve the extrapolation from

test results to performance in use. It offers direct observational information
on what is being tested and what is in the structure, giving another way of appraising the relation between samples being tested and materials in use and
judging how similar the two are.
Communication Problems
A petrographic examination of concrete ordinarily begins and ends with a

problem of communication between the person who requests the examination (usually an engineer) and the person who makes it (usually a
petrographer). Unless the two succeed in producing a clear, mutually
understood statement of the problem, they cannot expect a clear, useful
answer to be obtained economically. In nearly all cases, the engineer who
asks for an examination of a particular concrete suspects that the concrete is
unusual; the more clearly he defines the features prompting his interest, the
more he directs the petrographer toward the important aspects. The engineer
may not be familiar with the techniques that the petrographer may use or
with his approach; the petrographer may not realize the engineer's responsibility for decision and action, may not find out all the engineer could tell
him about the concrete, and may not realize which petrographic findings are
useful and relevant. The petrographer should not expect petrographic results
to be taken on faith; the rationality of the techniques producing them should
be demonstrable. Both should remember that the essentials of petrographic
examination of concrete are practiced anytime anyone looks intelligently at
concrete either in a structure or as a specimen and tries to relate what he can
see to the past or future performance of the concrete. On this basis, it is, clear
that many of the most useful petrographic examinations are made by inspectors, engineers, chemists, physicists--anyone concerned with the production
or use of concrete. No one should hesitate to examine concrete with all
available means; all, from novice to expert, should question the indicated
conclusions and verify them in as many ways as possible.
Methods--Standardized and Described
There is a recommended practice for petrographic examination of concrete

aggregates, namely ASTM Recommended Practice for Petrographic Examination of Aggregates for Concrete (C 295). ASTM Recommended Practice for Examination and Sampling of Hardened Concrete in Constructions
(C 823), which was adopted in 1975, [2] gives guidance for the examination
of concrete in constructions for many purposes in addition to petrographic
examination of the samples taken. Steps to be taken before examination and
preliminary investigations are outlined; the desirability of assembling reports
and legal documents concerning the construction is pointed out, as well as
134
the usefulness of interviews with contractors and others connected with the
construction and with the owners, occupants, and users of the constructions.
Procedures for detailed investigations of the concrete in place in constructions are described. Sampling hardened concrete is discussed, with the
preparation of appropriate sampling plans and selection of the number and
size of samples. Information needed to accompany samples is described. It is
the hope of the subcommittee by which the recommended practice was
prepared that it will be useful not only to petrographers but also to engineers
and others who have reason to examine constructions.
ASTM Recommended Practice for the Petrographic Examination of
Hardened Concrete (C 856) was adopted in 1977. It turned out to be a more
complicated document than was expected, because the purposes differ for
examining concrete from constructions, from specimens in field exposure,
from specimens in simulated service, from specimens in laboratory tests, and
from specimens after laboratory curing. What may be recognized as features
produced in field exposure, simulated service, laboratory tests, and
laboratory curing may be features interesting and revealing in themselves in
laboratory experiments. The features may also make more comprehensible
similar or related features of concrete from constructions. The subcommittee
hopes that this recommended practice will serve the engineers who buy a
petrographic examination and then want more understanding of what they
are getting and why; that it will also serve petrographers approaching the examination of concrete for the first time; and that it will remind concrete
petrographers of things that they may have forgotten or neglected. The illustrations appear in ASTM Adjunct C 856.
In 1976, Evaluation of Methods of Identifying Phases of Cement Paste [1] 2
was published. It will be helpful to those who wish to determine alite and
belite residues in paste, examine hydrated cement by X-ray diffraction, differential thermal analysis, or infrared absorption spectroscopy. Pastes of
three cements, a Type I, a white cement, and a cement containing no
tricalcium aluminate (C3A), were examined at ages to one year; a number of
pure phases were synthesized and examined by the methods listed above.
Another publication relevant to the petrographic examination of hardened
concrete is Guide to Compounds of Interest in Cement and Concrete
Research [2]. Other publications of interest are found in the volumes of the
Fifth International Symposium on the Chemistry of Cement [3] and those of
the Sixth International Symposium on the Chemistry of Cement [4]. One of
the most stimulating serial publications came into being in 1971, Cement
and Concrete Research, an International Journal [5]; it appears bimonthly
and covers a wide field of cement and concrete topics and has the advantage
of being truly international since the editorial board and the contributing
K, MATHER ON PETROGRAPHIC EXAMINATION
135
authors do come from most of the centers in various parts of the world where
cement and concrete research is carried out.
Before the production of ASTM R e c o m m e n d e d Practice C 856,
petrographic techniques for the examination of hardened concrete had been
described and discussed but not standardized. In 1966 1 wrote:
At present I know of no laboratory where petrographic examinations of concrete are
made that is equipped to use all the methods that have yielded useful information and no
one person who has digested the available approaches and developedthe ability to choose
the particular combination of techniques best suited to each problem encountered. Concretes are more complexthan most rocksused as aggregates; their constituents are lesswell
known; concretes change through time more rapidlythan most aggregates. All of these circumstances combine to make each petrographic examination of concrete unique and thus
to make the methods harder to generalize and standardize. Each examination presents
some new facet for the petrographer who is willing to learn.
In 1977 there is more sophisticated equipment in more laboratories but the
number of laboratories may have diminished. More people are using
sophisticated equipment and techniques in the study of cement hydrates and
even of concrete, and that is good and highly desirable, and may bring about
a breakthrough by orders of magnitude in our ability to interpret the
behavior of concrete in constructions. This real and exciting probability
brings with it dangers of losing touch with the importance of the very great
range in scale, from constructions with dimensions in hundreds of metres
(kilometres in the case of pavements), to core samples with dimensions in
hundreds of millimetres that can be examined with great advantage at low
magnification by a stereomicroscope on a large stand, to thin sections with
scales ordinarily about 800 mm 2 by 15 to 30/zm thick, to X-ray diffraction
samples which may be milligrams of material hand-picked under the
stereomicroscope or a few grams of paste concentrated by hand-picking the
aggregate from carefully broken concrete, to scanning electron microscope
specimens that may be 200 mm a or much smaller and thinner, to single
crystals a few micrometres cubed that are examined by microprobe and to
the nanometre-sized material examined by the transmission electron
microscope. Few cases will be found in which all the steps in scale listed
above will be used, but the determining steps are looking at the constructions, looking at the samples from it with the eye and the stereomicroscope,
and usually then forming one or more working hypotheses that indicate to
the petrographer how to make the transition from the macroscale to the
microscale most economically in knowledge gained for each unit of effort,
and preferably using more than one method so as to obtain an independent
verification. Proper caution is needed so that proper sampling is done at all
levels of scale. For example, samples should be taken from deposits in several
cracks and voids in several cores, paste concentrates are made from several
cores, and so a sensible sampling plan is carried from the macro- to the
microscale.
136
One should also bear in mind the straw that broke the camel's back theory
of concrete deterioration, which says that unless the concrete was hit by a
truck or a large object from outer space, it is highly probable that evidence of
more than one deteriorative mechanism (or what appears to be such
eyidence) will be present in any sample of deteriorated concrete. Sorting out
the major cause or causes from the minor causes or secondary effects of
deterioration requires the exercise of the petrographer's best judgment.
The history of the construction and its behavior is of great importance, and
for this kind of information the petrographer should consult engineers, inspectors, and contractor's personnel who have been present during construction and those who have later inspected or occupied the construction.
The newer techniques offer an opportunity to understand in much more
detail than previously the chemical reactions that have gone on in a concrete,
because it will be possible to characterize the hydration and reaction products more clearly and consequently the chemical reactions of normal hydration or abnormal deterioration, or both, in the concrete, and to evaluate,
with much more certainty than is possible without identification of the reaction products, the relative roles of chemical attack and physical attack in
producing deterioration. The goal of relating better established and more
familiar techniques in petrographic examination of hardened concrete to the
more intimate and detailed insights made possible by the newer techniques
remain to be achieved completely and while progress since 1966 is considerable, the task is not yet complete.
Purpose and Approach
Purpose
A petrographic examination attempts to answer two general objective
questions: "What is the composition?" and "How is it put together?" The
first question refers to the recognizable individual constituents present on the
scale at which they are considered. The second question refers to structural
fabric, that is, the articulation or packing in space of the component
elements making up any sort of external form [6] or heterogeneous solid
body. Both questions may be answered on any useful scale by choice of
technique or techniques of appropriate resolving power. The resolving power
needed differs depending on the specific questions to be answered.
Approach
Step one, in any case, is to define the problem in order to find and ask the
right questions, those that need to be answered to solve the problem that
caused the examination to be requested. These right questions should be
answered insofar as they can be in the context, limited as it will be by money,
137
time, instrumentation, and the state of the art. The best petrographic examination is the one that finds the right questions and answers them with
maximum economy in minimum time, with a demonstration clear to all concerned that the right questions were answered with all necessary and no
superfluous detail. The approach to the ideal varies depending on the problem, the skill with which the questions are asked, and the skill of the petrographer. One measure of the petrographer's skill, is knowing when to stop,
either because the problem is adequately solved, or, in some cases, because
it has been shown to be unsolvable under the circumstances.
Fabric and Composition

Fabric--the packing of component elements in space--is the heterogeneity
obvious as one looks at a weathered concrete structure or at a broken or
sawed surface of concrete. Fabric includes all of the structural elements,
ranging in scale from gross to atomic, and comprises both structure and texture as those terms are used in rock description. The fabric appears on the
scale of the lift or course or batch or the structural crack, on the scale of the
coarse aggregate, on the scale of the sand grains or the air voids in the mortar
or the "microcracks," on the scale of the residual unhydrated cement or the
calcium hydroxide crystals, on the scale of the hydrous calcium aluminates
and the scale of the almost amorphous crystalline hydrous calcium silicates
in the hydrated cement paste, or on the scale of the atomic structure of any
crystal forming a part of any of the structural components. The closest
naturally occurring analogue among rocks to the fabric of concrete is
graywacke conglomerate with abundant matrix. The closest naturally occurring analogue to hardened cement paste is silty clay.
Fabric and composition together define, characterize, and form the basis
for descriptive classification of solid multicomponent substances. Composition and fabric are so closely interrelated in concrete that they cannot be
separated clearly.
One important decision in a petrographic examination is the decision
whether the request for the petrographic examination was made because of
problems created by the physical structure of the concrete or because of
problems created by its chemical composition. Problems arising from the
physical structure or fabric include inadequacies in mixing or consolidation
and inadequacy of the air-void system to provide frost resistance. Problems
arising from the chemical composition include those resulting from errors in
batching, reactions between cement and aggregate, reactions between a contaminant and cement paste, and reactions between cement paste and solutions from external sources. Were the construction practices employed
suitable for producing concrete capable of giving satisfactory service in the
particular environment and exposure? Were the materials that were chosen
susceptible to participation in chemical reactions that have deleterious conseCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:38 EDT 2014
138
quences? Was there a failure to modify the environment, for instance, by improving the drainage so as to increase the ability of the concrete to survive the
chemical or physical character of the environment? Usually several causes
have interacted, but one is probably the originator; if it can be identified, the
appropriate techniques are pointed out more clearly.
Investigating composition and fabric provides a specific, unique definition
of what is being examined. The standard tests do not always supply information that permits discrimination between one piece of concrete and another,
but direct observation on the relevant scale does. There are n possible concretes all having S0-mm (2-in.) slump, with air content of 5 percent, with
31-MPa (4500-psi) compressive strength at 28 days, but the No. 2 cylinder in
the set of three broken on Day A in Laboratory B is unique and different,
perceptibly and logically, from Nos. 1 and 3 and from all the members of the
other possible sets, and its top is different from its bottom as cast. The salient
lesson from the study of composition and fabric of concrete is the individuality and uniqueness of each structure or part of a structure, of each specimen
and each thin section; this individual combination of fabric and composition
reflects the history and forecasts future performance of the concrete. What is
investigated at any time is particular concrete, not concrete in general. Each
structure and each part of a structure is unique in terms of composition,
fabric, history, and exposure.
Comparisons
To say that each structure and specimen is unique does not mean that
comparisons are useless or impossible; they are essential, and concretes can
be grouped rationally and compared usefully within classes and between
classes, if the basis for the grouping is objective. Each comparison leaves out
of account some characteristics of the things compared, so that it is necessary
to bear in mind that the accidentally or deliberately omitted factors may
prove to be important. Paste, mortar, or concrete of known proportions,
materials, age, and curing history offers the logical basis for comparison and
extrapolation; laboratory specimens made to be examined or salvaged just
after having been tested for strength provide a good source of such comparative material. Specimens exposed to laboratory air outside the moist
room or curing tank for more than a few hours are much less suitable,
because specimens that are cracked or that have slender cross sections
sometimes carbonate very rapidly. Specimens exposed to simulated weathering tests, or wetting and drying cycles, or prolonged drying are not to be considered as representative of normally cured or of naturally weathered concretes. Natural weathering differs from part to part of a structure, as well as
from climate to climate, elevation to elevation, and subgrade to subgrade.
139
Interpretation of Observations
Normal Concrete
For this discussion, " n o r m a l " constituents and fabrics are defined as those
present in serviceable concrete of the class and age in the region. "Serviceable" is used instead of "undeteriorated" because it is possible to tell
whether concrete in a structure is serving as it was intended to but the criteria
that distinguish inevitable chemical and physical changes from deterioration
in concrete 20 or 50 years old have not been well established.
The most valuable information that can be obtained by petrographic examination of concrete comes from the examination of normal concrete; only
by comparison with the range of constituents and fabrics in normal satisfactory concrete can that which differs from the normal be recognized and its
differences specifically defined. Unless it can be demonstrated that the constituents, or the proportions of constituents, or the fabric, depart from those
found in serviceable concrete of the age and class in the region, there is no
logical basis for assuming any connection between constituents, or proportions, or fabric, and service behavior. Even when it can be shown that a concrete has a peculiar service record and some unique feature or features not
shared by a dozen others of comparable class, age, and provenance, it remains to be seen whether the known unique feature and the peculiar service
record are connected causally, or whether both are related to some third or
nth factor that is the effective cause of the abnormal behavior.
Class of Concrete
The restriction to concrete of one class is necessary because changes in cement content, water/cement (w/c) ratio, and m a x i m u m size of aggregate
large enough to change the class entail such large changes in properties that
no close comparison will be significant. If, for example, the criteria for paving concrete are applied to mass concrete, all mass concrete appears very inferior, which it is not for the purpose it is intended to serve. Class of concrete
is important in the definition as it implies relative homogeneity in mixture
proportions, particularly in w/c ratio, cement content, and m a x i m u m size of
aggregate. It is possible by microscopic methods to sort mass concretes that
are fairly homogeneous in cement content and w / c ratio in order of increasing age, or it is possible to sort mass concretes fairly homogeneous in age in
order of increasing cement content. In terms of the ability to sort mass concrete microscopically, "fairly homogeneous" in cement content and w/c ratio
means a m a x i m u m difference in cement content between concretes of about
30 k g / m 3 (50 lb/yd3), and 0.1 by weight in w/c ratio.
140
Age of Concrete
Some restriction on the ages of concretes compared is necessary unless age
is the variable being studied. Unless the age is known or unless one has
younger and older concretes of otherwise comparable characteristics so that
the age of the unknown may be estimated in relation to the knowns, it may be
impossible to judge the significance of observations. For example, in one case
calcium sulfoaluminate was found in many voids as far as 130 mm (5 in.)
from the outer surfaces of a concrete pavement of high flexural and compressive strength and of unknown age. In other field concrete from the
region, calcium sulfoaluminate is commonly present in concrete over 5 years
old made with Type I or Type III cement, but it is not abundant and is confined to voids near outer surfaces. If the concrete of unknown age is in fact 5
or 7 years old, it differs conspicuously from others of comparable age and
class in the region and the difference probably justifies some concern about
its future; if it is 15 years old, it is peculiar, but the peculiarity is probably of
less practical importance.
Provenance of Concrete
Restriction of an investigation to one region assists in rational comparison
from several points of view. The aggregates economically available in an area
are determined by the regional geology and consequently show some
homogeneity of composition resulting from similarity of origin and history.
In a particular region, cements and aggregates economically available are
used in making concrete which is exposed to the climate characteristic of the
region--the prevailing temperature range and temperature frequency
distribution and the characteristic amount and sequence of precipitation.
The extent of a region of comparable concrete may vary from a few square
miles to many thousands, depending on variation in: (a) regional geology-as it determines quantity and uniformity of aggregate supply; (b) topography--a region of low relief and generally uniform slope such as the
Great Plains, or the Atlantic or Gulf Coastal Plain, has widespread, essentially comparable range and distribution of temperature and precipitation,
but in a region of high relief and broken slopes, temperature varies considerably with altitude, and precipitation with orientation to prevailing
winds, making important differences in exposure over short distances; and
(c) patterns of distribution of aggregates and cement from competitive
sources--in some areas, only one type of natural sand and gravel is available;
no manufactured aggregate is produced, and synthetic aggregate sources are
not common. Metropolitan marketing areas served by water transportation
usually have available a selection of natural coarse and fine aggregates,
manufactured coarse aggregates, and synthetic aggregates. All variations
141
between the two extremes just mentioned can be found in availability of aggregates.
Cement plants per state range from none in New Hampshire and Vermont
to 19 in Pennsylvania [7]. Ports and coastal areas may be served by overseas
and domestic cement sources.
An additional influence that may appear is a prevailing engineering opinion, in an organization placing concrete in a large area, on what is desirable
in mixture proportions or methods of placing or consolidation. The existance
of satisfactory structures built in many different ways underlines the need to
define "normal" concrete in objective and restricted terms.
Normal and Unusual Concrete

Although the most important kind of petrographic examination of concrete is the examination of normal concrete, usually the concrete that a
petrographer is asked to examine has behaved in an unexpected way. Before
and during the early stages of the examination, the information on the
history and behavior should be considered and the following questions asked:
1. What process or processes could produce the described results? 2. What
observable traces could the process or processes leave in the concrete? 3.
Would such traces be unique and specific evidence of what is supposed to
have happened?
Reconstruction of History of Field Concrete

To pass from consideration of simple petrographic examination to the
petrographic examination of concrete that has aged and perhaps
deteriorated in service introduces two important new unknowns--time and
the precise environment of the structure. The effects of the passage of
relatively short periods of time on the constituents present in several cement
pastes of known w/c ratio stored under laboratory conditions have been investigated, but anomalies remain in the results, although both the composition of the pastes and the nature of the environments were known and controlled far more thoroughly than the composition and environment of any
field concrete.
Composition
If the changes in composition of cement paste with time in laboratory conditions were known for a representative number of cement compositions and
w/c ratios, effects of both cement-aggregate interactions and of environmental influences would be easier to recognize and could be interpreted more
usefully.
142
Environment
Why do exposed vertical walls of chert-gravel concrete in the vicinity of St.

Louis, Mo., generally have fewer popouts than apparently similar walls in the
vicinity of Memphis, Tenn.? The winters are colder in St. Louis, but the
mean annual rainfall is lower; and in Memphis a larger proportion of the
higher mean annual rainfall occurs in winter. The difference probably is that
the chert gravel in the Memphis walls is more likely to be critically saturated
when it freezes. The Weather Bureau's climatological data for the location
are a valuable source of information that can assist in many petrographic examinations of hardened concrete.
The discovery in Mississippi of several highway pavements and associated
structures affected by sulfate attack and by combined sulfate and acid attack
[8] emphasizes the need to make use of available information on the composition of foundations and subgrades.
Thus, the examination of samples of field concrete after extended service
involves an increase in complexity, a decrease in available information, and a
decrease in the confidence that may be placed in the answer, as compared to
examinations of laboratory test specimens of hardened concrete.
It does not belittle the petrographer to admit that he cannot make bricks
without straw; neither can the chemist, the physicist, or the engineer;
sometimes the petrographer can recover evidence not accessible by other approaches. Several reasons make reconstruction of the history of deteriorated
field concrete difficult; in any particular instance it may be important and yet
impossible in the present state of our knowledge to decide what weight
belongs to each.
Deteriorated field concrete that is submitted to a laboratory or to a
petrographer is almost never concrete that has performed abnormally for one
single obvious cause; such simple cases usually can be and are explained on
the spot to the satisfaction of those concerned. The field concrete that is examined by a petrographer is concrete that has worried some responsible person enough to make the effort and expense of sampling and testing appear
justified. There is thus a build-in bias in the sampling process. Normally the
concrete that a petrographer sees as part of his assigned duties is controversial concrete sent in by organizations with alert conscientious concrete
technologists or concrete that has become the subject of controversy under
other situations. In practice, this generally means that he sees only the poor
concrete produced under conditions where a degree of control was intended.
Concrete produced where there was little intent for control to be exercised,
that is, the worst concrete, rarely is sampled and sent to a petrographer; good
concrete is rarely controversial.
Furthermore, the older the concrete the less information is likely to be
available about materials, proportions, conditions of placing, and the
characteristics that undeteriorated comparable concrete would have.
143
Although one can deduce from the concrete that w/c ratio was high or low,
and usually that cement content was high or low or medium, and the general
quality of the workmanship, one cannot reconstruct the alkali content of the
cement. Aggregate sources, particularly of natural sand and gravel, can be
located from their composition--the constituents present and their size
distribution are diagnostic of the region and sometimes of the particular
source.
Finally, deteriorated field concrete usually shows superimposed traces of
several processes, with at least one in an advanced stage. The most advanced
process may conceal the evidence of others that was more important in effect. Frequently the most conspicuous process is carbonation of other surfaces and along the borders of old cracks.
Laboratory test exposures are simplified compared to natural exposures by
the exclusion of some factors and the regulation of those retained and often
are "accelerated" by altering some factor so as to remove it from the range
possible in nature. Consequently, a laboratory procedure often results in
symptoms different from symptoms encountered in a field example of the
process the test is intended to simulate.
Samples of field concrete, when examined using light microscopy, frequently are found to contain secondary calcium carbonate near their outer
surfaces, along old cracks, and sometimes in the interior. Such calcium carbonate, when examined by optical methods, is generally found to be calcite,
rarely aragonite, and almost never vaterite, the form-birefringent spherulitic
calcite with interstitial water. Vaterite, however, was found by optical
methods to be common on mortar bars that had been tested according to
ASTM Test for Potential Alkali Reactivity of Cement-Aggregate Combinations (C 227) and had been found on concrete specimens tested for resistance
to freezing and thawing according to ASTM Test for Resistance of Concrete
Specimens to Rapid Freezing and Thawing in Water (C 666). The use of
X-ray diffraction to examine cement-paste concentrates from field concrete
has revealed that vaterite, not recognized by optical methods, is frequently a
major constituent of the secondary calcium carbonate [9], especially on
samples from seawater exposures or from other wet environments. Vaterite is
known to persist for several months in laboratory specimens stored in room
conditions. The sequence from poorly crystallized vaterite, calcite, and
aragonite to well-crystallized calcite in the carbonation of pastes and mortars
has been clarified by Cole and Kroone [10], and vaterite is now known as a
natural mineral [11].
Accelerated freezing and thawing in water according to ASTM Method C
666, Procedure A, produces a characteristic loss of surface skin and loss of
mortar, which is not like the condition of specimens exposed on the meantide rack at Eastport, Maine [12]. Field concrete that is not air-entrained and
is deteriorated by natural freezing and thawing develops sets of subparallel
cracks normal to the placing direction of the concrete or deteriorated regions
144
parallel to the nearest free surface. These p h e n o m e n a are not reproduced in

accelerated freezing and thawing in water.
Field concrete that has deteriorated primarily because of the alkali-silica
reaction usually has much more advanced and conspicuous internal symptoms of this reaction than are f o u n d in m o r t a r bars o f expansive combinations examined after test according to A S T M M e t h o d C 227. O n the other
hand, some field concrete regarded as undeteriorated has shown a range of
evidence of alkali-silica reaction.
Alkali-carbonate and alkali-silica reactions exist together in varying
degrees of development in some concretes, and inconspicuous degrees of
reaction may be the only recognizable peculiarities in cases of unsatisfactory
service with possibly expensive consequences. The several alkali-carbonate
reactions are described in other papers in this volume, as is alkali-silica reaction.
Closure
" T h i n g s which are s e e n " - - c o n c r e t e and m o r t a r - - " w e r e not made of
things which do a p p e a r " [13] to the eye and to the light microscope. The use
of X-ray diffraction, electron microscopy, differential thermal analysis, and
thermogravimetry, with the electron probe and infrared spectroscopy, in conjunction with the observing eye and the light microscopy, still offers the
chance of sorting out the qualitative and quantitative differences in hydration products and in submicroscopic fabric that are related to serviceable and
deteriorated concrete.
References
[1] "Evaluation of Methods of Identifying Phases of Cement Paste," Transportation Research
Circular No. 176, W. L. Dolch, Ed., Transportation Research Board, National Research
Council, Washington, D.C., June 1976.
[2] "Guide to Compounds of Interest in Cement and Concrete Research," Special Report
No. 127, Highway Research Board, National Research Council, Washington, D.C., 1972.
[3] Proceedings, Fifth International Symposium on the Chemistry of Cement, Tokyo, 1968,
Vols. I-IV, 1969.
[4] Proceedings, Sixth International Symposium on the Chemistry of Cement, in four volumes,
1974; Moscow Stroyizdat 1976; in part available in English preprints which were distributed before the meeting to registrants.
[5] Della M. Roy, Ed., Cement and Concrete Research, an International Journal, bimonthly,
Pergamon Press, 1971.
[6] Structural Petrology, Memoir 6, Geological Society of America, 1938, p. 12, translation
of B. Sander, Gefiigekunde der Gesteine.
[7] "Portland Cement Plants, U.S., Canada and Mexico," Pit and Quarry Publications, Inc.,
Chicago, Ill., 1973.
[8] Lossing, F., "Sulfate Attack on Concrete Pavements in Mississippi," Symposium on
Effects of Aggressive Fluids on Concrete, Highway Research Record No. 113, Highway
Research Board, National Academy of Sciences, 1966, pp. 88-102.
[9] Mather, K. and Mielenz, R. C., "Cooperative Examination of Cores from the McPherson
Test Road," Proceedings, Highway Research Board, 1960, pp. 205-216.
145
[10] Cole, W. F. and Kroone, B., "Carbonate Minerals in Hydrated Portland Cement,"
Nature, Vol. 184, 1959, p. B.A.57.
[11] McConnell, J. D. C., "Vaterite from Ballycraigy, Larne, Northern Ireland," Mineralogical
Magazine, Vol. 32, No. 250, Sept. 1960, pp. 535-544.
[12] Kennedy, T. B. and Mather, K., "Correlation Between Laboratory Accelerated Freezing
and Thawing and Weathering at Treat Island, Maine," Proceedings, American Concrete
Institute, Vol. 50, 1953, pp. 141-172.
[13] St. Paul, "Epistle to the Hebrews," Chapter 11, Verse 3, The Holy Bible, New Testament,
King James Version, originally published 1611, American Bible Society, New York.
K . N. D e r u c h e r 1
Chapter 12 Strength
Introduction
Whether used in buildings, bridges, pavements, or any other of its
numerous areas of service, concrete must have strength, the ability to resist
force. The forces to be resisted may result from applied loads, from the
weight of the concrete itself, or, more commonly, from a combination of
these. Therefore, the strength of concrete is taken as an important index
of its general quality. Hence, tests to determine strength are the most
common type made to evaluate the properties of hardened concrete, because
(a) the strength of concrete, in compression, tension, shear, or a combination
of these, has in most cases a direct influence on the load-carrying capacity
of both plain and reinforced structures; (b) of all the properties of hardened
concrete, those concerning strength usually can be determined most easily;
and (c) by means of correlations with other more complicated tests, the
results of strength tests can be used as a qualitiative indication of other
important properties of hardened concrete [1]. 2
The results of tests on hardened concrete usually are not known until
it would be very difficult to replace any concrete which is found to be faulty.
These tests, however, have a policing effect on those responsible for construction and provide essential information in cases where the concrete
forms a vital structural element of any building. The results of tests on
hardened concrete, even if they are known late, help to disclose any trends
in concrete quality and enable adjustments to be made in the production
of future concrete [2].
1professor, Department of Civil Engineering, University of Maryland, College Park,

Md. 20742.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
DERUCHER ON STRENGTH
147
Significance of CompressiveStrength
Concrete is used in many ways and is subject to a variety of different
loading conditions, and so different types of stress develop. Very often
the dominant stress is compressive in nature, since this material has long
been known to exhibit its best strength characteristics when subjected to
compressive loading. The compressive strength of concrete, one of its most
important and useful properties and one of the most easily determined, is
indicated by the unit stress required to cause failure of a specimen. Concrete also exhibits tensile and shear strength, in which compressive strength
is frequently used as a measure of these properties. The tensile strength
of concrete is roughly 10 to 12 percent of the compressive strength, and the
flexural strength of plain concrete, as measured by the modulus of rupture,
is about 1S to 20 percent of the compressive strength.
In addition to being a significant indicator of load-carrying ability,
strength is also indicative of other elements of quality concrete in a direct
or indirect manner. In general, strong concrete will be more impermeable,
better able to withstand severe exposure, and more resistant to wear. On
the other hand, strong concrete may have greater shrinkage and susceptibility to cracking than a weaker material.
Finally, the concrete-making properties of the various ingredients of the
mix are usually measured in terms of the compressive strength.
Specimens
Specimens to determine the compressive strength of concrete are obtained
generally from four different sources: (a) cylinders made in the laboratory,
(b) cylinders made in the field, (c) cores of hardened concrete cut from
structures, and (d) portions of beams broken in flexure. Each type of
specimen has a specific purpose or purposes.
ASTM Making and Curing Concrete Test Specimens in the Laboratory
(C 192) describes in detail methods for preparation and examination of the
constituent material; proportioning and mixing or concrete; determining
the consistency of the mix; and molding, curing, and capping of the
specimens. Cylinders made in the laboratory constitute a large portion of
the compression specimens, for the following reasons: (1) in research, to
determine the effect of variations in materials or conditions of manufacture
storage, or testing on the strength and other properties of concrete; (2) as
control tests in conjunction with (a) tests on plain or reinforced concrete
members or structures, or (b) tests to determine other properties of hardened
concrete; and (3) to evaluate mix designs for laboratory or field use. In the
making of such specimens large variations can be introduced into the
148
results of the compression test if great care is not taken in the manufacture
of the specimen. These variations may be attributed to the character of the
cement, conditions of mixing, character and grading of the aggregate, size
of the aggregate, size and shape of the specimen, curing and aging, temperature, and moisture content at the time of testing.
ASTM Making and Curing Concrete Test Specimens in the Field (C 31)
describes the detailed method of making standard cylinders in the field.
When cylinders are prepared in the field, they should be made from the
same concrete used on the job. In addition, the same curing process, or
as closely as possible, should be used. The purpose of cylinders made in
the field may be to check the adequacy of the laboratory mix design, to
determine when a structure may be put in service, or to measure and
control the quality of the concrete.
ASTM Test for Compressive and Flexural Strength of Concrete Under
Field Conditions (C 683) describes the methods by which concrete may be
tested if made in the field. The accuracy by which the test is made is much
less than a laboratory test ___3 percent but is considered appropriate.
Compressive test results of cored hardened concrete usually result in
lower compressive strength than anticipated. The ASTM Obtaining and
Testing Drilled Cores and Sawed Beams of Concrete (C 42) covers the
procedure for securing and testing the cylindrical cores which are used
most commonly for determining compressive strength. Cores are drilled
only when results of the standard cylinder tests are questionable or when
investigations are made of old structures.
Finally, ASTM Test for Compressive Strength of Concrete Using Portions
of Beams Broken in Flexure (C 116) describes the procedure and apparatus
necessary to determine compressive strength of concrete from broken portions of beams tested in flexure. This test is extremely useful where beam
specimens are made to determine the modulus of rupture, as in highway
construction, of which one would like an approximate value of the compressive strength. The method is not meant to be used as a comparison
with laboratory cylinder tests. When the method is used and a correlation
attempt is made, a correction factor must be applied [3].
Making Specimens
The compressive strength of concrete depends primarily on the water/
cement (w/c) ratio. However, other factors, such as character of the cement, conditions of mixing, character and grading of the aggregate, size
of the aggregate, size and shape of the specimen, curing and aging, temperature, and moisture content at time of testing also have a bearing on
the compressive strength [4].
The characterization of the cement for a given w/c ratio plays an important role in the early compressive strength development of concrete. All
149
portland cements behave more or less similarly, although the gain in strength
with age is not always the same. Some cements gain their strength more
rapidly at first, whereas others show greater increase at later periods. This
applies not only to the five types covered by ASTM specifications, but to
some extent to different cements within a single group. Tests have shown
that the strength for a given w/c ratio show the greatest difference among
cements at the early ages. For 90 days and later the differences are much
less [5, 6].
The importance of thorough mixing for the development of strength
and for uniformity throughout the batch has long been recognized. The
earliest studies in concrete showed increases in strengths with continued
mixing, but the increase became slight after a first rapid rise. The size of
the batch, the type and consistency of the concrete, and the type of mixer
are all involved in fixing the period in which gain in strength with time
of mixing is significant. The time of mixing is governed by the ASTM
Specification for Ready-Mixed Concrete (C 94).
Surface conditions, the size and shape of the particles, and the gradation
are the characteristics of the aggregate which are of principal concern to
the strength of concrete. The surface conditions of the aggregate affect
the adhesion of the cement paste to the aggregate particles. The presence
or absence of adherent dirt or clay, the roughness, and the texture affect
the adhesion. These characteristics have a greater effect on flexural strength
than on compressive strength [7]
The shape of the particles influence the strength of the concrete by
affecting the quality and the amount of paste that is required for workability with a given mixture. Also, the bond with the cement paste may be
weakened where relatively large surface areas of the flat pieces of aggregate
occur, especially when they happen to be or combined in planes of shear
and tension.
When the w/c ratio is the same and mixtures are plastic and workable,
considerable changes in grading will affect the strength of the concrete
only to a small degree [8]. The principal effect of changing the aggregate
grading is to change the amount of cement and water needed to make the
mixture workable with the desired w/c ratio.
In general, as the maximum size of the aggregate is increased, lower w/c
ratios can be used for suitable workability and, therefore, greater strengths
are obtained for a given cement content., In the high strength range, over
31 MPa (4500 psi), higher compressive strengths usually are obtained at
a given w/c ratio with smaller maximum sizes of aggregates. Data from
compression tests of concrete containing very large aggregates, 10.16 cm
(4 in.), are conflicting because of limitations in the size of specimens [9].
Compression tests of concrete are conducted ordinarily on cylindrical
specimens with height equal to twice the diameter so that surface rupture,
produced upon fracture, will not intersect the end bearings. The ends of
150
the cylinders should be formed carefully to give parallel, smooth surfaces

so as to obtain uniform distribution of stress. A uniformly stressed cylinder
which has been molded properly will break in the shape of a double cone
with vertex in the center of the cylinder [I0].
It is accepted generally that the diameter of the specimen should be at
least three times the nominal size of the coarse aggregate. A 15.24 by 30.48
cm (6 by 12 in.) cylinder is the standard for aggregate smaller than 5.08 cm
(2 in.). If the aggregate is too large for the size of mold available, the oversize aggregate may be removed by wet screening [1I ,12]. If a mold having a
diameter less than three times the maximum size of the aggregate is used,
the indicated compressive strength will be lowered. However, a larger mold
may be used; in some cases molds as large as 91.44 cm (36 in.) in diameter
have been used for concrete containing very large aggregates, such as those
used in dam construction. The size of the cylinder itself affects the observed
compressive strength, for example, the strength of a cylinder 91.44 cm (36
in.) in diameter by 182.88 cm (72 in.) high may be only about 82 percent
of that of a standard 15.24 by 30.48 cm (6 by 12 in.) cylinder. A reduction
in the size of the specimen below that of the standard 15.24 by 30.48 cm
(6 by 12 in.) cyclinder will yield a somewhat greater indicated compressive
strength [13].
Unless the specimens are molded carefully, erratic and irregular results
will be obtained. Generally, a cylinder of poorly compacted concrete will
have a lower strength than one that is compacted properly. Thus, it is
necessary for the standards for making specimens, C 31 and C 192, to
specify procedures for compacting, rodding, or vibrating the concrete in
the mold. If the specifications under which the work is being performed do
not state the method of consolidation, then the choice is determined by
the slump. Concrete with a slump greater than 7.62 cm (3 in.) should be
rodded. If the slump is between 2.54 and 7.62 cm (1 and 3 in.), the concrete may be either vibrated or rodded. When the slump is less than 2.54 cm
(1 in.), the specimens must be consolidated by vibration. When the concrete is to be rodded, it should be placed in the cylinder in three layers and
rodded 25 strokes per layer if the cylinder is 7.62 to 15.24 cm (3 to 6 in.)
in diameter, 50 if 20.32 cm (8 in.), and 75 if 25.40 cm (10 in.). When
vibrating, the mold is filled and vibrated in two layers [14]. Care must be
exercised to vibrate only long enough to obtain proper consolidation. Overvibration tends to cause segregation. These methods are specified in order
to permit reproducibility of results by different technicians [15].
Cylinder molds should be of nonabsorbent material and are generally
of steel; however, cardboard molds are used quite often in the field. The
ASTM Specification for Molds for Forming Concrete Test Cylinders
Vertically (C 470) defines adequate paper molds as well as lightweight
sheet steel molds. Although the cardboard is paraffinned heavily, in most
cases it absorbs part of the water in the concrete mixture. The use of cardCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:42 EDT 2014
151
board molds may lower the observed compressive strength on the average
about 3 percent, and reduction as great as 9 percent have been noted.
Capping procedures are standardized under ASTM Capping Cylindrical
Concrete Specimens (C 617). Any material which is sufficiently strong and
can be molded can be used for capping. The most common capping materials
are neat portland cement paste, high-strength gypsum plaster, and sulfur
compounds [16-18]. The cap should be as thin as possible and the plane
surface of the cap at either end of the specimen should be truly at right
angles to the axis of the cylindrical specimen. All surfaces that depart from
a plane by more than 0.005 cm (0.002 in.) should be capped, and all caps
should be checked for this by means of a feeler gage and steel straightedge.
Slight irregularities can be made good by scraping if the capping material
has not set too hard [19]. Further, according to ASTM Method C 617 the
surface of the capping must not depart by more than 0.5 deg from perpendicularity with the axis of the cylinder or a cant of I in 96.
If neat portland cement is used for capping, it should be mixed about 3
h before use to minimize any harmful effect due to shrinkage. The capping
procedure should be carried out at least 3 days before testing [20].
According to ASTM Method C 617, gypsum plaster can be used for capping provided it has a strength of over 31 MPa (S000 psi) when tested as a
5.08 cm (2 in.) cube. Suitable mixtures of sulfur (melted) and granular
materials applied about 2 h prior to testing are recommended also, but care
is necessary to avoid overheating the melted mixture to prevent loss of
rigidity.
Concrete can gain in strength only as long as moisture is available and
used for hydration. The term "curing" is used in reference to the maintenance of a favorable environment for the continuation of the chemical
reactions which take place. It is through the early curing process that the
internal structure of the concrete is built up to provide strength and water
tightness. While simply retaining moisture within the concrete may be
sufficient for low to moderate cement contents, mixes that are rich in
cement generate considerable heat of hydration which may expel moisture
from the concrete in the period immediately after setting. The standard
curing conditions require that the specimen be held at a temperature of 23
+ 1.7~ (73.4 + 3~ and in the "moist conditions" until the time of the
test (ASTM Method C 192). Any variation from this procedure may produce
a specimen having a different strength from that which would be produced
under standard conditions.
A method of accelerated curing concrete is given by ASTM Making,
Accelerated Curing, and Testing of Concrete Compression Test Specimens
(C 684). In this method, three procedures for making, curing, and testing
specimens of concrete stored under conditions intended to accelerate the
development of strength are covered. The choice of which procedure to use
should be made by the user on the basis of his experience and local conCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:42 EDT 2014
152
ditions. The three procedures are: warm water, boiJmg water, and autogenous curing.
Cylinders to be used for quality control should be cured according to the
standard conditions; however, cylinders made in the field and tested to
measure the strength of the concrete in the structure should be cured in
the same manner as the structure. Two methods of field curing presently
are used: (a) one which interposes a source of water, in the form of ponding,
or a wet material to prevent or counteract evaporation a n d , (b) one which
minimizes loss of water by interposing an impermeable medium or by other
means. A t h i r d method of curing which is used in the manufacture of
concrete products is the artificial application of heat while the concrete is
maintained in a moist condition [21].
Curing and aging cannot be separated; an increase in age provides for
further chemical combinations if the conditions are favorable for continued
reaction. Provided favorable conditions are met, concrete gains in strength
with age.
Temperature also plays an important role in the curing process of concrete. The chemical reactions proceed more rapidly at higher temperatures
[22]. Tests of specimens sealed against loss of moisture show higher early
strength but lower strengths at later ages as the temperature is increased in
the range of 4.4 to 26.1~ (40 to llS~
The U. S. Bureau of Reclamation
has found that for job control, specimens cured at 21.2~ (70~
lower
temperatures at the time of casting and for a few hours thereafter, give
higher strengths at 1 to 3 months [23]. The rapid stiffening in the first few
hours under the higher temperatures apparently is detrimental to the later
development in strength.
Test Procedure
Once the specimen is made, the method by which it is tested may further
affect the strength obtained. Two of the more important influences are
the rate of loading and the eccentricity of loading.
The rate of loading has a definite effect on compressive strength, although
the effect is usually fairly small over the ranges of speed used in ordinary
testing. The results of tests on concrete indicate that the relationship between strength and rate of loading is approximately logarithmic; the more
rapid the rate, the higher the indicated strength [24,25]. A rapid rate of
loading may indicate as much as a 20 percent increase in the apparent
compressive strength. For this reason, ASTM Test for Compressive Strength
of Cylindrical Concrete Specimens (C 39), which applies also to testing of
cores, specified that the rate of loading for screw-powered machines shall
be 0.11 c m / m i n (0.05 in. 1 min) and for hydraulic machines 138 to 345
kPa/s (20 to 50 psi/s).
The effect of eccentric loading is obvious and the alignment of all machines should be checked. Any eccentricity will tend to decrease the strength
153
of the specimen, the amount of decrease being greater for low-strength

than for high-strength concrete [26]. Therefore, to ensure that a concentric
and uniformly distributed load is applied to the specimen, a spherically
seated bearing block is required on one end, and the specimen should be
centered carefully on this bearing block. The object of the block is to overcome the effect of a small lack of parallelism between the head of the machine and the end face of the specimen, giving the specimen as even a
distribution of initial load as possible. It is desirable that the spherically
seated bearing block be at the upper end of the specimen. In order that the
resultant of the forces applied to the end of the specimen should not be
eccentric with the axis of the specimen, it is important that the center of the
spherical surface of this block be in the flat face that bears on the specimen, and that the specimen itself be centered carefully with respect to the
center of this spherical surface. Owing to increased frictional resistance as
the load builds up, the spherically seated bearing cannot be relied upon to
adjust itself to bending action that may occur during the test [27].
Significance of Results
The results of a compression test are essentially only comparative as the
value of the compressive strength obtained cannot be regarded as equal
to the strength of the concrete deposited in the work. The value will give
only an indication of the quality of concrete. The reason for this is due to
the factors affecting the mix which have already been discussed. This lack
of knowledge regarding the relationship between the strengths of concrete
in a cylinder and in a structure requires the use of a larger factor of safety
than would otherwise be necessary.
Compressive strength may be used as a qualitative measure of other
properties of hardened concrete. No exact relationship exists between compressive strength and flexural strength, tensile strength, modulus of elasticity, wear resistance, fire resistance, or permeability [28]. Only an approximation can be made of these properties. Nevertheless, this approximation
is very useful to the engineer.
Compressive tests further aid in the selection of ingredients that may be
used in making concrete. Compressive strength is a measure of the indirect
effect of admixtures which may be beneficial for one purpose, but detrimental to another property [29].
Tensile and Flexural Strength
Significance of Tensile and Flexural Strength

Flexural tension is developed most commonly in beams and slabs as the
result of loads, temperature changes, shrinkage, and in some cases moisture
changes. The case of simple uniaxial tension rarely is encountered in strutCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:42 EDT 2014
154
tures or members and can be obtained in laboratory tests only with care.
However, significant principal tension stresses may be associated with
multiaxial states of stress in walls, shells, or deep beams.
As concrete which has to withstand tensile stresses normally is reinforced,
its tensile strength has not received much attention, although it is of great
importance in determining the ability of concrete to resist cracking due
to shrinkage on drying and thermal movements. The tensile strength
develops more quickly than the compressive strength and is usually about
one-tenth the compressive strength at ages up to about 14 days, falling
to about 5 percent at later ages. Cracking of concrete is usually a tensile
failure and this alone makes the tensile strength of concrete quite important.
Specimens
It is not easy to perform an axial tension test on concrete. It is difficult

to apply the load truly axially and to grip the ends of the specimen without
imposing high local stresses; a relatively large specimen must be used if a
measurable load is to be applied and a large number of specimens are
required to ensure a reliable average. The beam test for flexural tension
and split cylinder test are the simplest procedures for determining the tensile strength and are the tests usually performed.
Flexural tension tests may be made in several ways, the most common
being ASTM Test for Flexural Strength of Concrete (Using Simple Beam
with Third-Point Loading) (C 78). ASTM Test for Flexural Strength of
Concrete (Using Simple Beam with Center-Point Loading) (C 293) also is
used to a limited extent. The second method is intended for use with small
specimens and not as an alternative to the test with third-point loading.
The results of the flexural tests are expressed by the formula
R -= M c / I
where in consistent units

R
M
c
I
= modulus of rupture,
= maximum bending moment,
= one-half the depth of the beam, and
---- moment of inertia of the cross section.
Another procedure for obtaining an indication of tensile strength is given

in ASTM Test for Splitting Tensile Strength of Cylindrical Concrete Specimens (C 496). In this test a standard cylinder is loaded in compression
on its side. Fracture occurs along the plane which includes both lines
through which the load is applied. While high compressive stresses occur
at the lines where load is applied, the plane on which fracture occurs is
155
subjected largely to a uniform tensile stress. The splitting strength is calculated as follows
T = 2P/Trld
T
P
l
d
=
=
=
=
splitting tensile strength,

maximum applied load,
length, and
diameter.
Making Specimens
ASTM Methods C 192 and C 31 describe the procedures for making
flexural specimens and also cylinders for the splitting tensile strength test
in the laboratory and in the field. ASTM Method C 31 stipulates that the
length of the beam should be at least 5.08 cm (2 in.) longer than three times
its depth and that its width should be not more than one and one-half
times its depth. The minimum depth or width should be at least three times
the maximum size of aggregate. A typical speciman used would be 15.24
by 15.24 by 53.34 cm (6 by 6 by 21 in.), and is tested under third-point
loading on a span of 45.72 cm (18 in.). As the depth of the beam is increased, there is a decrease in the modulus of rupture [30, 31].
Many of the parameters which affect the compressive strength of concrete also apply to the flexural strength as well.
Test Procedure
Flexural strength measurements are extremely sensitive to all aspects
of specimen preparation and testing procedure [32].
The principal requirements of the supporting and loading blocks of the
apparatus for the flexural strength of concrete are as follows: (1) they
should be of such shape that they permit use of a definite and known
length of span; (2) the areas of contact with the material under test should
be such that unduly high stress concentrations (which may cause localized
crushing around bearing areas) do not occur; (3) there should be provision
for longitudinal adjustment of the position of the supports so that longitudinal restraint will not be developed as loading progresses; (4) there should
be provision for some lateral rotational adjustment to accommodate beams
having a slight twist from end to end, so that torsional stresses will not be
induced; (5) the arrangement of parts should be stable under load.
Apparatus for measuring deflection should be so designed that crushing
at the supports, settlement of the supports, and deformation of the supportCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:42 EDT 2014
156
ing and loading blocks or of parts of the machine do not introduce serious
errors into the results. One method of avoiding these sources of errors is
to measure deflections with reference to points on the neutral axis above
the supports.
The conduct of routine flexure tests is usually simple. Ordinarily only
the modulus of rupture is required; this is determined from the load at
rupture and the dimensions of the specimen. When the modulus of elasticity is required, a series of load-deflection observations are made.
The dimensions of the concrete specimen should be measured to the
nearest 0.025 cm (0.01 in.). The supporting and loading blocks should be
located with a reasonable degree of accuracy (0.2 percent of the span length).
The assembly of supports and specimens should be placed centrally in the
testing machine and should be checked to see that they are in proper alignment and can function as intended. Deflectometers and strainometers
should be located carefully and checked to see that they operate satisfactorily and are set to operate over the range required [33].
The rate of load application, unless standardized, may cause considerable variation in the results of flexure tests, the variation being as much
as 15 percent for the range of rates that may be obtained in the average
laboratory [34]. Concrete beams may be loaded rapidly at any desired rate
up to 50 percent of the breaking load, after which loads should be applied
at a rate such that the extreme fibers are stressed at 0.93 MPa (150 psi)
or less per minute.
Beams may be tested under either center-point or third-point loading.
Third-point loading invariably gives lower strengths than center-point
loading. Tests indicate the following order of decreasing magnitude of the
strength obtained: (1) center loading, with moment computed at center; (2)
center loading, with moment computed at point of fracture; and (3) thirdpoint loading [35]. Third-point loading probably gives lower strengths because the maximum moment is distributed over a greater length of the
beam; since the concrete is not homogeneous, this loading method seeks
the weakest section.
Curing effects the tensile strength in much the same manner as it affects
the compressive strength [36]. A beam that has been allowed to dry before
testing will yield lower flexural strength than those tested in a saturated
condition. Consequently, in tests to determine or control the quality of concrete, uniformity of results will be assured only if the beams are cured in a
standard manner and tested while wet.
The temperature of a beam at the time of testing will also affect the resuits. As the temperature increases, the strength decreases [37].
ASTM Test C 78 is also prescribed for tests of beams sawed from hardened concrete. When such beams are used primarily as a control of concrete quality, they should be turned on their sides before testing and will
157
usually require capping because of the irregularity of the sawed surface

(sides are not plane and parallel).
The test for splitting tensile strength described in ASTM Test C 496 is
simple to make. The effectiveness with which the material in the bearing
strips is able to conform to the irregularities of the specimen surface and
distribute the load affects the results. For uniformity ASTM Test C 496
specifies that 0.32 cm (1/8 in.) thick plywood shall be used for bearing
strips. Care must be taken to apply the load through a diametrical plane.
The load should be applied such that the stress increases between 689 and
1379 kPa/min (100 and 200 psi/min.)
Significance of Results
In the case of pavement construction it appears that the flexural strength
of concrete is equally important (or more so) as the compressive strength.
It has been suggested that pavements be designed on the basis of flexural
strength [38]. The design procedure would be exactly the same except that
the tables would have to be prepared on a basis of fiexural strength instead of compressive strength. Many agencies which are involved in pavements only make flexural specimens. The results of the test on these specimens given an indication when the concrete has gained sufficient strength
that load may be applied or the forms removed.
The splitting tension test has been used to evaluate bond splitting resistance of concrete.
With increased emphasis on control of cracking in reinforced concrete
an appreciation of the tensile strength has become more important. However, the three general methods of estimating tensile strength give slightly
different results. The splitting tensile strength is the easiest to perform l~nd
gives the more uniform results. The tensile strength from the splitting tests
is about one and one-half times greater than obtained in a direct tension
test and about two thirds of the modulus of rupture [39].
Shearing and Torsional Strength
Significance of Shearing Strength

The shearing strength of concrete is a most important property of the
material since it is the real determining factor in the compressive strength
of short columns. The strength of concrete beams depends also, under
certain conditions, upon the shearing strength of the material.
The average strength of concrete in direct shear varies from about 0.5 of
the compressive strength for rich mixtures to about 0.8 of the compressive
strength for lean mixtures [40].
158
Torsion
The application of torsion alone to a concrete specimen produces pure
shearing stresses on certain planes. However, the case of pure shear acting
on a plane is seldom if ever encountered in actual structures. Further, failure under these conditions will occur in tension rather than shear. Therefore, the strength of concrete subjected to torsion is related to its tensile
strength rather than to its shearing strength.
Combined Stresses
Significance of Combined Stresses

Concrete in structures is almost never subjected to a single type of stress.
Just as nearly all structural members are acted upon by various combinations
of moments, shear, and axial load, the concrete in them is usually subjected
to some combination of compressive, tensile, and shearing stresses.
ASTM Recommended Practice for Determining Mechanical Properties
of Hardened Concrete Under Triaxial Loads (C 801) is useful in determining
the strength and deformation characteristics of concrete such as shear
strength at various lateral pressures, angle of shearing resistance, strength
in pure shear, deformation modulus, and creep behavior. Prior to the
application of this relatively new standard extensive research was conducted
on cylinders with combinations of axial tension and lateral compression,
and torsion and axial compression [41,42].
Results of Tests
There is no universally accepted criteria of failure of concrete; hence,
there is no single correct way of discussing the behavior of concrete under
combined stresses and many ways in presenting the results. ASTM Recommended Practice C 801 allows four ways in which the data may be presented: (1) graphical plots of the following equations
fl =fi'
+K(f3)"
or, for the strength increase beyond the uniaxial strength

f~ - - f ~ ' = K ( f 3 ) o
jq =
=
f~ =
K, a =
largest principal stress,

smallest principal stress,
unconfined compressive strength, and
empirical coefficients.
159
(2) A graphical plot of the stress difference versus axial strain. Stress difference is defined as the maximum principal axial stress minus the minimum principal stress. The value of the minimum principal stress should
be indicated on the curve. (3) A graphical plot of axial stress versus axial
strain for different confining pressures. (4) Mohr stress circles constructed
on an arithmetic plot with shear stresses as ordinates and normal stresses
as abscissas. At least three triaxial compression tests, each at a different
confining pressure, should be made on the same material to define the
envelope to the Mohr stress circle.
Fatigue Strength
Concrete will, when subjected to repeated load, fail at a load smaller
than its static ultimate strength. Early work on fatigue was conducted on
low strength concrete [43]. Thus in the early classic work of Probst [44]
the strength of the concrete tested was only just over 12.4 MPa (2000 psi).
His results nevertheless agree with more recent work conducted on concretes of much higher strength [45-47]. He discovered that there was a
critical stress below which concrete subjected to repeated loading increased
in strength and elasticity. His minimum stress was 5 percent of the ultimate
and the maximum or critical stress ranged from 47 to 60 percent of the
ultimate. Further tests have shown the fatigue strength at ten million cycles
is about 55 percent of the static flexural strength. The fatigue strength in
axial compression is about 55 percent of the static compressive strength.
No data are available to establish the fatigue strength in axial tension, in
shear, or under combined stresses. Concrete does not have a fatigue or endurance limit at least at less than ten million cycles of load as most bodycentered cubic metal does. Failure under repeated loads is especially important in pavement design.
Frequent rest periods during a fatigue test may raise the fatigue strength
as much as 9 percent higher than if there had been no rest periods. The
fatigue strength increases as the rest periods are increased in duration to
5 min. No additional gain is obtained for longer rest periods [48].
Tests to determine the fatigue strength of concrete should be made on
specimens as large as possible in order to decrease the influence of lack
of homogeneity. The cross section of the specimen should be at least three
times the maximum nominal size of the aggregate, and even larger dimensions might be desirable in some cases. Because of the large size of specimens used, the large testing machines required are usually capable of
applying load at a rate ranging from only a few cycles a day to 500 cpm.
Thus, it takes a minimum of two weeks to several months to apply as many
as ten million cycles of load. Because of the time involved, the specimens
generally are aged and air dried before being tested in order to prevent
gain of strength during the test.
160
References
[1] Kesler, C. E., "Strength of Hardened Concrete," Significance of Tests and Properties
of Concrete and Concrete Making Materials, ASTM STP 169A, American Society for
Testing and Materials, 1966.
[2] Orchard, D. F., Concrete Technology, Wiley, New York, 1973.
[3] Mather, Bryant, "Effect of Type of Test Specimen on Apparent Compressive Strength
of Concrete," Proceedings, American Society for Testing and Materials, Vol. 45, 1945,
p. 802.
[4] Mills, A. P., Hayward, H. W., and Rader, L. F., Materials of Construction, Wiley,
New York, 1955.
[5] Lea, F. M., "Modern Developments in Cement in Relation to Concrete Practice,"
Journal of the Institute of Civil Engineers, Feb. 1943.
[6] Keil, F., "Cements," Third International Symposium on the Chemistry of Cement, London, Sept. 1952.
[7] Krebs, R. D. and Walker, R. D., Highway Materials, McGraw Hill, New York, 1966.
[8] Neville, A. M., Properties of Concrete, Wiley, New York, 1963.
[9] McMillan, F. R., "Suggested Procedure for Testing Concrete in Which the Aggregate
is More Than One-Fourth the Diameter of the Cylinders," Proceedings, American Society
for Testing and Materials, Vol. 30, Part I, 1930, p. 521.
[10] Kesler, C. E., "Effect of Length to Diameter Ratio on Compressive Strength--An ASTM
Co-operative Investigation," Proceedings, American Society for Testing and Materials,
Vol. 59, 1959, p. 1216.
[11] McMillan, F. R., "Suggested Procedure for Testing Concrete in Which the Aggregate
is More Than One-Fourth the Diameter of the Cylinders," Proceedings, American Society
for Testing and Materials, Vol. 30, Part I, 1930, p. 521.
[12] "Effect of Wet-Screening to Remove Large Size Aggregate Particles on the Strength
of the Concrete," Corps of Engineers, Ohio River Division Laboratories, Mariemont,
Ohio, Jan. 1953.
[13] Kesler, C. E., "Effect of Length to Diameter Ratio on Compressive Strength--An ASTM
Cooperative Investigation," Proceedings, American Society for Testing and Materials,
Vol. 59, 1959, p. 1216.
[14] Forssblad, L., "Investigations of Internal Vibration of Concrete," Acta Polytechnia
Scandinavia, Civil Engineering and Building Construction Series No. 29, Stockholm,
1965.
[15] Tuthill, L. H. and Davis, H. E., "Over-Vibration and Re-Vibration of Concrete,"
lndian Concrete Journal, Vol. 20, 1946.
[16] Kennedy, T. B., "A Limited Investigation of Capping Materials for Concrete Test Specimens," Proceedings, American Concrete Institute, Vol. 41, 1944, p. 117.
[17] Werner, George, "The Effect of Type of Capping Material on the Compressive Strength
of Concrete Cylinders," Proceedings, American Society for Testing and Materials, Vol.
58, 1958, p. 1166.
[18] Werner, George, "The Effect of Type of Capping Material on the Compressive Strength
of Concrete Cylinders," Report of the Bureau of Public Roads presented to the SixtyFirst Annual Meeting of the American Society for Testing and Materials, June 1958.
[19] Gonnerman, H. F., "Effect of End Condition of Cylinder in Compression Tests of Concrete," Proceedings, American Society for Testing and Materials, Vol. 24, Part II, 1924,
p. 1036.
[20] "Methods of End Conditions Before Capping Upon the Compressive Strength of Concrete Test Cylinders," Proceedings, American Society for Testing and Materials, Vol.
41, 1941, p. 1038.
[21] Bergstrom, S. G., "Curing Temperature, Age and Strength of Concrete," Construction
Review, Vol. 27, No. 3, July, 1954.
[22] Klieger, Paul, "Effect of Mixing and Curing Temperature on Concrete Strength,"
Proceedings, American Concrete Institute, Vol. 54, 1958, p. 1063.
[23] "Curing Concrete Specimens," U.S. Bureau of Reclamation, Special Report 16, 1954.
[24] Watstein, D., "Effect of Straining Rate on the Compressive Strength and Elastic Properties of Concrete," Proceedings, American Concrete Institute, Vol. 49, 1953, p. 729.
161
[25] Shideler, J. J. and McHenry, D., "Effect of Speed in Mechanical Testing of Concrete,"
Speed of Testing of Nonmetallic Materials, ASTM STP 185, American Society for Testing and Materials, 1955.
[26] Davis, H. E., Troxell, G. E. and Wiskocil, C. T., The Testing and Inspection of Engineering Materials, McGraw-Hill, New York, 1964.
[27] Shank, J. R., "Plastic Flow of Concrete at High Overload," Proceedings, American
Concrete Institute, Vol. 45, 1949, p. 493.
[28] Kesler, C. E., "Statistical Relation Between Cylinder Modified Cube, and Beam Strength
of Plain Concrete," Proceedings, American Society for Testing and Materials, Vol. 54,
1954, p. 1178.
[29] Nash, J. P., "Tests of Concrete Road Aggregates," Proceedings, American Society for
Testing and Materials, Vol. 17, Part If, 1917, p. 394.
[30] Wright, P. J. F. and Garwood, F., "The Effect of the Method of Test on the Flexural
Strength of Concrete," Magazine of Concrete Research, No. 15, March 1954, p. 105.
[31] Nielson, K. E. C., "Effect of Various Factors on the Flexural Strength of Concrete Test
Beams," Magazine of Concrete Research, No. 15, March 1954, p. 105.
[32] Walker, Stanton and Bloem, D. L., "Studies of Flexural Strength of Concrete," Part 2:
"Effects of Curing and Moisture Distribution," Proceedings, Highway Research Board,
Vol. 36, 1957, p. 334; "Part 3: Effects of Variations in Testing Procedures," Proceedings,
American Society for Testing and Materials, Vol. 57, 1957, p. 1122.
[33] Goldbeck, A. T., "Apparatus for Flexural Tests of Concrete Beams," Report of ASTM
Committee C-9 on Concrete and Concrete Aggregates, Appendix VIII, Proceedings,
American Society for Testing and Materials, Vol. 30, Part 1, 1930, p. 591.
[34] Talbot, A. N. and Richart, F. E., "The Strength of Concrete," Bulletin No. 137, Engineering Experiment Station, University of Illinois, Urbana, Ill., 1923.
[35] Kellerman, W. F., "Effect of Size of Specimen, Size of Aggregate, and Method of Loading Upon the Uniformity of Flexural Strength Tests," Public Roads, Vol. XIII, No. 11,
Jan. 1933, p. 177.
[36] Price, W. H., "Factors Influencing Concrete Strength," Proceedings, American Concrete Institute, Vol. 47, 1951, p. 417.
[37] Waters, T., "The Effect of Allowing Concrete to Dry Before It Has Fully Cured," Magazine of Concrete Research, Vol. 7, No. 20, July 1955.
[38] Wright, P. V. F., "The Design of Concrete Mixes on the Basis of Flexural Strength,"
Proceeding of a Symposium on Mix Design and Qualtiy Control of Concrete, Cement
and Concrete Association, London, May 1954.
[39] Wright, P. J: F., "Comments on an Indirect Tensile Test on Concrete Cylinders," Magazine of Concrete Research, No. 20, 1955, p. 87.
[40] Talbot, A. N. "Tests of Concrete: I. Shear; II. Bond," Bulletin No. 8, Engineering
Experiment Station, University of Illinois, Urbana, Ill., 1906.
[41] Richart, F. E., Brandizaeg, A., and Brown, R. L., "A Study of the Failure of Concrete
Under Combined Compressive Stresses," Bulletin No. 185, Engineering Experiment
Station, University of Illinois, Urbana, Ill., 1928.
[42] Smith, F. C. and Brown, R. Q., "The Shearing Strength of Cement Mortar," Bulletin
No. 106, Engineering Experiment Station, University of Washington, Seattle, Wash.,
1941.
[43] Probst, E., "The Influence of Rapidly Alternating Loading on Concrete and Reinforced
Concrete," The Structural Engineer, Vol. 9, No. 10, Oct. 1931 and No. 12, Dec. 1931.
[44] Probst, E., "Further Investigations of Alternating Loads on Concrete," Cement and
Concrete Research. Vol. 31, March 1942.
[45} Nordly, G. M., "Fatigue of Concrete--A Review of Research," Proceedings, American
[46] Murdock, J. W., "A Critical Review of Research on Fatigue of Plain Concrete," Bulletin No. 975, Engineering Experiment Station, University of Illinois, Urbana, Ill., 1965.
[47] Murdock, J. W. and Kesler, C. E., "Effect of Range of Stress on Fatigue Strength of
Plain Concrete Beams," Proceedings, American Concrete Institute, Vol. 55, 1959, p. 221.
[48] LeCamus, B., "Research of Fatigue Strength," Journal, American Concrete Institute,
Vol. 63, No. 1, Jan. 1966.
M . H. Wills ~
Chapter 13--Accelerated Strength

Tests
Introduction
Rapid construction practices throughout the concrete industry have
brought increasing pressure on specifying agencies to assess the quality of
concrete at an earlier age than 7 or 28 days after placement. Currently, the
later age is still specified for compression tests of 152 by 304-mm (6 by 12-in.)
cylinders of concrete as delivered to the job site in order to determine quality.
During the 28-day period, it is not unusual for a multistory building to rise
several floors before the strength tests are conducted. This situation is considered by many to be too precarious for construction to proceed on a sound
technical basis and with adequate assurance of safety. Furthermore, extensive and costly delays are encountered when 28-day test results are low, since
a field investigation may be necessary to verify the load-carrying capacity of
the structure. Further delay is certain if concrete must be reinforced or
replaced. Conversely, economics dictate that work continue without knowing
quality, thus setting the stage for a catastrophic collision between advancing
construction technology and the price owners are willing to pay for their
capital facilities. Surely, an earlier assessment of concrete quality is absolutely essential.
Committee C-9 on Concrete and Concrete Aggregates of the American
Society for Testing and Materials (ASTM) formed Subcommittee C09.02.09
on Accelerated Strength Testing in 1964 to canvass the needs of the concrete
industry and to study the suitability of several procedures being developed by
King [I,2] 2 and Akroyd [3] in Great Britain and Smith and Chojnacki [4l
and Malhotra and Zoldners [5] in Canada for possible standardization. A
positive response to this canvass led to a cooperative test program among
nine laboratories to evaluate three procedures involving the use of either hot
1Supervisor, Concrete and Aggregate Department, Cement Technical Center, Martin
Marietta Cement, Baltimore,Md. 21227.
2Theitalic numbers in brackets refer to the list of referencesappendedto this paper.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
lntcrnational
www.astm.org
WILLS ON ACCELERATED STRENGTH TESTING
163
or boiling water to accelerate the strength development of concrete. Four

additional procedures were considered, also. One of these employed autogenous curing in an insulated container wherein the heat of hydration of the
cement caused the acceleration.
Experimental Program
Each of the procedures was conducted with two types of cement: Type I,
general purpose (or Type II meeting Type I specifications) and Type III, high
early strength. Cement contents were 265, 325, and 385 k g / m 3 (450, 550,
and 650 lb/yd3). They were used in combination with sufficient air-entraining admixture and mixing water to produce an air content from 5.0 to
6.0 percent and slump from 51 to 76 mm (2.0 to 3.0 in.).
Concretes were mixed without a retarder and with a normal dosage of a
water-reducing retarder. Fine and coarse aggregates were graded to a No. 57
size (1 in. or 25.4 mm to No. 4). Each laboratory used materials available in
their locality without interchange between laboratories. In all cases, compressive strength was measured at 1 or 2 days of age and was compared to
strengths developed under standard curing conditions at ages of 28 days and
364 days. All specimens were prepared, cured, and tested according to
standard methods of ASTM. All materials conformed to appropriate ASTM
specifications.
Test Procedures
The nine participating laboratories were to conduct each of the following
procedures:
Procedure A--Hot Water, 35~ (95~

Cylinders immediately put in 35 ~ water for 24 h.
Procedure B--Modified Boiling
At 23-h age, molded cylinders boiled 3.5 h.
Procedure C--Fixed Set Followed by Boiling
At final Proctor set, molded cylinders boiled 15 h.
Additionally, they had the option of conducting one or more of the following
procedures:
Procedure D--Autogenous Curing
At 1-h age, place in insulated container for 46 h.
Procedure E--Fixed Set Followed by Hot Water, 55~ (130~
At initial Proctor set, molded cylinders immersed 15 h.
Procedure F--Fixed Set Followed by Hot Water, 75 ~ (I 75 ~
Procedure G--Fixed Set Followed by Hot Water, 90~ (195~
Despite the plans of the subcommittee, it was necessary to curtail some
164
aspects of the experimental program. Foremost, Procedure C, involving the

measurement of final Proctor set followed by boiling, required too many
overtime man-hours to conduct and was abandoned. Consequently, Procedures E, F, and G were not conducted since they also required measurement of setting time and overtime. Finally, only four of the nine participating
laboratories conducted Procedure D, which was optional. However, all
laboratories conducted the remaining aspects of the program for Procedure
A (Hot Water) and Procedure B (Modified Boiling).
Accelerated Curing Apparatus

The six accelerated procedures involving the use of either hot or boiling
water were conducted in a thermostatically controlled tank, as shown in Fig.
1. The tops of the cylinders were covered by about 102 mm (4 in.) of water.
Water volume and heater capacity were sized to prevent an appreciable reduction in the water temperature desired. Specimens were not placed in a
tank already containing those being cured.
Details of the autogenous curing container are shown in Fig. 2. Polyurethane foam was used to retain the heat of hydration of the cement which,
in turn, accelerated the strength development of the concrete. Once curing
was initiated, the container was not opened.
Procedure A--Hot Water, 35 ~ (95 ~

Immediately after casting, the tightly closed cylinder molds were immersed
INSULATED
rATER TANK
l _ WAT_EE
LE_V_EL__
WATER LEVEL
f,CYLINDERSNN '
5]
51 :: ] 5 ~ m m
5~L
]02ram
mm
mm
311El
METAL PLATE
lO00 WATT HEATING ELEMENT
THERMOSTAT
FIG. 1--Hot or boiling water accelerated curing tank.
C O N C R E T E ~
CYLINDER
165
INNERCONTAINER-SHEETMETAL
162mm iNSIDE die.
B
E L~( _ _ . ~BT H E R
V" ~ . _ ~
=B
ISALMOSTNOAIRSPACE
/ BETWEENCYLINDERAND
,NSULAT,O.
- ~ POLYURETHANEINSULATION
FOAMEDIN PLACE ( k _.O.12)
SECTION A-A
LID FILLEDWITH
POLYURETHANE
RECESSIN LID ~
381turn
FOR MAX.-MIN.~.J
\
FT')
THERMOMETER~
~
OUTERCONTAINER ]~ /
PLASTICGARBAGE IL/
"~/
~
i/ / 1 /
STEEL COVERPLATE
~E
k../
E ___
E
E
>/WLK
O.~LE
318rnm
=[
SECTION B-B
FIG.2--Autogenous curing container.
in 33 to 37~ (92 to 98~ water, where they remained for a period of 24 h _

15 rain. Sulfur mortar caps were applied to the cylinders and aged at least I h
prior to measuring compressive strength at an age of 26 h + 15 min. The hot
water procedure accelerated concrete strengths from 1.1 to 1.6 times those
achieved after a day of standard moist curing.
Graphical comparisons between accelerated, 28-day, and 364-day data are
shown in Figs. 3 and 4 for concretes made with Type I cement. Obviously,
as accelerated strength increased both 28 and 364-day values also increased,
but it appeared that each laboratory was obtaining significantly different
results. Indeed, accelerated strength did seem to relate to 28 and 364-day
strength equally and with good correlation within a given laboratory. Therefore, Procedure A can be used with a high degree of confidence in assessing
concrete quality when the tests are made by the same laboratory.
Although not shown, the same general trends were observed with concretes
made with Type III cement. The relationships were at a higher strength level,
however. Since the test procedures were carefully controlled, they emphasized the impact of materials on the results, particularly the type of cement.
166
(8000}
PROCEDURE
I0
20
HOT WATE'R
/ /
(6)
(4)
10
TYPE'I CEMENT
50
40
30
20
LAB. NO. ]
_>
Go
Go
nn
0
(J
>(:3
~o
(6)
(41
50
40
30
20
LAB. NO. 6
(61
LAB. NO. 5
/ /
"1-
t,-c~
Z
b.I
n,I-u~
LAB. NO. 4
//
LAB. NO. 9
LAB. NO. 8
50
(41
LAB. NO. ]0
(19)
(0)
LAB. NO. ] ]
i
(2)
ACCELERATED
40
50
20
LAB. NO. ]2
i
(0)
STRENGTH,
(2)
(4.000)
MPo (psi)
FIG. 3--Relation of accelerated to 28-day strength acceleration by immersion in 35~
(95~
water.

After casting the cylinders, the sealed molds were placed in a standard
moist room. At 23 h _ 15 min from the time of casting, the cylinders, including molds and covers, were immersed in boiling water. Reduction in
water temperature was limited to 3 ~ (5 ~ and was required to recover to
the boiling point in no more than 15 min. After 3.5 h _ 5 min, cylinders
were removed from the molds and allowed to cool for about 45 min. They
were capped with sulfur mortar at least 1 h before strength was measured, at
which time, their age was between 28.5 4- 15 min. Modified boiling accelerated concrete strength between 1.1 to 2.1 times that measured after a day
of standard moist curing.
Results for concrete cured by modified boiling are shown graphically in
Figs. 5 and 6 for Type III cement. Trends discussed previously are reemphaCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:43 EDT 2014
WILLS ON ACCELERATED
I0
/.
(8000:
(6)
20
~. (4)
lO
20
STRENGTH
/ /
]o
]
TESTING
167
20
i
50
40
3O
20
(z)
I,-CD
(6)
(,'3
(4)
(E
0...
~E
oL ) (2)
>-
,=I
o
4,'.D
(67
to
LAB. NO. I
LAB. NO. 4
LAB. NO. 5
/
-/
/ / //
LAB, NO. 6
LAB. NO. 8 / '
./..
5O
3O
20
LAB. NO. 9
5O
3O
(4:
PROCEDURE A
HOT WATER
LAB. NO. lO
(2{'0)
(2)
LAB. NO. l l
(0)
TYPE TCEMENT z o
LAB. NO, 12
(2)
(0)
(2)
ACCELERATED STRENGTH, MPo (psi)
(40001
FIG. 4--Relation of accelerated to 36d-day strength acceleration by immersion in 35~

(95~ water.
sized. Most important was that the laboratories still obtained what seemed to
be significantly different results. Again, this is attributed to each using
locally available aggregate and cement. However, within a given laboratory,
good correlations of accelerated strength to 28 and 364-day values were
obtained. Therefore, Procedure B was also seen to have equal merit in assessing concrete quality.
Procedure C--Fixed Set Followed by Boiling

Although the fixed set-boiling procedure was abandoned by the subcommittee during the course of the cooperative test program, six laboratories
did conduct it. The procedure's main disadvantage was overtime man-hours,
particularly with the retarded concrete. On the plus side, accelerated comCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:43 EDT 2014
168
TESTS A N D PROPERTIES
20
,
(8000)
30
,
40
OF C O N C R E T E
I/o I
20
50
(6)
40
20
i
30
40
50
f.,
oo
30
~. (4)
no
2O
2~
:ff
LAB. NO. 1
(2',
I-z
o~
(61
>_
m
o')
W
141
LAB. NO. 4
LAB. NO. 5
50
,o
3O
20
0
LAB. NO. 6
(2
~b
('4 (6)
-/
LAB. NO. 8
L A B NO. 9
50
.o
:30
(4)
PROCEDURE B
(2)
{2)
FIG.
LAB. NO. IO
i
MODIFIED BOILINC TYPETn"CEMENT

LAB. NO. ] l
i
20
LAB. NO. 12
I
(4)
(2)
(4)
(2)
(4)
(6000)
S--Relation of accelerated to 28-day strength acceleration by boiling for 3.5 h.
pressive strengths obtained were the highest among the three procedures-A, B, and C--and ranged from 1.6 to 2.1 times the concrete strength measured after a day of standard moist curing. It still appeared that differences
between laboratories were significant but that within a given laboratory,
correlation with later age strength was quite good. This served to build
confidence in the basic premise that the accelerated strength test could be
used on a par with the standard, 28-day strength assessment of concrete
quality.

One hour after mixing, molded cylinders were sealed in plastic bags and
placed inside the autogenous (insulated curing) chamber. They remained
2o
(8000
30
40
20
30
40
20
30
~. (4)
g_
LAB. NO. 1
:I:" (21
(.9
Z
uJ
n,..
I..G)
u,J
>
o9
o3
uJ
(6)
40
30
20
LAB. NO. 5
0
50
40
30
(4)
20
13,.
0(..)
>.<
t-,.
40
(61
169
LAB. NO. 6 '
(2)
LAB. NO. 8
LAB. NO. 9
50
,,.o
ro
(6) -
(4)
PROCEDURE B
(2)
(2)
LAB. NO. lO
i
40
$0
MODIFIED BOILING TYPE ]IT CEMENT. 20

LAB. NO. ] l
LAB. NO. ]2
I
(2)
(4)
(2)
(4)
(4)
(6000)
FIG. 6--Relation of accelerated to 364-day strength acceleration by boiling for 3.5 h.
there for 46 h, after which they were removed from the molds and allowed to
cool for 45 min. Sulfur mortar caps were applied and aged at least 1 h before
strength was measured at an age of 49 h _+ 15 rain.
Only four laboratories performed this procedure but their data were
sufficient to justify including autogenous curing in the ASTM standard as
were Procedures A and B. Accelerated 2-day strengths ranged from 1.4 to 2.5
times those obtained after an equal length of standard moist curing--the
highest level of acceleration measured.
Considering that high concrete temperatures are detrimental to the early
hydration reactions of cement, Procedure D should be most conducive to
producing the highest accelerated strengths. With this procedure, the heat of
hydration of the cement causes the acceleration but at a low rate of temperature increase, which is beneficial to early hydration reactions. Concrete
quality can be assessed with a high degree of confidence using this procedure
170
since 28-day strength levels will be approached closely at 2-days age, compared to 1 day with the other procedures.
Significance of Test Procedures
Being convinced that laboratories, test procedures, and cement types had
significantly different effects on the data, the subcommittee analyzed the
data from that viewpoint [6-8]. The statistical model involved the assumption of separate regression equations for each cement type in the form of
Y = B0 + B~X)
where B0 was the intercept coefficient and B1 was the slope coefficient. Least
squares fits, correlation coefficients, and standard errors are summarized in
Tables 1 through 4 for Procedures A, B, C, and D, respectively.
Based on their analysis, the regression equation for Laboratory 10 using
Procedure B and Type I cement was
Y -----19.48 + 1.195X1MPa
= (19.48 + 1.195X1) (6895 10 -6) psi
TABLE 1--Linear regression analysis for Procedure A.
Simple Linear Regression a
Lab
No.
Cement
Type
B0, MPa
B0, (psi)
BI
SE, MPa
SE, (psi)
I
III
13.54
15.65
(1965)
(2270)
1.120
0.960
0.910
0.975
2.10
0.86
(305)
(125)
I
III
19.10
18.96
(2770)
(2750)
1.515
1.095
0.955
0.925
1.96
2.76
(285)
(400)
I
III
15.27
17.86
(2215)
(2590)
1.375
0.905
0.960
0.980
1.86
1.07
(270)
(155)
I
III
I
III
I
III
14.17
14.31
9.10
15.44
16.31
23.48
(2055)
(2075)
(1320)
(2240)
(2365)
(3405)
1.525
1.290
1.985
1.285
1.320
0.540
0.985
0.940
0.905
0.935
0.765
0,870
0.86
1.93
2.90
2,31
3.69
1.59
(125)
(280)
(420)
(335)
(535)
(230)
10
I
III
19.13
24.20
(2775)
(3510)
1.440
0.770
0.915
0.815
1.62
1.28
(235)
(185)
11
I
III
14.34
13.76
(2080)
(1995)
1.475
1.090
0.905
0.725
2.83
3.62
(410)
(525)
12
I
III
17,65
15.65
(2560)
(2270)
1.440
0.580
0.915
0.825
1,96
2.07
(285)
(300)
8
9
aAccording to the model Y = Bo + BIX, the terms of which are defined in the text.
171
Note t h a t the correlation coefficient was extremely high, 0.950, whereas

1.000 is perfect, a n d s t a n d a r d error c o r r e s p o n d i n g l y low at 1.24 M P a (180
psi). L a b o r a t o r y 8 h a d a m u c h h i g h e r s t a n d a r d error. Therefore, its d a t a d i d
not fit the a s s u m e d m o d e l well. A f t e r studying the s t a n d a r d errors, it was
concluded t h a t all three p r o c e d u r e s were e q u a l in c o r r e l a t i n g a c c e l e r a t e d a n d
later-age strengths. F u r t h e r , it was f o u n d t h a t a c c e l e r a t e d compressive
strength c o r r e l a t e d with 364-day strength as well as 28-day strength. In fact,
the 28-day s t a n d a r d test was j u s t as variable, a situation presently ignored in
assessing concrete quality within a certain l a b o r a t o r y . C o r r e l a t i o n coefficients were quite high in most cases, as w o u l d be expected by t h e close fit of
the d a t a to the linear r e l a t i o n s h i p s shown in Figs. 3 t h r o u g h 6 for each laboratory. Subsequently, the s u b c o m m i t t e e e m b a r k e d u p o n the p r e p a r a t i o n of a
tentative test m e t h o d t h a t i n c l u d e d P r o c e d u r e s A, B, a n d D.
Precision Statements
F r o m the c o m p u t e r analysis, w i t h i n - b a t c h a n d b a t c h - t o - b a t c h s t a n d a r d
deviations a n d coefficients of variation were o b t a i n e d which e n a b l e d the
p r e p a r a t i o n of precision s t a t e m e n t s . W h i l e a c c e l e r a t e d strengths m e a s u r e d
by the nine l a b o r a t o r i e s were significantly different, the variances in m a k i n g
TABLE 2--Linear regression analysis for Procedure B.

Simple Linear Regressiona
Lab
No.
Cement
Type
B0, MPa
B0, (psi)
B1
SE, MPa
SE, (psi)
I
III
I
III
I
III
I
III
I
III
I
III
I
III
I
III
I
III
12.82
16.55
17.79
13.96
9.86
14.96
11.58
13.13
15.62
14.86
16.89
19.55
19.48
17.79
9.52
13.41
18.13
18.27
(1860)
(2400)
(2580)
(2025)
(1430)
(2170)
(1680)
(1905)
(2265)
(2155)
(2450)
(2835)
(2825)
(2580)
(1380)
(1945)
(2630)
(2650)
1.060
0.910
1.290
1.145
1.225
0.840
1.280
1.050
1.280
1.220
1.015
0.780
1.195
1.000
1.515
1.095
1.020
0.430
0.970
0.985
0.965
0.955
0.975
0.930
0.970
0.965
0.810
0.950
0.865
0.960
0.950
0.975
0.975
0.800
0.860
0.530
1.34
0.66
1.55
2.17
1.21
1.28
0.79
1.07
4.03
2.03
1.58
1.24
1.24
0.90
1.86
3.48
2.48
3.07
(195)
( 95)
(225)
(315)
(175)
(185)
(115)
(155)
(585)
(295)
(230)
(180)
(180)
(130)
(270)
(505)
(360)
(445)
4
5
6
8
9
10
11
12
aAccording to the model Y : B0 + B1X, the terms of which are defined in the text.
172
TABLE 3--Linear regression analysis f o r Procedure C.

Lab
No.
Cement
Type
Bo, MPa
Bo, (psi)
Bl
SE, MPa
SE, (psi)
I
III
I
III
I
III
I
III
I
III
I
III
16.38
18.48
10.79
12.72
10.20
16.89
20.54
30.06
7.48
21.24
16.06
14.80
(2375)
(2680)
(1565)
(1845)
(1480)
(2450)
(2980)
(4360)
(1085)
(3080)
(2330)
(2145)
0.900
0.870
1.190
0.990
1.120
0.845
1.105
0.415
1.300
0.705
0.800
0.560
0.970
0.970
0.850
0.980
0.965
0.975
0.955
0.830
0.925
0.775
0.910
0.850
1.52
1.55
2.31
0 62
1.90
1.41
1.34
1.07
3.58
2.45
2.03
1.90
(220)
(225)
(335)
( 90)
(275)
(205)
(195)
(155)
(520)
(355)
(295)
(275)
6
8
10
11
12
aAccording to the model Y =- Bo + BjX, the terms of which are defined in the text.
TABLE 4--Linear regression analysis f o r Procedure D.

Lab
No.
Cement
Type
B0, MPa
B0, (psi)
Bl
SE, MPa
SE, (psi)
I
III
I
III
I
III
I
III
11.45
3.79
22.03
22.65
7.17
9.96
16.86
9.82
(1660)
( 550)
(3195)
(3285)
(1040)
(1445)
(2445)
(1425)
1.000
1.230
1.315
0.875
1.400
1.125
0.930
0.810
0.990
1.000
0.780
0.960
0.950
0.970
0.900
0.930
0.96
0.41
4.07
1.48
1.72
1.93
2.10
1.45
(140)
( 60)
(590)
(215)
(250)
(280)
(305)
(210)
5
8
12
~According to the model Y = Bo + BIX, the terms of which are defined in the text.
these strength measurements seemed compatible and were therefore pooled

across l a b o r a t o r i e s a n d c e m e n t types. T h i s r e s u l t e d in t h e f o l l o w i n g p r e c i s i o n
s t a t e m e n t s , w h i c h w e r e p r e p a r e d a c c o r d i n g to A S T M R e c o m m e n d e d P r a c tice for P r e p a r i n g P r e c i s i o n S t a t e m e n t s for T e s t M e t h o d s for C o n s t r u c t i o n
M a t e r i a l s (C 670). B e c a u s e of the s m a l l d i f f e r e n c e s b e t w e e n t h e m , t h e y w e r e
c o m b i n e d i n t o a single p r e c i s i o n s t a t e m e n t w h e n A S T M M a k i n g Accelerated Curing, and Testing of Concrete Compression Test Specimens
(C684) was p u b l i s h e d .
173
Procedure A--Hot Water, 35 ~ (95 ~

1. The single-laboratory coefficient of variation (1S percent) has been
determined as 2.9 percent for a pair of cylinders cast from the same batch.
Therefore, results of two properly conducted strength tests by the same
laboratory on the same materials should not differ more than 8.2 percent
of their average (D2S percent).
2. The single-laboratory, multiday coefficient of variation (1S percent) has
been determined as 7.2 percent for the average of pairs of cylinders east from
single batches mixed on 2 days. Therefore, results of two properly conducted
strength tests by the same laboratory on the same materials should not differ
by more than 20.4 percent of their average (D2S percent).

2. The single-laboratory, multiday coefficient of variation (1S percent)
has been determined as 8.7 percent for the average of pairs of cylinders
cast from single batches mixed on 2 days. Therefore, results of two properly
conducted strength tests by the same laboratory on the same materials
should not differ by more than 24.6 percent of their average (D2S percent).

2. The single-laboratory, multiday coefficient of variation (1S percent)
has been determined as 8.5 percent for the average of pairs of cylinders
cast from single batches mixed on 2 days. Therefore, results of two properly
conducted strength tests by the same laboratory on the same materials
should not differ by more than 24.1 percent of their average (D2S percent).
Prediction of Later-Age Strength

Prediction of later-age strength involves the use of statistics to estimate
the size of confidence intervals about both the accelerated compressive
strength and the regression lines determined from the statistical analysis of
174
the data. These confidence intervals are necessary to predict 28 or 364-day

compressive strength realistically, depending upon a certain laboratory's
control of the strength measurement as indicated by its standard deviation,
variance, or standard error. Confidence intervals can be calculated from
the statistical analysis according to the procedure shown in NBS Handbook
91 [9] and then, predictions can be made.
An example of this is shown for Laboratory 10 using Procedure B to
accelerate concrete made with Type I cement at three concentrations with
and without a normal dose of retarder [10]. There are twelve pairs of data
points, each the average of two cylinders from the same batch of concrete.
The regression equation is
Y = 19.48 + 1.195Xl MPa
= (19.48 + 1.195X0 (6895 10 -6) psi
with a standard error of estimate of 1.25 MPa (180 psi). The plot of the
equation with its 95 percent confidence interval is shown in Fig. 7. The
confidence interval was constructed by calculating Y for several selected
values of X1 and plotting Y ___ W1 where
WI =(SE)(2F)~/2( 1 + (X1S
-
~)2)1/2
and
F =
=
n =
XI =
,~ =
SE
F ratio for 2 and 10 deg of freedom,

standard error of estimate,
number of data points,
selected value of accelerated strength,
average value of accelerated strength, and
S = E X12 - - ( E
Xl)2/n
Assume the average strength of two cylinders tested at 1 day of age by

Procedure B in Laboratory 10 on the same materials was 17 MPa (2460
psi). From the regression equation this yields an estimated average 28-day
strength of 39.8 MPa (5760 psi). The 95 percent confidence interval is 39.0
to 40.7 MPa (5650 to 5900 psi) when the accelerated strength is without
error.
However, the accelerated strength has a sampling distribution whose
variance is described in the precision statement for Procedure B. The singlelaboratory coefficient of variation for cylinders cast from the same batch is
3.0 percent and the D2S limit is 8.5 percent. For an average accelerated
(7000)
12
14
16
18
i
!
I
I
&
v
2..2
vi
46
~.~
44
Q.
-r
I-(.9
Z
hl
20
175
(6)
(6090)
i --
~900)
. . . .
42
40
I--"
(5650)
CO
(5)
-4
I
(1.5)
FIG.
38
36
co
p,l
----4
0..
54
I
f
32
I
(2)
(2.5)
(3)
ACCELERATED STRENGTH, MPa (psi)
7--Prediction of 28-day strength for Laboratory 10 using Procedure B with Type I
cement.
strength of 17 MPa (2460 psi) the D2S limit is 1.44 MPa (210 psi). Dividing
by (2) ~/2 to obtain the limit for the average of two cylinders yields _ 1.0
MPa (4-145 psi) for the 95 percent confidence interval. Projecting the limits
of this interval to the confidence interval for the regression line results in
37.8 to 42.0 MPa (5480 to 6090 psi) for the approximate 95 percent confidence interval for 28-day strength.
Hence, in 95 tests out of 100, the 28-day strength wilt lie within this
interval when the accelerated strength is 17 MPa (2460 psi). Each different
measurement of accelerated strength produces a new confidence interval
for 28-day strength. Furthermore, a new regression line is obtained when
different materials or accelerated test methods are employed. Therefore,
each laboratory must conduct enough tests with a given set of materials
and a certain procedure to establish the regression line and measure the
standard error before predictions of later-age strengths are possible.
176
Effect of Cement Chemistry

In order to explain the differences between laboratories, a program was
conducted to determine the chemical or physical properties, or both, governing the 1-day compressive strength of concrete cured in 35~ (95~
water--Procedure A. Similar concrete batches were mixed utilizing eight
Type I cements mixed with the same source of sand and gravel. An attempt
was made to hold the slump constant at 76 to 102 mm (3 to 4 in.). Four
152 by 304-mm (6 by 12-in.) test cylinders were molded from each batch.
Two cylinders were cured according to ASTM Method C 684, Procedure A,
for 1-clay accelerated strengths and two cylinders were moist-cured at 23 ~
(73 ~ in 100 percent relative humidity for 28-day strengths.
After all cylinders were tested, the physical and chemical properties of
each cement along with the compressive strength relative to that cement
were compiled and computer analyzed initially for individual correlations.
The sodium alkali (Na20) content exhibited a high coefficient of correlation
with a low standard error relative to the 1-day accelerated strengths. The
data are shown in Fig. 8 and are represented by the equation
Y = 10.34 (1 + Na20) MPa
= 10.34 (1 + Na20)(6895 X 10 -6) psi
(50OO)
50
no {4ooo)
ff
(3000)
20
oo (2000)
...-..w--
]0
Y = ~034 ( I + N o 2 0 )
C~
oooo)
u =(1500) (1+ Ne20)
Ld
(0
O.l
02
03
0.4
o
0.5
No20, %
FIG. 8 - - E f f e c t o f alkali content on accelerated strength.
In view of the individual correlations, several multiple correlations were

attempted. For two independent variables, only combinations involving
Na20 produced a useful correlation with 1-day accelerated strength; moreCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:43 EDT 2014
177
over, only loss on ignition (LOI) coupled with the Na20 produced a better
multiple correlation than Na20 alone and is represented by the equation
Y = 10.34 + 9.45 Na20 + 0.28 LOI MPa
= (10.34 + 9.45 Na20 + 0.28 LOI)(6895 X 10 -6) psi
Therefore, the variation between the different Type I cements was mainly
attributable to variations in alkali content. Undoubtedly, this caused the
principal variation between laboratories since each used a different Type I
cement.
Summary and Conclusions

ASTM Subcommittee C09.02.09 conducted a cooperative test program
between nine laboratories to study the merits of seven test procedures for
accelerating compressive strength development of concrete with heat. Four
of these procedures involved the use of hot water, two boiling water, and
one the heat of hydration of the cement to cause the acceleration. The
amount of acceleration produced by these procedures ranged from 1.1 to
2.5 times the compressive strength developed after 1 or 2 days of standard
moist curing. With the autogenous procedure, 2-day strengths as much as
86 percent of the 28-day strength were developed.
During the course of the program it was determined that one of the boiling water procedures and three of the hot water procedures were not practical because of large overtime expenditures necessary to conduct them,
since each of these involved the measurement of Proctor time of set before
heat was applied to the concrete. Consequently, most of the data were
obtained on Procedure A, Hot Water, 35~ (95~
Procedure B, Modified Boiling; Procedure D, Autogenous Curing. An analysis of variance of
these data substantiated the following conclusions.
1. Significantly different results were obtained in the nine laboratories
due to the use of local materials including cements.
2. Significantly different results were obtained between laboratories because of variations in alkali content of the cements.
3. The omission or addition of a normal amount of chemical retarder in
the concrete produced the same results.
4. Batches of concrete mixed on different days did not affect the results
significantly.
5. Correlation of accelerated strength with 28 and 364-day, standard
strength ranged from good to excellent within each laboratory.
6. When cement types were separated, accelerated compressive strength
related to 364-day strength as well as 28-day strength, with equal variance
in the data.
178
Therefore, the results of this p r o g r a m have justified the p r e p a r a t i o n of a

standardized m e t h o d for accelerated strength tests which can be used with
as much confidence as the 28-day test in assessing concrete quality. A
method was adopted by A S T M a n d recently advanced to a s t a n d a r d designated A S T M Method C 684. It includes Procedure A - - H o t Water, Procedure B - - M o d i f i e d Boiling, a n d Procedure D - - A u t o g e n o u s Curing.
W h e n the m e t h o d is limited to a single laboratory-procedure-materials set
of conditions, it results in a n assessment of concrete quality at 1 or 2 days
of age as reliable as that o b t a i n e d after 28 days of s t a n d a r d moist curing.
References
[1] King, J. W. H., "Accelerated Testing of Concrete," Proceedings, 50th AnniversaryConference of the Institution of Structural Engineers, London, Oct. 1958, pp. 376-381,
386-387.
[2] King, J. W. H., "Accelerated Test for 7 and 28-day Compressive Strengths of Concrete,"
Journal of Applied Chemistry, Vol. 10, June 1960, pp. 256-262.
[3] Akroyd, T. N. W., "The Accelerated Curing of Concrete Test Cubes," Journal, Institution of Civil Engineers, Vol. 19, No. 6441, May 1961, pp. 1-22.
[4] Smith, P. and Chojnacki, B., "Accelerated Strength Testing of Concrete Cylinders,"
DHO Report RB 35, Ontario Department of Highways, Downsview,Ont., May 1963.
[5] Malhotra, V. M. and Zoldners, N. G., "'Some Field Experience in the Use of an Accelerated Method of Estimating 28-day Strength of Concrete," Internal Report MPI
68-42, Minerals Processing Division Department of Energy, Mines and Resources,
Ottawa, Ont., Aug. 1968.
[6] Natrella, M. G., private communication, 20 Nov. 1968.
[7] Hopper, J. S., Bodenschatz, J. S., and Hendrix, C. D., "Accelerated Testing of Concrete," unpublished report, Research and Development, Union Carbide Corp., Charleston, W. Va., 1969.
[8] Miller, R. W. and Chamberlin, W. P., "Prediction Equations for Accelerated Strength
Testing of Concrete," unpublished report, Engineering Research and Development
Bureau, New York State Department of Transportation, Albany, N.Y., Nov. 1970.
[9] Natrella, M. G., Experimental Statistics, NBS Handbook 91, National Bureau of Standards, Washington, D.C., Aug. 1963.
[10] Arni, H. T., private communication, 30 Dec. 1974.
Bibliography
Brandenberger, E., "A New Method for the Accelerated Testing of the Strength Development
of Portland Cements," Report No. 188, Federal (Swiss) Materials Testing Laboratory,
Zurich, 1957.
Concrete Manual, 3rd ed., Bureau of Reclamation, U.S. Department of the Interior, 1941,
p. 170.
Dixon, W. J. and Massey, F. J. in Introduction to Statistical Analysis, McGraw-Hill, New
York, 1951, pp. 209-220.
Hudson, S. B. and Steele, G. W., "Prediction of Potential Strength of Concrete from the
Results of Early Tests," Highway Research Record, No. 370, Highway Research Board,
1971, pp. 25-36.
Malhotra, V. M., "An Accelerated Method of Estimating the 28-Day Splitting-Tensile and
Flexural Strengths of Concrete," Report MRP/MSL 76-188, Mineral SciencesLaboratories,
Department of Energy, Mines and Resources, Ottawa, Ont., July 1976.
179
Malhotra, V. M. and Bauset, R., "Rapid Estimation of Concrete Strength Potential for
Hydro-Quebec Dams with Special Reference to Modified Boiling Method," Report MPI(A)
69-16, Minerals Processing Division, Department of Energy, Mines and Resources, Ottawa,
Ont., March 1969.
Malhotra, V. M., Zoldners, N. G., and Lapinas, R., "Accelerated 28-Day Test for Concrete,"
Canadian Pit and Quarry, March 1966, pp. 51-54.
McGhee, K. H., "Water Bath_Accelerated Curing of Concrete," VHRC 70-R8, Virginia
Highway Research Council, Charlottesville, Sept. 1970.
Orchard, D. F., Jones, R., and AL-Rawi, R. K., "The Effect of Cement Properties and the
Thermal Compatability of Aggregates on the Strength of Accelerated Cured Concrete,"
Journal of Testing and Evaluation, Vol. 2, No. 2, March 1974, pp. 95-101.
Smith, P. and Biekley, J. A., "Accelerated Testing of Concrete--The Sky's The Limit,"
Ministry of Transportation and Communications, Downsview, Ont., Canada, Aug. 1976.
Smith, P. and Tiede, H., "Earlier Determination of Concrete Strength Potential," DHO Report RR 124, Ontario Department of Highways, Downview, Ont., Canada, Jan. 1967.
R. E. Philleo
Chapter 14--Elastic Properties and

Creep
Introduction
When a load is applied to a body, the body is deformed. For a particular body loaded in a particular environment the amount of the deformation depends upon the magnitude of the load, the rate at which it is applied, and the elapsed time after the load application that the observation
is made. Different materials vary widely in their response to load. This
response is known as rheological behavior. While instantaneous effects
and time-dependent effects are not entirely separable, it is common to
consider them separately as elastic properties (instantaneous) and creep
(time-dependent).
A knowledge of the rheological properties of concrete is necessary to
compute deflections of structures, to compute loss of prestress in prestressed
structures, to compute stresses from observed strains, and, when working
stress design procedures are employed, to proportion sections and to determine the quantity of steel required in reinforced concrete members.
Although a vast amount of work has been done on the rheology of materials,
much of it is not applicable to concrete. Because of the peculiar "gel"
structure of cement paste, concrete behaves quite differently under applied
load than does a crystalline material.
Elastic Properties
A body which returns to its original dimensions after enduring stress is
elastic. A quantitative measure of elasticity is the ratio of stress to corresponding strain. Robert Hooke in 1678 discovered that for many materials
this ratio is constant over a fairly wide range of stress. This ratio is termed
the modulus of elasticity, and it has become one of the most commonly
1Chief, Structures Branch, Engineering Division, Civil Works, Office of Chief Engineers,
Department of the Army, Washington, D.C. 20314.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
PHILLEO ON ELASTIC PROPERTIES AND CREEP
181
used parameters to describe material properties even though many materials

do not exhibit a linear stress-strain relationship. Two additional terms are
used to describe limits of elastic behavior: (1) proportional limit and (2)
elastic limit. The proportional limit is defined in ASTM Definitions of
Terms Relating to Methods of Mechanical Testing (E 6) as "the greatest
stress which a material is capable of sustaining without any deviation from
proportionality of stress to strain (Hooke's law)." The elastic limit is "the
greatest stress which a material is capable of sustaining without any permanent strain remaining upon complete release of the stress."
Concrete has neither a definite proportional limit nor elastic limit.
Therefore, the manner in which its modulus of elasticity is defined is somewhat arbitrary. Various forms of the modulus which have been used are
illustrated on the stress-strain curve in Fig. 1. They are defined in ASTM
Definitions E 6 as follows.
Initial Tangent M o d u l u s - - T h e slope of the stress-strain curve at the origin.
Tangent M o d u l u s - - T h e slope of the stress-strain curve at any specified
stress or strain.
Secant M o d u l u s - - T h e slope of the secant drawn from the origin to any
specified point on the stress-strain curve.
Chord M o d u l u s - - T h e slope of the chord drawn between any two specified points on the stress-strain curve.
Unit Strain
FIG. 1-- Variousforms of static modulus of elasticity.

182
Modulus of elasticity may be measured in tension, compression, or

shear. The modulus in tension is usually equal to the modulus in compression and is frequently referred to as Young's modulus of elasticity.
The shear modulus, also called the modulus of rigidity or torsional modulus,
is the ratio of shear stress to shear stain. Shear stress is defined in ASTM
Definitions E 6 as "the stress or component of stress acting tangential
to a plane," and shear strain is defined as "the tangent of the angular
change between two lines originally perpendicular to each other."
When stress is applied in a given direction, there are changes in dimension in directions perpendicular to the direction of the applied stress, as
well as in the direction of the stress. The magnitudes of the lateral strains
are different for different materials. Thus, two parameters are required
to describe the elastic behavior of a material. The parameters may take
many forms, but the two most commonly used are Young's modulus of
elasticity and Poisson's ratio. Poisson's ratio is defined in ASTM Definitions
E 6 as "the absolute value of the ratio of transverse strain to the corresponding axial strain resulting from uniformly distributed axial stress
below the proportional limit of the material." The transverse strains are
opposite in direction to the axial strains. For a material obeying Hooke's
law, Poisson's ratio is constant below the proportional limit. It can be
shown that the following relationship exists among Young's modulus of
elasticity, modulus of elasticity in shear, and Poisson's ratio
--
E
2G
where
/~ = Poisson's ratio,
E = Young's modulus of elasticity, and
G = modulus of elasticity in shear.
Thus, if any two of these quantities are determined, the third can be calculated.
It can be shown that the natural frequency of vibration of an elastic
body is proportional to the square root of either Young's modulus or the
shear modulus, depending on the mode of vibration. Likewise, the velocity
with which a compressional shock wave travels through an elastic body is
proportional to the square root of Young's modulus. Since these phenomena
and their application to concrete are discussed elsewhere in this volume,
only the measurements of modulus of elasticity by "static" test methods
will be described here.
Modulus of Elasticity in Compression

Since structural concrete is designed principally for compressive stresses,
183
by far the greatest amount of work on the elastic properties of concrete

has been done on concrete in compression. The only ASTM standard
test method for static modulus of elasticity of concrete, ASTM Test for
Static Young's Modulus of Elasticity and Poisson's Ratio in Compression
of Cylindrical Concrete Specimens (C 469), is a compressive test method.
It stipulates a chord modulus between two points on the stress strain curve
defined as follows: the lower point corresponds to a strain of 50 millionths
and the upper point corresponds to a stress equal to 40 percent of the
strength of concrete at the time of loading. The lower point is near the
origin but far enough removed from the origin to be free of possible irregularities in strain readings caused by seating of the testing machine
platens and strain measuring devices. The upper point is taken near the
upper end of the working stress range assumed in design. Thus, the determined modulus is approximately the average modulus of elasticity in
compression throughout the working stress range.
The 150 by 300-ram (6 by 12-in.) cylinder is the specimen size most commonly used for the determination of the modulus of elasticity in compression. In order to compensate for the effect of eccentric loading or nonuniform response by the specimen, strains should be measured along the
axis of the specimen or along two or more gage lines uniformly spaced
around the periphery of the cylinder. The selection of the gage length is
important. It must be large in comparison with the maximum aggregate
size so that local strain discontinuities do not unduly influence the results, and it must be large enough to span an adequate sample of t h e
material. It must not, however, encroach on the ends of the specimen.
Because of restraint where the specimen is in contact with the steel platens
of the testing machine, strains near the ends may differ somewhat from
strains elsewhere in the specimen. ASTM Method C 469 specifies that the
gage length shall be not less than three times the maximum size of aggregate nor more than two thirds the height of the specimen. Half the
specimen height is said to be the preferred gage length. A convenient device for measuring the strains is a compressometer, such as the one pictured in Fig. 2. The lower yoke is rigidly attached to the specimen, whereas the upper yoke is free to rotate as the specimen shortens. The pivot
rod and dial gage are arranged so that twice the average shortening of
the specimen is read on the di~l. This type of device was used in the first
comprehensive investigation of modulus of elasticity by Walker [1],2 and
it is cited in ASTM Method C 469 as an acceptable device. Electrical
strain gages have also been used successfully. These include both gages
embedded along the axis of the specimens and those bonded to the surface. There are also available compressometers in which the strain, instead of being observed on a dial gage, is indicated and recorded by an
electrical device such as a linear differential transformer.
2The italic numbers in
brackets refer to the
list of references appended to this paper.
184
FIG. 2--Compressometer.
Because the test is intended to measure only time-dependent strains, it

is important that the specimen be loaded expeditiously and without interruption. For this purpose an automatic stress-strain recorder is helpful but not essential. Since it is desired that only those length changes due
to load be measured, temperature and moisture conditions should be controlled during the test.
A large numer of results have been reported in the literature [2-4].
The range of results has been from 7 000 to 20 000 MPa (1 000 000 to
3 000 000 psi) for structural lightweight concrete and from 14 000 to
35 000 MPa (2 000 000 to 5 000 000 psi) for normal weight concrete. A
simple relationship between modulus of elasticity and other easily measured
properties of concrete, such as strength and unit weight, would be useful.
While no theoretical relationship exists, an approximate equation adopted
by the American Concrete Institute (ACI) Building Code [5], which is
discussed later, has practical value.
Although the standard method of test is not concerned with the behavior
of concrete at stresses above 40 percent of the strength, the shape of the
stress-strain curve at high stresses is of significance in determining the
ultimate load-carrying capacity of a concrete member. In most testing
machines, concrete cylinders, except those with very low strength, fail
suddenly shortly after the maximum load has been attained. It has been
demonstrated [6-10, 15] that such failures are related to the properties of
the testing machine rather than the properties of the concrete. By using
a suitably stiff testing machine or by artificaUy stiffening a testing machine
185
by surrounding the concrete specimen with steel springs, it is possible

to obtain stress-strain curves covering a strain range several times as great
as that required to attain m a x i m u m stress. Figure 3 illustrates such stressstrain curves.
Modulus of Elasticity in Tension

A limited amount of work has been done on the determination of Young's
modulus in tension of concrete [11-13]. The test is complicated by the
limited stress range of concrete in tension and the problems associated
with gripping the tension test specimens. Recent work has been concerned with the development of specimen shapes which will ensure uniform
stress distribution throughout the section in which measurements are
made. Since Young's modulus in tension does not appear to differ from
Young's modulus in compression at low stresses and since there is no strain
range beyond maximum tensile stress to be investigated, there is relatively
little stimulus for the development of data on the tensile modulus.
Modulus of Elasticity in Flexure

Since a principal use of reinforced concrete is in flexural members,
several investigators have determined Young's modulus on specimens loaded
as beams. An obvious approach is to measure deflections caused by known
loads and to calculate the modulus of elasticity from well-known beam
deflection formulas. Unfortunately, the depth-to-span ratios of concrete
beams normally used for such tests are so large that shear deflection comprises a significant part of the total deflection. In applying shear corrections, certain other corrections must be made to take care of discontinuities
in the shear deflection curves at load points. The corrections most com-
25
,,
_z
g.
3
-20~
w
r
15
~
z
10 w
1
9 DAYS ~ . _ ~
~:
c~
5
0
I
0.002
I
I
0.004
0.006
CONCRETE STRAIN
I
0.008
0.010
FIG. 3--Complete stress-strain curves (taken from R e f 15).

186
monly used are those of Seewald [14]. For center-point loading he gives
the following deflection formula
-- 48----~- 1 + (2.4 + 1.5/~)
-- 0.84
where:
6
P
l
E
I
/~
h
=
=
=
=
=
=
=
m a x i m u m deflection,
applied central load,
distance between supports,
modulus of elasticity,
moment of inertia of the section with respect to the centroidal section,
Poisson's ratio, and
depth of the beam.
The portion of the expression outside the brackets is the simple beam
formula.
In some tests, strain gages have been placed on the tensile and compressive faces to determine strain as a function of applied load. The load may
be converted to fiber stress by standard beam formulas. For such tests the
beam is usually loaded at two symmetrical points so that there is a constant
stress condition between the two load points. Such tests have the advantage
of indicating the position of the neutral axis, but if the stress-strain curve is
nonlinear the computed stresses and, therefore, the computed modulus of
elasticity are in error.
In a comprehensive investigation of the compression side of a concrete
beam Hognestad et al [15] utilized specimens of the types shown in Fig. 4.
The central prismatic portion of the specimen is loaded both concentrically
and eccentrically. The eccentric load is continuously adjusted so that the
HYDRAULIC
I~ '~.~ ~,=
~__~
a2
cT
t
'
'
F///~!-k2~c
. P1
"
NEUTRAL SURFACE
al
Ii
lbll
~,E
Ill
"SR-4 GAGES
FIG. 4--Test specimens for stress distribution in a beam (taken from Ref 15).
187
strain on one surface remains zero. Thus, the section simulates that portion
of a beam between the neutral axis and the extreme compressive fiber. The
complete stress-strain curve in flexure can be determined from the observed loads and strains if the following two assumptions are made:
1. The distribution of strain across the section is linear.
2. All fibers follow the same stress-strain curve. The first has been
amply verified in these and other tests. The second depends on the speed of
testing. If the specimen is loaded expeditiously, time-dependent effects are
held to a minimum. In these tests the stress-strain curves in flexure agreed
very well with stress-strain curves obtained on companion cylinders con
centrically loaded in compression both below and above the maximum
stress.
Modulus of Elasticity in Shear

The shear modulus, or modulus of rigidity, is most often determined dynamically or by calculation from Young's modulus and Poisson's ratio.
When a direct static determination is desired, a torsion test is the common
procedure. When a given torque is applied to a body the angle of twist in a
given length is inversely proportional to the shear modulus as indicated by
the following equation
L
G~-----
where
G
L
4~
I
---- modulus of elasticity in shear,

= torque,
= angle of twist per unit length, and
= polar moment of inertia of the cross section.
A method for testing concrete cylinders has been described by Andersen

[16]. He used a level bar with micrometer adjustment for measuring differences in angular change of two radial arms placed a fixed distance apart
on a horizontally positioned specimen.
Poisson's Ratio
Static determinations of Poisson's ratio are made by adding a third
yoke and second dial gage to a compressometer so that a magnified transverse strain may be measured, as well as a magnified axial strain or by
mounting strain gages on the surface of a specimen perpendicular to the
direction of loading. The same considerations apply to gage length for
lateral strain measurement as for longitudinal strain measurement except
188
that for lateral strains there is no upper limit since end restraint is not a
factor. Procedures for determination of Poisson's ratio are included in
ASTM Method C 469. Poisson's ratio is also commonly computed from
results of Young's modulus and shear modulus determined dynamically.
The static value at stresses below 40 percent of the ultimate strength is
essentially constant; for most concretes the values fall between 0.15 and
0.20. The dynamic values are usually in the vicinity o f 0.25. At high
stresses or under conditions of rapidly alternating loads a different picture
emerges. Probst [17] has shown a systematic increase in the value of
Poisson's ratio with stress repetition and Brandtzaeg [18] has shown a
marked increase at very high stresses. When the value is below 0.50 there
is a decrease in volume of the body as a compressive load is applied.
Brandtzaeg's work indicates that above about 80 percent of the strength
there is an increase in volume as additional compressive loads are applied.
Creep
Creep is defined in ASTM Definitions E 6 as "the time-dependent increase
in strain in a solid resulting from force." All materials undergo creep
under some conditions of loading. In metals and other crystalline materials
creep has been attributed to slip in crystals. While slip of this nature undoubtedly occurs in aggregate particles and within crystalline particles that
are part of hydrated paste, there is ample evidence that these are only
secondary factors in the creep of concrete. Crystalline slip is normally
detectable only above some threshold level of stress. Creep of concrete is
observed at all stresses. Furthermore, creep of concrete is approximately
a linear function of stress up to 35 to 40 percent of its strength. This behavior is not associated with crystalline slip. Finally, the order of magnitude of concrete creep is much greater than that of crystalline materials
except for metals in the final stage of yielding prior to failure. Thus, concrete creep is considered to be an isolated rheological phenomenon associated with the gel structure of cement paste. This unique aspect of the
problem has the advantage that creep may be measured without extremely sensitive equipment but the disadvantage that little of the research
which has been done on other materials is applicable to concrete. A review of current knowledge of creep is to be found in the 1964 ACI Sym-
posium on Creep of Concrete [19].

The particular aspect of the gel structure of concrete which causes its
unusual behavior is the accessibility of its large internal surface to water.
In fact, Mullen and Dolch [20] found no creep at all in oven-dried pastes.
The movement of water into and out of the gel in response to changes in
ambient humidity produces the well known shrinking and swelling behavior of concrete. A principal view among investigators [21-25] is that
creep is closely related to shrinkage. In creep, gel water movement is
189
caused by changes in applied pressure instead of differential hygrometric

conditions between the concrete and its environment. This concept is supported by the similar manner in which creep and shrinkage curves are
affected by such factors as water-cement ratio, mix proportions, properties
of aggregate, compaction, curing conditions, and degree of hydration.
Another explanation of the effect of gel water [26,27] is delayed elasticity.
If a load is suddenly imposed on a body consisting of a solid elastic skeleton
with its voids filled with a viscous fluid, the load will be carried initially
by the fluid and will gradually be transferred to the skeleton as the fluid
flows under load. This is the behavior exhibited by the theological model
known as a Kelvin body, which consists of a spring and dashpot in parallel.
The concept of delayed elasticity has been chiefly responsible for the widespread attempts to reproduce the theological behavior of concrete by means
of theological models.
Creep of concrete has been attributed by some [28-31] to viscous flow of
the cement paste. The reduction in strain rate with time has been attributed
by these investigators both to the increasing viscosity of the paste and to the
gradual transfer of load from the cement paste to the aggregate. This concept
is supported by the concept that creep strain is proportional to the applied
stress over a wide range of stress. A convincing argument against it is the
fact that the volume of concrete does not remain constant while it creeps. In
fact Poisson's ratio for creep has usually been found to be less than for elastic
stress. In at least one investigation [32] Poisson's ratio was found to be
zero. Another argument against the concept is the partial recovery of creep
when the load is removed.
The fact that creep is associated primarily with the cement paste phase
of concrete produces a serious difficulty in the interpretation and application of creep data. Unless the work is restricted to very mature concrete,
the specimens do not maintain constant physical properties throughout the
test. Creep measurements must necessarily be made over a considerable
period of time and during that time the cement paste continues to hydrate.
Frequently information is desired'at early ages when the cement is hydrating
relatively rapidly.
Measurement of Creep
The use to be made of creep measurements usually determines the age
at which creep tests are begun and the stress level to which specimens are
loaded. A test procedure has been standardized in ASTM Test for Creep o.f
Concrete in Compression (C 512). The method stipulates loading moistcured specimens at an age of 28 days to a stress not exceeding 40 percent of
the strength of the concrete at the time of loading, although provision is
made for other storage conditions or other ages of loading. The stress is
restricted to the range throughout which creep has been found to be proCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:44 EDT 2014
190
portional to stress. Limitations on gage lengths similar to those in the test for
modulus of elasticity apply. The age of loading for a standard test is necessarily arbitrary. The method is intended to compare the creep potential of
various concretes. Testing at a single age of loading is satisfactory got this
purpose. It is required in the test method that the stress remain constant
throughout the one-year duration of the test within close tolerances. The load
may be applied by a controlled hydraulic system or by springs, provided in
the latter case the load is measured and adjusted frequently. It is also required that there be companion unloaded specimens. Length changes of
these specimens are measured and subtracted from the length changes of the
loaded specimens to determine creep due to load. This correction is intended
to eliminate the effects of shrinkage and other autogenous volume change.
While this correction is qualitatively correct and yields usable results, most
modern theories deny the independence of shrinkage and creep and thus
indicate that the two effects are not additive as assumed in the test. It is
now common to label creep which occurs in the absence of drying "basic
creep" and to label the additional deformation not accounted for by
shrinkage "drying creep" [33]. Thus, the total shortening at any time may be
considered the sum of elastic strain, basic creep, drying creep, and shrinkage.
As with testing for modulus of elasticity, most creep testing has been
concerned with specimens subjected to compression. Several investigators
[11,34-36] have studied creep in flexure because of the obvious application
to beam deflections. A limited amount of work has been done in tension
[32,37,38] and in torsion [16, 39, 40].
Creep Equations
Creep under constant stress always proceeds in the same direction at a
rate which is a decreasing function of time. It is, therefore, common to
plot creep test results on a semilogarithmic graph in which the linear axis
represents creep strain and the logarithmic axis represents time. Many sets
of data show an approximately straight line over a considerable period
of time. This has led to the development of numerous logarithmic equations for creep. Such an equation is cited in ASTM Method C 512. In that
test method the results are reported by listing the strains at specified
ages up to a year. In addition it is suggested that the report contain the
parameters of the following equation
= ---ff + F ( K )
lOge (t + 1)
191
where
= total strain per unit stress,
E = instantaneous elastic modulus,
F ( K ) = creep rate, calculated as the slope of a straight line representing
the creep curve on the semilog plot, and
t : time after loading, in days.
While it is not intended that a theoretical logarithmic law should be inferred from the equation, the slope of the least squares line is a convenient
parameter for comparing the creep characteristics of different concretes.
Most of the other equations which have been used to describe creep have
been based on the assumption that there is a limiting value of creep.
Typical of these is the Lorman [24] equation
C
mt
m
n+t
(7
where
c = creep strain after time, t, for a sustained stress, o,
m = ultimate creep strain per unit of stress, and
n = the time at which half the ultimate creep is attained.
McHenry [41] added coefficients to take into account the effects of hydration during the loading period with the following triple exponential equation
ec :
c~(1 -- e - n ) + 13e-Pk(1 -- e - " t )
where
ec
k
t
a, /3, r, p, and m
=
=
=
=
specific creep,
the age at time of loading,
time after loading, and
empirical constants.
A summary of creep equations is given in Concrete Society Technical

Paper No. 101 [42].
Rheological Models
Many investigators have used mechanical models as' an aid in setting up

mathematical equations. The elements normally used are the ideal spring,
in which force is proportional to strain, and the ideal dashpot, in which
force is proportional to the rate of strain. They are usually grouped together
in pairs. A spring and dashpot in parallel form a Kelvin unit, while a
192
spring and dashpot in series form a Maxwell unit. By selecting an appropriate group of elements it is possible to produce a system empirically
which can reproduce any given set of creep data. The models are frequently
simplified somewhat if some of the spring or dashpot constants are made
time-dependent or stress-dependent. While the model is an aid in visualizing
creep behavior and in writing equations for that behavior, it should not be
considered as representing the structure of the real material. Typical of
models which have been proposed is that of Freudenthal and Roll [43],
which is pictured in Fig. 5.
Principle of Superposition
Of considerable practical value is a knowledge of the response of concrete
to changing loads. While many of the investigations have included complete unloading as a portion of the study, there has been only a limited
study of other forms of variable loading. McHenry [41] proposed a theory
of superposition whereby the effect of each change in load is assumed to
be added algebraically to the effects of all previous loads. The material is
assumed to have a perfect memory. The effect of each increment of load
lasts forever. Figure 6 shows how the theory applied to creep recovery.
Three sets of specimens are required to produce such a figure. Two are
a m
(a)
ak
~k(b)
a2~r
~~(d)
FIG. S--Typical rheological model for representing creep (taken from Ref 43).
PHILLEO
ON
ELASTIC PROPERTIES AND CREEP
193
loaded at an age of 28 days. One of these is unloaded at 90 days. A third

set is loaded at 90 days to the same stress level as the original sets. The
principle of superposition is valid if the strain in the last set equals the
difference in strains of the first two sets. The principle is consistent with
the preponderance of observations that only partial recovery occurs when
a specimen is unloaded. While there is not unanimous agreement [44] on
the principle of superposition, it is used regularly in computing stresses in
mass concrete from measured strains.
Effect of Paste Content

Creep is influenced by mixture proportions, type of aggregate, conditions
of storage, and age at loading. Most of the creep studies which have been
conducted have been for the purpose of determining the effect of one or
more of these variables on creep. One aspect of the effect of mixture
proportions is of continual interest in the field of mass concrete. That is
the relationship between creep and paste content. Mass concrete commonly contains 150 mm (6-in.) aggregate. Fabrication and testing of concrete specimens containing such large aggregate is expensive. Work at the
Waterways Experiment Station of the Corps of Engineers [45] and at the
University of California [46] has demonstrated that for paste contents
normally used in concrete the creep of sealed specimens, having identical
water-cement ratios and air contents in the mortar phase but different
amounts and sizes of a given coarse aggregate, is proportional to the
paste content. This finding has made it possible to develop significant data
for mass concrete from small specimens. A general treatment of the effect
of aggregate properties and paste contents covering the entire range of
paste contents from 0 to 1 has been given by Neville [47].
Effect of Specimen Size

It has been demonstrated [45,48] that creep of sealed specimens is in0.3
~MEASURED CREEl" RECOVERY
Q28
t~.
a_
0.2
Z
COMPUTED CREEP RECOVERY-uzb - Qgo
.~ o.1
LU
t)
0.0
40
FIG.
80
120
AGE AFTER CASTiNg, DAYS
6--Principle of superposition (taken from Ref 41).
194
dependent of specimen size. This observation plus the observation concerning mass concrete in the preceding paragraph indicate that the techniques and specimens of ASTM Method C 512 are applicable to all types
of concrete sealed to prevent loss of moisture. For unsealed specimens exposed to a drying atmosphere it is evident that there must be a size effect
associated with the moisture gradients within the specimen. The creep of
a structure may be only a fraction of that in a test specimen. Ross [49]
investigated the effect of specimen size on shrinkage and found at each age
an excellent correlation between shrinkage and the ratio of exposed surface
to volume of specimen independent of the shape of the specimen. Hansen
and Mattock [50], in an investigation of both size and shape of specimens,
found that shrinkage and creep were dependent only on the ratio of surface
to volume. Information of this sort may make it possible to apply correction factors to the'data obtained from ASTM Method C 512 to determine the creep in any size and shape of structure.
Creep Prediction Equations

A series of creep prediction equations has been published by Committee
209, Creep and Shrinkage in Concrete, of the American Concrete Institute
[51]. The equations take into account the effects of type and quantity of
cement, consistency of concrete, air content, age at loading, relative
humidity, and size of member.
Significance of Properties
Deflection of Flexural Members
Concrete flexural members undergo deflection upon application of load
and continue to deflect with the passage of time. It is not uncommon for a
reinforced concrete flexural member eventually to reach a deflection three
times as great as its initial deflection. While a precise prediction of these
deflections is possible only if the elastic and creep properties are known,
the required precision is usually not great. Relatively little creep testing is
directed to predicting deflections of specific structures.
Loss of Prestress
In contrast to the lack of precision needed for deflection measurements,
an accurate knowledge of the early-age rheological properties of concrete is
valuable to the prestressed concrete industry. After the prestress is applied,
there is a loss of prestress resulting from creep of the concrete, shrinkage
of the concrete, and relaxation of the steel. Since the initial pretress is
limited by the strength of the steel and the load-carrying capacity of the
195
member is limited by the residual prestress, a knowledge of the factors

governing loss of prestress has important economic implications.
Structural Design
When reinforced concrete is designed by working stress theory, perfect
bond between the concrete and steel is assumed under design load conditions. Therefore, the load carried by the steel is a function of the ratio of
the modulus of elasticity of steel to that of concrete. Design is impossible
unless these quantities are known or assumed. For many types of structures such as arches, tunnels, tanks, and flat slabs, a knowledge of Poisson's ratio as well as modulus of elasticity is needed.
As discussed previously, a knowledge of the complete stress-strain curve
for concrete has assisted in the refinement of strength design theory.
Stress Calculations
A prerequisite for the calculation of stresses from measured strains is a
complete knowledge of a material's rheological behavior. When stresses
result from nonuniform temperature or humidity as well as from applied
load, the thermal coefficient of expansion and shrinkage data must also be
available. Perhaps the most extensive use of creep data has been in the
stress analysis of mass concrete structures [52]. These structures are unreinforced and contain complicated stress distributions because of temperature
conditions resulting from the heat of hydration of cement. The procedure
requires the installation of a large number of strain meters within the
structure during construction. It is necessary to run creep tests on sealed
specimens in the laboratory at enough ages of loading so that by interpolation and extrapolation a complete knowledge of rheological behavior is
available. Such data can be represented pictorially by a creep surface such
as that shown in Fig. 7. To convert strain data to stress, it is assumed that
creep is a linear function of stress and that the principle of superposition
applies. At the location of each strain meter, stress calculations must be
made continually from the time the concrete hardens to the end of the
period for which results are desired. The time is divided into small intervals. If during any interval the change in strain is different from that which
would be expected from the creep produced by the sum of all the stresses
present at the point at the start of the interval, the change in strain must
be accounted for by adding or subtracting an amount of stress sufficient to
produce the desired strain increment at the appropriate age of loading. By
this technique it has been possible to obtain an accurate assessment of the
safety of dams and to evaluate the effectiveness of various methods of
temperature control. A discussion of the accuracy of the method has been
given by Pirtz and Carlson [53].
196
~.,,
l.Or .........................
o.
o
6
,o
iiiii----,,,,
~';~
~0\\~\\
o\\\\\\
\"..\
".
\'..\
50'OJ5102030
\
50
-.. ,,
~o.~
",,\I
75
~
a
100
Time After Loading, Days
FIG. 7--Typical creep surface.
A method developed by Houk et al [54] measures the strain capacity

in slow loading directly by laboratory flexural testing of relatively small
beams. The technique takes into account the effects of mixture proportions and aggregate properties, and it has been used to establish construction controls required to reduce cracking in mass concrete structures [55].
Specifications
Limits on rheological properties are almost never included in specifications. In some structures in which designers wished to minimize deflections,
contractors have been restricted to those aggregates among the economically
available materials which have been demonstrated to produce concrete
having the lowest creep.
It has been pointed out that a knowledge of modulus of elasticity is required for some aspects of structural design, and where building codes
impose limits on deflections a knowledge of creep behavior is also necessary. Specifications for structural concrete are primarily concerned with
strength. It has not been considered feasible to complicate the specifications by additional rheological requirements. The problem has been ap-
197
proached in the ACI Building Code [5] by a s s u m i n g that the m o d u l u s of

elasticity is related to strength a n d u n i t weight as follows
E = W l's 4 3 f c '
where
E = m o d u l u s of elasticity, in Pa,
W = u n i t weight, in kilograms per cubic metre, a n d
f c ' = specified compressive strength, in Pa.
The ratio of long-time deflection to i m m e d i a t e deflection is assumed to
vary from 1.8, when the a m o u n t of compression r e i n f o r c e m e n t equals
the tension r e q u i r e m e n t , to 3.0, when there is no compression reinforcement. Neither of these estimates can claim a high degree of precision; both
are p r o b a b l y a d e q u a t e for the i n t e n d e d purpose. However, as structural design increases in sophistication a n d as applications of prestressed concrete increase, the interest in rheological properties is increasing. This
interest is evident in the steadily increasing a m o u n t of creep testing b e i n g
performed.
References
[1] Walker, Stanton, "Modulus of Elasticity of Concrete," Proceedings, American Society
for Testing and Materials, Vol. 21, Part 2, 1919, p. 510.
[2] Teller, L. W., "Digest of Tests in the United States for the Determination of the
Modulus of Elasticity of Portland Cement Mortar and Concrete," Proceedings, American
Society for Testing and Materials, Vol. 30, Part 1, 1930, p. 635.
[3] "Bibliographies on Modulus of Elasticity, Poisson's Ratio, and Volume Changes of
Concrete," Proceedings, American Society for Testing and Materials, Vol. 28, Part 1,
1928, p. 377.
[4] Philleo, R. E., "Comparison of Results of Three Methods for Determining ~oung's
Modulus of Elasticity of Concrete," Journal, American Concrete Institute, Jan. 1955,
Proceedings, Vol. 51, pp. 461-469.
[5] Building Code Requirements for Reinforced Concrete (ACI 318-77.), American Concrete
Institute, Oct. 1977.
[6] Whitney, C. S., discussion of a paper by V. P. Jensen, "The Plasticity Ratio of Concrete
and Its Effect on the Ultimate Strength of Beams," Journal, American Concrete Institute,
Nov. 1943, Supplement, Proceedings, Vol. 39, pp. 584-2 to 584-6.
[7] Saliger, R., "Bruchzustand und Sicherheit im Eisenbetonbalken," Beton und Eisen,
Berlin, Vol. 35, No. 19 and 20, Oct. 1936, pp. 317-320 and 339-346.
[8] Kiendl, O. G. and Maldari, T. A., "A Comparison of Physical Properties of Concrete
Made of Three Varieties of Coarse Aggregate," B. S. thesis, University of Wisconsin,
Madison, Wis., 1938.
[9] Ramaley, D. and McHenry, D., "Stress-Strain Curves for Concrete Strained Beyond the
Ultimate Load," Laboratory Report No. SP-12, U.S. Bureau of Reclamation, Denver,
Colo., March 1947.
[10] Blanks, R. F. and McHenry, D., "Plastic Flow of Concrete Relieves High-Load Stress
Concentrations," Civil Engineering, Vol. 19, No. 5, May 1949, pp. 320-322.
[11] Davis, R. E., Davis, H. E., and Brown, E. H., "Plastic Flow and Volume Changes in
1937, p. 317.
198
[12] Johnson, A. N., "Tests of Concrete in Tension," Public Roads, Vol. 7, No. 4, June 1926.
[13] Johnson, J. W., "Relationship Between Strength and Elasticity of Concrete in Tension
and in Compression," Bulletin No. 90, Engineering Experiment Station, Ames, Iowa,
1928.
[14] Seewald, F., "Abhandlungen," Aerodynamischen Inst. an der Technischen Hochschule,
Aachen, Vol. 7, 1927, p. 3.
[15] Hognestad, E., Hanson, N. W., and MeHenry, D., "Concrete Stress Distribution in
Ultimate Strength Design," Journal, American Concrete Institute, Dec. 1955, Proceedings, Vol. 52, pp. 455-479.
[16] Andersen, P., "Experiments with Concrete in Torsion," Transactions, American Society
of Civil Engineers, Vol. 100, 1935, p. 949.
[17] Probst, E., "The Influence of Rapidly Alternating Loading on Concrete and Reinforced
Concrete," The Structural Engineer, (British), Vol. 9, No. 12, Dec. 1931.
[18] Brandtzaeg, A. "The Failure of Plain and Spirally Reinforced Concrete in Compression,"
Bulletin No. 190, University of Illinois Engineering Experiment Station, Urbana, Ill.,
April 1929.
[19] Symposium on Creep of Concrete (ACI SP-9), American Concrete Institute, 1964.
I20] Mullen, W. G. and Dolch, W. L., "Creep of Portland-Cement Paste," Proceedings,
Vol. 64, 1964, pp. 1146-1171.
[21] Lynam, C. G., Growth and Movement in Portland Cement Concrete, Oxford University Press, London, 1934.
[22] Lea, F. M. and Lee, C. R., "Shrinkage and Creep in Concrete," Symposium on Shrinkage and Cracking of Cementive Materials, The Society of Chemical Industry, London,
1947, pp. 17-22.
[23] Seed, H. B., "Creep and Shrinkage in Reinforced Concrete Structures," The Reinforced Concrete Review, London, Vol. l, No. 8, Jan. 1948, pp. 253-267.
[24] Lorman, W. R., "Theory of Concrete Creep," Proceedings, American Society for Testing
[25] Hansen, T. C., "Creep of Concrete," Swedish Cement and Concrete Research Institute,
Royal Institute of Technology, Stockholm, 1958, pp. 48.
[26] Freyssinet, E., "The Deformation of Concrete," Magazine of Concrete Research, Vol. 3,
No. 8, Dec. 1951, pp. 49-56.
[27] Torroja, E. and Paez, A., "Set Concrete and Reinforced Concrete," Chapter 8, Building
Materials, Their Elasticity and Plasticity, editor, M. Reiner, North-Holland Publishing
Co., Amsterdam, 1954, pp. 290-360.
[28] Freudenthal, A. M., The Inelastic Behavior of Engineering Materials and Structures,
Wiley, New York, 1950, pp. 587.
[29] Reiner, M., Deformation, Strain and Flow, H. K. Lewis and Co., Ltd., London, 1960.
[30] Thomas, F. G., "Conception of Creep of Unreinforced Concrete and an Estimation of
the Limiting Values," StructuralEngineering, London, Vol. 11, 1933, p. 69.
[31] Glanville, W. H. and Thomas, F. G., "Further Investigations on the Creep or Flow of
Concrete Under Load," Building Research Technical Paper 21, Department of Scientific
and Industrial Research, London, 1939.
[32] Ross, A. D., "Experiments on the Creep of Concrete Under Two-Dimensional Stressing," Magazine of Concrete Research, Vol. 6, No. 16, June 1954, pp. 3-10.
[33] Neville, A. M., Properties of Concrete, 2nd Ed., Wiley, New York, 1973.
[34] Chang, T. S. and Kesler, C. E., "Correlation of Sonic Properties of Concrete with Creep
and Relaxation," Proceedings, American Society for Testing and Materials, Vol. 56,
1956, pp. 1257-1272.
[35] Vaishnav, R. N. and Kesler, C. E., "Correlation of Creep of Concrete with its Dynamic
Properties," T. & A. M. Report No. 603, University of Illinois, Urhana, Ill., Sept. 1961.
[36] Chang, T. S., "Prediction of the Rheological Behavior of Concrete from Its Sonic
Properties," Ph.D. thesis, Theoretical and Applied Mechanics, Report No. 522, University
of Illinois, Urbana, Ill., 1955, pp. 207.
[37] McMillan, F. R., "Shrinkage and Time Effects in Reinforced Concrete," Bulletin No. 3,
University of Minnesota, Minneapolis, Minn., 1915, p. 41.
[38] Wajda, R. L. and Holloway, L., "Creep Behavior of Concrete in Tension," Engineering,
London, VoL 198, No. 5130, Aug. 14, 1964.
199
[39] Duke, C. M. and Davis, H. E., "Some Properties of Concrete Under Sustained Combined Stresses," Proceedings, American Society for Testing and Materials, Vol. 44,
1944, pp. 888-896.
[40] Glucklich, J. and Ishai, O., "Creep Mechanism in Cement Mortar," Journal, American
Concrete Institute, July 1962, Proceedings, Vol. 59, pp. 923-946.
[41] McHenry, D., "A New Aspect of Creep in Concrete and Its Application to Design,"
Proceedings, American Society for Testing and Materials, Vol. 43, pp. 1069-1084.
[42] "The Creep of Structural Concrete," Report of a Working Party of Materials Technology Divisional Committee, Concrete Society Technical Paper No. 101, Jan. 1973.
[43] Freudenthal, A. M. and Roll, Frederic, "Creep and Creep Recovery of Concrete Under
High Stress," Journal, American Concrete Institute, June 1958, Proceedings, Vol. 54,
pp. 1111-1142.
[44] Davies, R. D., "Some Experiments on Applicability of the Principle of Superposition to
the Strain of Concrete Subjected to Changes of Stress with Particular Reference to Prestressed Concrete," Magazine of Concrete Research, Vol. 9, No. 27, 1957, pp. 161-172.
[45] "Creep of Mass Concrete," Miscellaneous Paper No. 6-132 Report 3, U.S. Army
Engineer Waterways Experiment Station, Jan. 1958, p. 22.
[46] Polivka, Milos, Pirtz, David, and Adams, R. F., "Studies of Creep in Mass Concrete,"
Symposium on Mass Concrete: ACI Special Publication SP-6, American Concrete
Institute, 1963, pp. 257-283.
[47] Neville, A. M., "Creep of Concrete as a Function of its Cement Paste Content," Magazine of Concrete Research, VoI. 16, No. 46, March 1964, pp. 21-30.
[48] Karapetrin, K. S., "Influence of Size upon Creep and Shrinkage of Concrete Test Specimens". I. Tekhiniceskie Nauki, Armenian Academy of Sciences (Yerevan), Akad. Nauk
Armianskoi SSR, Fiziko--Mathematicheskie (Estestvennye), Vol. 9, No. 1, 1956,
pp. 87-100 (in Russian).
[49] Ross, A. D., "Shape, Size, and Shrinkage," Concrete and Constructional Engineering,
Vol. 39, No. 8, Aug. 1944.
[50] Hansen, T. C. and Mattock, A. H., "The Influence of Size and Shape of Member on the
Shrinkage and Creep of Concrete," American Concrete Institute, Feb. 1,966; Proceedings, Vol. 63, pp. 267-290.
[51] "Prediction of Creep, Shrinkage, and Temperature Effects in Concrete Structures,"
Reported by ACI Committee 209, American Concrete Institute, Special Publication 27,
1971.
[52] Raphael, J. M., "The Development of Stresses in Shasta Dam," Transactions, American
Society of Civil Engineers, Vol. 118, 1953, pp. 289-321.
[53] Pirtz, David and Carlson, R. W., "Tests of Strain Meters and Stress Meters under
Simulated Field Conditions," Symposium on Mass Concrete, ACI Special Publication
SP-6, American Concrete Institute, 1963, pp. 287-299.
[54] Houk, Ivan E., Jr., Paxton, J. A., and Houghton, D. L., "Prediction of Thermal Stress
and Strain Capacity of Concrete by Tests on Small Beams," Journal, American Concrete
Institute, March 1970; Proceedings, Vol. 67, pp. 253-261.
[55] Houghton, D. L., "Concrete Strain Capacity Tests, Their Economic Implications,"
Economical Construction of Concrete Dams, Proceedings of the Engineering Foundation Conference, American Society of Civil Engineers, 1972, pp. 75-99.
E. A. W h i t e h u r s t ~ a n d V. M. M a l h o t r a 2
Chapter 15
Nondestructive Tests
Introduction
For many years one of the goals of those engaged in the control of concrete
quality and in the service behavior of concrete has been the development of
suitable nondestructive tests to supply the information desired. To be of
greatest usefulness, suchtests should be applicable to concrete in the structure.
The term "nondestructive tests" has been assigned various meanings
throughout the years. For example, it has been defined as dynamic testing in
which the load is applied and removed in a manner such that the effects of
creep during testing are negligible, and which does not result in destruction
or damage to the concrete. In general, the results of such tests have led to the
computation of values of Young's modulus of elasticity that are somewhat
higher than those determined from slower applications of load in which both
elastic and plastic deformations may occur.
During recent years, however, there has been a growing tendency to assign
to the family of nondestructive tests a number of tests which measure in situ
some parameter which may be related theoretically or empirically to compressive strength. Not all such tests meet the requirements of the aforementioned definition, since some produce sufficient damage to the concrete
tested in the structure to require repair of the surface. However, the damage
in these cases is generally agreed to be more cosmetic than structural and
repairs normally are made for aesthetic reasons.
For purposes of this discussion, tests which are identified generally as
nondestructive may be subdivided into three groups: (a) those identified as
sonic tests, generally involving determination of the resonant frequency of a
specimen, (b) those identified as pulse velocity tests, generally involving
measurement of the velocity of a compressional pulse travelling through the
IProfessor of Civil Engineering and director, Transplex/OSU, The Ohio State University,
Columbus, Ohio. 43210.
2Head, Construction Materials Section, Mineral Sciences Laboratories, CANMET, Department of Energy, Mines, and Resources, Ottawa, Canada.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
WHITEHURST AND MALHOTRA ON NONDESTRUCTIVE TESTS
201
concrete, and (c) those involving measurement in situ of a parameter to be

related to compressive strength and generally consisting of surface hardness,
rebound, penetration, pull-out, and maturity concept techniques.
Sonic Tests
The expression "sonic testing" is generally considered to include all testing
of concrete which involves the generation of a sustained vibration in the concrete. By far the majority of such tests involves the determination of the
fundamental resonant frequency of a specimen.
Equipment used to perform tests of this nature varies from the hammer
and home-made sonometer reported by Powers [1] 3 to electronic signal
generators for driving the specimen and highly complex electronic counters
for measuring the actual number of vibrations per unit of time. The hammersonometer method has been largely superseded by other methods, primarily
because of the difficulty encountered by many operators in matching the tone
emitted by the concrete specimen to the tone of the sonometer and because,
in some cases, the striking of the specimen resulted in a change in its
characteristics [2]. The use of the complex counter is generally restricted to
research laboratories. The large bulk of the apparatus falls between these extremes.
The components required for performing a test of this nature are an audio
signal generator, an amplifier, a driving unit, a pickup unit, another
amplifier, and a metering device. These are available individually from a
number of sources and are also available in a combined form suitable for
direct use in testing of this nature. The driving unit is frequently a
permanent-magnet speaker with a rod attached to the speaker coil. The
pickup is usually a piezoelectric crystal, often a phonograph pickup cartridge.
In earlier tests, the specimen was supported on knife-edges located at the
nodal points for flexural vibration (a distance of 0.224 of the specimen length
from each end). The driver was placed in the center of the specimen and the
pickup at one end. Sufficient power was applied to the driver to cause mild
vibration of the specimen. The frequency of the oscillator was then varied.
When the frequency approached that of resonance for the specimen, the
amplitude of specimen vibration increased considerably. The oscillator was
tuned until a maximum indication was observed on the metering device. The
frequency at which this occurred was recorded as the fundamental transverse
frequency of the specimen.
Subsequent investigations indicated that if the specimen was allowed to
rest uniformly on a sheet of soft sponge rubber the restraint on the specimen
would be sufficiently low to remove the necessity for mounting it at its nodal
202
points. This has now become a generally accepted practice. Considerable attention has also been directed to the use of torsional vibration for testing concrete specimens. In this technique, generally applied to prisms, the driving
unit is placed against one corner of the specimen and the pickup against a
diagonally opposite corner. The specimen is thus vibrated with a twisting motion, the node occurring at the center of the beam. Further investigations
have included the use of longitudinal vibrations in which the driving unit is
placed against one end of the specimen and the pickup unit against the
other. In this case, also, the nodal point occurs at the center of the specimen.
ASTM Test for Fundamental Transverse, Longitudinal, and Torsional
Frequencies of Concrete Specimens (C 215) makes provisions for testing
concrete specimens for fundamental transverse, longitudinal, and torsional
frequencies, and provides a schematic diagram illustrating use of the components enumerated above. The relationships between the several fundamental frequencies and other properties of the concrete are given as follows
Dynamic E = C W n 2
(1)
Dynamic E = D W ( n ' ) 2
(2)
Dynamic G = B W ( n " ) 2
(3)
where
Dynamic E = dynamic Young's modulus of elasticity, psi,
Dynamic G = dynamic modulus of rigidity, psi,
W = weight of specimen, lb,
n =
fundamental transverse frequency, cPs,
n r
fundamental longitudinal frequency, cPs, and
n t~
fundamental torsional frequency, cPs.
The constants C, D, and B are factors depending for their values upon the
shape of the specimen tested and Poisson's ratio. Detailed instructions for
their calculation may be found in ASTM Test C 215.
Where E and G are determined as outlined above, Poisson's ratio may be
calculated from the relationship
-- ~
/~ - - 2 G
--
(4)
Several investigators [3,4] have shown that a somewhat different

characteristic of concrete, its damping capacity, may be determined from
studying the behavior of a specimen vibrating at or near resonance. Two
measures of this characteristic are suggested, the damping constant and the
logarithmic decrement of a free vibration.
203
The damping constant is given by

f0
0
f i -- f2
(s)
where
f0 = resonant frequency of vibration, cPs, and
f l , f2 = frequencies on either side of resonance at which the amplitude is
1 / ~ times the amplitude at resonance.
These values may be determined if a meter is used as the indicating device
for determining resonance, provided that a sufficiently precise method of
determining vibration frequency is available. This becomes extremely important since the frequency range betweenf~ and f2 is very small indeed. A frequency counter has been found to be satisfactory for this type of work.
The logarithmic decrement is given by
Zl
6 = I n - A2
(6)
where
b = logarithmic decrement, and
A1, A z ----- amplitudes of two successive vibrations after the driving force
has been removed from the specimen.
To determine A1 and A 2, a cathode-ray oscillograph may be used as the indicator. After resonance has been located, the driving oscillator is turned off
and the decay of the specimen vibration recorded on a moving film strip.
When the film has been developed, the amplitudes of successive cycles may
be measured accurately.
The damping constant and logarithmic decrement are related by
Q-
7~
(7)
Both of the aforementioned methods have a disadvantage in that the

damping effects of the specimen supports must be extremely low. In an effort
to minimize these effects, Obert and Duvall [3] supported their specimens on
piano wires accurately located at the nodal points. For some reas o n - p e r h a p s this difficulty in obtaining supports that exercise sufficiently low restraint on a specimenmthese test methods have not been used
204
widely. Kesler and Higuchi [4], however, have reported tests in which the
logarithmic decrement, in combination with the dynamic modulus of elasticity determined from the transverse resonant frequency of a specimen, was
used in predicting the compressive strength of concrete. They report an accuracy of prediction generally within five percent for the limited tests made.
In more recent reports, Chang and Kesler [5,6] have reported studies of
short-time creep in compression, long-time creep in compression, relaxation
in compression, creep in flexure under uniform loads, and creep in flexure
under concentrated loads in which reasonably good correlations were obtained between these measured creep values and values predicted on the
basis of dynamic modulus of elasticity, logarithmic decrement, and compressive strength of the concrete studied.
Mention should also be made of the possibility of using sustained vibrations for testing concrete in place. Long and Kurtz [7] have reported such
tests in which a large auditorium-type loudspeaker was attached rigidly to a
concrete wall and driven at a fairly high power level. A pickup was then
moved about on the surface of the wall to determine points of maximum
vibration. The quantity measured was the velocity of the standing wave, since
the frequency of vibration was known and the wave length of the vibration
within the concrete could be determined. This and similar techniques have
not been widely used, although Pickett [8] has suggested that tests of a
similar nature might be useful in testing concrete pavements.
The greatest use of sonic techniques has been made in evaluating the performance of concrete specimens subjected to natural or artificial weathering.
They have also been used to study the effect of moisture content and to compare different mixes [3], to investigate the efficiency of various curing compounds [7], and for other purposes [9]. It has been suggested that they may
be used to study the setting characteristics of concrete, but such usage has
generally proved unsuccessful. With the exception of the work of Kesler and
Higuchi [4], most efforts to relate the elastic properties of concrete to its
strength have been largely unsuccessful. Somewhat better results have been
obtained in efforts to correlate changes in dynamic modulus of elasticity to
changes in strength, and in weathering studies it is fairly common to equate a
30 percent decrease in dynamic modulus to a 50 percent decrease in flexural
strength.
Unfortunately all of the techniques discussed above, with the possible exception of the last, are subject to two major limitations. The methods are
basically applicable to specimens of relatively small size and are of little value
in studying the behavior of concrete in place. Further, because of the complexity of the calculations involved in computing the constants D, C, and B,
specimens must be either cylinders or prisms, of uniform cross section
(square or rectangular). It is perhaps due largely to these restrictions that recent attention has been directed toward the development of devices for determining pulse velocity in concrete.
205
Pulse Transmission Tests
The application of pulse transmission techniques to the testing of concrete

is believed to have had its origin with Obert [10]. Tests were made on concrete replacement pillars in mines and involved the use of two geophones, two
high-gain amplifiers, and a camera with a moving film strip. Two holes, approximately 6.09 m (20 ft) apart vertically, were drilled into the pillars. The
geophones were placed in the backs of the holes and the holes filled with cotton waste. A hammer blow was struck at the base of the pillar, and at the
same time the camera lens was opened and the moving film strip exposed.
After the film was developed, the transit time of the impulse in traveling from
one geophone to the other was determined by measuring the distance between the two signals on the film, the speed of motion of the film having been
controlled carefully. The velocity could then be calculated.
Long and Kurtz [7, p. 1067] reported performing somewhat similar experiments with a Shepard seismograph in which the longitudinal velocity of
the pulsation created by a single impact was measured between arbitrarily
placed geophones. They stated that only very limited experiments of this
nature had been conducted but that the method appeared to hold great
promise providing the apparatus could be adapted to the measurement of
much shorter time intervals than those of which the seismograph was
capable.
Long et al [11] undertook further investigations along these lines and in
1945 reported on the instrument and technique that resulted from their
work. The apparatus consisted of two vibration pickups (in the form of
phonograph cartridges), two amplifiers, two thyratron tube circuits, and a
ballistic galvanometer circuit. The impact of a hammer blow was impressed
upon the concrete in a horizontal direction, approximately in line with the
two pickups. The energy impulse thus generated actuated the first pickup,
the voltage from which energized the first thyratron and started a flow of current through the galvanometer. When the energy impulse reached the second
pickup, the voltage from its amplifier ionized the second thyratron and cut
off the flow of current. The deflection of the galvanometer was directly proportional to the time required for the wave to travel the distance between the
two pickups.
In a discussion of this paper, the substitution of an electronic interval
timer for the ballistic galvanometer was suggested. This device consists of a
capacitor which begins to charge when the first thyratron is ionized and stops
charging when the second is ionized and a vacuum-tube voltmeter which
measures the charge. The meter may be calibrated directly in units of time,
thus eliminating the necessity for computations involving the magnitude of
the current flowing through the galvanometer. This device was found to be
more reliable than the ballistic galvanometer for field use.
Subsequent investigations in this country and abroad have resulted in the
206
development of a number of other devices quite similar in most respects.

These include the Micro-timer developed by the U.S. Bureau of Reclamation, the Condenser Chronograph developed by the Danish National Institute
of Building Research, and devices developed at the National Physical
Laboratory, South African Council for Scientific and Industrial Research,
and the Laboratoires du Batiment et des Travaux Publics, France. All make
use of either hammer blows or small explosive charges to generate the impulse.
In 1946 the Hydro-Electric Power Commission of Ontario, Canada, in an
effort to develop a technique for examining cracks in monolithic concrete
structures, began a series of studies which resulted in the construction of an
instrument known as the Soniscope. The device consists basically of a pulse
generator using piezoelectric crystals, a similar pulse receiver, and electronic
circuits which actuate the pulse generator, provide visual presentation of
transmitted and received signals on a cathode-ray tube, and accurately
measure the time interval between the two. Development of this instrument
was first reported to Committee 115, Research, of the American Concrete Institute in 1948. A more complete report was published by Leslie and
Cheesman [12] in 1949.
The physical and electrical features of the Soniscope have passed through
several stages of improvement since 1947, and a number of these instruments have been built by various laboratories in the United States and
Canada. Considerable use of the instrument has been reported in Canada
[13] and the United States [14].
During approximately the same time that the Soniscope was being
developed in Canada and the United States, work of a similar nature was being conducted in England. These investigations resulted in the development
of an instrument known as the Ultrasonic Concrete Tester. This instrument
and the uses to which it has been put have been described at length by Jones
[15,16]. The Ultrasonic Concrete Tester differs from the Soniscope primarily
in the much higher frequency used within the transmitted pulse and in the
repetition rate which is about three times as great as that of the Soniscope.
These changes improve the accuracy of measurement on very small
specimens but limit the usefulness of the instrument for field testing, since
the high frequencies suffer much greater attenuation in passing through concrete than do the lower ones. The maximum range of the Ultrasonic Concrete
Tester is believed to be about 2 m (7 ft), whereas that of the Soniscope in
testing reasonably good concrete is 15 m (50 ft) or more.
Several devices have been reported which incorporate some of the features
of the Soniscope and some of the Electronic Interval Timer. In these instruments the impulse is generated by a mechanical blow, frequently that of
a spring-loaded hammer operated by a motor-driven cam at the rate of about
five blows per second. The receiver may be any one of a number of types of
pickup. Transmitted and received signals are displayed on a long persistance
cathode-ray oscilloscope containing a calibrated time base for determining
207
the transit time between signals. Development of one such device has been
reported in the United States and another in France.
The devices mentioned above and the uses to which they have been put
have been described in detail by Whitehurst [17] and in several publications
of the Rdunion des Laboratoires d'Essais et de Recherches sur les Materiaux
et les Constructions [18].
Whereas the use of the sonic tests has been restricted primarily to the
evaluation of specimens undergoing natural or artificial weathering and the
techniques for such use have been largely standardized, pulse transmission
techniques have been applied to concrete for many purposes and, in most
areas of investigation, only limited agreement has been reached concerning
the significance of test results. The quantity actually measured by all of these
instruments is the transmission time of an impulse passing through the concrete under test. If the path length between generator and receiver is known
or can be determined, the velocity of the pulse may easily be computed. It is
in the interpretation of the meaning of this velocity and in its use for determining various properties of concete that agreement is incomplete. The technique is as applicable to concrete in place as to laboratory-type specimens,
and results appear to be unaffected by the size and shape of the concrete
tested, within the limits of transmission of the instrument employed, provided care is taken when testing very small specimens. This, of course, is a
highly desirable attribute and, in many respects, makes the pulse transmission techniques more useful than those involved in sonic testing.
Because of the fundamental theoretical relationship between pulse techniques and resonant frequency techniques, there is a strong inclination for
users of the pulse technique to endeavor to compute the dynamic modulus of
elasticity from the results of the tests. Theoretically, such values of modulus
should be the same as those determined by resonant frequency tests upon the
same specimens. The experience of several investigators [19,20], however,
has shown that on some occasions this is true and on others it is not. Because
of these presently unexplainable differences, most of those experienced in the
use of pulse velocity techniques are inclined to leave their results in the form
of velocity without attempting to calculate moduli therefrom.
If the modulus of elasticity is to be computed from the pulse velocity, the
relationship generally recommended is
(1 + ~)(1 -- 2v)
E = V2p
(8)
(1 - - # )
where
E
V
p
/z
= dynamic modulus of elasticity,

= longitudinal pulse velocity,
= mass density, and
---- Poisson's ratio.
208
This equation relates modulus to pulse velocity and density in an infinite

medium and presumably should apply only to mass concrete. However, the
experience of most investigators has been that even for very small laboratory
specimens this relationship gives better results than do those applying to
either slabs or long slender members. Leslie and Cheesman [12] have suggested that best results are obtained if, for concretes having unit weights in
excess of approximately 2240 k g / m a (140 lb/fta), the value of Poisson's ratio
is assumed to be 0.24. If this is done, Eq 8 is reduced to
E = 0.000216 V2d(0.848)
(9)
where d = weight of concrete, k g / m a (lb/fta).

The factor 0.848 represents the correction for Poisson's ratio, taken in this
case to be 0.24.
Fundamentally all of the pulse velocity techniques and equipment
developed for use on concrete result in the measurement of the compressional, or longitudinal, pulse velocity. In rare cases, with the Soniscope-type
instrument, it is possible to measure the transverse pulse velocity and the
Rayleigh pulse velocity as well, when the test is being made across a corner of
a structure and the path length involved is fairly long (in the order of 4 to 6 m
(15 to 20 ft)). Under these circumstances it is theoretically possible to
calculate Poisson's ratio directly from any two of the measured velocities.
The occurrence is sufficiently rare with present equipment, however, to merit
no more than mention at this time.
The use of pulse velocity techniques for testing concrete has been suggested for evaluating the strength of concrete, its uniformity, its setting
characteristics, its modulus of elasticity, and the presence or absence of
cracks within the concrete. There appears to be little question of the
suitability of such techniques to determine the presence, and to some extent
the magnitude, of cracks in concrete, although it has been suggested that if
the cracks are fully water-filled their locations may be more difficult to ascertain. In all of the other fields of investigation, independent investigators have
reported widely different degrees of success through the use of these techniques [I 7].
The most complete report of the experiences of users of pulse velocity
techniques in the United States and Canada may be found in "Effects of
Concrete Characteristics on the Pulse Velocity--A Symposium," Bulletin
No. 206, Highway Research Board, 1958.
Experiences in the use of pulse velocity measurements for evaluating concrete quality have been reported in many other countries, notably Great
Britain and Russia. In some instances investigators in these countries have
appeared to have greater confidence in the use of such techniques for acceptance testing than has been the case in the United States.
Jones [16] has reported values of minimum velocity considered acceptable
209
for concrete for certain specific structural purposes in Great Britain as

follows:
prestressed concrete--T sections
prestressed concrete--anchor units
reinforced concrete frame building
suspended floor slab
15 000 ft/s
14 300 ft/s
13 500 ft/s
15 500 ft/s
In a later publication [21], however, he stated that "pulse-velocity has rarely

been used as an acceptance criterion for structural concrete because it is
usually too difficult in site-work to ensure that the control of mix proportions, size and type of coarse aggregate remains uniform."
Several reports are available of the use of pulse velocity techniques in
Russia. Filina [22] has reported in some 80 various items of equipment for
testing concrete which are in use in the USSR, including those making use of
ultrasonics and vibrations.
Zashchuk and Nefedova [23] have described the design and operation of
ultrasonic testing equipment used in connection With the construction of the
Moscow Ring Road. They state that toward the completion of the road
ultrasonic methods replaced all other techniques.
It is generally agreed that very high velocities are indicative of very good
concrete and that very low velocities are indicative of poor concrete. It is further agreed that periodic, systematic changes in velocity are indicative of
similar changes in the quality of the concrete. Beyond these areas of agreement, however, it appears that the investigator must have a rather intimate
knowledge of the concrete involved before attempting to interpret velocities
as measures of strength or other properties of the concrete. This is particularly true if the aggregate involved is a lightweight aggregate.
In-Situ Testing for Strength Properties
In current concrete practice, the accepted methods of evaluating the
quality of concrete in structures consist of testing simultaneously cast companion specimens in compression and flexure. The main disadvantages of
this approach are: the delay in obtaining test results; the specimens may not
be truly representative of concrete in a structure because of different placing,
compacting, and curing conditions; the necessity of testing the specimens to
failure; the lack of reproducibility in test results; and the high cost of testing.
These, combined with the fact that structural units are considerably larger
and more massive, casts further doubts as to the validity of such
measurements of the strength ofin-situ concrete. As a result, there have been
a large number of attempts over the past 40 years to develop quick, inexpensive, and reproducible methods for testing concrete in structures.
Because the direct determination of strength implies that concrete
specimens must be loaded to failure, it becomes abundantly clear that
210
nondestructive methods of testing concrete cannot be expected to yield absolute values of strength. These methods, therefore, attempt to measure
some other property of concrete from which an estimate of its strength is obtained. Some such properties of concrete are its hardness, its resistance to
penetration by projectiles, and its rebound number.
The following sections describe in some detail the surface hardness, rebound, penetration, and pull-out techniques. Also included is a brief
description of the maturity concept. Although these tests are relatively simple
to perform, the analysis and interpretation of the test data are not so easy
because concrete is a complex material. Therefore, engineers are cautioned
that interpretation of the test data must always be carried out by specialists
in this field rather than by technicians performing the tests.
Surface Hardness Methods

The known surface hardness methods are of the identation type, and these
consist essentially of impacting the surface of concrete in a standard manner,
using a given mass activated by a given energy, and measuring the size of indentation.
The basic principles of various indentation devices to measure the hardness of concrete have been outlined by Gaede [24] and Vassitch [25]. There is
little apparent theoretical relationship between the strength of concrete and
its hardness as so measured. However, within limits, empirical correlations
have been established between strength properties and the data obtained
from surface hardness tests.
The three known methods employing the identation principle are: (a)
Williams testing pistol, (b) Frank spring hammer, and (c) Einbeck pendulum hammer.
Williams Testing Pistol--In 1936, Williams [26] reported the development
of a testing pistol weighing about 0.9 kgf (2 lb) that uses a ball as an indentor. The diameter of the impression made by the ball is measured by a
magnifying scale or other means. The depth of indentation is only about 1.5
mm (0.06 in.) for concrete with compressive strengths as low as 7 MPa (1000
psi).
On the basis of several hundred tests, Williams established the following
relationship: fc is proportional to 1/Z' wherefc is compressive strength, Z is
the curved surface area of indentation. Because of the difficulty of precisely
measuring the curved surface area, this method was not accepted widely.
Frank Spring HammerwThe Frank spring hammer consists of a springcontrolled mechanism housed in a tubular frame. The tip of a hammer can
be fitted with different diameters of balls, and impact is achieved by placing
the hammer against the surface under test and manipulating the spring
mechanism. Generally, about 20 impact readings are taken at short
distances from one another, and the mean of the results is considered as one
211
test value. The diameter or depth of identation (or both) is measured and
this, in turn, is correlated with the compressive strength of concrete.
Einbeck Pendulum H a m m e r w T h i s hammer consists of a horizontal leg,
at the end of which is pivoted an arm with a pendulum head weighing about
2.3 kgf (5 lb). The indentation is made by holding the horizontal leg against
the concrete surface under test and allowing the pendulum head to strike the
concrete. The diameter and depth of indentation are measured, and these
are then correlated with the compressive strength of concrete.
This hammer can be used on vertical surfaces only and is, therefore, less
versatile then the Frank spring hammer.
The surface hardness tests are simple to use and provide a large number of
readings in a short time. According to Jones [27], the impact hammers have
been standardized in a German standard [28,29]. To interpret the data correctly, it is desirable to know the mix proportions, type of coarse aggregate
used, age, and moisture condition of the concrete under test.
According to Weil [30] and a RILEM 4 working group [31], the strength of
concrete under investigation can be predicted with an accuracy of 20 to 30
percent by the use of test hammers. Williams has claimed somewhat better
accuracy with the testing pistol.
Rebound Method
In 1948, a Swiss engineer, Ernst Schmidt [32], developed a test hammer
for measuring the hardness of concrete by the rebound principle. There appears to be little apparent theoretical relationship between the strength of
concrete and the rebound number of the hammer. However, within limits
empirical correlations have been established between strength properties and
the rebound number. Kolek [33] also has attempted to establish a correlation
between the hammer rebound number and the hardness as measured by the
Brinnel method.
The Schmidt hammer consists of a spring-controlled hammer mass that
slides on a plunger within a tubular housing. When the plunger is pressed
against the surface of the concrete, it retracts against the force of the spring;
when completely retracted, the spring is released automatically. The hammer
impacts against the concrete and the spring-controlled mass rebounds, taking a rider with it along the guide scale. By pushing a button, the rider can be
held in position to allow readings to be taken. While the hammer is still in its
testing position, the sliding index is read to the nearest whole number. This
reading is designated as the hammer rebound number.
The detailed procedure for calibrating the hammer has been described
elsewhere [34,35]. A typical relationship between compressive strength and
4Reunion Internationaledes Laboratoiresd'Essais et de Recherchessur les Materiaux et les
Constructions.
212
TESTS
AND
PROPERTIES
OF CONCRETE
rebound number for limestone aggregate concrete, obtained by Zoldners

[36], is shown in Fig. 1. In 1975, ASTM issued a tentative Test for Rebound
Number of Hardened Concrete. (C 805).
,
55OO
Hammer type : N-2
Hammer number : 3 0 8 0
W . 5000
rp
Z
i
04
p
"
Totolnumberofcy,inflereteeted =460
t .~/~,'" [~//
" " "'/
]'
" 't
9
. +
I / " -. "." *" .:, " 9
40O0
' ~ - , , "
I
//
"/
,.." .
>t)
" I .///
,,'o\o
.
350 _z
..J
,,". "-.C}- ::.-:/~.. "

9.. =: -.-. ...
d
~J
Q
r/
"" ~'" "" f'~"

9
~87
" o\o t /
Coarse agqregate ; crus.~e, ,,mestone

Free aggregate : natural eand
4500
+ /
,,/
50<)
if)
/
X
550C
8 Nu
ZO(?
5000
UJ
~.1
250C
,," " /
t0j
/ / ~ /
200C
0
~'~
ill;~
/
+o
~5
.-/,/
~ r a b o u n d
~," "
".',' !
#
;'v ~
'
20
number is overage of iO hammer reoclings
On each one of the test cylinders

.
.
.
.
.
.
All cylinders were tested in SSD condition
25
REBOUND N U M B E R , ( h a m m e r
50
150
0
>-
35
40
le5 ~
C~
in horizontal p o s i t i o n )
FIG. 1--Relationship between compressive strength and rebound number for limestone
aggregate concrete obtained with Type N-2 hammer. From Ref 35.
Although the rebound hammer provides a quick, inexpensive means of

checking uniformity of concrete, it has many serious limitations and these
must be recognized. For example, the results of the Schmidt rebound hammer are affected by the following: (a) smoothness of surface under test, (b)
size, shape and rigidity of the specimen, (c) age of the specimen, (d) sm'face
and internal moisture condition of the concrete, (e) type of coarse aggregate,
( f ) type of cement, (g) type of mold, and (h) carbonation of concrete surface.
These limitations are discussed in detail elsewhere [35,37-39].
According to Kolek [33] and Malhotra [35], there is a general correlation
between compressive strength of concrete and the hammer rebound number.
However, there is a wide degree of disagreement among various research
workers concerning the accuracy of the estimation of strength from the rebound readings. By consensus, the accuracy of estimation of compressive
strength of specimens cast, cured, and tested under laboratory conditions by
a properly calibrated hammer lies between _+ 15 and -+-20 percent. However,
the probable accuracy of prediction of concrete strength in a structure is +_ 25
213
percent. It cannot be overemphasized that the rebound hammer must not be

regarded as a substitute for standard compression tests but as a method for
determining the uniformity of concrete in the structure and comparing one
concrete against another.
Penetration Techniques
The evaluation of hardness by probing techniques was first reported by
Voellmy [40] in 1954. The use of two techniques was reported. In one case, a
hammer known as Simbi was used to perforate concrete, and the depth of
borehole was correlated to the compressive strength of cubes. In the other
technique, the probing of concrete was achieved by blasting with spit pins,
and the depth of the penetration of the pins was correlated with the compressive strength of concrete.
Apart from the data reported by Voellmy, there is little published work
available on these tests, and they appear to have received little acceptance in
Europe or elsewhere.
In 1966, a new technique known as the Windsor probe was introduced in
the United States for in-situ testing of concrete [41]. Since its introduction, a
number of organizations in both the United States and Canada have carried
out studies with this technique. Arni [42] has reported results of a detailed
investigation on the evaluation of the Windsor probe, and Malhotra [43, 44]
has reported results of his investigations on both 152 by 305-mm (6 by 12-in.)
cylinders and 610 by 610 by 200-mm (24 by 24 by 8-in.) concrete slabs. In
1975, ASTM issued Test Penetration Resistance of Hardened Concrete (C
803). At the present time, the only equipment which is known to be available
and meets the requirements of ASTM Test C 803 is the Windsor probe.
The Windsor probe equipment consists of a powder-activated 1gun or
driver, hardened alloy probes, loaded cartridges, a depth gage ibr measuring
penetration of probes, and other related equipment. The probes have a
diameter of 6.3 mm (0.25 in.), a length of 79.5 mm (3.125 in.), and a
frustoconical point on the front. The rear of the probe is threaded and screws
into a probe driving head which is 12.6 mm (0.5 in.) in diameter and fits
snugly into the bore of the driver.
The method of testing is relatively simple. Briefly, the powder-activated
driver is used to fire a probe into the concrete, using a single probe templet.
The exposed length of the probe is measured by a calibrated depth gage. The
exposed length of the probe is taken as a measure of the compressive strength
of concrete.
A relationship between the exposed probe length and 28-day compressive
strength of 152 by 305-mm (6 by 12-in.) cylinders is shown in Fig. 2. The
relationships established by Arni [42], Malhotra [44], and Law and Butt [45]
are shown in Fig. 3.
The Windsor probe is basically a hardness tester and provides an excellent
214
TESTS A N D PROPERTIES OF CONCRETE
mm
30
40
~2
60
50
......... ] ..........................
Y= - 5 3 3 3 + 5385 X psi - IS.E= 327 psi
/
3000
o ./~950/.
41
Confidence Umits
....
21
>-
.oi
Y/'J .......... t
c~
oo
cv
~
I
~0
'"
Aggregate type = limestone

Maximum size = I in.(25mm)
Size of slabs probed 2 4 X 2 4 X ~ in. (610 X 6 1 0 X 2 0 0 mm)
1.2
1.4
].6
'~
2.0
2.2
2.4
01.
-Jr',
2-6~
Exposed Probe Length, In.- X
FIG. 2--Relationship between exposed probe length and 26-day compressive strength of
concrete. From Ref 35.
mm
30
ff 600(: .
40
50
60
4.000
~
~
(/I
AI:~,Molhotro I
. . . . . . . . . . . .
,,~
L ._.}
5oo . . . . . . . . . .
z,
LO
12
28
14
16
18
20
:~
14
22
Exposed Probe Length, in.
FIG. 3--Relationship between exposed probe length and 28-day compressive strength of
concrete as published by different investigators. From Ref 35.
215
means of determining the relative strength of concrete in the same structure

or relative strengths in different structures, without extensive calibration
with specific concretes. Because of the very nature of the test equipment, it
cannot and should not be expected to yield absolute values of strength.
The Windsor probe equipment is simple, rugged, and needs little
maintenance except for occasional cleaning of the barrel of the gun. The
system has a number of built-in safety features that prevent accidental
discharge or escape of the projectile from the gun. However, wearing of
safety glasses is strongly recommended.
The penetration of the probe into the concrete is affected by the hardness
of the aggregate as measured on Mohs' scale of hardness. 5 Therefore, it is
desirable for each user of the Windsor probe to prepare his own calibration
charts for the type of concrete under investigation; with change in source of
aggregates, new calibration charts become mandatory.
The published data by Arni [42], Malhotra [44], and Gaynor [46] indicate
that the variation in the probe test results is large compared with the variation in compressive strength on companion specimens. The within-batch
standard deviation and coefficient of variation, as reported by various investigators, are shown in Table 1.
Pull-Out Tests
Utilizing a special dynamometer, a pull-out test measures the force required to pull out from concrete a specially shaped steel rod whose enlarged
end has been cast into the concrete (Fig. 4). Because of its shape, the steel
rod is pulled out with a cone of the concrete. The concrete is simultaneously
in tension and in shear, the generating lines of the cone running at approximately 45 deg to the direction of pull (Fig. 5). The pull-out force is then
related to compressive strength, the ratio of the pull-out to compressive
strength being between 0.1 and 0.3.
The pull-out techniques, though in use in the Union of Soviet Socialist
Republics [47] since 1935, are relatively new elsewhere. In 1944, Tremper
[48] in the United States reported results of laboratory studies dealing with
pull-out tests covering strengths up to 35.2 MPa (5000 psi). In 1968, Tassios
[49], in Greece, reported the development of a test in which a standard nail,
34 mm (1.37 in.) long and 4 mm (0.16 in.) in diameter, is driven into a concrete surface using a gun. Ten minutes after driving, the nail is extracted, the
necessary pull-out force being measured on a manometer. In recent years,
Richards [50], in the United States, has advocated these tests on structural
concrete members. Gaynor [51], Malhotra [52], and Rutenbeck [53] have
reported data on the pull-out tests proposed by Richards.
s Named after mineralogist Mohs who deviseda scale of hardness in ~r
talc, the softestof
all minerals, is given No. 1 and diamond, the hardest of all known substances, is No. 10.
semilightweight (Expanded shale as

coarse aggregate)
25
25
19
3/4
quartz
19
19
3/4
3/4
19
25
3/4
50
mm
in.
16 by 20 by 8-in. slab
(410 by 510 by 200-mm)
16 by 20 by 8-in. slab
(410 by 510 by 200-mm)
6 by 12-in. cylinders
(152 by 305-mm)
24 by 24 by 8-in. slabs
(610 by 610 by 200-ram)
6 by 6 by 66-in. prisms
(150 by 150 by 1690-mm)
24 by 24 by 8-in. slabs
(610 by 610 by 200-ram)
6 by 23 by 48-in. walls
(150 by 580 by 1210-mm)
6 by 23 by 48-in. walls
(150 by 580 by 1210-mm)
Type of
Specimens
Tested
256,I
384e
48a
28 c
48a
20r
189b
136b
Total
Number
of
Probes
3 and 91
3 and 91
7 and 28
35
7 and 29
7 and 28
3, 7, and 28
3, 7, and 28
Age of
Test,
days
0,17
0.16
0.087
0.062
0.054
0.124
0.105
0.143
in.
6.8
5.5
7.7
3.4
3.4
5.5
mm
3.62
2.66
3.14
1.37
1.57
2.21
4.30
4.05
Coefficient
of
Variation,
%
Standard
Deviation
NoTE--Compressive strength of 6 by 12-in. (152 by 305-mm) cylinders ranged from 2000 to 6000 psi (13.8 to 41.3 MPa).
a From Ref 35.
b4 probes per test.
c2 probes per test.
a3 probes per test.
el6 probes per test.
f9 probes per test.
Gaynor
limestone
Malhotra
gravel
gravel, limestone,
trap rock
Type of
Aggregate
Used
Arni
Investigation
Reported
by
Maximum
Aggregate
Size
TABLE 1--Within-batch standard deviation and coefficient of variation of probe measurements, a
m
--4
m
0
z
0
"1"1
--4
"o
m
"o
217
I
/
Bearing Ring
\\
Threaded Shaft
Embedded Head
--------A
\
//
//
\\
d I =5.00in. (12ZOrnrn)
//
d2=2.25 in. (57.1rnrn)

h =2.08in (52.8mm)
d2 -dl
0r =67 degrees
FIG. 4--Sketch showing the relative posttion and dimensions o f the ring-bearing plate and
the embedded head. From R e f 52.
5.O-in. (127.0 turn)
t_
'
2o.o. o,7o,. *a,,.,,

{51xg.Smrnl
Shaft
.j
'
'
:./
"
E
E
'
: F,
~
'1
e"~'
2.25-in. x 0.375-in. W a e h e r - - / 1
(,57.3 x 9.Srnm)
/Threaded
I"~1
l I --
T-
~1"-- I
I~
[~
// ~Steel
II/
Nuts
.........
FIG. 5 - - A line drawing of the pull-out assembly, together with the cone of concrete, immediately after the test. From R e f 54.
A relationship between the pull-out strength and 28-day compressive

strength of 152 by 305-mm (6 by 12-in.) cylinders is shown in Fig. 6. The
results obtained by Malhotra [52], and Malhotra and Carette [54] have been
compared with those reported by Rutenbeck [53] and are shown in Fig. 7.
Though these data cannot be compared directly because of the use of difCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:31:12 EDT 2014
218
TESTS A N D
PROPERTIES OF CONCRETE
MPo
I
. 8 0 0 o 2'
IM
O
Z
2.8
3,4
4.1
4.8
5,5
62
6.9
i
7.6
8.3
9.0
552
2B-DAY TEST RESULTS
-~ 700(]
J48.3
>t.)
6000
x
x
N 5ooo
z
Y = 6.24X-791 p s i - SE.= 684 psi

Z
Correlation coefficient = 0.9/
,~ 4 0 O 0
9
tO
---~
/ /
nt
Coorle ogg
Fine ogg
41.1
34.5
j
x
276
,NocmalPortlond ASTM TypeI

9 :5/4-in.(19-mm)n-~t. size crush~J grovel
=Noturol$ond
T 5000
Z
~ 2000 - -
138
Compresl~ve strength volues ore the overoge of three test results.
Pull-out slrenglh vOluesom oneltwo tilt results.
uJ I 0 0 0
0
300
400
500
600
700
800
900
I000
I100
1200
1:300
28-DAY PULL-OUT STRENGTH,P S I - - X

F I G . 6--Relationship between pull-out and 28-day compressive strength of 152 by 305-ram
cylinders. From Ref 54.
ferent materials and pull-out assemblies, the correlations appear to be much

alike.
The main advantages of the pull-out tests are that they do measure the insitu strength of concrete. The technique is simple, effective, and inexpensive.
The testing can be done in the field in a matter of minutes and the test results
are reproducible.
The pull-out strength is of the same order of magnitude as the direct shear
strength of concrete, indicating that the strength value obtained in the pullout test is probably a measure of the direct shear strength.
The major disadvantage of the pull-out test is that damage to the concrete
surface must be repaired. However, if a pull-out force of a given minimum
strength is applied without failure, it may be assumed that a minimum
strength has been reached for in-situ concrete and the structural unit need
not be stressed to failure. Another disadvantage is that the pull-out tests being used in North America have to be planned in advance and, unlike the
other in-situ tests, cannot be performed at random after the concrete has
hardened.
WHITEHURST
AND
MALHOTRA
1.4
2.8
ON
NONDESTRUCTIVE
TESTS
219
MPo
0
~176176
4.1
5.5
6.9
8.3
9.7
11'~52
7O00
=,
48.3
IK~.HOTRA and CARETTE (1975) -
~oo
x
41.1
_z 5000
34.5
400o
276
-r
(1972)
2017
~2ooo
i
I000
13.8
/~-~
2OO
RUTENBECK (1973)
400
600
800
I000
1200
28-13AY PULL-OUT STRENGTH,PSI
69
0
1600
1400
FIG. 7--Comparison of relationships obtained in this investigation with those obtained

by others. From R e f 54.
.==
!
|
8
300O
2500
E
i
2000
1500
./
K
~o
~4f~IC ROtiO=0-41
[
Corm ~rl~lte = ~
~4
LCal~otte
/~r Entroin~gand WoterReducing
. . . . . I0
Admlxturu Wenl Lh;ed
1000
500
400
600
800
I 0 0 0 1200 t400 1600 1800 2000
Moludfy Meter Reodlng-(C,entigrede Degree Houri )
FIG. 8--Relationships between maturity of concrete and its compressive strength.
220
Maturity Concept
It is well known that the compressive strength of well cured concrete increases with time. However, the increase in strength is governed by many factors other than curing time, the most important being the temperature of
curing. The combined effect of time and temperature has been the subject of
study by several investigators since 1904 but no hypothesis was formulated in
early years. Then in the 1950s, the concept of maturity was advanced by
Mclntosh [55], Nurse [56], Saul [57], and others [58,59], and strengthmaturity relationships were published. Maturity was defined as the product
of time and temperature with a datum temperature of --10~ (14~
In 1956, Plowman [60] examined relationships between concrete strength
and its maturity, and attempted to establish a rational basis for datum
temperature for use in maturity calculations. He defined the datum
temperature for maturity as the temperature at which the strength of concrete remains constant. As a result of his investigations, he concluded that
the datum temperature was -- 12.2 ~ (10 ~
The maturity of in-situ concrete can be monitored by thermocouples or by
instruments called maturity meters. The strength of in-situ concrete is then
estimated using prior relationships established between maturity and compressive strength of test cylinders. One such relationship is shown in Fig. 8.
Swenson [61] has used the maturity concept to estimate the strength gain
of concrete in structures. In recent years, the maturity concept has been used
to estimate the in-situ strength of concrete during construction stages in a
number of buildings in the Toronto area, the CN Tower being one of them
[62, 63]. Hudson and Steele [64] have used the maturity approach to predict
potential strength of concrete on highway projects in West Virginia using the
equation
logS28 = 2.9844 + 0.75 logSe -- 0.51 logm
(10)
where
S2s = predicted 28-day compressive strength, psi
Se = compressive strength (psi) of specimens tested at an early age and
having a maturity, m, and
m = degree hours of maturity at the time of test (~ h).
Malhotra [65] and others [66] have attempted to relate compressive
strengths obtained using accelerated-strength tests with the maturities for
these tests.
The advantages of the use of the maturity concept are obvious; with proper
use of in-place thermocouples, strength of concrete in structures can be
monitored successfully. However, this concept has serious limitations and
these must be recognized. It is generally agreed that the maturity concept is
applicable only if: (a) testing of concrete is confined within the range of
221
maturity represented by about 3 to 28 days at normal temperatures, (b) the

initial temperature of concrete is between 15.6 and 26.8~ (60 and 80~
(this is a rather limited range) and (c) no loss of moisture by drying occurs
during the curing period.
Summary
It has been the experience of users of dynamic testing that many factors influence the results in varying degrees. Recognizing these influences and making proper allowances for them are basic requirements in evaluating such test
data.
Usually the most pronounced influence on results is the type of aggregate
used in the concrete because of the wide ranges possible in elasticity and density. Frequently the maximum size of coarse aggregate must be considered in
comparing different series of tests. Other factors influencing results include
moisture content, temperature, density, reinforcement, type of cement,
aggregate-cement and water-cement ratios, and admixtures. The most exhaustive investigation of these factors reported to date is that of Jones [16].
It appears that the sustained frequency tests have their main application in
tracing the course of deterioration in specimens subject to weathering or exposure tests. In general, application of these techniques to structures, including pavements, is hampered by boundary effects, power requirements,
and difficulties in the interpretation of measurements on any but the most
elementary forms.
The major applications of pulse velocity tests on concrete are to establish
the degree of uniformity, or lack thereof, throughout a structure, to follow
progressive changes in the quality of concrete in either specimens or structures, and to determine the presence or absence of cracking in monolithic
concrete. Velocity tests on concrete are not hampered significantly by size
and shape effects. The level of effective pulse transmission is the limiting
condition that will govern the operating range, depending on the
characteristics of instrumentation and on the inherent property of concrete to
attenuate the impulse. This maximum range may vary from perhaps a few
centimetres (inches) in unset concrete to 15 to 18 m (50 to 60 ft) in sound concrete of high quality. The attenuation of the impulse may lead to important
errors, particularly in the case of those devices which do not provide for
visual examination of transmitted and received signals.
With respect to the prediction of values for other properties of concrete on
the basis of the results of dynamic tests, it is both desirable and proper to
compute the dynamic modulus of elasticity or rigidity from the appropriate
resonant frequency. Such computation is necessary if results of tests on
specimens of different sizes and shapes are to be compared. Since both the
resonant frequency and the weight of the specimen can be measured directly
and Poisson's ratio does not enter heavily into the computations (if
222
longitudinal resonance is measured), it is believed that there is little danger

of the introduction of significant error in making these computations. The
use of resonant frequency techniques for predicting other properties of concrete does not appear to be well supported by data presently available,
although the work of Kesler [4] and Chang and Kesler [5, 6] indicates that
two dynamic parameters may define the strength of concrete and its creep
characteristics.
With respect to pulse velocity techniques, there seems to be very little
reason for computing anything other than pulse velocity from the results of
such tests. The computation of dynamic modulus of elasticity from pulse
velocity requires a knowledge of both the unit weight of the concrete and
Poisson's ratio, both of which values at least in the case of tests on structures,
must be estimated. Since results of these tests are not dependent upon size
and shapes of the concrete tested, direct comparisons may be made between
tests made on different concretes or different sections of the same concrete.
Batchelder and Lewis [20] have shown that the velocity itself correlates better
with the results of resonant frequency tests on laboratory specimens than
does the modulus of elasticity computed from velocity tests. No evidence has
yet been presented to suggest that any better relationship exists between
pulse velocities and other properties of the concrete than between resonant
frequencies and these properties.
It is emphasized that all the in-situ tests and concepts discussed in the
paper cannot and do not yield absolute values of compressive strength of concrete in a structure and must not be considered as a substitute for standard
compression tests. The pull-out test does measure strength but this is
probably the shear strength of concrete from which an estimate of compressive strength may be made. However, the techniques discussed are
satisfactory for determining relative strengths of concrete in different parts of
the same structure or relative strengths in different structures.
References
[1} Powers, T. C., "Measuring Young's Modulus of Elasticity by Means of Sonic Vibration,"
Proceedings, Vol. 38, Part II, 1938, p. 460.
[2] Thomson, W. T., "Measuring Changes in Physical Properties of Concrete by the Dynamic
Method," Proceedings. American Society for Testing and Materials, Vol. 40, 1940, p.
1113.
[3] Obert, L. and Duvall, W. I., "Discussion of Dynamic Methods of Testing Concrete with
Suggestions for Standardization," Proceedings, American Society for Testing and
Materials, Vol. 41, 1941, p. 1053.
[4] Kesler, C. E., and Higuchi, Y., "Determination of Compressive Strength of Concrete by
Using Its Sonic Properties," Proceedings, American Society for Testing and Materials,
Vol. 53, 1953, p. 1044.
[5] Chang, T. S. and Kesler, C. E., "Prediction of Creep Behavior in Concrete from Sonic Properties," Proceedings, Highway Research Board, Vol. 35, 1956, p. 436.
[6] Chang, T. S. and Kesler, C. E., "Correlation of Sonic Properties of Concrete with Creep
and Relaxation," Proceedings, American Society for Testing and Materials, Vol. 56, 1956,
p. 1257.
223
[7] Long, B. G. and Kurtz, H. J., "Effect of Curing Methods Upon the Durability of Concrete
as Measured by Changes in the Dynamic Modulus of Elasticity," Proceedings, American
[8] Pickett, G., "Dynamic Testing of Pavements," Journal, American Concrete Institute,
April 194S; Proceedings, Vol. 41, p. 20.
[9] Stanton, T. E., "Tests Comparing the Modulus of Elasticity of Portland Cement Concrete
as Determined by the Dynamic (Sonic) and Compression (Secant at 1000 psi) Methods,"
Bulletin No. 131. American Society for Testing and Materials, Dec., 1944, p. 17.
[10] Obert, L., "Measurement of Pressures on Rock Pillars in Underground Mines," R. L
3521, U.S. Bureau of Mines, 1940.
[11] Long, B. G., Kurtz, H. J., and Sandenaw, T. A., "An Instrument and a Technique for
Field Determination of the Modulus of Elasticity of Concrete (Pavements)," Journal,
American Concrete Institute, Jan. 1945; Proceedings, Vol. 41, p. 11.
[12] Leslie, J. R. and Cheesman, W. J., "An Ultrasonic Method of Studying Deterioration and
Cracking in Concrete Structures," Journal, American Concrete Institute, Sept. 1949; Proceedings, Vol. 53, t953, p. 1043.
[13] Parker, W. E., "Pulse Velocity Testing of Concrete," Proceedings, American Society for
Testing and Materials, Vol. 53, 1953, p. 1043.
[14] Whitehurst, E. A., "Soniscope Tests Concrete Structures," Journal, American Concrete
Institute, Feb. 1951; Proceedings, Vol. 47, p. 433.
[15] Jones, R., "The Non-Destructive Testing of Concrete," Magazine of Concrete Research,
No. 2, June, 1949.
[16] Jones, R., "The Testing of Concrete by an Ultrasonic Pulse Technique," Proceedings,
Highway Research Board, Vol. 32, 1953, p. 258.
[17] Whitehurst, E. A., "A Review of Pulse Velocity Techniques and Equipment for Testing
Concrete," Proceedings, Highway Research Board, Vol. 33, 19S4, p. 226.
[18] Reunion des Laboratoires d'Essais et de Recherches sur les Materiaux et les Constructions
(Union of Testing and Research Laboratories for Materials and Structures), Bulletins Nos.
13-18, March 19S3-June 1954.
[19] Cheesman, W. J., "Dynamic Testing of Concrete with the Soniseope Apparatus," Proceedings. Highway Research Board, Vol. 29, 1949, p. 176.
[20] Batchelder, G. M. and Lewis, D. W., "Comparison of Dynamic Methods of Testing Concrete Subjected to Freezing and Thawing," Proceedings, American Society for Testing and
Materials, Vol. 53, 19S3, p. 1053.
[21] Jones, R., Non-Destructive Testing of Concrete, Cambridge University Press, 1962.
[22] Filina, T. A., "Pribory dlya Kontrolya Kachestva betonnykh i zhelezobetonnykh
konstruktsii," Beton i Zhelezobeton, No. 3, March 1961.
[23] Zashchuk, I. V. and Nefedova, E. F., "The Control of Concrete Quality in Road
Pavements Without Destructive Tests," Avtom. Dorogi, Vol. 26, 1963, (in Russian).
[24] Gaede, Kurt, "Ball Impact Testing for Concrete (Die Kugelschlagpriifung yon Beton),"
Bulletin No. 107, Deutscher Ausschuss for Stahlheton, Berlin, 1952, pp. 15-30.
[25] Vassitch, Pavle, Proceedings, International Symposium on Nondestructive Testing of
Materials and Structures, Vol. 2, RILEM, Paris, 1954, pp. 301-306.
[26] Williams, J. F., The StructuralEngineer, Vol. 14, No. 7, London, July 1936, pp. 321-326.
[27] Jones, R., Proceedings, Symposium on Nondestructive Testing of Concrete and Timber,
Institution of Civil Engineers, London, June 1969, pp. 1-7.
[28] "Ball Impact Test for Normal Concrete (Kugelschlagpriifung yon Beton mit dichtem
Gef0ge)," Standard Code DIN 4240, No. 6, German Committee for Reinforced Concrete,
1966, pp. 311-319.
[29] "Ball Test for Cellular Concrete (Kugelschlagprtlfung von Gasund Schaumbeton)," Draft
Code of Practice (June 1955), Beton und Stahlbetonbau (Berlin), Vol. 50, No. 8, Aug.
1955, pp. 224-225.
[30] Well, Gustav, Proceedings, International Symposium on Nondestructive Testing of
[31] "Report of the RILEM Working Group on the Nondestructive Testing of Concrete,"
RILEMBulletin, New Series No. 27, Paris, June 1965, pp. 121-125.
[32] Sehmidt, Ernst, Proceedings, International Symposium on Nondestructive Testing on
224
[33] Kolek, J., Magazine of Concrete Research, Vol. 10, No. 28, London, March 1958, pp.
27-36.
[34] Malhotra, V. M., "Nondestructive Methods for Testing Concrete," Mines Branch
Monograph No. 875, Department of Energy, Mines and Resources, Ottawa, 1968.
[35] Malhotra, V. M., "Testing Hardened Concrete: Nondestructive Methods," ACI
Monograph No, 9, Detroit, 1976.
[36] Zoldners, N. G., Journal, American Concrete Institute, Proceedings, Vol. 54, No. 2, Aug.
1957, pp. 161-165.
[37] Greene, G. W., Journal, American Concrete Institute, Proceedings, Vol. 51, No. 3, Nov.
1954, pp. 249-256. Discussion pp. 256-1 to 256-20.
[38] Mitchell, L. J., and Hoagland, G. G., Bulletin No. 305, Highway Research Board, 1961,
pp. 14-27.
[39] Grieb, W. E., Public Roads, Vol. 30, No. 2, June 1958, pp. 45-50.
[40] VoeUmy, A., Proceedings, International Symposium on Nondestructive Testing of
[41] Kopf, R. J., SP 22, American Concrete Institute, Detroit, 1969, pp. 55-68.
[42] Arni, H. T., Highway Research Record. Highway Research Board, No. 378, 1972, pp.
55-67.
[43] Malhotra, V. M., "Preliminary Evaluation of Windsor Probe Equipment for Estimating
the Compressive Strength of Concrete," Mines Branch Investigation Report IR 71-1,
Department of Energy, Mines and Resources, Ottawa, Dec. 1970.
[44] Malhotra, V. M., Materials and Structures, Vol. 7, RILEM, No. 37, Jan.-Feb. 1974, pp.
3-15.
[45] Law, S. M., and Burt, W. T., III, "Concrete Strength Study," Research Report No. 44,
Research Project No. 68-2C(B), Louisiana HPR (7), Louisiana Department of Highways,
Dec. 1969.
[46] Gaynor, R. D., "In-Place Strength of Concrete--A Comparison of Two Test Systems,"
presented at 39th Annual Convention of the National Ready Mixed Concrete Association
(New York, Jan. 28, 1969). Published with NRMCA Technical Information Letter No. 272,
4 Nov. 1969.
[47] Skramtajew (sic) (also Skramtaev), B. G., Journal, American Concrete Institute, Proceedings, Vol. 34, No. 3, Jan.-Feb. 1938, pp. 285-303; Discussion, pp. 304-305.
[48] Tremper, Bailey, "The Measurement of Concrete Strength by Embedded Pull-Out Bars,"
Proceedings. American Society for Testing and Materials, Vol. 44, 1944, pp. 880-887.
[49] Tassios, T. P., "A New Nondestructive Method of Concrete Strength Determination,"
Publication No. 21, National Technical University, Athens, 1968.
[50] Richards, Owen, "Pull-Out Strength Tests of Concrete," Paper presented at the Research
Session, Annual Meeting of American Concrete Institute, Dallas, Texas, 1972.
[51] Gaynor, R. D., Internal Laboratory Reports issued by National Ready-Mixed Concrete
Association, Washington, D. C.
[52] Malhotra, V. M., Materials and Structures. Vol. 8, No. 43, RILEM Jan.-Feb. 1975, pp.
19-31.
[53] Rutenbeck, Todd, "New Developments in In-Place Testing of Shotcrete," Paper presented
at the 1973 Engineering Foundation Conference on Use of Shotcrete for Underground
Structural Support, Berwick Academy, South Berwick, Maine.
[54] Malhotra, V. M. and Carette, G., "Comparison of Pull-Out Strength of Concrete With
Compressive Strength of Cylinders and Cores, Pulse Velocity, and Rebound Number,"
CANMET Report 76-8, Department of Energy, Mines and Resources, Ottawa, Canada,
1976.
[55] Mclntosh, J. D., Magazine of Concrete Research, Vol. 1, No. 1, 1949, pp. 21-28.
[56] Nurse, R. W., Magazine of Concrete Research. Vol. 1, No. 2, 1949, pp. 79-88.
[57] Saul, A. G. A., Magazine of Concrete Research, Vol. 2, No. 6, pp. 127-140.
158] Bergstrom, S. C., Magazine of Concrete Research, Vol. 14, No. 14, pp. 61-66.
[59] Rastrup, Erik, Magazine of Concrete Research, Vol. 61, No. 17, pp. 79-92.
[60] Plowman, J. M., Magazine of Concrete Research, Vol. 8, No. 22, pp. 13-22, Discussion,
Vol. 8, No. 24, pp. 169-183.
[61] Swenson, E. G., Engineering Journal (Canada), Vol. 50, No. 9, pp. 27-32.
[62] Bickley, J. A., Transportation Research Record 558, Transportation Research Board,
1975, pp. 45-49.
225
163] Mukherjee, P. K., Transportation Research Record 558, Transportation Research Board,
1975, pp. 87-92.
[64] Hudson, S. B. and Steele, G. W., Transportation Research Record 558, Transportation
Research Board, 1975, pp. 1-12.
[65] ~Malhotra, V. M., "Maturity Concept and the Estimation of Concrete Strength--A Review,
Mines Branch Information," Circular IC 277, Canada, Department of Energy, Mines and
Resources, Ottawa, Canada, 1971.
[66] Ramakrishnan, V. and Dietz, J., Transportation Research Record 558, Transportation
J. L. S a w y e r 1
Chapter 16 Volume Change
Introduction
Concrete is subject to changes in volume during and after the hardening
period. This dimensional instability is of considerable importance to the construction industry, and the understanding of the nature of these changes is
important for those using concrete as a construction material. These volume
changes, if excessive, can result in high stresses and cracking, with resulting
poor performance of the concrete. It is an important consideration whether
the volume changes occur in hardened concrete or prior to hardening.
Normal volume changes in hardened concrete can result from changes in
temperature and moisture content, and due to applied loads. Many factors
affect the magnitude of these volume changes. Excessive volume change,
with resulting high stresses and cracking, can be prevented or minimized by
controlling the variables that affect volume changes.
Normal volume changes would be of little concern if the concrete were free
to deform. However, concrete usually is restrained by foundations,
subgrades, reinforcement, or connecting members, and stresses can be produced in the concrete which may cause distress and even failure. Due to concrete being weaker in tension than in compression, restained shrinkage causing tensile stresses is usually more important than expansion that causes
compressive stresses. Volume changes due to variations of temperature,
humidity, and loads are partly or entirely reversible, but volume changes due
to unsound materials or chemical and mechanical action are not reversible
and are cumulative as long as the action continues.
Reliable information is available on the major factors affecting the
magnitude of volume changes. When proper attention is given to these factors, concrete can be produced that is relatively free of cracks, with satisfactory dimensional stability.
The magnitude of volume changes generally is stated in linear rather than
volumetric units. This is for convenience, as these linear changes can be
measured easily and are of primary interest. The volume changes that or1Chief, Research and Development, Lone Star Industries, Houston, Tex. 77001.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
SAWYER ON VOLUME CHANGE
227
dinarily occur are small, ranging from negligible up to 0.1 percent change in
length. These length changes are often expressed in "millionths." For example, a change in length of 500 millionths of a metre per metre may also be expressed as 0.05 percent, or 0.05 m in 100 m (0.60 in. per 100 ft).
Volume Changes in Fresh Concrete

Freshly mixed concrete remains plastic for a relatively short time. During
this period, volume changes can occur due to cement hydration, thermal
change, absorption, and bleeding. Bleeding is the term used to describe the
accumulation of water at the surface of fresh concrete due to sedimentation
of solids. This begins shortly after the concrete is placed and continues until
maximum compaction of the solids, particle interference, or loss of plasticity
of the concrete due to setting stops further sedimentation. Reactions between
the water and the cement, absorption of water by the aggregates, and
evaporation of bleed water tend to decrease the volume of the concrete.
When the surface of the fresh concrete is subjected to rapid drying, surface
cracking of the plastic concrete can occur. This "plastic cracking" can occur
soon after the concrete has been placed. These cracks appear mostly on
horizontal surfaces and may have considerable depth. They may be practically eliminated by taking appropriate measures during construction to
reduce the rate and total amount of water evaporation from the surface of the
plastic concrete. Small changes in relative humidity, temperature, and wind
velocity can have a pronounced effect on rate of evaporation and, therefore,
cracking can occur one day but not the next.
Volume Changes in Hardened Concrete

Hardened concrete undergoes small changes in volume due to sustained
stress and changes in temperature and moisture content. Also, concrete may
be subjected to potentially destructive forces due to chemical and mechanical
attack.
Continuously Applied Stress

When a sustained load is applied to concrete, deformation caused by this
load can occur immediately. Also, a time-dependent deformation which
begins immediately but continues for years can result from this sustained
load. This latter deformation is called "creep." This change in volume or
length due to creep is largely unrecoverable in newly placed concrete.
However, it is largely recoverable when it occurs in old or dry concrete.
Although this creep continues for years, the rate decreases with time.
Creep is proportional to stress within normal stress ranges. However, concretes of equal strength but of different ages will have different creep
228
characteristics. Concrete loaded at a late age will creep less than a concrete of
equal strength loaded at an early age. The method of curing prior to loading
has an influence on the amount of creep of the concrete. Concrete that has
been cured with high-pressure steam (autoclaving) has less creep than atmospheric steam-cured concrete. Normal moist-cured concrete has the most
creep.
Therefore, creep is dependent upon the magnitude of the stress, the age
and strength of the concrete, and the duration of the sustained stress. Other
factors affecting creep are: type, amount, and maximum size of the aggregate; type of cement; amount of cement paste; size and shape of the concrete mass; amount of steel reinforcement; and curing conditions.
Temperature Changes
As the temperature of concrete rises and falls, concrete expands and contracts. This is a physical phenomenon common to all materials. However, for
concrete it is complicated by differential expansion of its components which
produces high internal stresses. These changes in volume and the internal
stresses produced by these changes have a strong influence on all types of
concrete structures. Consideration of these changes, and the application of a
proper thermal coefficient of expansion for concrete in structural design,
help to avoid damage to structures from internal stresses caused by the thermal volume changes of the concrete. These changes are largely dependent on
the volume changes of the concrete's two principal components, the cement
paste and the aggregate [1]. 2 The thermal coefficient of concrete can be
calculated by using the weighted average of the coefficients of its ingredients
[2,31.
The average value that is frequently used for this length change is around
10 10-6/~ (5.5 10-6/~ or 0.10 m for 100 m of concrete subjected to
a rise or fall of 100~ (0.66 in. for 100 ft for 100~
The coefficient of expansion is essentially constant over the normal temperature range and is usually
between 6.3 and 11.7 10-~176 (3.5 to 6.5 10-6/~ when the effect of
moisture change is taken into account [4]. This varies with factors such as aggregate type, mix proportions, water-cement (w/c) ratio and concrete age.
The coefficient of thermal expansion of a given concrete can vary with the
degree of water saturation of the concrete [5]. The mineral composition and
structure of aggregates are the major factors determining the coefficient of
thermal expansion of concrete [6].
Fortunately, these volume changes are on the same order of magnitude as
observed for steel and, consequently, reinforced concrete structures perform
satisfactorily over a wide range of temperature variations.
Thermal expansion of concrete is influenced by the type of aggregate, due
229
primarily to the large differences m the thermal coefficient of expansion of

the various types of aggregate, and even for aggregates of a given type but
from different sources. The coefficient for various rocks that are used for
concrete aggregates varies from 3.6 to 12.6 X 10-6/~ (2.0 to 7.0 X
10-6/~
Data indicate that the thermal expansion of rock and, therefore,
concrete is mainly influenced by the proportion of quartz present. Rocks,
such as quartzite and sandstone, have a high quartz content and, therefore,
the highest coefficient. Rocks with little or no quartz, such as limestone, have
the lowest coefficient. Igneous rocks with medium quartz content, such as
granite, rhyolite, basalt, etc., have intermediate coefficients [7].
Thermal changes in mass concrete are important, as the heat liberated due
to cement hydration can result in high maximum concrete temperatures, and
cracking can occur as the concrete cools from this maximum temperature.
Factors affecting the maximum concrete temperatures are: heat of hydration
of the cement, initial concrete temperature, ambient temperature, rate of
construction, and the specific heat and thermal conductivity of the concrete.
These thermal changes in mass concrete usually are minimized by one or
more of the following techniques: using low-heat portland cements (Types II
and IV); replacing part of the cement with a pozzolan; chilling the mix
water; chilling the aggregates; cooling the in-place concrete.
Thermal changes in pavements can be important as temperature gradients
can cause curling of the slab, with the possibility of cracking under traffic
loads [8].
Moisture Changes
Concrete expands due to a gain in moisture, and contracts with a loss in
moisture. Expansion of the moist concrete begins after the initial setting
shrinkage has taken place and, if kept continuously moist, it will orpand
slowly for years. However, the total expansion is normally small enough to be
unimportant. This total expansion is usually less than 0.25 percent. The
amount of cement is of greater importance than the type of cement, although
cements with more than the optimum sulfur trioxide (SOa) content can cause
more than normal expansion in moist concrete. The rate of expansion
decreases with time and is very small after several years.
Concrete that is allowed to dry will contract or shrink. Concrete that is exposed to the atmosphere loses some of its moisture content and, therefore,
concrete normally exists in a somewhat contracted state. Drying shrinkage
can cause larger volume changes than normally are observed due to expansion of moist concrete.
The amount of shrinkage observed as concrete dries is dependent upon
several factors, such as: amount and type of aggregate, amount of mixing
water, properties and amounts of admixtures, ambient temperature and
relative humidity, and the method of curing. The differences in shrinkage
230
due to cement type and period of moist curing, prior to drying, appear to be
of little practical significance [9]. It cannot be said that a cement, because it
conforms with the chemical and physical requirements of any one type of cement, will have therefore more or less shrinkage than a cement meeting
the requirements for another type [IO]. However, a cement that is deficient in
gypsum can result in more shrinkage than would be the case if the cement
contained the optimum amount of gypsum [11]. ASTM Measuring the Drying Shrinkage of Mortar Containing Portland Cement (C 596) can be used to
determine the effect of portland cement on the drying shrinkage of a graded
Ottawa sand mortar subjected to stated conditions of temperature, relative
humidity, and rate of evaporation of the environment. The drying shrinkage
of mortar as determined by this method has been shown to have a linear relation to the drying shrinkage of concrete made with the same cement and exposed to the same drying conditions. However, it seems inconsistent to rely
on the behavior of either neat pastes or rich mortars to predict the ultimate
shrinkage of concretes, especially if the tests are concluded at early ages [12].
Shrinkage of plain concrete in the range of 0.04 to 0.08 percent has been
observed when the concrete was exposed to air at 50 percent relative humidity. This means that concrete when dried from a saturated condition to a
state of equilibrium with air at 50 percent relative humidity can shrink about
0.06 m/100 m (0.72 in. for 100 ft). This is about the same as the thermal contraction caused by a decrease in temperature of 56~ (100~
For concrete
exposed to weather, the volume changes due to temperature and moisture
variations frequently tend to offset each other.
Carbonation shrinkage, due to the reaction between carbon dioxide and
cement constituents, takes place concurrently with drying shrinkage. Much
of the reported data have not differentiated between the two and combined
shrinkage is reported as drying shrinkage.
Oven-dry concrete or saturated concrete is not very susceptible to carbonate shrinkage. However, at intermediate humidities, carbon dioxide produces a gradual irreversible shrinkage. The amount of shrinkage varies with
the humidity and reaches a maximum at around 55 percent relative humidity.
Concrete that has been subjected to carbonation shrinkage will still shrink
and swell with changes in moisture content; however, the amplitude of these
changes is likely to be smaller than before carbonation.
Alternate cycles of wetting and drying cause cycles of swelling and shrinking. However, the swelling that occurs during continuous wet storage over a
period of several years is only about one-third of the shrinkage observed in
concrete air-dried for the same period. Shrinking and swelling of concrete
due to wetting and drying is almost exclusively a property of the cement paste
[131.
Normally aggregates act as a restraint on concrete volume changes due to
231
wetting and drying; however, it has been found that some aggregates can
have dimensional changes greater than 0.04 percent when dried and subsequently wetted [14]. This can contribute to excessive volume changes and
disintegration of concrete. Fortunately this is rare, as inferior aggregates are
generally rejected on the basis of tests other than dimensional change
characteristics.
The amount of water per unit volume of concrete is the most important
controllable factor affecting shrinkage. Even though the relationship between change of water content and change of volume is not the same for all
pastes [15], shrinkage can be minimized by keeping the water content of the
paste as low as possible. Also, keeping the total aggregate content of the concrete as high as possible will help to minimize shrinkage. The use of high
slump concrete, high fresh concrete temperatures, or any practice that increases the water requirements can increase shrinkage.
Aggregates embedded in the cement paste restrain the shrinkage of the cement paste; therefore, an increase in the amount of aggregate tends to decrease shrinkage. As concrete with large size aggregates tends to have a
greater total aggregate content, the use of the largest practical size minimizes
shrinkage. However, it has been reported that in practical concrete mixes the
size and gradation of the aggregates have little affect, as the effect on
shrinkage is a function only of the total quantity of the aggregates per unit
volume of the concrete [16].
The total amount and properties of the aggregates have an influence on
shrinkage. Some aggregates are hard and difficult to compress and,
therefore, are more resistant to shrinkage of cement paste. Aggregates with a
high modulus of elasticity and those with rough surfaces offer more restraint
to shrinkage. Higher values for dynamic Young's modulus of elasticity of
concrete have been found to be associated with lower shrinkage and expansion values [17]. Concrete made with sandstone, slate, hornblende, and
pyroxene may shrink up to two times as much as concretes made with
granite, quartz, feldspar, dolomite, and limestone [18]. The use of dirty
sands or unwashed coarse aggregate containing clay can cause excessive
shrinkage. Shrinkage of structural lightweight concrete varies from equal to
about double that of normal weight concrete [19].
Admixtures have a varying effect on the drying shrinkage of concrete. The
reported data are sometimes conflicting but, in general, an admixture that
increases the unit water content may be expected to increase drying
shrinkage. Water reducing admixtures generally have a tendency to reduce
drying shrinkage due to reduced water requirements [20]. However, all admixtures that reduce the unit water content do not reduce drying shrinkage.
Accelerators such as calcium chloride and triethanolamine tend to increase
shrinkage. Some water-reducing admixtures increase shrinkage although
they reduce water content. This is particularly true for those that contain an
232
accelerator to counteract the retarding effect of the admixture. Some

chemical admixtures can increase shrinkage by altering the SO3 content required for minimum shrinkage from that determined when the cement was
manufactured [21]. ASTM Specification for Chemical Admixtures for Concrete (C 494) limits the increase in drying shrinkage due to the admixture to
135 percent of the control concrete when the length change of the control is
0.030 percent or greater or an increase in shrinkage of not over 0.010 percent
when the length change of the control concrete is less than 0.030 percent.
Within normal limits, entrained air has little effect on drying shrinkage.
Pozzolanic materials, such as pumicite or raw diatomaceous earth, when
used as a replacement for cement, tend to increase shrinkage. However, it is
reported that the use of up to 30 percent fly ash--used as a replacement for
cement--did not increase shrinkage [22]. ASTM Specification for Fly Ash
and Raw or Calcined Natural Pozzolans for Use in Portland Cement Concrete (C 618) limits the increase in drying shrinkage due to these materials to
0.03 percent at 28 days with the tests being made on mortar bars.
Reinforced concrete shrinks less than plain concrete, with the difference
depending on the amount of reinforcement. The steel reinforcement tends to
restrain but does not prevent drying shrinkage. Structures with reinforced
concrete having normal amounts of reinforcement usually have 0.02 to 0.03
percent drying shrinkage.
The rate and total amount of shrinkage is affected by the size and shape of
the concrete unit. The rate and total shrinkage is less for a large mass of concrete than for a small size concrete unit; however, shrinkage continues over a
longer period for the larger unit. Some data indicate a surface volume to
shrinkage relationship that can be expressed as a volume to surface ratio.
Some care must be taken in relating shrinkage observed on small laboratory
specimens to shrinkage expected in large concrete members.
Volume changes in hardened concrete, particularly when the concrete is
restrained, produce tensile stresses and can result in cracking of the concrete.
Cracks in a concrete member are potentially weak spots with respect to applied load. Also, they offer avenues of egress and lower the resistance of the
concrete to attack by aggressive solutions. However, it can be argued that
although cracks are serious defects for some uses, in other uses cracks are little more than interesting phenomena [23].
The susceptibility of concrete to cracking is influenced by several factors.
Factors such as degree of restraint, amount of shrinkage (drying, carbonation, thermal), and tensile strength of the concrete effect the ability of concrete to resist cracking. Some of the stress produced in the concrete by these
factors can be relieved by creep of the concrete. A combination of the least
amount of shrinkage and the highest tensile strength should give good
resistance to cracking. It has been observed that shrinkage should be considered a rate process rather than in absolute terms when its effect on cracking of the concrete is considered [24].
233
Control of Volume Changes

Many factors influence volume changes, and it appears that several factors
can act at the same time to increase drying shrinkage. The net volume
change can be the product, rather than the sum, of the individual effects. A
combination of wrong choices could result in seven times as much shrinkage
as would result from the best choices [25].
When materials are such that the concrete has inherently high drying
shrinkage, the effects of additional adverse factors can be critical. The
following methods help to minimize volume changes [26].
I. Use the minimum water content required for proper placement.
2. Avoid high slumps and high concrete temperatures which tend to increase the water demand of the concrete.
3. Use the largest total amount of aggregate, as this tends to reduce water
requirements.
4. Where available, use fine and coarse aggregate that exhibit low
shrinkage characteristics when used in concrete.
5. Avoid aggregates that contain excessive amounts of clay, as clay functions like gel instead of restraining shrinkage [27].
Chemical Attack and Unsoundness

Volume changes can occur in concrete due to the concrete being subjected
to chemical attack. Signs of attack are generally first evident as expansion;
and, if the action continues, this expansion can increase until disintegration
occurs. This attack can be due to internal causes, such as unsound cement or
chemical reactions between cement and aggregates.
The undesirable effects of unsound cement have been recognized and
reflected in specifications over the years. Excessive concrete expansion due to
the hydration of uncombined or free calcium oxide (CaO) and hardburned,
crystalline magnesium oxide (MgO) (periclase) is referred to as unsoundness.
(It should be noted that the term unsoundness as used here is not the same as
unsoundness when used for aggregate.) The presence and amount of these
materials in cements depends on several factors such as burning and cooling
conditions during production of the cement. The hydration of these materials
is slow and their potential expansion is delayed [28]. Tests made during the
manufacture of portland cements and for acceptance testing of cement include tests for free CaO and MgO as well as for SO3 and alkalies, as these can
affect the volume stability of concrete under certain conditions. For these
chemical tests, ASTM Chemical Analysis of Hydraulic Cement (C 114) can
be used. The above plus the other constituents of the cement are determined
by this method.
This potential unsoundness is normally detected by the ASTM Test for
Autoclave Expansion of Portland Cement (C 151), and the amount of expanCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:45 EDT 2014
234
sion is limited in cement specifications. ASTM Specification for Portland Cement (C 150) limits the maximum autoclave expansion to 0.80 percent for all
types of cement. This plus the 6,0 percent limit placed on MgO content by
ASTM Specification C 150 generally is considered to provide protection
against unsoundness. There is some question as to whether the autoclave test
provides adequate protection against unsoundness with blended cements
I291.
Calcium sulfate when present in excessive amounts can eause expansion
due to the continued formation of calcium sulfo-aluminate. This type of expansion is not detected by the autoclave test [30]. However, this is not a problem with commercial cements since specifications restrict the SO3 contents
to safe levels.
ASTM Specification C 150 limits the amount of SO3 based on the type of
cement and the tricaleium aluminate (C3A) content of the eement. For Type
I and IA cements having a C3A content of 8 percent or less, the maximum
allowed is 3.0 percent SO3. If the C3A is above 8 pereent, then up to 3.5 percent SOa is allowed. For Type II and IIA, a maximum of 3.0 percent SO3 is
permitted. For Type III and IIIA cements having 8 percent or less C3A, the
maximum is 3.5 percent SO3; but if the C3A is over 8 percent, then 4.5 percent SO3 is allowed. For Types IV and V, a maximum of 2.3 percent SO3 is
permitted. However, it is reeognized that this is more restrictive than
necessary for some cements; and for these cements ASTM Specification C
150 permits additional SO3 in amounts up to 0.5 percent above the specified
limits. However, this higher SO3 limit is only permitted when it has been
demonstrated by the ASTM Test for Optimum SO3 in Portland Cement (C
563) that this higher SO3 is required to provide the optimum SO3 level for
this particular cement. An additional safeguard is provided in that this
higher level ean only be used if, when the cement is tested with the additional
SO3 by the ASTM Test for Caleium Sulfate in Hydrated Portland Cement
Mortar (C 265), the ealeium sulfate in the hydrated mortar at 24 _+ 0.25 h,
expressed as SO3, does not exeeed 0.50 g/litre.
ASTM Test for Potential Expansion of Portland Cement Mortars Exposed
to Sulfate (C 452) sometimes is utilized when it seems advisable to determine
the potential sulfate resistanee of portland cements. This method is primarily
for use in research and not for use as a basis for acceptance testing.
Excessive expansion, loss of strength, and disintegration can occur due to
reactions between the alkalies in cement and constituents of aggregates.
These siliceous constituents or argillaceous dolomitic limestones can react
with alkalies, and under moist conditions this reaction can cause excessive
expansion and cracking of concrete. These alkali-aggregate reactions can be
controlled or reduced by using nonreactive aggregates or by using cements
with alkali contents less than 0.60 percent expressed as sodium oxide (Na20)
equivalent. ASTM Specification C 150 has this limit of 0.60 percent for all
types of cement as an optional chemical requirement that may be specified
235
when the cement is to be used in concrete with aggregates that may be

deleteriously reactive. ASTM Specification for Concrete Aggregates (C 33)
states that fine and coarse aggregate for use in concrete that will be subject to
wetting, extended exposure to humid atmosphere, or contact with moist
ground shall not contain any materials that are deleteriously reactive with the
alkalies in the cement in an amount sufficient to cause excessive expansion of
mortar or concrete; however, if such materials are present in injurious
amounts, the aggregate may be used with a cement containing less than 0.6
percent alkalies calculated as sodium oxide or with the addition of a material
that has been shown to prevent harmful expansion due to alkali-aggregate
reaction. Appendix XI of ASTM Specification C 33 gives various methods for
evaluating the potential reactivity of an aggregate.
This appendix lists a number of methods that have been proposed for
detecting this potential reactivity. However, these methods do not provide
quantitative information on the degree of reactivity to be expected or
tolerated in service. Therefore, evaluation of potential reactivity of an aggregate should be based upon judgement, interpretation of data, and examination of concrete structures containing a combination of fine and coarse
aggregates and cements for use in new work. Data obtained by the following
test methods will assist in making this evaluation.
ASTM Recommended Practice for Petrographic Examination of Aggregates for Concrete (C 295) outlines procedures for petrographic examination by means of optical microscopes of materials proposed for use as aggregates in concrete. Certain materials are known to be reactive with the
alkalies in cements. Determination of the presence and quantities of these
materials is helpful in evaluating the potential reactivity. Some materials
render an aggregate deleteriously reactive when present in quantities of 1.0
percent or even less.
ASTM Test for Potential Reactivity of Aggregates (Chemical Method) (C
289) covers the chemical determination of the potential reactivity of an aggregate with alkalies in portland cement concrete. This test can be made
quickly and, while not completely reliable in all cases, provides helpful information. Results obtained by this test may not be correct for aggregates containing carbonates of calcium, magnesium, or ferrous iron, such as calcite,
dolomite, magnesite or siderite, or silicates of magnesium such as antigarite
(serpentine) [31,32]. The error introduced by calcium carbonate is not
significant unless the test values indicate that the potential reactivity is
marginal.
ASTM Test for Potential Alkali Reactivity of Cement-Aggregate Combinations (Mortar-Bar Method) (C 227) covers the determination of the susceptibility of cement-aggregate combinations to expansive reactions involving
the alkalies by measuring the increase (or decrease) in length of mortar bars.
Alkalies participating in the expansive reactions usually are derived from the
cement. However, under some circumstances they may be derived from other
236
constituents of the concrete or from external sources. This test, when made
with a high-alkali cement, furnishes information on the likelihood of harmful
reactions occurring. However, this method is not recommended as a means
to detect alkali-carbonate reactions involving dolomite in certain calcitic
dolomities and dolomitic limestones. Expansions produced in the mortar-bar
test by the alkali-carbonate reaction are generally much less than those produced by the alkali-silica reaction for combinations having equally harmful
effect on concrete in service.
ASTM Test for Potential Alkali Reactivity of Carbonate Rocks For Concrete Aggregates (Rock Cylinder Method) (C 586) covers the determination
of the expansive characteristics of carbonate rocks when immersed in a solution of sodium hydroxide (NaOH) at room temperature. These length changes
indicate the general level of reactivity of rocks and whether tests should be
made to determine the effect of aggregate prepared from the rocks upon the
volume change in concrete. This test is useful as a research or preliminary
screening method rather than for specification enforcement. This method is
intended to supplement data from field service records, petrographic examinations, and tests of aggregate in concrete.
ASTM Test for Potential Volume Change of Cement-Aggregate Combinations (C 342) covers the determination of the potential expansion of cementaggregate combinations by measuring the linear expansion developed in
mortar bars subjected to variations of temperature and water saturation during storage. This method is particularly applicable to certain cementaggregate combinations common in the central part of the United States.
Cement-aggregate combinations that expand 0.200 percent or more at an age
of 1 year may be considered unsatisfactory for use in concrete exposed to
wide variations in temperature and degree of saturation with water. The incorporation of fine ground pozzolanic materials in the concrete can provide
some protection, and the use of lithium and barium salts has been reported
as a means of reducing expansion due to the alkali-siliceous aggregate reaction [33].
ASTM Test for Effectiveness of Mineral Admixtures in Preventing Excessive Expansion of Concrete Due to the Alkali-Aggregate Reaction (C 441)
covers the determination of the effectiveness of mineral admixtures in
preventing excessive expansion caused by the reaction between aggregates
and alkalies in portland cement mixtures. This evaluation is based on the expansion developed by cement-admixture combinations in mortar bars made
with reactive aggregates (borosilicate glass). A minimum value of 75 percent
for reduction of mortar expansion is an optional requirement in ASTM
Spec!fication C 618 when the pozzolan is proposed for use in combination
with a high-alkali cement and an aggregate known to be potentially
deleteriously alkali reactive. If the cement to be used on the job is available,
then it is preferred to use the quantities of cement and admixture proposed
for use on the job and the maximum mortar expansion at 14 days be limited
237
to 0.020 percent. The test for mortar expansion should be performed with
each cement to be used on the job.
It should be noted that a mineral admixture that is effective in preventing
excessive expansion caused by the alkali-aggregate reaction is not necessarily
suitable for use as an admixture in concrete. In addition, it must meet the requirements of the specifications covering the class of mineral admixture to
which it belongs.
If hardened concrete is subjected to water or soils containing soluble
sulfates, expansion can occur due to the formation of calcium sulfoaluminate (ettringite). This can cause cracking of the concrete by developing
tensile stresses that are high enough to overcome the tensile strength of the
concrete. [34]. Various researchers have shown that the use of some pozzolans can increase the resistance of concrete to sulfate attack. Others have
shown that for long-time resistance the concrete should be made with high
quality pozzolans and a sulfate resisting cement [35].
There is some controversy regarding the effect of calcareous filler on
sulfate resistance. However, some researchers have shown improvement in
sulfate resistance when calcite filler was used [36].
ASTM Type II cement can be used for moderate sulfate resistance, and
Type V or cements with little or no C3A can be used when high sulfate
resistance is desired. ASTM Test C 150 limits the amount of CaA in Type II
and IIA cement to a m a x i m u m of 8 percent. Type V cement limits the CaA to
a m a x i m u m of 5 percent and the tetracalcium aluminoferrite plus twice the
tricalcium aluminate [4 CaO.Al2Oa.Fe203 + 2(3 CaO.AI203)] or solid
solution (4 CaO.Al203.Fe203 + 2 CaO.Fe2Oa) to a m a x i m u m of 20.0 percent. These limits provide moderate (Type II) or high (Type V) sulfate
resistance when the concrete is subjected to water containing sulfate.
In general, if soils contain less than 0.10 percent sulfate ion, or water con-,
tains less than 150 p p m sulfate ion, no damaging sulfate reaction would be
expected. For soils containing between 0.10 and 0.20 percent sulfate ion, or
water containing between 150 p p m and 1000 p p m sulfate ion, Type II cement
should prevent damaging sulfate reactions. For soils containing more than
0.20 percent sulfate ion, or water containing more than 1000 p p m sulfate
ion, Type V cement should be used. For soils containing more than 0.50 percent sulfate ion, or water containing more than 2000 p p m sulfate ion, severe
damage can be expected unless Type V cement or other means are taken to
prevent or reduce damaging sulfate reactions [37,38].
Expansive and Nonshrink Concrete

Several types of expansive cements have been formulated, usually based on
the formation of hydrated calcium sulfoaluminates to give controlled expansion. In concrete made with these expansive cements, the magnitude of the
expansion is reduced by restraint and a self-stress is developed. The level of
238
this stress can be varied to give either shrinkage-compensating concrete

or--with higher stress levels--chemically prestressed concrete [39].
ASTM Specification for Expansive Hydraulic Cement (C 845) has been
developed to cover hydraulic cements that expand during the early hardening
period after setting. Three kinds of expansive cements are identified, and it is
noted that expansive cements also may be prepared in other ways.
For testing these expansive cements, ASTM Test for Restrained Expansion of Expansive Cement Mortar (C 806) can be used to determine the
length changes of expansive cement mortar while under restraint due to the
development of internal forces resulting from hydration of the cement.
ASTM Test for Restrained Expansion of Shrinkage-Compensating Concrete (C 878) provides a method for determining the expansion of concrete
made with shrinkage-compensating cement. This method also can be used to
study expansion involving degrees of restraint or variables in cements and cement contents, mix designs, schedules, or environmental treatments. Since
the potential for expansion under conditions of controlled restraint of concrete made using shrinkage-compensating cement cannot always be
predicted satisfactorily from tests of mortars made in accordance with ASTM
Test C 806, this test using concrete specimens is of value.
Expansive or nonshrink materials are available commercially for use in
grouting, patching, and other applications where it is desirable that the concrete fill a given volume after hardening. ASTM Test for Early Volume
Change of Cementitious Mixtures (C 827) covers the determination of early
volume change in cementitious mixtures, including cement paste, grout,
mortar, and concrete. This method is of particular value in the determination of volume changes that occur during the plastic state, including
shrinkage or expansion due to subsidence, hydration, evaporation, and other
physical and chemical effects. These tests may be continued beyond the time
of hardening. The specimens used in this method are not completely
unrestrained so that the measurements are primarily useful for comparative
purposes rather than as absolute values.
Preplaced aggregate concrete (intrusion or grouted concrete) can be used
where concrete of low volume change is required. Preplaced aggregate concrete differs from conventional concrete in that it contains a higher percentage of coarse aggregate; and due to the point to point contact of the coarse
aggregate, drying shrinkage is about one-half the magnitude of that occurring in conventional concrete [40].
Test Methods Used for Determining Volume Changes

Of a more general nature than methods previously mentioned and for
testing hardened concrete for volume change, ASTM Test for Length Change
of Hardened Cement Mortar and Concrete (C 157) can be used to determine
the length changes of cement mortar and concrete due to causes other than
239
externally a p p l i e d forces a n d t e m p e r a t u r e changes. This test is sufficiently

restrictive to be used as a basis for a c c e p t a n c e testing a n d can be a d a p t e d for
studies o f volume c h a n g e involving different schedules or e n v i r o n m e n t a l
treatment.
A S T M Test for Length Change of Drilled or Sawed Specimens o f C e m e n t
M o r t a r a n d Concrete (C 341) can be used to d e t e r m i n e the length changes of
drilled or sawed specimens due to causes other t h a n externally a p p l i e d forces
a n d t e m p e r a t u r e changes. These specimens can be t a k e n from existing concrete structures b u t m u s t be free f r o m reinforcing steel, visible cracks, or
o t h e r s t r u c t u r a l defects. The a p p a r a t u s r e q u i r e d for testing by these m e t h o d s
is given in A S T M Specification for A p p a r a t u s for Use in M e a s u r e m e n t of
Length Change of H a r d e n e d C e m e n t Paste, M o r t a r , a n d Concrete (C 490).
To d e t e r m i n e the volume changes in concrete p r o d u c t s , tests such as
A S T M Test for D r y i n g S h r i n k a g e o f Concrete Block (C 426) can be used as a
r o u t i n e s t a n d a r d i z e d p r o c e d u r e for d e t e r m i n i n g the drying shrinkage of concrete b l o c k , brick, or o t h e r concrete p r o d u c t s u n d e r specified accelerated
drying conditions.
References
[1] Zoldners, N. G., "Thermal Properties of Concrete Under Sustained Elevated Temperatures," Temperature and Concrete, SP-2S, American Concrete Institute, Detroit, 1971,
p. 3.
[2] Walker, Stanton, Bloem. D. L., and Mullen, W. G., "Effects of Temperature Changes
on Concrete as Influenced by Aggregates," Journal. Proceedings, American Concrete
Institute; Vol. 48, No. 8, April 1952, pp. 661-680.
[3] Mullen, W. G., "A Study of Thermal Properties of Concrete and Concrete Aggregates,"
Masters thesis, University of Maryland, 1951.
p. 7.
[5] Powers, T. C., "Causes and Control of Volume Change," Journal, Portland Cement
Association Research and Development Laboratories, Jan. 1959, p. 29.
p. 4.
[7] Zoldners, N. G., "Thermal Properties of Concrete Under Sustained Elevated Temperatures," Temperature and Concrete, SP-25, American Concrete Institute, Detroit, 1971,
p. 5.
[8] Hveem, F. N., "Slab Warping Affects Pavement Joint Performance," Proceedings,
American Concrete Institute, Vol. 47, June 1951, pp. 797-808.
[9] "Effect of Initial Moist Curing on Drying Shrinkage of Concretes Made With Different
Types of Cement," TSA-87, Portland Cement Association, Skokie, Ill., Aug. 1976, p. 9.
[10] Blaine, R. L., Arni, H. T., and Evans, D. N., "Shrinkage of Hardened Portland Cement
Pastes," Interrelations Between Cement and Concrete Properties, Part 4, Section 9, National Bureau of Standards, March 1969, p. 33.
[11] Lerch, William, "The Influence of Gypsum on the Hydration and Properties of Portland
Cement Paste," Proceedings, American Society for Testing and Materials, Vol. 46, 1946,
pp. 1252-1292.
240
[12] Swayze, M. A., "Volume Changes in Concrete," Materials Research and Standards, Vol.
1, No. 9, Sept. 1961, p. 703.
[13] Powers, T. C., "Causes and Control of Volume Change," Journal, Research and Development Laboratories, Jan. 1959, p. 29.
[14] Roper, H., "Volume Change of Concrete Affected by Aggregate Type," Journal, Portland
Cement Association, Sept. 1960, p. 19.
[15] Powers, T. C., "Causes and Control of Volume Change," Journal, Research and Development Laboratories, Jan. 1959, p. 33.
[16] Pickett, Gerald, "Effect of Aggregate on Shrinkage and a Hypothesis Concerning Shrinkage," Proceedings. American Concrete Institute, Vol. 52, 1956, pp. 581-590.
[17] Blaine, R. L. and Arni, H. T., "Shrinkage and Expansion of Concrete," Interrelations
Between Cement and Concrete Properties, Part 4, Section 10, National Bureau of Standards, March 1969, p. 33.
[18] Carlson, R. W., "Drying Shrinkage of Concrete as Affected by Many Factors," Proceedings, American Society for Testing and Materials, Vol. 38, Part II, 1938, pp. 419-437.
[19] "Guide for Structural Lightweight Aggregate Concrete," Journal. American Concrete
Institute, Aug. 1967, p. 457 (Report of ACI Committee 213).
[20] MacPherson, D. R. and Fischer, H. C., "The Effect of Water-Reducing Admixtures
and Set-Retarding Admixtures on the Properties of Hardened Concrete," Effect of WaterReducing Admixtures and Set-Retarding Admixtures on Properties of Concrete, A S T M
STP 266, American Society for Testing and Materials, 1960, p. 215.
[21] Verbeck, G. J. and Helmuth, R. H., "Structures and Physical Properties of Cement
Paste," Proceedings of The Fifth International Symposium on the Chemistry of Cement,
Tokyo, Vol. 3, 1968, p. 13.
[22] Abdun-Nur, E. A., "Fly Ash in Concrete," Bulletin 284, Highway Research Board, 1961.
[23] Fling, R. S., "Crack Control in Perspective," Journal, American Concrete Institute, Aug.
1976, p. N3.
[24] Roper, Harold, "Cement Paste Shrinkage-Relationship to Hydration, Young's Modulus
and Concrete Shrinkage," Proceedings of the Fifth International Symposium on the
Chemistry of Cement, Tokyo, Vol. 3, 1968, p. 99.
[26] "Volume Changes of Concrete," Design and Control of Concrete Mixtures. Portland
Cement Association, Skokie, I11., l l t h ed., Chapter 12, 1968, p. 86.
[28] Calleja, J., "Expansion of Cements and Methods To Determine It," The VI International
Congress on the Chemistry of Cement, Moscow, Sept. 1974, p. 3.
[29] Brown, P. W., Berger, R. L., Clifton, J. R., and Frohnsdorff, G., "Limitations To Fly
Ash Use In Blended Cements," Ash Utilization Symposium, St. Louis, Mo., March 1976,
p. 6.
[30] Gonnerman, H. F., Lerch, W., and Whiteside, T., "Investigation of the Hydration Expansion Characteristics of Portland Cements," Bulletin 45, Portland Cement Association,
Skokie, I11., June 1953, p. 5.
[31] Mielenz, R. C. and Benton, E. J., "Evaluation of The Quick Chemical Test for Alkali
Reactivity of Concrete Aggregate," Bulletin 171, Highway Research Board, 1958, p. 1.
[32] Chaiken, Bernard and Halstead, W. J., "Correlation Between Chemical and Mortar Bar
Tests for Potential Alkali Reactivity of Concrete Aggregates," Public Roads, Vol. 30,
1959, p. 177.
[33] McCoy, W. J. and Caldwell, A. G., "New Approach to Inhibiting Alkali-Aggregate Expansion," Journal, Proceedings, American Concrete Institute, Vol. 47, No. 9, May 1951,
pp. 693-708.
[34] Mehta, P. K., "Mechanism of Expansion Associated with Ettringite Formation," Cement
and Concrete Research, Vol. 3, No. 1, Jan. 1973, p. 2.
[35] Kalousek, L. C., Porter, Benton, E. J., "Concrete For Long-Time Service in Sulfate
Environment," Cement and Concrete Research. Vol. 2, No. 1, Jan. 1972, p. 88.
[36] Soroka, Itzhak and Stern, Naua, "Effect of Calcareous Fillers on Sulfate Resistance of
Portland Cement," Ceramic Bulletin, Vol. 55, No. 6, June 1976, pp. 594-595.
241
[37] Concrete Manual U.S. Bureau of Reclamation, Denver, 6th ed., 1956, pp. 10-12.
[38] Byers, W. G., "Fields Procedure for Evaluating Potential Sulfate Damage to Concrete,"
Journal American Concrete Institute, Aug. 1976, pp. 443-444.
[39] Mehta, P. K. and Polivka, Milos, "Expansive Cements," The VI International Congress
on the Chemistry of Cement, Moscow, Sept. 1974, p. 4.
[40] "Preplaced Aggregate Concrete for Structural and Mass Concrete," Journal American
Concrete Institute, Oct. 1969, pp. 785-797 (Report of ACI Committee 304).
J. A . R h o d e s ~
Chapter 17--Thermal Properties
Introduction
The thermal characteristics of concrete covered in this chapter are:

thermal conductivity, specific heat, thermal expansion (or contraction),
thermal diffusivity, and heat of hydration of cementitious materials. This
last property is included because of its significant influence on the thermal
and physical behavior of concrete in most of its uses.
One of the earliest and most comprehensive studies of the fundamental
thermal properties of concrete was carried out by the U.S. Bureau of Reclamation in conjunction with the design and construction of Boulder (Hoover)
Dam between 1930 and 1935. The investigations were stimulated by the
realization that in massive structures the heat generated by hydration of
the cement could be responsible for volumetric changes influencing the
integrity of the structure, and that dissipation of this heat under ordinary
circumstances would extend over as much as a century or more. Results
are given in the Boulder Canyon Reports [1]. 2
As with most construction materials, there is a direct relationship between a change in temperature of concrete and its change in length or
volume. This has been long recognized for such structures as highways,
bridges, walls, and buildings. Only at very high and very low temperatures
do the expansion characteristics vary from those under normal conditions.
The favorable thermal insulation characteristic of normal weight concrete,
and the even more favorable properties of light-weight concretes, have been
used effectively in building construction and for other applications where
resistance to steady-state heat flow is needed. In recent years the unique
ability of concrete to damp out annual and daily ambient temperature
variations (unsteady state) and to store or release heat over significant time
periods is being utilized by designers.
l Civil engineer, Officeof Chiefof Engineers, Department of the Army, Washington, D.C.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
RHODES ON THERMAL PROPERTIES
243
Thermal Conductivity
Definition and Units
Thermal conductivity is a measure of the ability of the material to conduet heat and may be defined as the ratio of the rate of heat flow to the
temperature gradient. In normal metric use it can be considered to be the
number of kilocalories passing between opposite faces of a 1-metre cube
per unit of time when the temperature difference is 1 ~ An alternate set
of dimensions is joules per second, square metre, degree Celsius per metre.
By dimensional manipulation and substitution, the approved SI (Syst~me
International) units are obtained, watts per metre. Kelvin, thus keeping
the unit of time at 1 s. In U.S. customary and British units, conductivity is
frequently expressed in Btu per h o u r . s q u a r e foot. degree Fahrenheit per
inch. Values of conductivity in these units may be converted to the SI units
by multiplying by 0.1441314. Largely by custom, engineering disciplines
frequently express temperature gradients, areas, and time in units most
useful to them, and users of conductivity values are cautioned to assure
that compatability and consistency exists, both within and between systems
of units [2].
Parameters and Values
Three principal conditions (water content, density, and temperature)
significantly influence the thermal conductivity of a specific concrete. The
mineralogical character of the aggregates largely determines the thermal
conductivity for normal weight concrete, while with lightweight concretes
the amount of air voids and moisture content mask the effect of aggregate
type. Other factors of slight or negligible importance in their effect on
conductivity are cement type and content, entrained air, water/cement
(w/c) ratio, and age.
Neat cement pastes, ranging from 0.3 to 0.6 w/c and ages from 3 days
to 1 year [3], exhibit a fairly constant thermal conductivity value of 1.2
W / m . K (8.0 B t u . i n . / h . f t 2. ~ at normal air temperatures and in a
moist condition. Thermal conductivity measured under reduced moisture
conditions has little meaning because the specimens suffer extensive cracking due to drying.
The amount of free water in concrete, regardless of density, is a major
factor influencing thermal conductivity. While water is a relatively poor
conductor of heat compared with rock, its thermal conductivity as shown
in Table 1 is many times that of air which it replaces in concrete [3].
Thermal conductivity of concrete varies directly with moisture content [4, 5].
The effect of moisture on thermal conductivity values from oven-dry to a
moist condition (not necessarily saturated) are given in Table 2.
244
TABLE 1--Thermal conductivity of water.
Water Temperature,
~
20
0
18
--
59
-101
-- 157
W/m. K
68
32
Conductivity,
Btu .in./h-ft 2. ~
0.59
0.56
2.3
2.6
3.3
5.2
75
-150
-250
-
4.1
3.9
16.0
18.0
23.0
36.0
TABLE 2--Typical variations in thermal conductivity with moisture at normal temperatures.

Conductivity,
Btu 9in./h 9ft2 9~
Moisture Condition
W/m 9K
Moist
50% relative humidity
Dry
2.2
1.7
1.4
15.0
11.0
10.0
Moist
Dry
2.9
2.2
1.4
Sandstone Concrete
20.0
15.0
10.0
Moist
Dry
Quartz Gravel Concrete

3.3
23.0
2.7
19.0
2.3
16.0
Moist
Dry
Expanded Shale Concrete

0.85
5.9
0.79
5.5
0.62
4.3
Limestone Concrete
For heavy-weight, normal-weight, a n d structural light-weight concretes

the mineralogical characteristics of the aggregate m a r k e d l y affect the conductivity of the concrete, as shown in T a b l e s 3 a n d 4. I n s u l a t i n g light-weight
concrete, densities less t h a n 960 k g / m 3 (60 l b / f t 3), may be aerated (foamed)
or c o n t a i n a very lightweight porous aggregate. T h e r m a l conductivities [6, 7]
given in T a b l e 5 are for air-dry or low moisture contents.
Over a t e m p e r a t u r e range from room t e m p e r a t u r e s to - - 1 5 7 ~ ( - - 2 5 0 ~
the t h e r m a l conductivity of oven-dry n o r m a l weight a n d lightweight concrete is essentially c o n s t a n t [3,8]. For moist n o r m a l weight concrete the
conductivity at - - 1 5 7 ~ ( - - 2 5 0 ~
has b e e n f o u n d to be a b o u t 50 percent
greater t h a n at n o r m a l t e m p e r a t u r e s [3]. At elevated temperatures, up to
about 7 5 0 ~ (1380~
conductivities of c e m e n t pastes, mortars, a n d concrete decrease in a consistently u n i f o r m m a n n e r . This has been a t t r i b u t e d
245
TABLE 3--Effect of aggregate type on conductivity of dry concrete at normal temperature.
Dry density,
lb/ft 3
kg/m 3
Aggregate type
Hematite
Marble
Sandstone
Limestone
Dolerite
Barite
Expanded shale
Expanded slag
Expanded slag
179
143
t 20
126
136
180
89
103
60
2870
2290
1920
2020
2180
2880
1430
1650
960
Conductivity,
Btu 9 i n . / h 9ft 2 9
~
W/m. K
18
12
10
10
8.6
8.5
4.3
3.2
1.5
2.6
1.7
1.4
1.4
1.2
1.2
0.62
0.46
0.22
TABLE 4--Effect of aggregate type of conductivity of moist concrete at normal temperatures.
Moist density,
Ib/ft 3
kg/m 3
Aggregate type
Hematite
Quartzite
Quartzite
Dolomite
Quartzite
Limestone
Quartzite
Sandstone
Sandstone
Granite
Limestone
Marble
Limestone
Basalt
Rhyolite
Barite
Dolerite
Basalt
Expanded shale
to disruption
190
150
152
156
3040
2400
2440
2500
153
147
133
150
151
151
152
152
157
146
190
147
158
99
2450
2350
2130
2400
2420
2420
2440
2440
2520
2340
3040
2350
2350
1590
of the intercrystalline
excessive thermal
expansion
bonds
Conductivity,
Btu 9 i n . / h 9f12.
~
W / m 9K
28
28
24
23
23
22
21
20
20
t8
18
15
15
14
14
14
14
13
5.9
in the aggregate
[9]. C o n d u c t i v i t y
4.1
4.1
3.5
3.3
3.3
3.2
3.1
2.9
2.9
2.6
2.6
2.2
2.2
2.0
2.0
2.0
2.0
1.9
0.85
resulting
values at about 400~
a r e a s f o l l o w s [5]:
cement paste
1:3 mortar
sandstone concrete
ilmenite concrete
W / m 9K
Btu 9 i n . / h 9ft 2 9 ~
0.56
0.75
1.6
1.2
3.9
5.2
10.9
8.2
from
(750~
246
T A B L E S--Thermal conductivity of insulating concrete.

Density,
kg/m 3
lb/ft 3
320
480
640
800
960
20
30
40
50
60
400
825
Thermal Conductivity,
W / m 9K
B t u 9 i n . / h 9 ft 2 9 ~
Aerated
0.07
0.11
0.14
0.20
0.26
0.5
0.75
1.0
1.4
1.8
25
0.10
0.72
52
Expanded Clay
0.17
Vermiculite
1.2
Above temperatures of 400~ gradual disintegration of the fully hydrated

cement paste occurs, resulting in further decreases in conductivity [9].
Test Methods
The test method for thermal conductivity developed for the Boulder
Canyon Reports [1] used 200 mm (8 in.) diameter cylinders subjected to
steady heat flow conditions. Use of water as the heating and cooling
mediums limited the results to a specific portion of the temperature range
between the freezing and boiling points of water, and to saturated or nearsaturated concrete.
The ASTM Test for Steady-State Thermal Transmission Properties by
Means of the Guarded Hot Plate (C 177) requires oven-dry specimens and
limits their thickness to 100 mm (4 in.). The procedure can accommodate
loose insulation materials, but will accept solid disks or slabs under rather
rigorous flatness tolerances.
The ASTM Test for Steady-State Thermal Transmission Properties by
Means of Heat Flow Meter (C 518) determines conductivity by comparison
with the conductance of oven-dry materials previously measured by ASTM
Test C 177. The method will accommodate flat disks or slabs only, with
conductivities not higher than those of the best insulating concrete.
The Corps of Engineers [10] Test for Thermal Conductivity of Lightweight Insulating Concrete (CRD-C 45) measures conductivity directly
under steady-state conditions in a manner similar to that of ASTM Test
C 177. Nominal specimen thickness is 25 mm (1 in.), and temperature differential through the oven-dry specimen is from 32 to 60~ (90 to 140~
Two indirect methods for determining thermal conductivity, Test for
Thermal Diffusivity of Concrete (CRD-C 36) and Test for Thermal Diffusivity of Mass Concrete (CRD-C 37), have been developed and used for
247
several years by the Corps of Engineers laboratories [10]. With these methods the thermal diffusivity property is measured, and conductivity calculated
therefrom. Test temperature range is 10 to 65~ (50 to 150~
and the
specimens have been conditioned by exposure to spray, soaking, or boiling
in water.
Lentz and Monfore, with the Portland Cement Association [8], developed
a nonsteady-state hot-wire method for determining thermal conductivity of
concrete or rock. A thermocouple cast along the axis of a prism (or in the
center of a split and lapped specimen) measures temperature response to a
measured alternating current input. There are no moisture or density
restrictions, and measurements are completed within a few minutes time.
Campbell-Allen and Thorne [5] developed, for experimental work up to
200~ (400~
a hollow-cylinder steady-state system for measuring conductivity. The specimen, at a selected moisture content, was coated with
mercury on all surfaces to retain moisture, and electrically heated internally
and cooled externally. Its acceptability or modifications are not known.
Specific Heat
The rigorous definition of specific heat is the ratio of the amount of heat
required to raise a unit weight of the material 1 deg to the amount of heat
required to raise the same weight of water 1 deg. In those systems of units
where the heat capacity of water is 1.0 (either Btu/lb 9 ~ or cal/g- ~
the specific heat (and heat capacity) values are the same. In SI units,
specific heat is expressed in joules/kilogram. K, which is obtained from
either the foot-pound or CGS values by multiplying by 4.1868 103.

The mineralogical differences among such aggregates as generally used
have little effect on specific heat of the concrete, with values at normal
temperatures ranging between 0.22 and 0.24 cal/g- ~ (Btu/lb. ~
Increased water content, according to the Boulder Canyon studies [1], tended
to increase the specific heat of the concrete. The actual values found were
0.22 cal/g. ~ at 4 percent of mixing water by weight of concrete and 0.24
cal/g. ~ at 7 percent mixing water by weight and higher.
Specific heat varies directly with concrete temperatures, as indicated in
Table 6. These values represent mass concrete mixes and several aggregate
types [11]. Specific heats of most rock types tend to increase up to at least
400~ (750~
and hydrated cement pastes up to 1000~ (1830~
Investigations by Harmathy and Allen [12] indicate the specific heat of expanded aggregate light-weight concrete differs little from that of normalCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:45 EDT 2014
248
TABLE 6--Typical specific heats of concrete.

Temperature, ~
10
38
66
SpecificHeat,
J/kg. K
cal/g. 9deg C
917
971
1038
0.219
0.232
0.248
weight concrete at ordinary temperatures and also increases up to at least

600~ ( l l l 0 ~
Testing Methods
Some type calorimeter apparatus is employed in all procedures to measure
the specific heat (heat capacity) of hardened concrete, aggregates, cement,
and cement pastes. Among the most suitable of the procedures is the Corps
of Engineers [10] Method Test for Specific Heat of Aggregates, Concrete,
and Other Materials (CRD- C 124), for materials which can be broken into
particles not larger than 25 mm (1 in.) in size. A test weight total of about
1 kg (2 lb) can be accommodated, although even with this volume a typical
or representative sample is not always assured. Normally the test is run at
about room temperature to minimize heat loss corrections. More precise
methods applicable to materials with which water may or would react, as
cement or cement pastes for example are the Corps of Engineers [10] Test
for Mean Specific Heat of Hydraulic Cements, Cement Pastes, and Other
Materials (CRD-C 242) and ASTM Test for Specific Heat of Liquids and
Solids (D 2766). These require smaller samples (up to 100 g weight) and
provide for cooling media whose physical or chemical properties differ from
those of water.
Thermal Diffusivity
The diffusivity property is described as a measure of the facility with
which temperature changes take place within a mass of material. Thermal
diffusivity is defined numerically as thermal conductivity divided by the
product of specific heat and density, or ot 1- k / c . p . Thus the diffusivity
results from the consolidation of three other properties which appear in
differential equations defining heat flow and heat storage under unsteady
state conditions.
When calculating diffusivity from its parts, care is required to assure
249
that the dimensional units of the three constituents are compatible. For
example:
British S y s t e m U n i t s
conductivity
specific h e a t
density
thus a
SI U n i t s
conductivity
specific h e a t
density
thus a
B t u / h 9 ft 2 9 ~ p e r ft
Btu/lb 9 ~
lb/ft 3
ft2/h
W / m 9K
J / k g 9K
kg/m 3
m2/s, where W = J/s
Note that in approved SI base units, diffusivity values are usually small. To
convert from British to more convenient SI units
ft2/h X 2.58064 X 10 -s = mVs
or
ft2/h 9.29030 X 10 -2 = m2/h

Those variables and conditions which influence thermal conductivity,
specific heat, and density also affect thermal diffusivity. Both thermal conductivity and density are sensitive to moisture content of the concrete, and
derived thermal diffusivity values should be based on conductivities and
densities which correspond to the condition of the concrete in service.
Within the same system of units, higher diffusivity values are associated
with concrete which heats or cools most easily.
Thermal diffusivity of concrete is determined largely by the mineralogical
characteristics of the coarse aggregate. Since specific heat varies directly
with temperature, diffusivity values for a specific concrete will decrease as
the concrete temperature increases.
Typical diffusivity values in Table 7, taken from [1,12,13], show the
general range to be expected for normal-weight and structural light-weight
concrete. The Bureau of Reclamation [1] reported diffusivities for neat
cement pastes ranging from 0.0012 to 0.0016 m2/h (0.013 to 0.017 W/h) at
and somewhat above normal room temperatures.
Test Methods
One practice, not necessarily always preferable, is to calculate thermal
250
TABLE 7--Typical thermal diffusivity values.

Type of aggregate in concrete
m2/h
Thermal diffusivity,
ft2/h
ft2/daya
Quartz
Quartzite
Limestone
Basalt
Expanded shale
0.0079
0.0061
0.0055
0.0025
0.0015
0.085
0.065
0.059
0.027
0.016
2.04
1.56
1.42
0.65
0.38
m2/day
0.190
0.146
0.132
0.060
0.036
Convenient when computing heat flow in large structures.
diffusivity from conductivity, specific heat, and density as determined from

laboratory tests. Test methods for conductivity are essentially limited to an
oven-dry condition, and diffusivity values will be applicable to dry environments only.
Two Corps of Engineers [10] test methods, CRD-C 36 and CRD-C 37,
determine diffusivity directly from a partially saturated concrete cylinder or
large moist concrete cube specimens. The cylinder specimen is heated in
boiling water and then cooled in a running bath of cold water. The 0.23 m 3
(8 ft a) cube specimen is heated to about 65~ and the surfaces then cooled
by a water spray. Thus a moisture gradient is created from the surface of
the specimen inward, and the concrete mass must be termed "moist"
rather than either "dry" or "saturated." Tables and charts for converting
the temperature change time history at the center of the specimen to the
thermal diffusivity which was responsible for this change over the observed
time duration. While the results are applicable only to moderate temperature ranges, they are quite reliable.
Thermal Expansion
As with most construction materials, concrete has a positive coefficient
of thermal expansion, which can be defined as the change in linear dimension per unit length divided by the temperature change. While a general value of 10 millionths/~ (5.5 millionths/~
has been widely used,
change in length is a complex process reflecting principally materials,
moisture, and temperature individually and together. The actual thermal
expansion is the net result of two actions occurring simultaneously. The
first is a normal expansion typical of anhydrous solids. Second, there is a
hygrothermal expansion or contraction associated with the movement of
internal moisture from capillaries to or from gel pores.
251

Since aggregate comprises from 80 to 85 percent of concrete, its thermal
properties greatly influence the behavior of the concrete. Thermal expansion
can vary. widely among aggregates because of differences in mineralogical
content. Quartzites and other siliceous aggregates exhibit high thermal
expansion properties, and concrete containing such aggregate frequently
show values up to 13 10-6/~ at normal temperatures. Some limestone
aggregate concrete exhibit expansion values of 5.6 X 10-6/~ for comparable conditions, and values with other natural rock types usually are between these two.
Cement paste occupies only about 15 to 20 percent of the concrete
volume, but its expansion coefficient ranges from 9 to 22 10-6/~
which may be several times that of the aggregate itself. In addition, most
of the capillary water and essentially all of the adsorbed water is contained
within the gel pore system, which makes the paste sensitive to water movements caused by temperature changes. Powers [14] showed one aspect of
the significance of the moisture content of paste specimens, where the expansion coefficients for oven-dry and saturated conditions were half the
maximum expansion coefficient of 11 10-6/~ which occurred at 75
percent relative humidity. A similar relationship was found for concrete,
except that the dry and saturated condition values were from 65 to 80 percent of the maximum value, which occurred at about 60 percent relative
humidity. Generally the coefficient of expansion increases with decrease in
the w/c ratio.
Some sources report a slight increase in coefficient of thermal expansion
with age up to 3 months, but a general decreasing trend thereafter. For
concretes used in Ilha Solteira Dam in Brazil, the thermal expansion coefficient was found to increase significantly with age when the aggregate
was all quartzite and less when only the fine aggregate was quartzite. See
Table 8.
For a given concrete mix, the magnitude of thermal expansion or contraction at temperatures from freezing to about 65~ is the same for each
unit temperature change, so the coefficient of thermal expansion is a
constant figure. Below the freezing point of water, the length changes are
smaller per unit of temperature, so the thermal expansion coefficient decreases. The slowing rate of contraction at low temperatures is continued
contraction of the concrete (and of the ice already formed) and an opposite
expansion resulting from the formation of additional ice. Berwanger and
Sarker [15] found the coefficient decreased from 7.7 to 5.1 millionths/~
for temperatures above and below 0~ respectively, and Monfore and
Lentz [16] showed a similar typical decrease from 9.4 to 6.8 millionths/~
There is some evidence that the coefficient continues to decrease with
further decreases in temperatures.
252
TABLE 8 - - E f f e c t o f age on t h e r m a l expansion. ~

Approximate
Age, Days
4
9
17
30
48
62
Quartzite Concrete/'
12.9
12.9
13.2
13.4
13.5
13.7
(7.14)
(7.18)
(7.34)
(7.42)
(7.51)
(7.62)
Basalt Concrete c
Units are millionths/K (~
10.2
10.2
10.2
10.7
10.8
10.6
(5.64)
(5.67)
(5.65)
(5.96)
(5.98)
(5.89)
Quartzite Mortar d
12.2
12.8
14.0
15.0
15.6
16.0
(6.75)
(7.12)
(7.75)
(8.33)
(8.69)
(8.90)
~Results are from University of California (Berkeley) tests with llha Solteira (Brazil) materials.
bAll natural quartzite aggregate.
CBasalt coarse aggregate and quartzite sand.
dQuartzite natural sand.
As stated previously, the expansion of concrete exposed to increasing

temperature is the net result of the inherent thermal expansion property
of the aggregate and a complex hygrothermal volume change of the cement
paste. Up to about 100~ the paste has achieved its natural expansion,
and at higher temperatures starts to shrink [I 7], continuing to do so up to
about 500 or 600~ At this level only the original dry ingredients remain.
For calcareous aggregate concrete, Philleo [17] reported thermal expansion
coefficient of 8.5 millionths/~ (4.7 millionths/~ below 260~ (500~
and 22.5 millionths/~ (12.5 millionths/~ above 425~ (800~
For
expanded shale aggregate concrete the values were 5.0 millionths/~ (2.8
millionths/~
and 8.8 millionths/~ (4.9 millionths/~
for similar
temperature levels.
Over a number of years the Bureau of Reclamation and Corps of Engineers have conducted thermal expansion coefficient tests of specimens
composed of concrete mixes having moderate to low cement factors and
large-size aggregates. The aggregates, from the actual job sources, are frequently complex in terms of type and mineralogical content; and the
expansion coefficients do not lend themselves to unique groupings by rock
type. Some typical results are listed in Table 9. The values therein are
applicable for an average temperature of about 38~
Test Methods
The significant effect that moisture and temperature have on behavior of

concrete, and the wide variety of environmental conditions to which concrete is exposed has precluded development of a standard test method for
general use. Physical measurement of specimen length change with the
necessary accuracy is not a major obstacle, but conditioning of the
specimen and control of the conditions while under test may be difficult.
253
TABLE 9--Thermal expansion coefficientsfor mass concrete.

Dam Name
Aggregate Type
Hoover
Hungry Horse
Grand Coulee
Table Rock
Greers Ferry
Dworshak
Libby
Jupia (Brazil)
limestone and granite

sandstone
basalt
limestone and chert
quartz
granite-gneiss
quartzite and argillite
quartzite
Coefficient of Thermal Expansion,

millionths/K millionths/~
9.5
11.2
7.9
7.6
12.1
9.9
11.0
13.6
5.3
6.2
4.4
4.2
6.7
5.5
6.1
7.5
Many test procedures, some quite sophisticated, have been developed in

conjunction with research studies [8,17].
The Corps of Engineers follow a standard, Test for Coefficient of Linear
Thermal Expansion for Concrete (CRD-C 39), in which a length comparator similar to that described in ASTM Specification for Apparatus for
Use in Measurement of Length Change of Hardened Cement Paste,
Mortar, and Concrete (C 490), is used to measure length changes of a
specimen alternately immersed in 60 and 40~ water baths. The operator
is free to change the temperature limits and substitute oven drying if this
better serves the data user.
Heat of Hydration
Definhion and Units
When water is added to cement the reaction is exothermic and a considerable amount of heat is generated over an extended period of time. The
heat liberated up to a specific time or age is measured in calories per unit
weight of cement (cal/g) or kilojoules per kilogram (kJ/kg). When the
amount of this heat and the heat capacity of the paste, mortar, or concrete
are known~ the resulting temperature rise can be calculated, assuming no
heat loss to the surroundings.
The hydration reactions of portland cement have been studied and

reported by many authors for many years. All the compounds present in
cement are anhydrous, and when in contact with water are all attacked at
various rates and at varying times. There is general agreement that the
tricalcium aluminate (Ca A) is the largest single compound contributor to
the evolved heat, followed by tricalcium silicate (C3S) and tetracalcium
254
aluminoferrite (C4AF) with about equal contributions, and finally dicalcium silicate (C2S). Approximate contributions of the four principal
calculated compounds in 0.40 w/c cement pastes cured at 21~ are given
in Table 10. Although the data apparently indicates progressive increases
TABLE l O--Typical compound contribution to heat of hydration[18].
Heat Evolved in cal/g
Compound
C3S
C 2S
C3A
C4AF
3 day
1 year
13 year
58
12
212
69
117
54
279
90
122
59
324
102
in the total heats generated by each compound, values at intermediate ages

for both CaA and C4AF deviate from this pattern. Possible causes and significance of the regression characteristics are discussed in Ref 18. When
water is first brought into contact with portland cement there is a very
rapid and very brief heat evolution, reaching possibly a rate of I cal/g/min.
There then occurs a 1 or 2 h delay, followed by a gradually increasing
hydration rate to about 6 or 8 h, and a slow decline in rate thereafter.
Relative proportions of the four major compounds, other minor
compounds, alkalies, and gypsum content cause variations in both early
and later age hydration rates. The nature and interrelationships of the
chemical reactions are complex, variable, and sometimes inconsistent or
controversial; the reader is referred to Lea [6] or other comparable texts for
a comprehensive presentation on the chemistry of cements.
As with most chemical reactions, the rate of cement hydration increases
with temperature. The accelerating effect of high curing temperatures is
limited to early ages, and subsequently the hydration process slows to a
rate less than that corresponding to normal placement and curing temperatures. Verbeck [19] suggests this is the result of a dense hydration
product created by high temperatures at early ages surrounding the cement
grains, which thereby substantially retards subsequent hydration. This
degradation in hydration is generally true for temperatures above about
24~ (75~
and becomes significant at 52~ (125~ and above. At and
below 0~ (32~ a sharp drop in rate of hydration occurs, and between
10 and -- 15~ hydration ceases.
The w/c ratio, in pastes, has a significant influence on the amount of
heat generated at ages of 3 days and greater. The differences over the
range 0.4 to 0.8 w/c ratios are greatest at intermediate ages, but continue
to be evident at up to 6 years. According to Verbeck [19], the hydration
product steadily increases in volume filling the capillary void space in the
-
RHODES ON THERMAL PROPERTIES 255
paste. If the capillary void space is small (low w/c ratio) the available
space will become filled completely with hydration products and hydration
of the remaining cement will cease. Thus a high w/c ratio will tend to
result in more complete hydration and in more heat developed. Data reported by Verbeck [18] show that an increase in w/c ratio from 0.4 to 0.8
had only a slight effect on heat liberation for Type IV (low-heat) cements,
but produced an increase of 10.8 cal/g, or 14 percent, for Type III (highearly) at 3 days of age. The increases in heat of hydration resulting from
an increase in w/c from 0.4 to 0.8 ranged from 11.4 cal/g for Type II
(moderate-heat) cement to 15.9 cal/g for Type III at 1-year of age.
Prior to about 1935, the fineness of a cement was a major factor in the
rate and amount of heat developed during hydration. The finer cements
presented much more surface area to be wetted, resulting in more rapid
and complete hydration. In recent years, especially after World War II,
all cements are so finely ground that the moderate variation in fineness of
different cements is no longer an important factor in cement hydration.
However, high-early strength cements are finer enough than other types
that their extra fineness works with their more active chemical compounds
to produce earlier strength and more rapid release of heat.
Pozzolans are defined in ASTM Specification for Fly Ash and Raw or
Calcined Natural Pozzolans for Use as a Mineral Admixture in Portland
Cement Concrete (C 618) as siliceous or siliceous and aluminous materials
which in themselves possess little or no cementitious value but, in finely
divided form and in the presence of moisture, will react chemically with
calcium hydroxide at ordinary temperatures to form compounds possessing
cementitious properties. Raw or calcined natural pozzolans (sometimes
used to replace a portion of cement in mass concrete) vary widely in their
composition, but will reduce the total heat evolved by a value up to onehalf of the replacement percentage figure [20].
Fly ash, also classified as a pozzolan, may range in fineness before processing from below to well. above that of portland cement. When used in
concrete, a minimum fineness limit considerably above the portland
cement value is imposed. The reaction between the glass in the fly ash and
the lime in the cement is particularly sensitive to heat, and adiabatic
curing of concrete containing a fly ash replacement will serve to increase
substantially the chemical activity of the fly ash component.
In a study of portland blast-furnace slag cements, Klieger and Isberner
[21] found there to be no consistent difference in heat of hydration up to 3
days of age between Type I and Type IS cements, and at later ages up to 1
year the Type IS cements exhibited slightly to moderately lower heat
generation characteristics.
Chemical admixtures, as the term is used here, refer to materials added
in relatively small amounts to mortar or concrete during mixing to modify
some characteristic of the product. Generally accelerators, retarders, and
256
to some extent, water reducers will effect the rate of hydration of the
cement, but the processes are not well delineated and will differ significantly
with different cements and proprietary products. There is little if any
evidence that air-entraining admixtures influence either the rate or the
total amount of heat generated for a specific cement or cement blend.
Test Methods
The most widely used method for determining the heat of hydration of a
hydraulic cement over long periods of time is by heat of solution, ASTM
Test for Heat of Hydration of Hydraulic Cement (C 186). This procedure
requires determination of the heat of solution of unhydrated cement and of
a cement sample hydrated for a specific period of time, the difference
being the heat of hydration up to that age. The procedure requires several
corrections and Lea [6] discusses a number of possible errors. Most significant to the user, other than laboratory techniques, is that the storage
temperature is standardized at 23~ (73~
which is almost never
representative of the environment experienced in a concrete construction.
A standard paste w/c ratio of 0.40 is required, but this will result in understating the heat generated where higher w/c ratios are encountered in
construction. The procedure is not recommended for pozzolan blends and
for slag cements where a portion of the test sample usually remains im
soluble at the end of the test procedure.
The Carlson-Forbrich vane conduction calorimeter, developed prior to
1940, determined the heat developed by a 0.40 w/c ratio cement paste by
measuring the rate of heat flow from the paste receptacle through cooling
vanes into a water bath. Because of difficulties in controlling extraneous
heat transfer, the test period usually was limited to 3 days. The method is
sensitive to the initial starting temperature, and is neither adiabatic nor
isothermal.
Monfore and Ost [22] developed a refined calorimeter for measuring
early rates of heat liberation which ideally is suited for laboratory research.
In this calorimeter the temperature rise of a sample of hydrating cement
weighing up to 8 g can be held to less than 0.5~ thus producing nearly
isothermal conditions at any temperature level. The test usually is limited
to 3-day duration, but the results include the immediate heat of hydration
(0 to 1 h). The method is well-suited to determining the effects of additions
and admixtures to the cements, as well as properties of specific cement
compounds or other similar materials.
A technique for measuring directly the temperature rise of a concrete
specimen and then calculating the accummulated heat responsible for the
temperature change is described in the Corps of Engineers Test for
Temperature Rise in Concrete (CRD-C 38) [10]. The specimen is an 8-ft 3
(0.22-m 3) insulated and sealed cube, maintained in a temperature environCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:45 EDT 2014
257
ment corresponding to its own temperature history for a period of 28 days.

Advantages are (a) cement-pozzolan blends may be used, (b) the
aggregates proposed for use in the prototype structure may be incorporated
in the test, and (c) the temperature rise approximates the nearly adiabatic
conditions existing in the interior of a massive structure. A minor difficulty is maintaining an adequate insulation condition at the later stages
when cement hydration rate is very low.
Significance of Thermal Properties

Heat Generation and Temperature Rise
The heat resulting from the exothermic water-cement reaction over a
given time interval is expressed easily in terms of temperature rise when the
specific heat value for the paste, mortar, or concrete mixture are known.
While the range of values for specific heat of concrete is relatively narrow,
the heat generation characteristics of the cementitious materials available
for use in concrete are quite broad. When positive control over
temperature rise is desired, upper limits on the heat of hydration of the
cement may be imposed through ASTM Specification for Portland Cement
(C 150) (Types II and IV) and ASTM Specification for Blended Hydraulic
Cements (C 595) (Types IS, IP and P). Specification limits for these types
and typical ranges for others not subject to restriction are indicated as
follows, with values in calories per gram.
Type
Type
Type
Type
Type
Type
Type
I (C 150)
II (C 150)
III (C 150)
IV (C 150)
IS (C 595)
IP (C 595)
P (C 595)
7 days
28 days
63 to 88
70
84 to 95
60
70
70
60
83 to 109
80
101 to 107
70
80
80
70
In lieu of established specification products, replacing up to 35 percent of

the cement in a mix by natural or processed pozzolans or one of the several
types of fly ashes will reduce significantly both the hydration rate and the
total amount of heat generated.
The importance of heat generation and disposal of that heat has long
been recognized by designers of large dams. Upon final completion, these
structures should be monolithic and in intimate contact with the foundation and abutments in order to achieve the design stress distribution
and stability. Cracks which disrupt the stress pattern and decrease stability
258
are caused principally by thermal tensile stress created by concrete volume

changes associated with a temperature decline as heat is dissipated. While
several design and construction practices are available to deal with temperature changes which tend to occur, restricting the amount of heat
that causes the temperature rise is a fundamental and certain scheme for
mitigating the thermal stress problem.
Waugh and Rhodes [23] have listed experiences at several large mass
concrete dams constructed by the Bureau of Reclamation, the Corps of
Engineers, a n d the Tennessee Valley Authority. The favorable trend
toward lower peak concrete temperatures has allowed the use of larger
monolith lengths without serious consequences. Recently, the temperature
rise at Glen Canyon and Dworshak Dams was limited to 14~ (25~
Biological shielding structures of concrete are widely used with nuclear
reactors. Such shields are up to 2 m (6 ft) in thickness, and the heavy
aggregates generally required are responsible for cement contents up to
350 k g / m a (590 lb/yd3). Peak temperatures in the concrete of as much as
36~ (65~ above the initial placing temperature may be expected.
Pier footings and heavily reinforced raft foundation slabs up to 2 m thick
and totaling 2000 m a of concrete with cement contents of 400 kg/m a (675
lb/yd a) have been successfully placed with no cracking detected [5]. Peak
interior concrete temperatures about 50~ (90~ above initial concrete
placing temperature required exterior insulation to control gradients and
cooling rates for prevention of thermal cracking.
Heat Flow
The rate at which heat flows into, through, or out of a concrete structure
is governed by the thermal conductivity of the concrete. The ease or
difficulty with which the concrete undergoes temperature change, as a
result of heat loss or gain, depends also on the heat capacity, and is
measured by the thermal diffusivity of the concrete.
Concrete in massive structures is usually placed in horizontal lifts from
0.75 to 2.3 m (2.5 to 7.5 ft) in thickness, the shallower lifts to facilitate the
loss of heat at locations where restraint is high. Control of thermal stresses
by restricting lift thickness and placement frequency is especially effective
for concrete which must be cast at a high initial temperature.
In exceptionally large concrete structures, principally arch and gravity
dams, embedded pipe cooling is used frequently to remove much of the
heat generated during hydration of the cement. Water is circulated
through pipe coils placed at the bottom of each new lift of concrete.
Spacing of the cooling pipes, initial temperature of the water, rate of water
circulation, duration of the cooling operation, and capacity of the refrigeration-pumping plant are determined primarily from the thermal
259
properties of the concrete, including heat generation characteristics of the

cement.
The practice of precooling concrete materials so as to depress the initial
temperature of the concrete at the time of placing is resorted to frequently
in massive concrete construction. This practice reduces the peak
temperature which subsequently will be attained. A maximum placing temperature of 10~ (S0~ is specified frequently for large concrete dams
which is usually well below the ambient air temperature during the normal
construction season. This low placing temperature will reduce the rate of
cement hydration initially, but during the period the concrete is at a
temperature below air temperature heat will be gained from the
surrounding air. However, this period of time is short compared with the
time the concrete is warmer than the surroundings.
Besides being important in the dissipation of internal heat, thermal
properties of concrete contribute to damping ambient cyclic variations.
Few buildings, residential or industrial, are designed without consideration
for thermal insulation. The light-weight aggregate concrete and foamed
or cellular concretes will provide a measurable degree of insulation for the
interior, in addition to their lower dead weight structural advantage. In
consideration of heat and energy costs, designers now utilize the damping
effect, expressed by thermal diffusivity, of leveling out the cyclic air temperature variation. Normal weight concrete ideally combines strength and
durability with a favorable thermal diffusivity property which is most
effective in reducing the heating and cooling requirements of the building
interior.
Volume and Length Changes

Linear and volumetic changes, when unrestrained, cause neither strain
nor stress. When these changes are not allowed to occur, either by internal
or external forces, strain will develop; magnitude of this strain is
determined primarily by the coefficient of thermal expansion and the
amount of temperature change.
In mass concrete structures, temperature decreases are most important
in that tensile strains will likely result. A low thermal expansion coefficient
is beneficial, since this will tend to reduce the total length or volume
decrease and lessen the magnitude of the potential thermal strain. For arch
dams and similar massive structures containing joints which subsequently
must be closed by grouting, the product of temperature drop and thermal
expansion (or contraction) coefficient should be such that joint openings of
at least 1 to 2 mm (0.04 to 0.08 in.) will occur. A minimum opening of this
magnitude is needed to permit successful and complete grouting.
The design of prestressed concrete reactor vessels should include
260
consideration of stresses or strains originating from temperature gradients

during operation as well as during shut-downs. Thermal conductivity at
higher temperature levels and the resulting thermal expansion or contraction can be determined analytically for introduction into the stress and
strain patterns.
The principal thermal property to be considered in rigid pavement design
is coefficient of thermal expansion under the cyclic variations in ambient
temperature. Spacing and width of expansion joints are determined mainly
by the amount of thermal expansion resulting from the amplitude of the
ambient temperature cylce and solar heating of the pavement slab.
Conventional building construction, principally rigid frame or continuous structures, warrant consideration of thermal expansion coefficients
as they may effect either joints or stresses.
Thermal expansion may be significant where epoxy mortars are used for
concrete repair. Epoxy resins, in general, have thermal coefficients greater
than 36 millionths/~ (20 millionths/~
and filled epoxy systems about
23 millionths/~ (13 millionths/~
These values, from two to four times
those for normal portland cement concrete, may create thermal incompatabilities and possible failures where the repairs are subjected to extremes
of temperature.
Winter Concreting and Insulation
Newly-placed concrete must be maintained at temperature levels which

will facilitate hydration and development of minimum strength requirements. Protection is required to prevent freezing of the uncombined free
water in saturated new concrete, and accelerating the hydration process by
the addition of calcium chloride or similar admixture will reduce the
duration of protection required. American Concrete Institute Recommended
Practice for Cold Weather Concreting (ACI 306) [24] set forth guidance
for concrete placement at low temperatures. Surface insulation is also
utilized to avoid development of steep thermal gradients which would result
in large differential length changes, tensile strain, and possible development
of cracks.
References
[1] "Thermal Properties of Concrete," Boulder Canyon Project Final Reports, Bulletin No.
1, Part VII, Bureau of Reclamation, 1940.
[2] Jakob, Max and Hawkins, G. A., Elements of Heat Transfer and Insulation, Wiley,
New York, 1950.
[3] Lentz, A. E. and Monfore, G. E., "Thermal Condnctivities of Portland Cement Paste,
Aggregate and Concrete Down to Very Low Temperatures," Research Dept. Bulletin
207, Portland Cement Association, 1966.
[4] Brewer, H. W., "General Relation of Heat Flow Factors to the Unit Weight of Concrete," Journal, Research and DevelopmentLaboratories, Portland Cement Association,
Vol. 9, No. 1, Jan. 1967.
261
[5] Campbell-Allen, D. and Thorne, C. P., "The Thermal Conductivity of Concrete,"

Magazine of Concrete Research, Cement and Concrete Association, Wexham Springs,
Slough, Vol. 15, No. 43, March 1963, and subsequent discussions.
[6] Lea, F. M., The Chemistry of Cement and Concrete, Chemical Publishing Co., New
York, 3rd ed., 1970.
[7] Neville, A. M., Properties of Concrete, Wiley, New York, 2nd ed., 1973.
[8] Lentz, A. E. and Monfore, G. W., "Thermal Conductivity of Concrete at Very Low
Temperatures," Journal, Research and Development Laboratories, Portland Cement
Association, Vol. 7, No. 2, May 1965.
[9] Temperature and Concrete, Publication SP-25, American Concrete Institute, Detroit, 1971.
[10] Handbook for Concrete and Cement, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, Miss., Aug. 1949, with quarterly supplements.
[11] ACI Committee 207, "Mass Concrete for Dams and Other Massive Structures," Journal,
American Concrete Institute, Vol. 67, No. 4, April 1970.
[12] Harmathy, T. Z. and Allen, L. W., "Thermal Properties of Selected Masonry Unit
Concrete," Journal, American Concrete Institute; Proceedings, Vol. 70, No. 2, Feb. 1973,
pp. 132-142.
[13] "Mass Concrete for Dams and Other Massive Structures," ACI Comm 207, Journal,
American Concrete Institute, Vol. 67, No. 4, April 1970.
[14] Powers, T. C., "The Physical Structure and Engineering Properties of Concrete," Research
Department Bulletin 90, Portland Cement Association, July 1958.
[15] Behavior of Concrete Under Temperature Extremes, Publication SP-39, American Concrete Institute, Detroit, 1973.
[16] Monfore, G. E. and Lentz, A. E., "Physical Properties of Concrete at Very Low Temperatures," Research Department Bulletin 145, Portland Cement Association, 1962.
[17] Philleo, R. E., "Some Physical Properties of Concrete at High Temperatures," Research
Department Bulletin 97, Portland Cement Association, Oct. 1958.
[18] Verbeck, G., "Energetics of the Hydration of Portland Cement," Proceedings, Fourth
Internatiorml Symposium on the Chemistry of Cement, National Bureau of Standards,
Monograph 43, Vol. 1, Washington, D.C., 1960.
[19] Verbeck, George, "Cement Hydration Reactions at Early Ages," Journal, Research and
Development Laboratories, Portland Cement Association, Vol. 7, No. 3, Sept. 1965.
[20] "Investigation of Cement-Replacement Materials, Report No. 1," Corps of Engineers,
Waterways Experiment Station, Vicksburg, Miss., MP No. 6-123, April 1955.
[21] Klieger, Paul and Isberner, A. W., "Laboratory Studies of Blended Cements-Portland
Blast-Furnace Slag Cements," Journal, Research and Development Laboratories, Portland Cement Association, Vol. 9, No. 3, Sept. 1967.
[22] Monfore, G. E. and Ost, B., "An 'Isothermal' Conduction Calorimeter for Study of the
Early Hydration Reactions of Portland Cements," Journal, Research and Development
Laboratories, Portland Cement Association, Vol. 8, No. 2, May 1966.
[23] Waugh, W. R. and Rhodes, J. A., "Control of Cracking in Concrete Gravity Dams,"
Journal of the Power Division, American Society of Civil Engineers, PO 5, Oct. 1959.
[24] ACI Committee 306, "Recommended Practice for Cold Weather Concreting (ACI 30666)," American Concrete Institute, Detroit, 1966.
George Verbeck i
Chapter 18
Pore Structure
Introduction
The pores, or voids, in concrete consist of pores in the hardened cement
paste, entrained or entrapped air voids, and voids in the pieces of aggregate. The porosity of the aggregate is treated extensively elsewhere in
this volume. Other void spaces such as honeycombing, which is the result
of gross failure properly to consolidate the concrete, and bleeding channels
and pockets, which result from excess water content or poor mix proportioning, are also important but are not treated in detail here, because
they are the result of poor practice and are not inherent to properly prepared concrete.
Most of the important properties of hardened concrete are related to the
quantity and the characteristics of the various types of pores in the concrete. The engineering properties, such as strength, durability, shrinkage,
creep, and permeability are directly influenced or controlled by the relative
amounts of the different types and sizes of pores.
The pores can exert their influence on the properties of the concrete in
various ways. As regards the strength and elasticity of the concrete, it is
primarily the total volume of the pores that is important, not their size
or continuity. The permeability is influenced by the volume, size, and
continuity of the pores. Shrinkage, at least that part of drying shrinkage
that is reversible, is largely a function of changes in surface energy at the
solid-to-pore interface and, therefore, depends upon the nature of the solid
surface and the total surface are of the pore system. Irreversible drying
shrinkage may involve capillary phenomena. The resistance of concrete to
freezing and thawing and deicer scaling is controlled by the volume and the
characteristics of entrained air voids. Therefore, it is not surprising that
there has been considerable interest in developing ways of measuring and
characterizing the different types of pores in concrete and in elucidating
I Deeeased. Formerly director of research, American Admixtures Company, Chicago, I11.
The revision of this chapter was completed by W. L. Dolch, Purdue University, Lafayette,
Ind. 47907. The assistance of R. F. Feldman and G. G. Litvan is gratefully acknowledged.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
lntcrnational
www.astm.org
VERBECK ON PORE STRUCTURE
263
the various mechanisms by which they influence the properties of the concrete.
Porosity of Concrete
The pores formed in the original plastic concrete are either water- or
gas-filled. After the concrete has hardened, the water-filled pores may tend
to dry, or the air-filled pores may tend to become water saturated, depending upon the history of the concrete, the external moisture conditions, and
the dimensions of the concrete member.
Hardened concrete that is properly proportioned, mixed, and placed
consists of hardened cement paste, aggregate, and relatively small air
voids. If these air voids are the result of the use of an added air-entraining
agent, they are termed entrained air. If they are the result of the inevitable
inability to effect complete consolidation of the plastic concrete, they are
termed entrapped air.
Pores in Hardened Cement Paste
The water-filled space in a freshly mixed cement paste represents space

that is available for the formation of cement hydration products. As hydration proceeds, the volume of this space, which initially was determined
by the water/cement (w/c) ratio of the paste, is continually reduced by
the formation of the hydrated gel, which has a bulk volume larger than
that of the original unhydrated cement. At any time, that part of the
original water-filled space not occupied by hydration products constitutes
part of the pore system of the paste, and generally speaking that part
with the largest pore sizes. At least at early ages, this part is interconnected.
Hydration reduces the size and volume of this capillary space. If the
original capillary space is small (w/c ratio about 0.35 by weight), the
bulk volume of the gel eventually will be sufficient to fill this space and
produce a paste free from capillary space. At higher w/c ratios, the gel
volume is not sufficient to fill the original water space in the paste, even
after complete cement hydration, and pastes having a relatively large
volume of capillaries will be produced. This capillary space can be visualized as a submicroscopic system of voids randomly distributed throughout
the hydrated cement paste matrix. These submicroscopic capillary spaces
include a wide variety of "sizes" and "shapes." Various experimental
methods have been used to investigate this portion of the pore system,
the most useful of which has been mercury intrusion porosimetry. In
this technique mercury is forced under pressure into the dried and evacuated pores. The volume that enters is the volume of the pores and the
pressure required to cause entrance is related to the size of the pores, or
264
more exactly, to the size of the entranceway of the pores. In this way a
pore size distribution can be determined. Studies on hardened cement
paste have shown that the size of its largest pores depends principally on
the degree of hydration, varying from about 1 #m for young pastes in
which the cement hydration has progressed only a little, to about 0.1 #m
in mature pastes that are more completely hydrated.
In a freshly-mixed or very young paste the capillary channels are, ipso
facto, interconnected. As the spaces become filled with hydration product
on continued curing, a time may come, if the original w/c ratio was small
enough, that most of the capillary spaces become isolated cavities separated
from each other by the gel that is the hydration product.
This gel is itself a porous solid. If a paste has such a low w/c ratio that
not all the cement can ever become hydrated, the gel will fill completely
all the space not occupied by unhydrated cement and will have a characteristic porosity containing, when saturated, water that can be removed by
heating to about l l 0 ~
or by vacuum pumping through a trap kept at
the dry-ice point (--78~
or by exposure to strong desiccants. This socalled evaporable water will have, in such a paste, a volume of about onefourth that of the bulk volume of the gel, and this volume equals that of
the pores from which the water came.
These pores that are part and parcel of the hydration product were
called gel pores by Powers; and if he did not precisely say so, his work
has been interpreted to mean that gel pores of a characteristic volume and
size are produced as part of the gel, and that the capillary pores are the
rest of the original water-filled space that has not become filled with gel.
Thus, essential elements of what has become known as the Powers model
for hardened cement paste are a gel containing about a quarter of its
volume of gel pores and a residual capillary pore space. The pores in the
gel are generally smaller than the capillary pores, and it has been inferred
that the pore space according to this model is bimodal, that is, there are
small gel pores and large capillary pores with a size region between the
two that is more or less missing.
The mercury porosimeter has been used to investigate the small-size
portions of the pore structure of cement paste, but a lower limit of about
3 nm (30 ,~,) diameter is imposed by the m a x i m u m pressure available to
the equipment. Therefore, the method most used to investigate the size
of the very small pores has been that of vapor adsorption. In this technique
the amount of a vapor adsorbed on or in the pores is determined as a
function of the ambient vapor pressure. Certain theories are then used
to relate these data to the surface area of the solids in the material and to
the pore size distribution of its pores. The most notable of these are the
Brunauer-Emmett-Teller theory using the relatively low vapor pressure
portion of the data to obtain the specific surface area and the capillarycondensation theory using the higher vapor pressure portion for the pore
size distribution.
265
Sorption studies have shown a size distribution for the small pores from
a lower limit of a few diameters of the water molecule throiagh a maximum
volume of about 5 nm (50 ,~) diameter, and then decreasing to the limit
of the method, which is a few hundred angstrom units.
Comparisons between pore size distributions obtained by sorption and
by mercury porosimetry, over the range where they both apply, have not
always shown good agreement. Mercury intrusion data have failed to show
any bimodality of the pore sizes in cement paste, although as mentioned
previously, the smallest pores cannot be so investigated.
Nevertheless it is reasonable to say that when the w/c ratio is higher
than a minimum (about 0.35), which is the case for almost all concrete,
a pore system is formed in the paste that is approximately continuous from
the smallest "gel" pores to the largest "capillary" pores, and that any such
distinction is arbitrary.
The earliest estimates of the surface area of the solids in hardened cement
paste were obtained by the vapor sorption method using water as the adsorbed species. The values obtained were approximately 200 m 2 /g of dry
solids. When the same experiment was performed with nitrogen as the
sorbed species, the results were lower, roughly a tenth that obtained with
water vapor. This discrepancy led to the advocacy of a different model for
the structure of hardened cement paste, which was set forth by Feldman
and Sereda.
They consider that even a dense, well-cured paste is capable of holding
relatively large quantities of H 2 0 molecules that exist in spaces between
layers of calcium silicate hydrate, which are on the average approximately
parallel and about one to two water molecules in width; this space is termed
"interlayer space." The water molecules in these small spaces will have
characteristics, such as vapor pressure, freezing characteristics, and mobility, notably different from those of free water in bulk. The presence of
this water increases the elastic modulus of the material by as much as
100 percent. Some of the same characteristics exist, but to a much lesser
degree, for water in the capillary spaces. The volume of the interlayer
space, in each quantity of paste, increases with continued hydration of
the cement, whereas the volume of the capillary space decreases with hydration.
The product of cement hydration that constitutes the solid fraction of
the hydrated cement material is composed of very fine particles. A calcium silicate hydrate ( " C - - S - - H gel") may constitute about 50 percent
of the hydrated cement. It is a poorly crystalline material consisting of
small, extremely thin sheets that may be crumpled or rolled as foils. Another product of hydration of the calcium silicates is calcium hydroxide,
which may form crystals large enough to be visible in an optical microscope and which may constitute 25 percent of the hydrated cement. The
overall average particle size of the elements of hydration is small. At low
w/c ratios the specific surface area is less than 1 m 2 / g of dry paste, while
266
at high w/c ratios, it is in the vicinity of 175 m 2 /g, as measured by nitrogen sorption. However, these values must be taken only as rough indicators;
the C - - S - - H gel is a material very sensitive to wetting and drying; major
changes occur to the layered structure as it is dried or wetted.
The sensitivity of the material makes it desirable to investigate the pore
structure in the wet state; this is extremely difficult. Low angle X-ray
scattering has indicated that the specific surface area, including that within the "interlayer" spaces, is about 700 m 2 / g for water-saturated pastes.
The same technique gave specific surfaces of dried pastes roughly the same
as those obtained with water vapor sorption, about 200 m 2 /g. Inelastic
neutron scattering has suggested that pore volumes measured by nitrogen absorption are realistic. So the Feldman-Sereda model considers that
the "water" area includes interlayer spaces not accessible to nitrogen because of the specific attraction of water molecules to the silicate surfaces,
and that the nitrogen, not the water, areas are the "true" values. The
Powers model regards the surface areas obtained by water vapor sorption
to be the "true" values and sees the nitrogen areas as lower because of
steric or thermodynamic hindrance to the access of nitrogen to the whole
surface.
As previously noted, water held either in the interlayer space or smaller
capillary pores does not behave as normal free water. The interlayer spaces
retain most of their water even at relatively low humidities, whereas most
of the capillary water is more volatile. It is the capillary spaces with which
the permeability of paste and concrete is associated most closely, for the
water is these pores can move much more freely under hydrostatic pressure than can the interlayer water.
Air Voids
Concrete normally contains air voids, accidently entrapped or purposely
entrained, dispersed throughout the paste component. Many of the accidentally entrapped voids can be seen with the unaided eye and may range
in size up to several millimetres.
Those purposely entrained voids that have a significant effect on the
resistance of the concrete to freezing and thawing and deicer scaling range
from a few to several hundred micrometres in size. The air voids may
constitute from less than 1 to more than 10 percent of the concrete volume,
the volume and size depending upon several factors including the amount
of air-entraining agent used, size distribution of fine aggregate, concrete
consistency, duration of mixing, and so on.
The small air voids having a significant effect on concrete durability
can be seen and studied using a microscope and a magnification in the
range of approximately 30 to 150 diameters. The volume of the air voids
in hardened concrete and the characteristics of the air-void system (void
267
size, voids per inch of traverse, and spacing factor) can be determined in
accordance with ASTM Recommended Practice for Microscopical Determination of Air-Void Content and Parameters of the Air-Void System in
Hardened Concrete (C 457).
In summary, concrete can be visualized as consisting of a heterogeneous
mixture of components, each component having its own characteristic
pores. In terms of the other pores in the concrete, the air voids, normally
the coarsest of all, may constitute from less than 1 to more than 10 percent
of the total volume of the concrete. Approximately 75 percent of the concrete is aggregate, frequently heterogeneous, with an internal pore volume
varying from almost 0 to 20 percent or more (most commonly about 1 to
5 percent), the pores ranging from relatively fine to coarse. The cement
paste component usually contains both extremely fine gel pores and the
coarser but submicroscopic capillary spaces.
Although these various pores and voids in concrete influence the physical
properties of water contained therein, the properties of the separate types
of pores are not sufficiently different to permit their complete indentification in concrete. The combined porosity of concrete (including the air
voids) can be represented by the total capacity for evaporable water between the stages of complete saturation and dryness--dry except for the
combined or nonevaporable water content of the cement hydrate.
Permeability of Concrete
Interest in the permeability of concrete arises in connection with some
proposed specific application of the concrete and, therefore, may result
in emphasis of some particular aspect of concrete permeability.
Movement of water through concrete can be produced by various combinations of air or water pressure differentials, humidity differentials,
and solutions of different concentrations (osmotic effects) or at different
temperatures. Although it is known that the observed rate of movement
is dependent upon the characteristics of the membrane and the permeating
materials, much more needs to be learned regarding these relationships.
Various tests have been devised to determine permeability. Although these
procedures may reveal the relative characteristics of the concretes involved,
the quantitative value obtained may depend considerably upon details of
the experimental conditions.
For example, considering the relatively simple case of a hydrostatic water
pressure differential across a membrane of concrete, it has been observed
that many factors influence the rate of flow of water through the membrane. The prior curing history of concrete, the direction of permeation in
relation to casting position (under-aggregate fissures), the treatment of the
surfaces (as-cast, sawed, or sand-blasted), carbonation, the down-stream
conditions (whether air or water), the measurement of inlet or outlet flow
268
or both, the nature and amount of solutes in water, and the electrical
conductivity between the membrane faces or factors that must be considered. However, the test procedures commonly used for determination of
the water permeability of concrete are probably sufficient to reveal relative
differences in the permeability of concretes made with different water
contents, admixtures, and so on.
At the present time the qualitative effects of many factors on the
water permeability of concrete are known. The permeability of concrete to
water under hydrostatic pressure will depend significantly upon the permeability of the cement paste component of the concrete, providing the concrete is intact--not previously damaged by frost or rapid drying and not
containing excessive under-aggregate fissures or honeycomb. All of the
permeating water must pass through the paste component of the concrete (the continuous phase), and if the paste is of low permeability, the
concrete will show similar characteristics.
The hydrostatic water permeability (expressible in terms of Darcy's law)
of a well-cured paste is reduced approximately a thousandfold by reduction in w/c ratios from 0.8 to 0.4 weight. This large reduction in permeability is due to the drastic reduction in capillary size and volume that
accompanies the decrease in w/c ratio. All the capillary volume in fresh
paste is capable of rapid transmission of water (as revealed by bleeding);
this "permeability" rapidly decreases with hydration. The permeability
of a paste of 0.8 w/c ratio may decrease a thousandfold between the curing
ages of 7 days and 1 year. Thus, there is millionfold difference between the
permeability of high-water-ratio paste at early age and that of well-cured
low w/c ratio paste.
The permeability of cement paste appears to undergo a relatively abrupt
reduction when, because of original w/c ratio and extent of cement hydration, it attains the condition at which the capillary system becomes
discontinuous due to blocking by increase in the concentration of the gel
component in the paste. This attainment of discontinuity in the capillary
system appears to occur when the total porosity (gel plus capillary) reaches
approximately 50 percent. It has been estimated that cement pastes of
0.40 0.50, 0.60, and 0.70 w/c ratio require approximately 3 days, 14 days,
6 months, and 1 year of normal hydration, respectively, in order to attain
this condition. There is evidence that even when the capillary system becomes "discontinuous" (as measured by water permeability) there still
remains some continuity in the large pore system. This seems necessary
in order to account for the spontaneous propagation of freezing in saturated
pastes. Apparently the large capillary spaces necessary for the propagation
of lee formation are not sufficiently numerous to influence significantly the
permeability measurements.
The permeability of dense, impervious aggregates, such as traprock or
marble, has been observed to be as low as that of mature pastes of about
269
0.40 w/c ratio. Some granites have a permeability comparable to that of a

mature paste of 0.70 w/c ratio, despite the enormous differences in porosity, namely, about 50 percent for the 0.7 w/c ratio paste and about I percent
for the granite. Obviously, if the aggregate is more or less permeable than
the paste matrix, the permeability of concrete will be changed accordingly.
Air voids in the concrete, because they are much larger than the capillary and gel pores and, hence, do not impede movement of water, should
increase the permeability of saturated concrete roughly in proportion to
their quantity, provided other factors remain constant. However, other
factors seldom do remain constant--it is commonly observed that air entrainment in most practical concretes will reduce segregation and bleeding
and permit reductions in the w/c ratio, with the result that the concrete
may actually be more impermeable despite the presence of the air voids.
Capillarity can also produce liquid movement in concrete. The presence
of water at one face of the concrete and unsaturated air at the other may
give rise at the air interface to large negative pressures or capillary tensions
(meniscus effects) that tend to draw the water through the concrete.
Absorption of Concrete
The term "absorption" is usually applied to concrete in regard to the
weight gain of partially dried specimens upon contact with or immersion
in water.
During absorption it may be considered that the larger capillary spaces
in the paste are the first to be wetted, with the finer gel pores perhaps
next, their rate of absorption being slow because of their low water permeability. The coarse pores in' aggregate can become nearly filled with water
only after a relatively high degree of saturation is established in the.paste
surrounding the aggregate. Because of its usually much finer pore structure, the paste, if it is much below saturation, can easily remove most of
the aggregate water. Upon long-continued exposure to water, the air voids,
particularly the smaller voids, may become filled with water. Such a process
requires that the air in the void compressed by the absorption process
must dissolve in the capillary water and slowly diffuse out of the specimen
to permit filling of the void with liquid.
The absorption of concrete is measured by ASTM Test for Specific
Gravity, Absorption, and Voids in Hardened Concrete (C 642). In this
method the water absorbed by an oven-dried specimen is measured after
48 h immersion or after such immersion followed by 5 h in boiling water.
The ratio of the water absorbed to the dry weight is the absorption. Factors
significantly influencing absorption are the curing history, w/c ratio, aggregate characteristics, air content, cement type and fineness (particularly
at early ages), specimen size and shape, method of surface preparation
(cast, broken) surface carbonation, and so on.
270
The absorption test is of value primarily as a basis of comparison of

different concretes, the absorption in a gross manner being a function
of the permeability and porosity of the specimen although influenced by
many factors of test procedure. With an appreciation of these factors, it
is not surprising that only rough correlations are observed between the
results and such parameters as w/c ratio, strength, durability (freezable
water content), and attack by aggressive solutions.
Frost Durability
Concrete that has in its pores a relatively large amount of water is often
damaged on exposure to cold temperatures by a process known as freezethaw action. While no agreement has been reached as to the cause of the
phenomenon, the existence of internal, probably nonuniform, pressures
cannot be questioned. The distress can arise in either the cement paste or
the aggregate component of the concrete. Only that in the paste will be
considered here.
The first analysis of the freeze-thaw process was that of Powers. His
description involved freezing of water in the comparatively large pores,
which could cause the unfrozen water to be forced ahead of the freezing
front due to the volume increase experienced by water on freezing. If such
movement was over a great enough distance, the so-called critical distance,
it would be associated with hydraulic pressures great enough to disrupt
the concrete and cause a crack. In terms of this hydraulic pressure hypothesis, the role of the entrained air is to provide escape voids into which
the unfrozen water can go without having traversed the critical distance
and, therefore, without generating disruptive hydraulic pressure.
Powers later advanced the concept of gel-water diffusion. The water
in the small pores cannot freeze at ordinary freezing temperatures, owing
to thermodynamic considerations based on the small pore size. This unfrozen water will then migrate to the ice in larger pores and cause expansive pressures. Osmotic pressures from differences in concentrations of
dissolved salts would be expected to augment these expansive pressures.
Again, entrained air voids offer an alternative destination for the gel water,
so that expansive pressures will not be generated.
More recently Litvan has proposed a mechanism wherein the water in
the small pores migrates to the ice already formed because of a vaporpressure differential that will develop between the two as the temperature
falls below freezing. Mechanical damage is caused by the inability of the
system to achieve an equilibrium moisture distribution and by subsequent
glassy ice formation.
The exact mechanism of freeze-thaw, or the extent to which each of the
above processes contributes to the difficulty, is still a matter of debate
and further research. Whatever the cause, moisture redistribution and,
271
consequently, pore pressure generated will depend upon the amount of

freezable water, the rate of freezing, the permeability of the paste, and
the distance the water has to migrate to achieve equilibrium. If the developed
stresses exceed the tensile strength of the concrete, deterioration ensues.
These stresses can be eliminated effectively by the presence of entrained
air voids of proper size, number, and spacing.
If the paste is so dense that only the smallest pores are present, it will
probably be durable to ordinary freezing exposure. But since practical
considerations of workability usually require higher w/c ratios, most nonair-entrained concrete is inherently susceptible to frost damage. Naturally,
if proper drainage, coatings, etc. keep the concrete relatively dry, no problem
will occur. But the only way to be sure that concrete exposed to moisture
and freezing temperatures will be durable is to entrain a proper air void
system by the use of an appropriate air-entraing agent.
To avoid the development of destructive internal pressures under most
practical conditions, the separation between entrained air bubbles should
be in the range of approximately 0.12 to 0.25 mm. This critical distance
has been verified experimentally and in practice. Entrained air has little,
if any, use in preventing freeze-thaw trouble arising in the aggregate component.
The test methods for freeze-thaw durability of concrete are ASTM Test
for Resistance of Concrete to Rapid Freezing and Thawing (C 666) and
Test for Critical Dilation of Concrete Specimens Subjected to Freezing
(C671). In either of these test methods the concrete is frozen and its changes
in dimension or in properties are measured. Excessive expansion or a
large decrease in sonic modulus of elasticity are the usual indications of a
nondurable concrete.
Mechanical Properties of Concrete

The important mechanical properties of concrete, such as strength,
stiffness, shrinkage, and creep, are all related to, and their magnitudes
are determined largely by, the porosity of the concrete. The effects of the
aggregate are important, but less so than those of the cement paste.
The well-known w/c ratio rule is merely an expression of this effect;
the higher the w/c ratio, the greater will be the capillary porosity of the
paste and, therefore, the lower the strength. This idea was first given
quantitative expression by Powers in his gel/space ratio rule, that is, the
higher the ratio of gel to space available for it, the greater the strength.
All porous materials have this same inverse dependency between strength
and porosity. Both theory and experiment lead to relationships that are
either parabolic or exponential between the two. Another result of such
studies is the existence of a critical total porosity of about 60 percent, at
which point the strength becomes zero. As a first approximation the rule
272
seems to hold as well for air voids as for the small pores; the total porosity is the important parameter.
The strength of concrete depends of course to some extent on other
factors--chemical composition of cement, aggregate properties, placement practices--but the main influence is the porosity of the paste. Both
Young's and the shear moduli of elasticity of pastes depend on the porosity, being greater with the cube of the gel/space ratio. Like the strength,
they are controlled by all those things that result in different paste porosities--principaUy w/c ratio and duration of proper curing.
Effective Pore Area

The magnitude of pore pressure that is required before destructive forces
develop in the concrete is related to the effective area of the concrete over
which they produce stress. Small and isolated pockets of high pressure
might be accommodated by the concrete, whereas lower pore pressures
operating over large areas of the concrete could produce forces above the
inherent strength of the concrete and could, therefore, cause damage.
This problem of the effective pore area in concrete is also of direct concern
to the designers of gravity dams. Various mechanical and analytical problems
confront investigators of this matter, and unanimous agreement has not
been reached concerning the effective pore area to water in mortar or.
concrete, reported values ranging from about 40 percent to almost 100
percent, depending upon the grade of materials studied and the method
of analysis applied. That the effective pore area is high is not surprising
when the high porosity and fine texture of the pores in concrete are considered.
Conclusion
The influence of porosity of the concrete, and in particular that of the
hardened cement paste component, on its important engineering and
technological properties is greater than that of any other property of the
material.
The exact nature of the pore system and of the solids that frame it
is still a matter of debate and ongoing research, as are also the quantitative
relationships between the parameters that describe the pore system and the
properties of the concrete it so strongly influences.
Bibliography
Symposia
Proceedings, Fifth International Symposiumon the Chemistry of Cement, Concrete Association of Japan, Tokyo, 1969.
273
"Pore Structure and Properties of Materials," Proceedings, International Symposium Rilem/

IUPAC, Academia, Prague, 1973.
Proceedings, Conference on Hydraulic Cement Pastes: Their Structure and Properties,
Cement and Concrete Association, Sheffield, 1976.
Durability of Concrete, SP-47, American Concrete Institute, Detroit, 1975.
Proceedings, Sixth International Congress on the Chemistry of Cement, Moscow, 1974.
General R e v i e w s
Copeland, L. E. and Verbeck, G. J., "Structure and Properties of Hardened Cement Paste,"
Proceedings, Sixth International Congress on the Chemistry of Cement, Moscow, 1974.
Diamond, S., "Cement Paste Microstructure--An Overview at Several Levels," Proceedings,
Conference on Hydraulic Cement P~istes: Their Structure and Properties, Cement and
Concrete Association, Sheffield, 1976, p. 2.
Kondo, R. and Daimon, M., "Phase Composition of Hardened Cement Paste," Proceedings,
Sixth International Congress on the Chemistry of Cement, Moscow, 1974.
Valenta, O., "Main Properties of the Pore System on Non-Metallic Structural Materials,
Their Significance and Experimental Determination," Proceedings, International Symposium
Rilem/IUPAC, Academia, Prague, 1973, p. E-75.
Verbeck, G. J. and Helmuth, R. H., "Structures and Physical Properties of Cement Pastes,"
Proceedings, Fifth International Symposium on the Chemistry of Cement, Cement Association of Japan, Tokyo, Vol. 3, 1969, p. 1.
Pore S t r u c t u r e o f H a r d e n e d C e m e n t Paste
Auskern, A. and Horn, W., "Capillary Porosity in Hardened Cement Paste," Journal of
Testing and Evaluation, Vol. 1, 1973, p. 74.
Brunauer, S., Skalny, J., and Odler, I., "Complete Pore Structure Analysis," Proceedings,
International Symposium Rilem/IUPAC, Academia, Prague, 1973, p. C-3.
Diamond, S., "A Critical Comparison of Mercury Porisimetry and Capillary Condensation
Pore Size Distributions of Portland Cement Pastes," Cement Concrete Research, Vol. 1,
1971, p. 531.
Diamond, S. and Dolch, W. L., "Generalized Log-Normal Distribution of Pore Sizes in
Hydrated Cement Paste," Journal Colloid Interface Science, Vol. 38, 1972, p. 234.
Feldman, R. F., "Volume Change, Porosity, and Helium Flow Studies of Hydrated Portland Cement," Proceedings, International Symposium Rilem/IUPAC, Academia, Prague,
1973, p. C-101.
Feldman, R. F., "Changes to Structure of Hydrated Portland Cement on Drying and Rewetting Observed by Helium Flow Techniques," Cement Concrete Research, Vol. 4, 1973,
p. 1.
Feldman, R. F., "Density and Porosity Studies of Hydrated Portland Cement," Cement
Technology, Vol. 3, Part 1, 1972, p. 5.
Feldman, R. F. and Sereda, P. J., "Discussions," Proceedings, Fifth International Symposium
on the Chemistry of Cement, Cement Association of Japan, Tokyo, 1969, Vol. 3, p. 36.
Powers, T. C., "Physical Properties of Cement Paste," Proceedings, Fourth International
Symposium on the Chemistry of Cement, National Bureau of Standards, Washington,
1960, p. 577.
Powers, T. C. and Brownyard, T. L., "Studies of the Physical Properties of Hardened Portland Cement Paste," Proceedings, American Concrete Institute, Vol. 43, 1947, p. 933.
Powers, T. C., Copeland, L. E., and Mann, H. M., "Capillary Continuity or Discontinuity
in Cement Pastes," Journal, Portland Cement Association, Research Division Laboratories,
Vol. 1, Part 2, 1959, p. 38.
Skalny, J. and Odler, I., "Pore Structure of Calcium Silicate Hydrates," Cement Concrete
Research, Vol. 2, 1972, p. 387.
Winslow, D. N. and Diamond, S., "A Mercury Porosimetry Study of the Evaluation of
Porosity in Portland Cement," Journal of Materials, Vol. 5, 1970, p. 564.
Winslow, D. N. and Diamond, S., "The Specific Surfaces of Hydrated Portland Cement Paste
as Measured by Low-Angle X-Ray Scattering," Journal Colloid Interface Science, Vol.
45, 1973, p. 425.
274
F r e e z e - T h a w Durability
Fagerlund, G., "The Significance of Critical Degrees of Saturation at Freezing of Porous
and Brittle Materials," Durability of Concrete, SP-47, American Concrete Institute,
Detroit, 1975, p. 13.
Larson, T. D., Cady, P. D., and Malloy, J. J., "The Protected Paste Volume Concept Using
New Air-Void Measurement and Distribution Techniques," Journal of Materials, Vol. 2,
1967, p. 202.
Litvan, G. G., "Phase Transitions of Adsorbates IV. Mechanism of Frost Action in Hardened
Cement Paste," Journal American Ceramic Society, Vol. 55, 1972, p. 38.
Litvan, G. G., "Pore Structure and Frost Susceptibility of Building Materials," Proceedings,
International Symposium Rilem/IUPAC, Academia, Prague, 1973, p. F-17.
Mielenz, R. C. et al, "Origin, Evoluation, and Effects of the Air-Void System in Concrete,"
Proceedings, American Concrete Institute, Vol. 55, 1958, pp. 95, 261,359,507.
Powers, T. C., "The Air Requirement of Frost-Resistant Concrete," Proceedings, Highway
Research Board, Vol. 29, 1949, p. 184.
Powers, T. C., "Freezing Effects in Concrete," Durability of Concrete, SP-47, American
Concrete Institute, Detroit, 1975, p. 1.
Powers, T . C. and Helmuth, R. A., "Theory of Volume Changes in Hardened Portland
Cement Paste During Freezing," Proceedings, Highway Research Board, Vol. 32, 1953,
p. 285.
MacInnis, C. and Beaudoin, J. J., "Pore Structure and Frost Durability, Proceedings, International Symposium Rilem/IUPAC, Academia, Prague, 1973, p. F-3.
M e c h a n i c a l Properties
Beaudoin, J. J. and Feldman, R. F., "A Study of Mechanical Properties of Autoclaved
Calcium Silicate Systems," Cement Concrete Research, Vol. 5, 1975, p. 103.
Fagerlund, G., "Strength and Porosity of Concrete," Proceedings, International Symposium
Rilem/IUPAC, Academia, Prague, 1973, p. D-51.
Helmuth, R. A. and Turk, D. H., "Elastic Moduli of Hardened Portland Cement and
Tricalcium Silicate Pastes: Effect of Porosity," Highway Research Record, Special Report No. 90, 1966, p. 135.
Hoff, G. C., "Porosity-Strength Relationships for Cellular Concretes," Cement Concrete
Research, Vol. 2, 1972, p. 91.
Popovics, S., "Effect of Porosity on the Strength of Concrete," Journal of Materials, Vol. 4,
1969, p. 356.
Popovics, S., "Method of Developing Relationships Between Mechanical Properties of Hardened Concrete," Proceedings, American Concrete Institute, Vol. 70, 1973, p. 781.
Sereda, P. J., "Significance of Microhardness of Porous Inorganic Materials," Cement
Concrete Research, Vol. 2, 1972, p. 717.
P. D. Cady '
Chapter 19--Corrosion of
Reinforcing Steel
Introduction
High quality portland cement concrete, properly placed, consolidated, and
cured, provides an environment that will normally prevent corrosion of
embedded steel. However, the technical literature abounds with evidence of
serious problems related to corrosion of reinforcing steel in concrete, both in
terms of frequency of occurrence and magnitude of resultant damage. Reinforced concrete highway structures (principally bridge decks) and structures
exposed to seawater or marine atmospheric environments are most prone to
this problem. The common denominator is the presence of chlorides which
result from the application of deicing salts in the former instance and from
the environment in the latter. The costs associated with repair or replacement of concrete bridge decks damaged by reinforcement corrosion is, and
shall be for years to come, a major maintenance expense for many highway
agencies [1, 2].2 While highway bridges are expected to last 30 years or more,
many show signs of deterioration in five years or less [3, 4].
The damage that ensues from corrosion of reinforcing steel is manifested
in the later stages as spalling of the concrete cover between the steel and the
nearest free surface. This is the familiar "pot hole" on a concrete bridge deck
(Fig. 1). It is the result of pressures generated in the concrete due to the
volume increase associated with the conversion of steel to corrosion products.
In addition to the loss of concrete cover, a reinforced concrete member suffers structural damage due to loss of bond and to loss of steel cross section,
sometimes to the extent that structural failure occurs (Fig. 2).
Mechanisms of Corrosion
The corrosion of steel in concrete results from the development of electrochemical corrosion cells. The electrical current involved may be induced,
IProfessor of civil engineering, The Pennsylvania State University, University Park, Pa.
16801.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
lntcrnational
www.astm.org
276
FIG. 1--Typical pot hole on a bridge deck.
that is by stray electrical currents (rarely), but more commonly results from
electrochemical potentials created within the concrete in accordance with the
Nernst equation [5]. For the chemical reaction
aA + b B - - c C
+ dD
the electrochemical potential

E=EoRT,
(ac)C (aD) d
where
E~=
R =
T =
n =
F =
a =
standard electrode potential,

universal gas constant,
absolute temperature,
n u m b e r of chemical equivalents reacting,
Faraday's constant,
activities of reactants (A and B) and products (C
and D), and
exponents a, b, c, d = molar quantities of reactants and products.
CADY ON CORROSION OF REINFORCING STEEL
277
FIG. 2--Structural failure o f a reinforced concrete column due to corrosion o f reinforcing

steel.
278
TESTS A N D
PROPERTIES
OF C O N C R E T E
The presence of an electrical potential is a necessary prerequisite for the occurrence of corrosion9 However, the corrosion rate is governed by the current
in accordance with Faraday's law
r=kI
where
r = corrosion rate in weight loss per unit time,
k = the electrochemical equivalent (a constant), and
I = current.
The electrochemical potentials that must be present to form the corrosion
cells may be created in several different ways. For the specific case of steel in
concrete, however, they are probably limited to the following:
(a) Differential aeration--differences in concentration of oxygen over the

surface of the steel.
(b) Differential ion concentration--differences in concentration of metal

ions, dissolved salts, and pH (alkalies and free lime) in the concrete in
the vicinity of the steel9
(c) Differential surface properties--variations in the surface characteristics of the steel such as mill scale or breaks in coatings.
The corrosion cell is shown schematically in Fig. 3, and a flow diagram showing the chemical reactions in the corrosion cell is shown in Fig. 4 [6]. It is evi9
".': "n:.: : " A " : . "~." :a'..'.:'.A:.A'..'..:.

9 h:." ~'. : ".'. " "" .: '" -'.Z. :''" :.":
:9. 9r
. ..".'.i .;.: ~" , ~

9
: Z'""--'"""i
9 ...:"
~""'-'"
~"
= OH
~ ."o ".
"'..A"" :" :.'.~"a :""
~..:".',.'.,-.'A
"'""
""'"
:. . :'.,:'."
"
"
"
"
"
-1
FIG. 3--Schematic diagram of the corrosion cell.
.
F:-]
F-q+?Tq.
+
FIG. 4--Flow diagram showing the basic chemical reactions in the corrosion cell (after Erlin
and Verbeck [6]).
279
dent from these figures that hydroxyl ions and other anions present (such as
chloride) will diffuse toward the anode during corrosion, while iron and other
cations (Ca + + and Na ) will diffuse toward the cathode. Since diffusion of
ions through the concrete is necessary for corrosion to occur, it follows that a
continuous conductive moisture path must exist in the concrete between the
anode and the cathode. Also, it will be observed that oxygen and water must
be available at the cathode. In the absence of oxygen, polarization of the
cathode by hydrogen released in electrolysis interferes with current flow and
corrosion ceases. Finally, conditions must exist that will tend to break down
the passivation of the steel that normally occurs in alkaline environments
(pH > 11.5)[6].
The normally protective environment provided by concrete relative to corrosion of steel results from the formation of gamma ferric oxide (Fe2Oa).
Any breaks that occur in this protective film are quickly repaired in the
presence of sufficiently high hydroxyl ion concentration, forming first ferrous
hydroxide which then reacts with oxygen to form predominantly cubic FeaO4
and gamma Fe203. Electrochemical production of gamma Fe203 then
builds up the protective film thickness until it becomes impervious to ions,
and passivity is restored [7].
Effect of Chloride Ion
One of the reaction products in the hydration of portland cement is free

lime which exists in concrete as crystals of calcium hydroxide. Therefore,
water contained within the pore structure of concrete becomes buffered at
least to the pH of saturated lime solution (about 12.5). The presence of small
quantities of alkalies (sodium and potassium oxides) in portland cement may
raise the pH even higher [8]. As mentioned previously, steel becomes
passivated at pH values above 11.5. However, the presence of chloride ion
changes the situation. The main function of the chloride ion in promoting
corrosion of reinforcing steel in concrete is to release ferrous ions at the
anode though the formation of ferrous chloride. Also, chloride salts being
strong electrolytes increase the conductivity of the ground path enhancing
the corrosion current. However, the influence of chloride content in reducing
the resistivity of concrete is minor, especially in comparison to the effect of
moisture content [5]. Finally, chloride ions also affect the pH which, as noted
previously, influences corrosion. Because a portion of the chloride ions enter
into chemical reaction with the alumina compounds in the hydrated cement
(to be discussed later), sodium chloride will tend to increase pH and calcium
chloride will tend to decrease it [6, 9].
The effects of chloride ion concentration on corrosion probability and rate
are difficult to quantify because of the strong dependence on moisture content and availability of oxygen to depolarize the cathode. However, Hausman
[8] probably established the lower limit for threshold chloride concentrations
280
through experiments with oxygenated, saturated lime solutions. He showed

that as the pH increases above 11.5, chloride ion threshold concentrations required to initiate corrosion increase, but chloride ion activity at the corrosion
threshold bears a constant ratio to hydroxyl ion activity of about 0.6. It
should be reemphasized that while the threshold chloride concentration increases with increasing pH, corrosion of steel in concrete might not occur at
the chloride ion threshold concentration predicted by Hausmann due to
somewhat more limited ion mobility and limitations on the rate of diffusion
of oxygen to the cathode in concrete as compared to oxygenated, saturated
lime solution. However, using Hausmann's results as lower limits, it may be
stated that the onset of corrosion may occur at chloride ion concentrations as
low as about 0.2 to 2.8 kg of chloride ion per cubic metre of concrete at pH
values of 12.5 to 13.2, respectively.
With regard to corrosion threshold concentration of chloride ion in concrete, work carried out by the Federal Highway Administration [10,11]
revealed a value of 0.20 percent C I - based on the weight of cement in the
concrete mixture. For the typical range of concrete mixtures normally encountered, this represents corrosion threshold chloride contents ranging
between 0.7 and 0.9 kg CI - / m 3 of concrete. These values are in good agreement with those predicted by Hausmann, as previously discussed.
Detailed study of 21 bridge decks of ages varying from 6 to 23 years, some
showing signs of deterioration and others not, led Stratfull et al [12] to conclude that the quantity of chloride ion associated with the incidence of active
corrosion of steel in concrete is about 0.6 kg/m 3 of concrete. Again, this
figure is in close agreement with the data cited previously. Other bridge deck
studies revealed similar corrosion threshold values [13].
It is generally contended that only ions of the halogen elements, of which
the chloride ion is the only one of practical significance with respect to concrete, are apparently capable of promoting corrosion of steel embedded in
concrete. However, Hausmann [8] also indicates that bisulfide ( H S - ) ions
can produce similar results.
Effect of Moisture Content
The presence of moisture is not only a necessary prerequisite for the initiation of corrosion, moisture content is the primary rate determining factor in
the corrosion process. This is largely due to the effect of moisture content in
reducing the resistivity of the concrete thereby increasing current flow--the
determinant of corrosion rate. Stratfull [14], in comparing electrical resistivity of concrete with concrete deterioration produced by corrosion of reinforcing steel, found a nearly linear inverse relationship up to a resistivity value of
60 000 fl-cm beyond which corrosion apparently does not occur. While
resistivity is influenced by chloride content, it is more sensitive to the
moisture content [5].
281
Effect of Concrete Mixture Variables

Properties of concrete that influence the alkalinity and the quantity or uptake and transmission of water, oxygen, and chlorides directly and profoundly affect corrosion of embedded steel. The concrete mixture variables
that affect these properties are water/cement (w/c) ratio, aggregates, cement, and admixtures.
Water~Cement Ratio
Water/cement ratio and curing (to be discussed later) are the primary
determinants of permeability of concrete. High permeability concretes permit ready access of moisture, chlorides, oxygen, and carbon dioxide to the
embedded steel. The functions of moisture, chloride ions, and oxygen have
already been discussed. Carbon dioxide reduces the pH by reacting with the
free lime to form calcium carbonate. The effect of reduction of pH on steel
corrosion also has been discussed previously.
Water/cement ratios normally used fall in the range 0.4 to 0.5 by weight,
with lower and upper practical limits of 0.35 and 0.6. The permeabilities for
well cured concretes with w/c ratios up to about 0.5 are very low and nearly
constant [9,15,16]. Above 0.5, permeabilities increase exponentially with
w/c ratio. This is illustrated in Fig. 5 [17].
The effect of the w/c ratio on permeability results from capillary porosity
produced by excess mixing water over and above that required for hydration
of cement particles. The coefficient of permeability of a mature paste with
zero capillary porosity has been determined to be about 7 X 10 - n darcys
[18].
The permeability of concrete to chloride ions also increases with increasing
w/c ratio. Verbeck [9] reported a threefold increase in permeability to
chloride ion as the w/c ratio was increased from 0.4 to 0.6. For chloride ions
that enter the concrete from a free surface, the ion concentration decreases
with depth of penetration in accordance with an exponential decay relationship. Spellman and Straffull [19] found that the average chloride content versus depth for 16 heavily salted bridge decks in California could be approximated by the following equation
C = 6.44 (0.724) s
where
C ---- chloride content of a 2.5 cm thick section, kg/m 3, and
S ---- maximum depth below the surface for the 2.5 cm thick section, cm.
This equation may be considered to represent the mean condition for bridge
decks subject to heavy salt applications for about a 10-year period.
282
6-~
5.
%
x
3-
1-
I
.40
I
.45
I
.50
Water / Cement
I
.S5
w
.60
!
.65
Ratio
FIG. S--Permeability versus w/c ratio for hydrated portland cement paste.
In experiments with simulated bridge deck slabs, Clear [11] found that the
chloride content at a given depth after 830 daily salt applications varied
directly with w/c ratio. For example, at a depth of about 5 cm, average
chloride contents of approximately 0.2, 1.0, and 2.4 kg C1 - / m a were found
for w/c ratios of 0.40, 0.50, and 0.60 respectively. The calculated value using
Straffull's equation for a 5-cm depth is 1.2 kg C l - / m a. Berman and
Chaiken [20] found the chloride concentration due to the penetration of a
1.18 N calcium chloride solution at all values of time and depth to be consistently greater for concrete of 0.5 w/c than for 0.4 w/c.
Aggregate
The primary effect of the aggregate in concrete on corrosion of embedded

steel relates to its effect on permeability. All other factors being constant,
permeability increases with increasing maximum size of the coarse aggregate
for most mineral aggregate materials [16]. This is due to the fact that most
mineral aggregates have permeability coefficients 10 to 1000 times greater
than the permeabilities of cement pastes in the usual range of w/c ratios.
283
Since aggregates typically constitute about 70 percent of the volume of concrete, it is evident that they play a major role in determining the permeability
of concrete. However, their effect in transmission of moisture, chloride ion,
oxygen, and carbon dioxide from free surfaces to the vicinity of embedded
steel is of considerably less importance than might be implied from their effect on permeability. The reason for this is that the more or less continuous
capillary porosity in the paste phase is generally composed of much smaller
pore diameters than those of the aggregates. Because of this, the aggregates
are generally unable to compete with the paste for available moisture on the
basis of capillarity. Thus, the capillary porosity of the paste phase constitutes
the conduit system for the movement of moisture and dissolved solids and
gases. About the only role played by the aggregates in the transport process is
to replenish moisture to the paste during periods of drying.
The gradation of the aggregates may have a significant effect on corrosion
of embedded steel if gradations are such as to result in segregation and
bleeding of the concrete mixture. Segregation and bleeding will produce
channels, promoting the ingress of water, chlorides, oxygen, and carbon
dioxide, and may produce voids in the vicinity of reinforcing bars. The occurrence of areas of active corrosion at the locations of these voids is commonly observed. Such mixtures will also frequently display a porous concrete
zone at the level of the reinforcing steel in heavily reinforced, horizontal concrete members (such as bridge decks) [21]. This is due to the interference by
the reinforcing mat to the sedimentation process that causes the segregation.
In rare instances, the aggregates may contribute to corrosion by supplying
a source of chloride ions.
Cement
The effect of cement content on corrosion of embedded steel is negligible
[11]. However, the physical and chemical properties of the cement can have
an effect, though it is generally minor in comparison to other variables.
Fineness of the cement, for example, can influence corrosion to the extent
that it affects bleeding. High alkali content can help to retard corrosion
through its effect on pH. However, high alkali contents may lead to difficulties with alkali-aggregate reactions. In a practical sense, the property of
the cement that has the greatest influence on corrosion is the tricalcium
aluminate content. All other factors equal, the tendency for corrosion to occur decreases with increasing tricalcium aluminate content of the cement.
This is due to the fact that chloride ions react with the hydrated tricalcium
sulfoaluminate hydrate in the hardened cement paste to produce tricalcium
chloroaluminate. Thus, normal Type I portland cement (typically 12 percent
tricalcium aluminate) is three to five times more effective in removing
chloride ion than sulfate resisting Type V portland cement (typically 4 percent tricalcium aluminate) [9,22].
284
Admixtures
A wide variety of admixtures are incorporated into concrete mixes for
various purposes including set retardation, set and early strength acceleration, water reduction, plasticizing, air entrainment, and lime-scavanging
(pozzolanic). Most of them are organic materials that have no effect, per se,
on corrosion. However, the water reducers, plasticizers, and air entraining
agents usually reduce the w/c ratio and may therefore be somewhat
beneficial in retarding corrosion. Since, as mentioned previously, there is
relatively little change in permeability with change in w/c ratio below about
0.5, it would be expected that their effect would be relatively minor.
Pozzolans, which combine with the free lime liberated during the hydration of portland cement would be expected to increase the tendency for corrosion due to reduction in alkalinity. However, this has not been found to be
the case [23]. On the contrary, it has been reported that pozzolans actually
reduce corrosion, especially in steam cured concretes [24]. This apparent
anomaly is probably due to the fact that pozzolans appreciably reduce the
permeability of concrete and the removal of soluble constituents by soft or
slightly acidic waters (effloresence).
Calcium chloride (CaC12) is sometimes used as an accelerator, especially
in cold weather concreting operations. Of course, the addition of chlorides
would be expected to increase corrosion. With ordinary reinforced concrete,
reports on the effect on steel corrosion of CaC12 additions at the recommended rate of 2 percent by weight of the cement content vary widely [5,25].
One source reported that CaCl 2 additions greater than 0.5 weight percent of
the cement are seriously detrimental to corrosion resistance [26]. Also, all investigators seem to agree that CaC12 additions produce serious corrosion problems in prestressed and steam cured concretes.
Effect of Construction Variables
Any activity in the concrete construction process that may produce conditions that facilitate the movement of water, chlorides, oxygen, and carbon
dioxide to the site of embedded steel has the potential for profound affect on
corrosion of the steel. As mentioned previously, water, chloride ions, and oxygen are prerequisites for establishing active corrosion. Carbon dioxide accelerates the process by combining with the free lime (carbonation), thereby
reducing the high pH necessary to maintain a passive environment. The three
construction activities that have been found to be most critical with regard to
corrosion are depth of cover of the reinforcing steel, consolidation, and curing.
Cover of Reinforcing Steel

At first thought, depth of cover of the reinforcing steel would seem to be a
design problem rather than construction related. Generally speaking, this is
285
true. However, the fact that design and actual depths of cover do not always
agree is illustrated in Fig. 6.
FIG. 6--Effect of insufficient reinforcing bar support on depth of cover.
The minimum required depth of clear cover with sound, uncracked concrete to prevent reinforcing bar corrosion is usually considered to be 5 cm
[15,21,27]. It is obvious that thicker covers should be more effective in
preventing corrosion, but beyond 5 cm protection is reportedly not
significantly improved [5]. However, the required minimum depth of cove~
depends on the permeability (that is, w/c ratio) of the concrete. Minimum
cover depths of about 5, 7.5, and 10 cm have been found to be necessary to
protect reinforcing steel for w/c ratios of 0.4, 0.5, and 0.6 respectively
[I1,281.
In an attempt to quantify the effects of w/c ratio, concrete cover, and
chloride ion availability on time to corrosion, Clear [tl] presented the following modification of a relationship originally developed by Stratfull
Rt--
4 1 S i 1.22
K ~ (w/c)
where
R, = time to deterioration of concrete continuously exposed to saline
water, years,
286
Si = depth of steel below surface, cm,

K = chloride concentration of water, ppm, and
w/c = water/cement ratio by weight.
Since the required depths of cover cited above are minimum values and
since variations in actual depths of cover occur in practice, as illustrated
previously, target or specified values must be somewhat greater than the
minimum values. Based on a study of 17 bridge decks in New Jersey [29], it
was found that values of depth of cover were distributed approximately normally, with the mean value close to the specified value and the standard
deviation about 1 cm. Based on these findings, one can readily calculate
specified target cover values to assure that a given percentage of the steel
will have concrete cover equal to or greater than the desired minimum value.
For example, if the minimum required cover is deemed to be 5 cm, the percentage of steel having a cover of at least 5 cm for various specified target
cover values will be as given in Table 1.
TABLE l--Percentage of steel protected (5 cm minimun cover) for
various target values of cover.
Target Value, cm
Steel Protected, %
S.0
5.5
6.0
50
69
84
6.5
7.0
7.5
93
98
99
The generally accepted 5 cm minimum depth of cover for sound, high

quality, uncracked concrete is based on permeability of chlorides. The effect
of carbonation in lowering pH, under these conditions, is not considered to
be a problem at depths greater than 25 m m [9].
The depth of cover of the reinforcing steel is of special significance in the
case of horizontal slab concrete members (for example, bridge decks). Here,
insufficient depths of cover have been observed to produce cracking of the
concrete immediately above the steel reinforcing bars. Even the highest
quality concrete is of no value in retarding corrosion if cracks are available to
admit chlorides, water, oxygen, and carbon dioxide. An example of cracking
over reinforcing resulting from shallow reinforcing bar cover is shown in Fig.
7. Cracking results from the interference to subsidence of fresh concrete
presented by the reinforcing bars. The existence of a zone of tensile stress in
the plastic concrete above the reinforcing steel has been illustrated in photoplastic studies [30]. It has been found that values of concrete cover of 5 cm or
more are sufficient to prevent subsidence cracking [21,27,30]. However, conCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:47 EDT 2014
287
FIG. 7--Subsidence crack in fresh concrete due to shallow reinJorcement.
crete slump and diameter of the reinforcing bar also have significant, though
considerably smaller, effects [30].
Consolidation
The degree of consolidation of the concrete, especially in the vicinity of
embedded steel, has a profound effect on corrosion. Insufficient consolidation provides channels of ingress for the ingredients of corrosion--water,
chlorides, oxygen, and carbon dioxide--and when occurring as "honey
combing" in the vicinity of the reinforcement promotes the formation of differential aeration corrosion cells. For concretes with slump values of 7.5 cm
or more, good consolidation is easily achieved by internal vibration. However,
with the lower slump concretes, normal construction practices cannot be
relied upon to produce adequate consolidation. In one series of tests [11] the
chloride content at a depth of5 cm in the concrete was about 12 times greater
for concrete that had been compacted from 92 to 94 percent of maximum
density than it was for maximum density concrete of the same mix design.
Curing
The primary influence of curing is the effect that it has on permeability.
For example, the permeability coefficient for cement paste after 1 day of
hydration will be more than 25 000 times that for paste hydrated for 7 days
288
[9]. Of course, the effect of curing on the permeability of concrete will not be
as pronounced as this because of the influence of bleeding, degree of consolidation, and permeability of the aggregates, and the fact that water will be
available in the capillary pores for some time after cessation of curing.
Spellman and Straffull [24] studied the effects of steam and water curing
on the time to corrosion of embedded steel. In both instances exponential
relationships were obtained, of the form
p
~_ ( I D b
where
P = days to active corrosion potential for partial immersion in saturated
sodium chloride solution,
D = days of underwater curing following initial curing (steam or water),
a = a constant, equaling 6.33 for steam curing, and 6.0 for water curing,
and
b ---- a constant, equaling 0.66 for steam curing, and 0.90 for water curing.
In general, it will be observed that water curing provides significantly greater
resistance to corrosion than steam curing.
Ferrocement
Ferrocement consists of very thin (1 to 4 cm) heavily reinforced concrete

sections. The reinforcement typically consists of steel reinforcing bars
(usually high strength) sandwiched between multiple layers of steel wire
mesh. Ferrocement sections are reinforced considerably more heavily than
ordinary reinforced concrete and the reinforcing steel has a much greater
specific surface. Also, minimum cover depths may be as small as 3 mm. It is
rather surprising, then, that ferrocement is applied so extensively and successfully in the construction of boat hulls, in view of potential corrosion problems. The apparent lack of corrosion problems is attributed to use of zinc
coated (galvanized) mesh, and high cement factor, low w/c ratio concrete
mortar in fabricating ferrocement sections. One corrosion related problem
has been encountered involving the evolution of hydrogen gas on plain steel
reinforcing rods sandwiched between layers of galvanized steel mesh. This
results from the galvanic cell created by the use of dissimilar metals (iron and
zinc), and the hydrogen gas produced appreciably reduces the bond of the
reinforcing bars and increases the porosity of the mortar in the vicinity of the
bars. The addition of chromic oxide at the rate of 300 ppm of the mixing
water apparently will eliminate this problem [17].
289
Steel Fiber Reinforced Concrete

Steel fiber reinforced concrete consists of concrete mixtures in which randomly oriented steel fibers, commonly about 0.25 to 0.50 mm in diameter by
2.5 to 6.4 cm long, are incorporated at the rate of 0.1 to 8 volume percent.
Like ferrocement, the specific surface of the reinforcing steel is considerably
greater than in conventional reinforced concrete. Fiber reinforced concretes
have only recently (since 1972) been put to use under conditions were corrosion might tend to be a problem (highway and bridge deck overlays), and it is
too early to make an assessment of their performance in this regard.
However, one laborabory study showed insignificant corrosion in 2 volume
percent steel fiber reinforced concrete specimens subjected to 90 daily cycles
of exposure to saturated salt water [31].
Prestressed Concrete
The factors that affect the corrosion of reinforcing steel in ordinary reinforced concrete are equally active in prestressed concrete except they are
somewhat more critical because of the smaller cross sections (larger specific
surface) of the prestressing tendons [15]. In addition, however, the high
strength steels used in prestressing tendons and the high stress levels encountered in prestressed members can lead to a form of corrosion not found
in ordinary reinforced concrete--stress corrosion cracking. Stress corrosion
cracking is a brittle form of failure that occurs under the conditions cited
above in the presence of certain ions. The chloride ion apparently does not
promote stress corrosion cracking [5, 9]. The most common ions known to
promote this kind of corrosion, of those that might be expected to come into
contact with prestressed concrete, are nitrate (NOa-) and bisulfide ( H S - )
[5]. Stress corrosion cracking has produced catastrophic failure of prestressed concrete sludge storage tanks, where hydrogen sulfide was present
[32]. However, reinforcement corrosion in prestressed concrete bridge
members, where ample chlorides from deicing salts are frequently available,
is practically nonexistent. Reportedly, less than 0.007 percent of the
prestressed bridges in the United States have shown any evidence of reinforcement corrosion and no catastrophic failures have occurred [5]. Of
course, this is vastly superior service to that obtained from the bridge decks.
The difference is probably attributable to the higher quality concrete,
somewhat lower levels of chloride exposure (except in marine environments),
and reduced tendency for cracking in prestressed members.
Assessing Corrosion Probabilities
Once corrosion of reinforcing steel has progressed to the stage where its effects are manifested by cracking and spalling of the concrete surface, it is
290
usually too late to take corrective measures that provide any degree of
permanency. Two techniques, often used in tandem, are currently employed
to assess the probability of active corrosion in the absence of physical
evidence at the surface of the concrete. These are determination of chloride
contents at the level of the reinforcing steel and the evaluation of electrical
half-cell potentials of the reinforcing steel. The Federal Highway Administration has adopted a procedure utilizing these two techniques as a
decision-making process for replacement or repair of concrete bridge decks
[33].
Chloride Determinations
As discussed earlier, threshold values of chloride content at which corrosion of embedded steel may occur are in the range of 0.6 to 1.2 kg/m 3 of concrete, depending on the pH. Furthermore, the probability of the presence of
active corrosion increases with increasing chloride content above the
threshold value. Therefore, determination of the chloride content in the
vicinity of the reinforcing steel provides an indicator of the probability of active corrosion.
Depth of the reinforcing steel is determined by means of a magnetic flux
device that is calibrated specially to indicate concrete cover (Pachometer).
The simplest and most economical method for acquisition of samples for
chloride analysis is the technique recommended by Clear [33]. It involves the
use of a rotary hammer to drill and pulverize the concrete to various depths.
Typically, samples are removed between depths of 0.6 to 1.9 cm (nominal
1.25 cm sample), 1.9 to 3.2 cm (nominal 2.50 cm sample), and 4.4 to 5.7 cm
(nominal 5.0 cm sample). A 5 cm outside diameter core bit will produce
about 50 g of powdered sample for each 1.25 cm sampling depth. Only 3 to 4
g are needed for chloride analysis.
The method recommended for analysis of chlorides was developed by Berman [34] and described in detail by Clear [33]. Briefly, it involves digesting a
weighed quantity of the pulverized sample in nitric acid, boiling, filtering,
and titrating the filtrate with silver nitrate solution of known normality. The
titration end point must be precisely determined and this is accomplished by
means of a potentiometric method utilizing a chloride ion specific electrode.
The combination solid state electrode (with built in reference electrode) is
recommended [33].
A number of precautions must be observed to assure the accuracy of the
results. These include running blank specimens for calibration purposes and
applying correction factors for variations in moisture and aggregate contents
of the test samples [33].
A rapid field technique for in situ analysis of the chloride content of
portland cement concrete has been described by Morrison et al [35]. The
method involves using a chloride ion specific electrode in a 1.9 cm diameter
291
hole drilled in the concrete, into which had been placed a specified quantity
of borate-nitrate solution.
Electrical Hal.f-Cell Potential

As discussed previously, the process of corrosion of steel in concrete involves the formation of galvanic corrosion cells due to the electrochemical
processes of oxidation at anodic areas and reduction at cathodic areas on the
embedded steel. Therefore, corrosion activity can be identified by its effect
on the electrical potential of a reference electrode (half-cell) inserted in the
galvanic corrosion circuit. A number of reference electrodes are suitable for
evaluating corrosion of steel in concrete, but either the copper-copper sulfate
(CSE) or the saturated calomel (SCE) half-cell are most commonly used. By
convention, potential readings are taken to be negative values.
In practice, half-cell potential measurements are obtained on a predetermined grid pattern over the surface of the concrete. Plots of equipotential
contours are then constructed to delineate areas of possible corrosion activity. Half-cell potential readings do not measure the structural or physical
condition of the concrete [36]. Furthermore, the readings are only statistically related to corrosion activity. Potential readings more negative than -- 0.35
V relative to the copper-copper sulfate half-cell (CSE) have been found to be
indicative of active corrosion 95 percent of the time [33]. Potentials in the
range --0.20 to --0.35 V (CSE) are inconclusive as regards corrosion, and
readings less negative than --0.20 V (CSE) are considered to be indicative of
passive conditions [28].
The detailed procedures involved in the half-cell method are presented in
the ASTM Test for Half Cell Potentials of Reinforcing Steel in Concrete
(C 876).
Protective Measures
The first and most important line of defense against corrosion of steel in
concrete is the exercise of sufficient care to assure that the steel has an adequate cover of high quality (high cement content), low permeability (low
w/c ratio) concrete that has been consolidated adequately around the reinforcement. However, for structures in particularly hazardous environments
relative to the potential for corrosion problems (for example, bridge decks
subject to deicer salts or structures in marine environments), additional protective measures may be warranted. Generically, these protective measures
may be classified into four categories: (a) those that remove chloride ions or
inhibit their corrosive action; (b) those that seal the concrete, preventing the
access of chlorides, water, oxygen, and carbon dioxide; (c) those that seal the
surface of the reinforcing steel with a material impervious to chloride ions,
water, and oxygen; and (d) those that impose a protective current flow in opCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:47 EDT 2014
292
position to the galvanic corrosion cell (that is, render the entire surface of the
steel cathodic).
Also, the protective measures can be relegated to two functional classes:
(a) those that must be instituted at the time the concrete is placed, and (b)
those that may be applied anytime after the concrete has hardened and
cured.
Obviously, the first of these two functional classes cannot be applied after
the fact to existing structures. In briefly describing the protective measures
that fall into each of the four generic categories, it will be readily evident to
which functional classification each belongs.
Removal or Inhibition of Chloride Ions

Two possible approaches exist for removal of chloride ions from concrete
to arrest or prevent corrosion. One consists of forcing water through the concrete to flush out the CI- ions. Attempts to do this have not been successful
[37]. The other possibility is to use electrochemical means, taking advantage
of the migration of ions in a direct current electrical field (electro-osmosis).
This technique has been successfully demonstrated in the field by at least two
investigative teams [37-39]. Simultaneous impregnation with a furfuralacetone mixture while removing chloride ions during electro-osmosis was
shown to be capable of removing up to about 12 kg/m 3 of chloride ion and
reducing the permeability of the concrete to about one-third that of untreated concrete [39]. The electrical cost to remove 6 k g / m 3 of C1- was
reported to be $0.48/m 2 on a concrete bridge deck.
The use of corrosion inhibiting admixtures in concrete has produced, at
best, uncertain results [9,26,37, 40-42]. Those that interfere with the corrosion process by stifling reactions at the cathode (for example, calcium carbonate, aluminium oxide, and magnesium hydroxide) do not tend to bond to
the steel very well. Those that stifle the anodic reactions (alkalies, phosphates,
and chromates), can produce intensified localized corrosion if used in insufficient quantities. Furthermore, the amount of inhibitor required increases
with increasing chloride content, and the effectiveness of inhibitors reportedly lasts for only a short time. Also, some inhibitors seriously reduce the
strength of concrete [41]. For these reasons, the use of corrosion inhibitors
generally is not recommended. Of the many potential corrosion inhibitors
tried, sodium benzoate and sodium nitrate generally are considered to be the
most promising.
Concrete Sealants
Concrete sealants are of two general classes--those that provide a barrier
at the surface to prevent, or at least seriously retard, the penetration of
water, chlorides, oxygen and carbon dioxide to the interior of the concrete,
293
and those that penetrate and seal the capillary pore system of the concrete,
thereby preventing ingress of corrosive elements and immobilizing any that
may already be present.
Waterproof membranes, protected from damage by asphaltic concrete
wearing surfaces, presently constitute the most widely used method of retarding corrosion of reinforcing steel in concrete bridge decks [38]. The earlier
waterproofing materials were applied as liquids and were usually of the
polyester resin or epoxy-modified coal tar variety. More recently, the
preformed sheet-type membrane has gained favor [43]. There are two problems associated with the use of waterproof membranes on bridge decks.
First, considerable care and attention must be focused on their installation to
assure proper sealing, especially at curbs and joints. Second, they constitute
only an intermediate time range solution since their serviceability is limited
to the life of the asphaltic concrete wearing surface (about 15 years) [37,38].
Thick (3.8 to 6.4 cm) overlays of impervious concrete constitute another
means of sealing structural concrete. Like waterproof membranes, overlays
are largely applicable to large, horizontal configurations and to date have
been used primarily on bridge decks. One technique calls for the use of a low
slump, low w/c ratio (0.32), high cement factor (486 kg/m 3) concrete mixture. The special construction techniques and equipment required to handle,
place, and finish this material were developed in Iowa where about 200
bridges have received overlays [37]. Excellent success in preventing bridge
deck damage due to rebut corrosion has been reported by the highway agencies in Iowa and Kansas, where low w/c ratio overlays have also been used
since the early sixties.
The other type of material used in thick layers of impervious concrete is
latex modified mortar. In this type of concrete mixture, capillary porosity, is
partially blocked with polymeric materials introduced to the concrete mixture as a water-styrene butadiene emulsion. The latex modified mortar
overlay has a relatively long and successful history in protection of bridge
decks [38]. However, it is expensive in comparison to the other techniques of
comparable service performance (low w/c ratio overlays and waterproof
membranes).
Low w/c ratio ("Iowa Mix") and latex modified mortar were found in tests
to substantially reduce (but not totally prevent) intrusion of chloride ions
[11].
Sealing of the structural concrete itself may be accomplished by taking
steps to reduce significantly the inherent capillary porosity of the concrete, by
incorporating materials in the concrete mixture to block the capillary pores,
or by impregnating the capillary pores of the hardened concrete with a
suitably inert material. The first of these techniques invokes the principle of
the "Iowa" overlay concrete mixture, (namely, low w/c ratio. The "Iowa"
mix lacks the workability needed to be useful in reinforced concrete
members. However, the recent appearance on the market of a class of conCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:41:47 EDT 2014
294
crete admixtures termed "super water reducers" or "super plasticizers" permits the formulation of workable concrete mixtures having w/c ratios of 0.30
or lower. No data are available on the chloride transmission rates of such
concretes, but, presumably, permeabilities to water and ions would be appreciably reduced at this level of w/c ratio [44].
The aforementioned second procedure is similar in principle to the latex
modified mortar overlay material, but is sufficiently less expensive to warrant its use in structural concrete. It is called "internally sealed" concrete and
consists of incorPorating about 3 weight percent of tiny (average 0.5 mm
diameter) wax spheres in the concrete mixture. The wax, which is a 25/75
blend of montan and paraffin, melts at about 85~ Thus, after the concrete has hardened, it is heat treated causing the wax to melt and flow into
the capillary pores where it solidifies upon cooling, blocking the pores. Tests
have shown that internally sealed concrete is highly impermeable to water
and salt and is capable of maintaining low corrosion potentials on encased
reinforcing steel when the concrete is subject to ponding with salt water [45].
The third method, impregnation of hardened concrete with inert
materials, also acts to block admission of water and chlorides to capillary
pores. In addition, if impregnation is accomplished to sufficient depth to encompass the reinforcing steel, chloride ions already present are immobilized.
Polymer impregnation involves removing the evaporable water from the concrete to the desired depth of impregnation, introducing a monomeric
material into the concrete by soaking or application of moderate pressure,
and polymerization of the monomer by the application of heat or gamma
radiation. The monomer most commonly used in the rather extensive
research carried out in this area since the late sixties is methyl methacrylate
[46]. In addition to phenomenal improvement in the mechanical properties
of the concrete, polymer impregnation has been shown to reduce intrusion by
water and chlorides to exceedingly low levels [10,11]. Field application of
polymer impregnation, however, has been limited pending development of
appropriate hardware and technique. Also, the economies of this method remain to be demonstrated in the field.
The potential application of lower cost impregnants, such as sulfur [37], is
currently under study, but development still lags well behind the polymers.
Reinforcing Bar Coatings

In terms of their function in retarding or preventing corrosion, reinforcing
bar coatings may be divided into two classifications--barrier coatings and
anodic coatings. Coatings of the barrier type interfere with the corrosion process by preventing the corrosive elements from coming into contact with the
steel. Organic materials of various types constitute the most extensively used
group of barrier coatings. Epoxies and chlorinated rubber are reported to be
suitable [47, 48], but one study indicated that only certain powdered epoxies
295
applied by electrostatic spray techniques provide the proper mechanical and

chemical properties to withstand fabrication and handling and possess sufficient protective qualities [49].
Those metals that are more noble than iron in the electrical potential scale
(cathodic) can be used as barrier coatings provided that they do not corrode
in the environment presented by salt contaminated concrete. This includes
such metals as nickel and copper. However, their use is not recommended
because very severe localized corrosion may occur at any breaks in the
coating due to the large cathode/anode areal ratio that exists under such circumstances.
Anodic coatings consist of those metals that are less noble than iron. They
provide protection primarily as barriers, but also through sacrificial action
where the coating presents a large anode-to-cathode ratio relative to breaks
in the coating. The most commonly used anodic coating is hot-dipped zinc
(galvanizing). The success achieved in preventing corrosion of reinforcing
steel in concrete by galvanizing is mixed. Zinc itself is not resistant to corrosion in the presence of chloride and hydroxyl ions, although the corrosion
products formed do not result in volume changes believed sufficient to cause
disruption. This has led to some concern regarding the long term durability
of galvanized rebars [38]. In spite of the fact that cases of satisfactory performance of galvanized reinforcement in concrete for 30 years or more have
been documented [5, 37], the Federal Highway Administration, after analysis
of data from many sources, in 1976 placed a moratorium on its use in bridge
decks until adequate time has elapsed to evaluate the performance of the
over 200 existing installations [50].
Cadmium, like zinc, has shown mixed results as an anodic coating for
reinforcing steel [5,47]. Besides, it is expensive and toxic [40].
Cathodic Protection
Cathodic protection prevents corrosion of steel by supplying current flow
that suppresses the galvanic corrosion cell through polarization of the steel
surface. This can be accomplished by applying direct electrical current or by
using sacrificial anodes. Sacrificial anodes generally are not considered to be
practical for the protection of steel in concrete because voltages and current
densities cannot be controlled [51]. Direct current cathodic protection of
reinforcing steel in concrete has produced mixed results [5,40,52,]. The
main problem involves controlling potentials and current densities over the
entire surface of the reinforcing steel. Excessive voltages will result in softening of the concrete in the vicinity of the steel and resulting reduction of bond
strength. However, these problems reportedly have been largely overcome by
using a conductive asphaltic concrete containing coke breeze to cover the surface of the concrete. This layer acts as a large anode to distribute uniformly
protective current to the reinforcement. Tests have shown that this technique
296
renders the reinforcement passive, even in chloride contaminated concrete

[51]. However, more time and experience are necessary to evaluate fully the
attributes of this method.
Summary of Protective Methods

The current status of potential corrosion preventive procedures, along with
pertinent selected references for each, is presented in Table 2.
References
[1] Dallaire, G., "Designing Bridge Decks that Won't Deteriorate," Civil Engineering,
Vo. 43, No. 8, Aug. 1973, pp. 43-48.
[2] Godfrey, K. A., Jr., "Bridge Decks," Civil Engineering, Vol. 45, No. 8, Aug. 1975,
pp. 60-65.
[3] Sandvig, L. D., "PennDOT Attacks Bridge Deck Durability Problem," Highway Builder,
Nov. 1974, pp. 16-17, 42.
[4] Carrier, R. E. and Cady, P. D., "Deterioration of 249 Bridge Decks," Highway Research
Record, No. 423, Highway Research Board, 1973, pp. 46-55.
[5] Moore, D. G., Klodt, D. T., and Hensen, R. J., "Protection of Steel in Prestressed
Concrete Bridges," Report 90, National Cooperative Highway Research Program.
[6] Erlin, B. and Verbeck, G. J., "Corrosion of Metals in Concrete--Needed Research,"
Corrosion of Metals in Concrete, SP-49, American Concrete Institute, 1975, pp. 39-46.
[7] Lewis, D. A. and Copenhagen, W. J., "Corrosion of Reinforcing Steel in Concrete in
Marine Atmospheres," Corrosion, Vol. 15, No. 7, July 1959, pp. 382t-388t.
[8] Hausmann, D. A., "Steel Corrosion in Concrete," Materials Protection, Vol. 6, No. 11,
Nov. 1967, pp. 19-23.
[9] Verbeck, G. J., "Mechanisms of Corrosion of Steel in Concrete," Corrosion of Metals in
Concrete, 5P-49, American Concrete Institute, 1975, pp. 21-38.
[10] Clear, K. C., "Reinforcing Bar Corrosion in Concrete: Effect of Special Treatments,"
Corrosion of Metals in Concrete, SP-49, American Concrete Institute, 1975, pp. 71-82.
[11] Clear, K. C., "Time-To-Corrosion of Reinforcing Steel in Concrete Slabs," Report No.
FHWA-RD-76-70, Federal Highway Administration, April 1976.
[12] Stratfull, R. F., Jurkovich, W. J., and SpeUman, D. L., "Corrosion Testing of Bridge
Decks," Transportation Research Record, No. 539, Transportation Research Board,
1975, pp. 50-59.
[13] Portland Cement Association, "Durability of Bridge Decks," Final Report, A Cooperative
Study by the Portland Cement Association and The Bureau of Public Roads in cooperation with 10 State Highway Agencies, 1970.
[14] Straffull, R. F. "The Corrosion of Steel in a Reinforced Concrete Bridge," Corrosion,
Vol. 13, No. 3, March 1957, pp. 173t-178t.
[15] Szilard, R. and Oddmund, W., "Effectiveness of Concrete Cover in Corrosion Protection
of Reinforcing Steel," Corrosion of Metals in Concrete, SP-49, American Concrete Institute, 1975, pp. 47-68.
[16] Bureau of Reclamation, U.S. Department of the Interior, Concrete Manual, 8th ed.,
Superintendent of Documents, Washington, 1975.
[17] Christensen, K. A. and Williamson, R. B., "Solving the Galvanic Cell Problem in FerroCement," Technical Report No. 2, Project NR032 522, Office of Naval Research, July
1971.
[18] Powers, T~ C., "Structure and Physical Properties of Hardened Portland Cement Paste,"
Journal, American Ceramic Society, Vol. 41, 1958, pp. 1-6.
[19] Spellman, D. L. and Straffull, R. F., "Chlorides and Bridge Deck Deterioration,"
Highway Research Record, No. 328, Highway Research Board, 1970. pp. 38-49.
[20] Berman, H. A. and Chaiken, B., "Techniques for Retarding the Penetration of Deicers
Into Cement Paste and Mortar," Public Roads, Vol. 38, No. 1, June 1974, 15p. 9-18.
1. Chloride removal
Electrochemical + ion exchange
2. Inhibitors (admixtures)
3. Seal concrete
(a) surface
(i) membranes
(ii) overlays
low w/c ratio
latex modified mortar
(b) low w/c ratio concrete
super water reducers
(c) internally sealed concrete
(d) impregnated concrete
(i) polymers
(ii) sulfur
4. Reinforcing bar coatings
(a) barrier--type
(b) anodic--type
5. Cathodic protection
impressed current
Method
X
X
X
X
X
X
X
X
10, 37, 38
44
37, 38, 45
10, 11, 37, 46

37
37, 40, 47, 48, 49

5, 37, 38, 40, 47, 50
5, 38, 40, 51, 52
X
X
X
X
10, 11, 37
Promising
37, 38, 43
PostConstruction
Construction
37, 38, 39
9, 26, 37, 40.41, 42
References
Apply
TABLE 2--Status of potential corrosion preventive procedures.
Doubtful
X
X
Field
Evaluation
Potential Needs
X
X
Research/
Development
r
..,4
I"11
I'--
u)
-n
m
z
-11
O
--n
O
O
o')
O
z
O
3o
20
O
z
(3
>.
-<
298
[21] Bukovatz, J. E., "Weathering Test of Reinforced Concrete Slab With Various Depths
of Steel," Bridge Deterioration Study, Part 7, Kansas State Highway Commission,
Topeka, 1968.
[22] Steinour, H. H., "Influence of the Cement on the Corrosion Behavior of Steel in Concrete,"
Bulletin 168, Portland Cement Association, Skokie, Ill., 1964.
[23] Committee MC-BS, Section 4, "Pozzolans in Highway Concrete," Admixtures in Concrete,
Transportation Research Board, Special Report 119, 1971, pp. 21-28.
[24] Spellman, D. L. and Stratfull, R. F., "Concrete Variables and Corrosion Testing,"
Highway Research Record. No. 423, Highway Research Board, 1973, pp. 27-45.
[25] Committee MC-BS, Section 1, "Accelerators in Highway Concrete," Admixtures in
Concrete, Transportation Research Board, Special Report 119, 1971, pp. 3-6.
[26] Gouda, V. K. and Monfore, G. E., "A Rapid Method for Studying Corrosion Inhibition
of Steel in Concrete," Journal, Portland Cement Association Research and Development
Laboratories, Vol. 7, No. 3, Sept. 1965, pp. 24-31.
[27] Stark, D., "Studies of the Relationships Among Crack Patterns, Cover over Reinforcing
Steel, and Development of Surface Spalls in Bridge Decks," Special Report 116, Transportation Research Board, 1971, pp. 13-21.
[28] Clear, K. C., "Time to Corrosion of Reinforcing Steel in Concrete Slabs," Transportation
Research Record, No. 500, Transportation Research Board, 1974, pp. 16-24.
[29] Weed, R. M., "Recommended Depth of Cover for Bridge Deck Steel," Transportation
Research Record, No. 500, Transportation Research Board, 1974, pp. 32-35,
[30] Dakhil, F. H., Cady, P. D., and Carrier, R. E., "Cracking of Fresh Concrete as Related
to Reinforcement," Journal, American Concrete Institute, Vol. 72, No. 8, Aug. 197.5,
pp. 421-428.
[311 Committee 544, "State-of-the-Art Report on Fiber Reinforced Concrete," Journal.
American Concrete Institute, Vol. 70, No. 11, Nov. 1973, pp. 729-744.
[32] Cornet, I., "Corrosion of Prestressed Concrete Tanks," Materials Protection, Vol. 3,
No. 1, Jan. 1964, pp. 90-100.
[33] Clear, K. C., "Evaluation of Portland Cement Concrete for Permanent Bridge Deck
Repair," Report No. FHWA-RD-74-5, Federal Highway Administration, Feb. 1974.
[34] Berman, H. A., "Determination of Chloride in Hardened Portland Cement Paste, Mortar,
and Concrete," Journal of Materials, Vol. 7, No. 3, Sept. 1972, pp. 330-335.
[35] Morrison, G. L., Virmani, Y. P., Ramamurti, K., and Gilliland, W. J., "Rapid In Situ
Determination of Chloride Ion in Portland Cement Concrete Bridge Decks," Report
No. FHWA-Ks-RD. 75-2, Kansas Dept. of Transportation, Topeka, Nov. 1976.
[36] Stratfull, R. F., "Half-Cell Potentials and the Corrosion of Steel in Concrete," Highway
Research Record, No. 433, Highway Research Board, 1973, pp. 12-21.
[37] Manning, D. G. and Ryell, J., "Durable Bridge Decks," Report RR 203, Ontario Ministry
of Transportation and Communications, April 1976.
[38] Hay, R. E., "The Bridge Deck Problem--An Analysis of Potential Solutions," Public
Roads. Vol. 39, No. 4, March 1976, pp. 142-147.
[39] Morrison, G. L., Virmani, Y. P., Stratton, F. W., and Gilliland, W. J., "Chloride
Removal and Monomer Impregnation of Bridge Deck Concrete by Electro-Osmosis,"
Report No. FHWA-Ks-RD. 74-1, Kansas Department of Transportation, Topeka,
April 1976.
[40] Gjr
O. E., "Control of Steel Corrosion in Concrete Sea Structures," Corrosion of
Metals in Concrete, SP-49, American Concrete Institute, 1975, pp. 1-9.
[41] Griffin, D. F., "Corrosion Inhibitors for Reinforced Concrete," Corrosion of Metals in
Concrete, 5P-49, American Concrete Institute, 1975, pp. 95-102.
[42] Craig, R. J. and Wood, L. E., "'Effectiveness of Corrosion Inhibitors and Their Influence
on The Physical Properties of Portland Cement Mortars," Highway Research Record,
[43] VanTil, C. J., Carr, B. J. and Vallerga, B. A., "Waterproof Membranes for Protection of
Concrete Bridge Decks," NCHRP Report 165, Transportation Research Board, 1976.
[44] Anonymous, "Superplasticizing Admixtures in Concrete," Cement and Concrete Association, Great Britain, 1976.
[45] Jenkins, G. H. and Butler, J. M., "Internally Sealed Concrete," Report No. FHWARD-75-20, Federal Highway Administration, Jan. 1975.
299
[46] DePuy, G. W. and Kukacka, L. E., "Concrete-Polymer Materials," Fifth Topical Report,
Brookhaven National Laboratory, Dec. 1973.
[47] Backstrom, T. E., "Use of Coatings on Steel Embedded in Concrete," Corrosion of Metals
in Concrete, SP-49, American Concrete Institute, 1975, pp. 103-113.
[48] Clifton, J. R., Beeghly, H. F., and Mathey, R. G., "Protecting Reinforcing Bars from Corrosion with Epoxy Coatings," Corrosion of Metals in Concrete, SP-49, American Concrete
Institute, 1975, pp. 115-132.
[49] Pike, R. G., Hay, R. E., Clifton, J. R., Beeghly, H. F., and Mathey, R. G., "Nonmetallic
Protective Coatings for Concrete Reinforcing Steel," Transportation Research Record, No.
S00, Transportation Research Board, 1974, pp. 36-44.
[50] Federal Highway Administration, "Use of Galvanized Rebars in Bridge Decks," FHWA
Notice N 5140.10, 9 July, 1976.
[51] Stratfull, R. F., "Experimental Cathodic Protection of a Bridge Deck," Transportation
Research Record, No. 500, Transportation Research Board, 1974, pp. 1-15.
[52] Robinson, R. C., "Cathodic Protection of Steel in Concrete," Corrosion of Metals in Concrete, SP-49, American Concrete Institute, 1975, pp. 83-93.
Bernard Erlin ~and H u b e r t Woods 2
Chapter 20--Corrosion of Embedded

Materials Other than Reinforcing
Steel
Introduction
This chapter deals with materials, other than conventional reinforcing
steel, which are sometimes used in conjunction with concrete. As the title
of this chapter implies, emphasis will be given to the possible degradable
aspects relative to their use, and conditions that may render them serviceable or unserviceable.
The materials to be described include metals, and inorganic and organic
substances. Among the metals are aluminum, lead, copper and copper
alloys, zinc, special alloys of iron, Monel metal, stellite, silver, and tin;
among the inorganic materials are glass, asbestos, and concrete; and among
the organic materials are a variety of plastics, and wood and similar cellulosic materials. Fiber-reinforced concrete is gaining in use, so information about some fibers has been included.
A general condition necessary for the chemical degradation of any
materials in concrete is exposure to moisture, and a short discussion on
that aspect precedes the discussions of material behavior.
General Condition
Moisture is usually necessary for the chemical degradation of any material.
That moisture can be water vapor, liquid water, or solutions of which
water is a component. Any concrete is never truly dry because it contains
air and that air will hold moisture in vapor form. For example, because
of the chemical nature of the cement paste, and the fact that most concrete contains some unhydrated cement, air in concrete will contain sufIPresident.petrographer, Erlin, Hime Associates, Inc., Northbrook, Ill. 60062.
2Consultant, Jalisco, Mexico.
Universidade
do Gois pursuant
Agreement.
authorized.
Www.astIII.OFg
Copyright9
1978tobyLicense
ASTM
International
ERLIN
AND WOODS
ON CORROSION
OF
EMBEDDED
MATERIALS
301
ficient moisture to maintain a relative humidity of about 80 percent, the

vapor pressure for the paste at equilibrium.
Furthermore, although a concrete may be relatively dry, carbonation of
the portland cement paste will generate moisture due to the release of
hydrate water when carbonation products are formed, such as the following
CaSiO2 9 x H 2 0 + CO2 -- CaCO3 + SiO2 + x H 2 0
Concrete is wet when mixed and for some time after setting, depending
principally on the quality of the concrete and the distance through which
free water must move to an external surface where it can evaporate. The
rate of drying also depends on external humidity, but even at low external
humidity a long drying time may be required to lower the internal humidity
to a point where corrosion is unlikely. Thus, concrete which will eventually
be dry may be internally wet long enough to promote serious corrosion of
susceptible metals, and, of course, concrete exposed continuously or
frequently to a damp environment may remain wet enough to support any
corrosion that may occur. The free moisture in wet concrete provides an
aqueous medium which facilitates transport of soluble chemical substances,
such as oxygen, calcium hydroxide, alkalies, and chlorides toward metal,
and of any soluble corrosion products away from the metal. It also increases the electrical conductivity of the concrete, thus aiding any tendency
for electrochemical corrosion.
The time required for concrete to dry to various internal humidity levels
is not commonly appreciated. The rates of drying of 15 by 90 by 90 cm
(6 by 36 by 36 in.) slabs of normal weight concrete are shown in Table 1
ll]. 3
This table shows that at an environmental relative humidity of 50 percent, 36 days drying time was required to lower the middepth relative
humidity to 90 percent, and 240 days drying time was required to reach 75
TABLE 1--Rates of drying 15 by 90 by 90-cm (6 by 36 by 36-in.)
slabs of normal weight concrete, a
Drying Time to Reach Various Relative

Humidities in the Concrete Slab at Middepth, days
Environmental
RH
10
35
50
75
a
90% RH
75% RH
50% RH
18
30
36
36
80
110
240
. .
620
840
...
.
From Ref 1.
302
percent. Additional tests showed that when the middepth relative humidity
reached 90 percent, the relative humidity at a point only 1.9 cm ( 90 in.)
from an exposed face was 87 percent. The moisture condition of concrete
required to support active electrochemical corrosion of aluminum and
other susceptible metals is not known with any accuracy, but it seems almost certain that such corrosion could proceed, other conditions being
favorable, at 90 percent relative humidity and probably at 75 percent
relative humidity, though at a much lower rate.
Concrete made with lightweight aggregates dries more slowly than that
made with normal weight aggregates, and thicker sections, such as beams
and columns, will dry more slowly than thin sections, such as walls, or
floor slabs not on the ground.
Materials in " d r y " concrete are relatively immune to chemical degradation; however, prolonged periods, as long as 50 years or more, may be
required before outward manifestations may become noticeable, if at all.
Aluminum
Aluminum has been given widespread attention because it has been the
cause of numerous problems in the past, which prompted laboratory
studies designed to describe conditions necessary for corrosion, and
methods for circumventing the corrosion.
Aluminum reacts in fresh concrete principally with alkali hydroxides
derived from the cement. One reaction product is hydrogen gas, and for
this reason aluminum powder sometimes is used in concrete to form extremely lightweight cellular concrete. In smaller amounts the powder provides a slight expansion of grout such as that used for bedding machinery
base plates. Aluminum in rod, sheet, or pipe form will react much less
vigorously than will the powdered metal, because of its lesser surface area
exposed to the alkalies.
Tests carried out by Jones and Tarleton [2] indicate that the corrosion of
aluminum embedded in plain concrete can crack the concrete under unfavorable circumstances. However, it has been shown that the situation can
be worse if the concrete contains calcium chloride and much worse if it also
contains steel, such as reinforcing steel which is connected (coupled) to the
aluminum.
Wright [3] has described a case of corrosion of sufficient severity to
cause collapse of aluminum conduit in reinforced concrete containing
calcium chloride, and an instance of extensive concrete spalling over
aluminum conduit in Washington Stadium has been published [4]. Corrosion of aluminum balusters embedded in concrete have been reported [5].
Several dozen additional cases of concrete cracking due to corrosion of
embedded aluminum conduit and posts, and window frames in contact
with the concrete have occurred. In every case chloride was present in
ERLIN AND WOODS
ON CORROSION
OF
EMBEDDED
MATERIALS
303
amounts of at least about 1A weight percent of cement. An example of concrete spalls over corroded aluminum conduit is shown in Fig. 1.
FIG. 1--Concrete spalled by corroding aluminum conduit (from Ref S).
The laboratory investigations by Monfore and Ost [6] show that calcium
chloride concentration, alkali content of cement, coupling of steel and
aluminum, and the ratio of steel area to aluminum area are all interrelated
in the corrosion of the aluminum and the subsequent cracking of concrete.
In their studies, 15-cm (6-in.) concrete cubes were prepared using cements
of high and low alkali contents, containing pieces of nominally t.25-cm
(1/2-in.) aluminum conduit embedded 1.25 cm (1/2-in.) from one face.
C-shaped sheets of mild steel were also embedded in the cubes, externally
connected for some tests and not connected for others. Various amounts of
calcium chloride were used. The cubes were removed from their molds at
Cement
Alkalies
as Na20,
%
0.24
Cement
Ratio of
Steel Area
to Aluminum
Area
28
14
28
7
14
28
28
14
28
7
14
28
0
0
CaCI2-2H20,
%
0
2
2
4
4
4
0
2
2
4
4
4
0
4
...
...
uncoupled
coupled
coupled
coupled
coupled
coupled
uncoupled
uncoupled
uncoupled
uncoupled
uncoupled
uncoupled
Electrodes
no crack
no crack
no crack
5
4
3
3
3
no crack
no crack
no crack
no crack
no crack
no crack
Days to
Cracking
0.09
0.04
0.16
0.92
1.2
1.6
2.3
2.4
0.09
0.07
0.10
0.07
0.07
0.04
Loss in
Surface
Thickness,
mils b
TABLE 2--Corrosion o f 6063 aluminum conduit embedded in 15-cm (6-in.) concrete cubes stored at 50 percent relative humidity f o r 28 days. a
=0
I'rl
---I
m
O
O
co
m
"rl
--4
--r
m
60
-4
r
),
Z
E~
"tJ
Go
0
Cj0~
97VlEI31VI~
O3OO:181~3
:IO
NOISOEIEIOO NO SOOOM ONV NI7EI3
~~o~888oo8~8o88
x
K
o~"
~o
II
-"
306
24 h, coated with a curing compound, and then stored at 23~ (73~

and S0 percent relative humidity for 28 days. They were observed regularly
for cracks. After 28 days the aluminum pieces were removed, cleaned, and
weighed. The principal results are given in Table 2.
Several important findings can be noted in these results.
1. All cubes that cracked contained calcium chloride.
2. The cubes containing 1 percent flake calcium chloride (CaCI2.2H20,
by weight of cement) did not crack, but metal losses in these cubes were as
high as, or higher than, those in other cubes which cracked. Exposure to a
damper atmosphere might have brought about sufficiently more corrosion
to have caused cracking.
3. With no calcium chloride and no coupling, the corrosion was greatest
with the high alkali cement.
4. With 2 or 4 percent calcium chloride present, and with metals coupled,
corrosion was a little greater with the higher alkali cement.
S. With calcium chloride present, and with metals coupled, corrosion
generally increased with the increasing ratio of steel area to aluminum
area, and invariably increased as the amount of calcium chloride increased.
6. The cubes that cracked did so within 7 days.
In the case of coupled metals, considerable galvanic currents were found
to flow in the circuit connecting the aluminum and steel.
Some measured currents as a function of time and calcium chloride
content are shown in Fig. 2. The effect of progressively increased calcium
chloride contents in increasing the current flow at all periods (up to end of
test at 28 days) is evident. When the total electrical flow during 28 days (in
ampere hours per unit area of aluminum) was plotted against the amount
of corrosion as loss in thickness, straight lines were obtained with slopes
I
~120
Steel oreo/oluminum oreo = 28

Cement D, 0.89~ olkolies
j\
80
LIJ
e:" ~_
a~ ~.
!~,
~ 40
"~ \,,
colciurn chloride 4%
,
9 2%
0
0
I0
20
50
TIME, DAYS
FIG.
2--Effect of calcium chloride on galvanic currents (from R e f 5).
ERLIN
AND
WOODS
ON
CORROSION
OF
EMBEDDED
MATERIALS
307
depending on the amount of calcium chloride used, the slope being greater
for the higher amounts of calcium chloride.
Tests somewhat similar to those reported by Monfore and Ost were
carried out on 31-cm (12-in.) concrete cubes by Wright and Jenks [7].
With coupled steel and aluminum (area ratio 10:1), the cubes did not
crack when no calcium chloride was used, but did crack at various ages
from 61 days with 1.1 percent flake calcium chloride to 8 days with 5.7
percent flake calcium chloride.
McGeary [8] also found that: (a) for aluminum conduit coupled to steel,
cracks did not occur in the encasing concrete when chlorides were not
present; (b) the cause of corrosion of the aluminum was due unquestionably to galvanic cells activated when chlorides were present in the concrete;
(c) the corrosion of the aluminum was of the cubic type, which is also referred to as intergranular corrosion; and (d) there was an enrichment of
chloride at the aluminum conduit surface.
Tests [6] were made to determine the effectiveness of several different
coatings applied to the aluminum before embedment in 1.2S-cm (lA-in.)
concrete cubes made with cement of 0.89 percent alkalies and containing
4 percent calcium chloride by weight of cement. The results are shown in
Table 3. These show that a silicone coating was ineffective and that
Lacquer B prevented cracking within 28 days but permitted some corrosion. Lacquer C and bituminous coatings A and D were all effective in
preventing both corrosion and cracking.
Other tests [8] indicated that resistance to galvanic corrosion was provided by the following coatings: certain bitumuls; epoxies: fluidized-bed
plastics; certain metallic pigmented coatings; and alkyd and phenolic
materials.
TABLE 3--Effect of protective coatings on corrosion o f aluminum conduit embedded for 28
davs in 6-in. concrete cubes containing 4 percent calcium chloride and steel coupled to the
aluminum, a
Protective Coating
Material
Thickness, mils
Days to
Cracking
Loss in
Surface
Thickness,
milsb
None
Silicone
Lacquer B
Lacquer C
Bitumen A
Bitumen D
...
...
1
2
5
15
3
2
no crack
no crack
no crack
no crack
2.5
2.1
0.47
nil
nil
nil
Conversion factors--1 mil = 2.54 10-s m; 1 in. = 2.5 cm.
aFrom Ref 5.
bCalculated from weight losses.
308
McGeary [8] found the following to be effective repair procedures for

concrete cracked because of corroded aluminum: (a) remove all loose
concrete; (b) remove all corrosion products adjacent to the embedded
aluminum surface; (c) patch with epoxy or epoxy polysulfide systems,
or both; and (d) engage an experienced patching specialist to complete
the repairs.
The results of these various investigations and field observations show
that reinforced concrete is likely to crack and spaU from corrosion of uninsulated aluminum embedded therein if an appreciable amount of calcium
chloride is present. Insulating coatings for aluminum to be used in concrete are commercially practical.
In view of chemical similarities between calcium chloride and sodium
chloride, it seems evident that the latter would also facilitate corrosion of
aluminum. Sodium chloride is the principal constituent of sea salt, and it
therefore seems prudent not to use aluminum in concrete in or near seawater. The role of the chloride appears to be that of an electrolyte, a
necessary part of the galvanic cell. Obviously, other salts which hydrolize
and provide a strong electrolyte can be damaging also.
Calcium chloride should not be used in concrete in which aluminum will
be used.
Lead
Lead has a high resistance to certain chemical actions, but in contact
with d a m p concrete it is attacked by the calcium hydroxide in the concrete
and becomes converted to lead oxide or to a mixture of lead oxides. If the
dampness persists, the attack will continue, and a lead pipe, for example,
may be destroyed in a few years. If the lead is coupled to reinforcing steel
in the concrete, galvanic cell action may accelerate the attack [9], in which
case the rate of corrosion may be several millimetres per year.
Lead partially embedded in concrete and thus partially exposed to the
air is susceptible to corrosion because of the differential electrical potential
that results. The air-exposed portion in the presence of water will become
cathodic with respect to the embedded portion which will become anodic,
and a galvanic cell is formed. Corrosion and deterioration of the embedded
lead will result.
A protective coating or covering always should be used when lead pipe or
cable sheaths are to be embedded in concrete. Bituminous coatings have
been used successfully. Synthetic plastic and other organic coatings or
sleeves, which are themselves unaffected by damp concrete, are suggested.
There appears little, if any, likelihood of concrete itself being damaged
by corrosion of lead because of the softness of the metal.
ERLIN
AND WOODS
ON CORROSION
OF EMBEDDED
MATERIALS
309
Copper and Copper Alloys

Copper will not corrode in concrete unless soluble chlorides are present.
Copper pipes are used successfully in concrete except under the unusual
circumstance where ammonia is present [15]. Very small amounts of ammonia and possibly nitrates can cause stress corrosion cracking. It is re,~
ported that brass wall ties have failed by stress corrosion and that manganese bronze bolts have sheared below their ultimate strength, but as such
phenomena can occur under circumstances unrelated to concrete, it is not
clear what role, if any, the concrete played in these cases. Very little
systematic work has been reported on the behavior of copper and its alloys
in or in contact with concrete, probably because these metals have given
satisfactory service under such conditions.
Copper, brass, red brass, bronzes, aluminum-bronze, and coppersilicon alloys have good resistance to corrosion in concrete [11].
When copper is connected to steel reinforcement, or they are close to
each other, and an electrolyte such as chloride is present, corrosion of the
steel due to galvanic action is likely to occur. Thus it is desirable to insulate
the copper with suitable coatings.
Zinc
Zinc reacts chemically with alkaline materials, but normally in concrete the reaction is superficial and may be beneficial to bond of zinccoated (galvanized) steel. The primary chemical reaction with calcium
hydroxide is
Zn + Ca (OH)2 -- CaZnO2 + H2
(1)
When zinc is used in concrete it is practically always as a coating for st'eel.

The products of reaction are not voluminous and, consequently, damaging
stresses are not created.
Although good concrete normally provides a nearly ideal environment for
protecting embedded steel from corrosion, galvanized steel reinforcement has been used in concrete exposed in marine and other environments
where chlorides are present (for example, bridge decks). Galvanized reinforcing steel has not always been successful in stopping corrosion of the
steel and thus conflicting reports result [12-14].
Galvanized corrugated steel sheets often are used as permanent bottom
forms for concrete roof or floor construction. Both satisfactory and unsatisfactory performances have been reported.
Figure 3 shows one instance of multiple perforation and corrosion of
such sheets under a roof slab [15].
310
FIG. 3--Corrosion and perJoration of galvanized steel from under concrete slab (from
R e f 15).
Most of the corroded spots were dry to the touch, but some were moist
and had pH values of 2.7 to 4.8, definitely in the acid range. Analysis of
the corrosion protuberances indicated the presence of iron, zinc, and
chloride. Further investigation of this and other similar cases showed that
in each instance calcium chloride was used in the concrete and may be presumed to have abetted the corrosion by chemical action on the zinc and by
increasing the electrical conductivity of the concrete to corrosion currents.
The chemical reaction is probably
Zn + CaC12 + 2H20 -- CaZnO2 + 2HC1 + FI2
(2)
As this action produces hydrochloric acid, it may well explain the observed acidity of the corrosion product. It is probable that this reaction
could not take place at first but only after local depletion of calcium hydroxide by the reaction expressed in Eq 1.
Admixtures containing chloride as a major active component should not
be used in concrete that will be exposed to moisture and contains or is in
contact with galvanic steel. It is also advisable to keep chloride-containing
solutions from permeating the concrete.
Corrosion of zinc can result when contact with relatively fresh concrete
occurs, and the corrosion will result in pitting of the surface. Thus, zinccoated steel used, for example, in conjunction with precast or cast-inplace concrete may show surface disfiguration due to the contact of the
zinc with alkalies from the portland cement paste.
Passivation of the zinc by use of chromate dips has been reported to be
effective in protecting galvanized products. The dips are solutions of
sodium or potassium dichromate acidified with sulfuric acid.
ERLIN AND WOODS
ON CORROSION
OF
EMBEDDED
MATERIALS
311
Concrete, in which galvanized reinforcing steel is located close to galvanized forms, has a tendency to stick to the forms. A chromate treatment,
such as previously described, has been described as a satisfactory method
for avoiding that problem.
Other Metals
The following metals have been reported to have good resistance to corrosion in concrete: stainless steels, chrome-nickel steels, chromiumaluminum-silicon steels, cast silicon-iron, alloyed cast iron, nickel, chromenickel, iron-chrome-nickel alloys, Monel metal, stellite, silver, and tin
[11]. The resistance of some of these metals to corrosion in concrete may
be affected seriously by the presence of "corrosion promoters" such as
soluble chlorides. Monel metal and Type 316 stainless steel are wellknown for their resistance to sodium chloride and other constituents of
seawater and should work well in concrete. Special circumstances might
justify the use of these more costly metals.
Nickel and cadmium-coated steel will not corrode in chloride-free concrete if the coatings are continuous [16]. However, the corrosion resistance
of these materials becomes questionable if chlorides are present in the concrete or in solutions that permeate the concrete [17].
Glass
Glass sometimes is embedded in mortar or concrete as artificial aggregate [18, I9] used for decorative or aesthetic purposes, as reinforcing as
a substitute for steel, as wall blocks or tile, and as frameless windows or
lights. Some glasses are reactive with alkalies in portland cement paste and
the resulting expansion may cause severe damage to the glass or the concrete or both [20].
Whenever glass is to be used in concrete, it is mandatory that the glass
be tested to ensure that it will be chemically stable in concrete and not
cause expansions due to alkali-silica reactivity. For example, waste bottle
glass used as aggregate in a decorative concrete facing of outside-exposed
precast concrete panels caused sufficient expansions to warp the panels.
The warping was due to the expansion of the facing against the nonexpansive concrete backing. Prisms cut from the panels and tested using
accelerated methods in the laboratory had expansions of 0.2 percent after 3
months of test. The glass aggregate removed from the concrete, tested
using the procedures in the ASTM Test for Potential Reactivity of Aggregates (Chemical Method) (C 289) was found to be deleteriously reactive.
Other reports [20] have also indicated the deleterious behavior of reactive
glass.
Before glass is used in portland cement concrete, it should be tested
312
using the methods provided in the Appendix of ASTM Specification for

Concrete Aggregates (C 33), and specifically those given in Paragraph
X1.1.2 of ASTM Test C 289 and Paragraph X1.1.3 of the ASTM Test for
Potential Alkali Reactivity of Cement-Aggregate Combinations (MortarBar Method) (C 227).
Nondeleteriously reactive glass for use in concrete is manufactured and
thus available for use in concrete.
Wood
Current trends in the use of new or unusual materials in concrete are

due to the emphasis on conservation of energy and the utilization of wastes
and by-products. Among materials prepared for use in concrete is wood
(includes bamboo, fibers, bark, jute, cotton, and rice stalks and hulls).
Problems incidental to the use of natural-cellulosic materials have included adverse effects of sugars on concrete setting and degradation of the
fibers due to the high alkalinity of concrete. Further, high differential
thermal coefficients of expansion of some of these materials and unaccommodative volume changes can cause cracks to develop. These factors
are perhaps foremost in precluding the use of many cellulosic materials.
Additional problems incidental to the use of natural-cellulosic materials
include swelling with moisture absorption and subsequent shrinkage, and
chemical degradation due to contact with calcium hydroxide. Prior treatment of the cellulosic materials such as impregnation or coating techniques, or carbonation of the portland cement paste [21] are possible ways
for improving their utilization.
Sawdust, wood pulp, and wood fibers have been incorporated in mortars
and concretes, and timbers have been embedded in or placed in intimate
contact with concrete in composite constructions. The use of fresh untreated
sawdust, wood chips, or fibers in concrete commonly results in very slow
setting and abnormally low strength because of interference with normal
setting and hardening processes by carbohydrates, tannins, and possibly
other substances in the wood. The amount of such substances differs with
wood species and from time-to-time and place-of-origin within a single
species. Softwoods generally give less trouble in this respect than hardwoods.
Many admixtures and wood treatments have been proposed or used to
circumvent the influence of wood constituents on setting and hardening.
Addition of hydrated lime to the mixture in an amount equivalent to onethird to one-half of the cement by volume has been found effective in overcoming this action [22]. The treatment is usually effective with mixed softwoods, except when a high proportion of larch or Douglas fir is present.
ERLIN AND WOODS
ON CORROSION
OF EMBEDDED
MATERIALS
313
Calcium chloride, to the amount of 5 percent of the cement, is sometimes

added as well as hydrated lime. With woods of high tannin or carbohydrate
content, the addition of lime with or without calcium chloride is not effective. Other treatments which have been suggested include soaking in
sodium silicate solution, moistening the wood with 1 percent sulfuric acid
for 4 to 14 h, then neutralizing with milk of lime; treating with 37 percent
aluminum chloride solution or 50 percent zinc chloride solution in a rotary
barrel with beater. A treatment found by Parker [22] to be effective with
all woods tried,consists of the following consecutive steps: (a) boiling sawdust in water, (b) draining and washing with water, (c) reboiling with a
solution of ferrous sulfate (2 percent) in water, and (d) draining and rewashing.
Concrete made with wood aggregate has considerably greater volume
change on wetting and drying, or simply on change in external humidity,
than concrete made with mineral aggregates. If the element is restrained,
drying may lead to cracking. If drying is not uniform, the element may
warp. The pretreatments mentioned previously have only a small influence
on these volume changes. Various methods have been employed to lessen
the changes in volume consequent upon changes in moisture. Some of
these methods involve encasement of the wood particles or of the finished
product in a material of low permeability to moisture, but the details of
such treatments and the results achieved have not been revealed in general.
Timbers embedded in concrete sometimes have been observed to deteriorate. It is said that the harm is done by calcium hydroxide which
causes dissolution of lipins and decomposition, chiefly of pentosans, to a
smaller extent of lignin, and least of all of cellulose. The most suitable
wood for embedment is said to be pine or fir, preferably of a type with
high resinous content [23].
Plastics
The use of plastics in concrete and concrete construction has increased

significantly. Plastic products are now being used as pipes, conduit shields,
sheaths, chairs, waterstops, and joint fillers, and their compatibility
with concrete is thus important.
The principal chemicals in concrete which could conceivably attack
plastics are calcium hydroxide, sodium hydroxide, and potassium hydroxide. The following plastic groups have excellent resistance to all three
of these alkalies at 24~ (75~
[24]: polyethylene, styrene copolymer
rubber-resin blends poly(vinyl chlorides), Types I and II, and polytetrafluoroethylene.
314
Another source [25] provides the following information on the resistance

of plastics to strong alkalies:
Class of Material
Resistance
Polyethylene
Polymethyl methacrylate
Polypropylene
Polystyrene
Polystyrene acrylonifrile
Polytetrafluoroethylene
Polytrifluorochloroethylene
Poly(vinyl chloride) and poly(vinylchloride) vinyl acetate
(rigid)
Polyvinylchloride and polyvinylchloride vinyl acetate
(plasticized)
Saran (monofilament grade)
Epoxy (unfilled)
Melamine (formaldehyde)
Phenol (formaldehyde)
Polyester styrene-alkyd
Urea (formaldehyde)
excellent
poor
excellent to good
excellent
good to excellent
excellent
excellent
excellent
fair to good
fair to good
excellent
poor
poor
poor
poor
Fibers
In a broad sense, fibers used for reinforcing concrete are small versions
of conventional steel reinforcement, and they provide a similar service. For
fibers to be useful and effective, they must enhance the physical attributes
of concrete, and be durable. Among the desirable characteristics that
fibers can impart to concrete are flexural strength, resistance to fatigue
and impact, and increased fracture resistance. Applications of fiberreinforced concrete are shown in Table 4. A state-of-the-art report on
TABLE 4--Some applications of fibers in concrete.
Type of Use
Cast-in-place and precast bridge deck units;
overlays; structural bridge deck elements;
pavement; dolosse; boats; poles; tunnel linings;
rock slope stabilization; highway, street, and
airfield pavements; sluices; industrial floors
Maintenance and repairs to dams, slabs,
pavements, bridges, culverts, etc.
Industrial floors, slab overlays, pile caps,
pavements
Miscellaneous small precast items (burial vaults,
steps, garden units, etc.)
Pipes, sheets, boards, fence posts, panels, piles,
building panels
Type of Fiber
steel
steel
steel, glass
glass
asbestos, glass, cellulose,
polypropylene
ERLIN AND WOODS
ON CORROSION
OF
EMBEDDED
MATERIALS
315
fiber-reinforced concrete was prepared by the American Concrete Institute [26], but little is given regarding corrosion characteristics of embedded materials.
Among the materials that have been used as fibers in concrete are steel,
glass, polyethylene, polypropylene, nylon, asbestos, and carbon. Organic
fibers are considered to rely upon mechanical interlock, and steel and
glass upon chemical interactions, to develop adhesion to the cement
matrix. Steel fibers have also been produced as crimped or deformed
fibers so that mechanical interlock is present in addition to chemical bond.
Steel
Good durability of steel-fiber reinforced concrete has been reported
[27,28]. However, steel fibers are subject to the same type of corrosion as
reinforcing steel, and thus the durability of concrete made with steel fibers
and used in environments chemically aggressive to the steel may be poor.
The deterioration of the steel is enhanced particularly by chlorides. However, it has been reported that outside-exposed, steel-reinforced concrete
used for pavements, bridge decks, and other similar usages where chloride
deicing agents are employed, can suffer reductions in strength [29.30].
Rusting of fibers and the surface region of concrete generally occurs, and
the rusting is not progressive [31]. If cracks are present, rusting will be
initiated at locations where steel is exposed in the cracks. Obviously,
chlorides should not be a component of admixtures used in fiber-reinforced concrete, and the concrete should have as low a permeability as
workability and w/c ratio will permit.
The results of unpublished studies of steel-fiber reinforced beams in a
simulated seawater environment for 8 years have revealed that rusting of
fibers to depths of 0.34 cm (1/16 in.) have occurred [31].
Steel fibers and steel filings are used in proprietary floor toppings to enhance wear characteristics, such as is desirable in warehouses. The amount
of metal used is significantly greater than for fiber-reinforced concrete. Instances have occurred where deterioration of the surfaces have resulted because of rusting of the steel. In one such case, chloride in an amount
equivalent to a purposeful chloride addition (for example, 1 percent
calcium chloride by weight of cement) was present in the concrete to which
a topping shake had been applied. This situation points out the desirability
for avoiding the use of chloride either as a component of the original
concrete mix, or as an application to concrete surfaces.
Glass
Glass fibers can be sensitive to alkalies in portland cement paste (see
316
discussion under "Glass"), and early work using glass fibers reflected the
adverse effects of alkali-silica reactivity [32].
Glass fibers have been developed that are resistant to alkali degradation
and are now available [33,34]. These new fibers are formulated with zirconia and are particularly resistant to alkali attack, and concretes made
with these fibers are reported to retain their strength properties when
stored in a variety of environments for periods of about 5 years (the tests
are still in progress).
These tests are gratifying because they also demonstrate that the loss of
fracture resistance caused by chemical changes within glass-fiber reinforced concrete (which occurs prior to 5-year exposures in wet environments) may be as dramatic as previously experienced. The embrittlement
associated with that exposure is thought to result because of reformation of
cement hydration products in and tightly around the fibers so that fiber
encasement is enhanced and failures are due to fracture rather than pullout of the fibers [35]. That explanation does not entirely satisfy all aspects of the situation, however, and further research is needed [36,37].
Organic Materials
Polymer fibers such as nylon and polypropylene improve impact strength
of concrete but not tensile or flexural strength because they have a low
modulus of elasticity. Other polymer fibers having a high modulus of
elasticity are apparently becoming available. These fibers by themselves are
reported to perform satisfactorily in concrete; however, reformation of the
portland cement matrix may cause the concrete to become brittle with
age [38].
Asbestos
Asbestos has been used in conjunction with portland cement since about
1900. Asbestos minerals are naturally occurring and include a variety of
different materials that fall into two groups: amphibole and chrysotile. The
asbestos minerals may differ in composition but have the common composition of magnesium silicate, and alkalies, magnesium, and iron which
may substitute freely for each other.
Although the durability of the embedded asbestos fibers is not considered a problem, some corrosion of the fibers may occur after prolonged periods because of chemical reactions with calcium hydroxide in
the portland cement paste. The chemical reactions have been found to be
surfacial and occur on fiber surfaces and on cleavage planes. Corrosion
products are probably magnesium hydroxide, magnesium carbonate, magnesium silicate hydrates, and low lime calcium silicate hydrates [39]. The
alteration of the asbestos fibers does not adversely affect properties of the
concrete because the alteration of the fibers is restricted to only its surface.
ERLIN
AND
WOODS
ON
CORROSION
OF
EMBEDDED
MATERIALS
317
Concrete
The c u r r e n t e m p h a s i s on recycling of m a t e r i a l s , a n d the raising of old
concrete structures a n d pavements, has resulted in the use o f old concrete
as aggregates for new concrete. F o r such use, the old concrete is crushed
and g r a d e d , a n d used either alone or b l e n d e d with other aggregates [40-42].
T h e concrete m a d e with concrete aggregates has h a d p r i n c i p a l use as a
base for concrete m a d e with conventional aggregates.
A l t h o u g h the w o r k a b i l i t y of concrete m a d e with recycled aggregates is
a b o u t t h a t of concrete m a d e with conventional aggregates, its compressive
strength is a b o u t 75 percent, a n d its m o d u l u s of elasticity is a b o u t 60 percent, of t h a t for conventional concrete [43].
Buck r e p o r t e d t h a t concrete m a d e with recycled aggregate h a d resistance
to cyclic freezing [44]. Aside from t h a t work, nothing has been r e p o r t e d
r e g a r d i n g the physical or c h e m i c a l stability of such aggregates.
Studies of recycled concrete aggregate is thus m e a g e r a n d the subject
certainly merits c o n s i d e r a b l e a n d extensive work. F o r e x a m p l e , aggregates
for use in concrete m u s t possess certain necessary physical a n d chemical
characteristics, such as those r e q u i r e d in A S T M Specification C 33. Because the concrete has seen prior service in a given e x p o s u r e a n d environm e n t for lengthy periods does not necessarily m e a n t h a t as concrete aggregate in o t h e r service a n d exposures it will p e r f o r m similarly.
Concrete from an elevated structure where the exposure has been dry,
and where chemically u n s t a b l e aggregates are present, when used in a
moist e n v i r o n m e n t , will p r o b a b l y r e s p o n d differently. C o n c r e t e m a d e with
a chloride a d d i t i o n , if used in reinforced n o n c h l o r i d e - c o n t a i n i n g concrete,
could p r o m o t e c o r r o s i o n of reinforcing steel. These are b u t two examples
t h a t d e m o n s t r a t e the need for establishing new specifications a n d requirements to ensure t h a t concrete m a d e with recycled concrete will provide
a d e q u a t e service.
References
[1] Abrams, M. S. and Orals, D. L., "Concrete Drying Methods and Their Effect on Fire
Resistance," Portland Cement Association, Research and Development Bulletin 181,
1965.
[2] Jones, F. E. and Tarleton, R. D., "Effect of Embedding Aluminum and Aluminum
Alloys in Building Materials," Research Paper 36, Great Britain Department of
Scientific and Industrial Research, Building Research Station, London, 1963.
[3] Wright, T. E., "An Unusual Case of Corrosion of Aluminum Conduit in Concrete,"
Engineering Journal, Canada, Vol. 38, No. 10, Oct. 1955, pp. 1357-1362.
[4] Anonymous, "Spalled Concrete Traced to Conduit," Engineering News-Record, Vol.
172, No. 11, 12 March 1964, pp. 28-29.
[5] Copenhagen, W. J. and Costello, J. A., "Corrosion of Aluminum Alloy Balusters in a
Reinforced Concrete Bridge," Materials Protection and Performance, Vol. 9, No. 9,
Sept. 1970, pp. 31-34.
[6] Monfore, G. E. and Ost, Borje, "Corrosion of Aluminum Conduit in Concrete,"
Journal, Portland Cement Association, Research and Development Laboratories, Vol. 7,
No. 1, Jan. 1965, pp. 10-22.
318
[7] Wright, T. E. and Jenks, I. H., "Galvanic Corrosion in Concrete Containing Calcium
Chloride," Proceedings, American Society of Corrosion Engineers, Journal Structural
Division, Vol. 89 (ST 5), Oct. 1963, pp. 117-132.
[8] McGeary, F. L., "Performance of Aluminum in Concrete Containing Chlorides,"
Journal, American Concrete Institute; Proceedings, Vol. 63, Feb. 1966, pp. 247-264.
[9] Anonymous, "Corrosion of Lead by Cement," Concrete Construction Engineering,
Vol. 44, No. 11, 1949, p. 348. Dodero, "Accelerating Reactions of the Corrosion of
Lead in Cement," Metaux & Corrosion, Vol. 25 (282), 1949, pp. 50-56.
[I0] Halstead, P. E., "Corrosion of Metals in Buildings. The Corrosion of Metals in Contact
with Concrete," Reprint No. 38, Great Britain Cement and Concrete Association. Reprinted from Chemistry and Industry, 24 Aug. 1957, pp. 1132-1137.
[11] Rabald, E., Corrosion Guide, Elsevier Publishing Co., Inc., New York, N.Y., 1951.
[12] Stark, David and Perenchio, William, "The Performance of Galvanized Reinforcement
in Concrete Bridge Decks," Portland Cement Association, Oct. 1975.
[13] Griffin, D. F., "Effectiveness of Zinc Coatings on Reinforcing Steel in Concrete Exposed to a Marine Environment," Technical Note N-1032, U.S. Naval Civil Engineering
Laboratory, Port Hueneme, July 1969, June 1970, and June 1971.
[14] Anonymous, "Use of Galvanized Rebars in Bridge Decks," Notice, No. 5, 140.10,
Federal Highway Administration, Washington, D.C., 9 July 1976.
[15] Mange, C. E., "Corrosion of Galvanized Steel in Contact with Concrete Containing
Calcium Chloride," Preprint, National Association of Corrosion Engineers, 13th Annual
Conference, St. Louis, Mo., t957.
[16] Freedman, Sidney, "Corrosion of Nonferrous Metals in Contact with Concrete," Modern
Concrete, Vol. 36, Feb. 1970.
[17] ACI Committee 201, "Guide to Durable Concrete," Journal. American Concrete Institute, Dec. 1977, pp. 591-592.
[18] Anonymous, "Waste Materials in Concrete," Concrete Construction, Vol. 16, No. 9,
Sept. 1971, pp. 372-376.
[19] Johnson, C. D., "Waste Glass as Coarse Aggregate for Concrete," Journal of Testing
and Evaluation, Vol. 2, No. S, 1974, pp. 344-350.
[20] Waters, E. H., "Attack on Glass Wall Tiles by Portland Cement," Report S-41,
Australia Commonwealth Scientific and Industrial Research Organization, Division of
Building Research, 1956.
[21] Anonymous, "Fiber Reinforced Concrete," Centre de Recherehes de Pont-a-Morisson,
French Patent No. 1,369,415, 14 Aug. 1964.
[22] Parker, T. W., "Sawdust-Cement and Other Sawdust Building Products," Chemistry
and Industry, 1947, pp. 593-596.
[23] Dominik, W. and Hans, M., "Action of Cement on Wood," Przemsyl Cheniczny,
Vol. 22, 1938, pp. 74-82.
[24] Seymour, R. B. and Steiner, R. H., Plastics for Corrosion-Resistant Applications, Reinhold, New York, 1955.
[25] Anonymous, Handbook of Chemistry and Physics, 41st ed. Chemical Rubber Company,
Cleveland, Ohio, 1959-1960.
[26] ACI Committee 544, "State-of-the-Art Report on Fiber Reinforced Concrete," Journal,
Title No. 70-65, American Concrete Institute, Nov. 1973, pp. 729-744.
[27] Shroff, J. K., "The Effect of a Corrosive Environment on the Properties of Steel Fiber
Reinforced Portland Cement," MS thesis, Clarkson College of Technology, Potsdam,
New York, Sept. 1966.
[28] Anonymous, "A Status Report on Fiber Reinforced Concretes," Concrete Construction,
Jan. 1976, pp. 13-16.
[291 Bateson, G. B. and Obszarski, J. M., "'Strength of Steel Fiber Reinforced Concrete
Exposed to a Salt Water Environment," Materials Performance, July 1977, Technical Note, p. 48; Paper No. 27, NACE Annual Conference, Houston, Tex., March 1976.
[30] Project OK1, Task 03, Work Unit 006, "Corrosion Behavior of Cracked Fibrous Concrete," Materials Systems and Science Division of the Construction Engineering Research Laboratory, 1975 (unpublished).
[31] Lankard, D., private communications, formerly Battelle Memorial Institute, now with
Materials Laboratory, Inc., 1977.
ERLIN
AND
WOODS
ON
CORROSION
OF
EMBEDDED
MATERIALS
319
[32] Marek, R. et al, "Promising Replacements for Conventional Aggregates for Highway
Use," HRB NCHRP, Final Report, Project 4-10, University of Illinois, Jan. 1971.
[33] Ironman, Ralph, "Stronger Market Seen for Glass-Figer Concrete," Concrete Products,
Jan. 1976, pp. 42-44.
[34] Tallentire, A. G., "Glass Fibre Cement Applications," Precast Concrete, Feb. 1977,
pp. 95-97.
[35] Grimer, F. J. and Ali, M. A., "The Strengths of Cements Reinforced with Glass Fibers,"
Magazine of Concrete Research, Vol. 21, No. 66, March 1969, pp. 23-30.
[36] Majundar, A. J., "Properties of Fiber Cement Composites," Symposium on Fiber
Reinforced Cement and Concrete, RILEM, 1975, pp. 279-313.
[371 Cohen, E. B. and Diamond, Sidney, "Validity of Flexural Strength Reduction as an
Indication of Alkali Attack on Glass in Fibre Reinforced Cement Composites," Symposium on Fiber Reinforced Cement and Concrete, RILEM, 1975, pp. 315-325.
[38] Anonymous, "Developments in Fiber Composites," Precast Concrete, Vol. 6, No. 10,
Oct. 1975, pp. 549-551.
[39] Opoczky, Ludmilla and Pentek, Lazlo, "Investigation on the 'Corrosion' of Asbestos
Fibers in Asbestos Cement Sheets Weathered for Long Times," Symposium on Fiber Reinforced Cement and Concrete, RILEM, 1975, pp. 269-276.
[40] Anonymous, "Rubble Recycling Saves Time, Energy, and the Environment," Rock
Products, Vol. 80, No. S, May 1977, pp. 107-108.
[41] Anonymous, "Reusing Concrete," Engineering News-Record, Vol. 186, No. 22,
June 1971, p. 15.
[42] Anonymous, "Recycled Slab is New Runway Base," Highway and Heavy Construction,
Vol. 120, No. 7, July 1977, pp. 30-33.
[43] Frondistou-Yamas, S., "Waste Concrete as Aggregate for New Concrete," Journal,
American Concrete Institute, Vol. 74, No. 8, Aug. 1977, pp. 373-376.
[44] Buck, A. D., "Recycled Concrete," Highway Research Record, No. 430, 1973, pp. 1-8.
L.A.
Lu~
Chapter 21 Bond with

Reinforcing Steel
Introduction
Early experimenters and designers of reinforced concrete recognized that
slip of the reinforcement had to be prevented in order to develop adequate
bond and shear strengths and maintain the integrity of reinforced concrete
members.
The first comprehensive series of bond tests was conducted by Abrams [112
who employed both pullouts and beam specimens and measured the slip
with care. This investigation laid the foundation for many subsequent
studies which finally culminated in the adoption of ASTM Tentative Specifications for Minimum Requirements for the Deformations of Deformed
Steel Bars for Concrete Reinforcement (A 305) in 1956. This standard
defined the minimum requirements for deformed reinforcing bars as we
know them today.
Tests for Flexurai and Anchorage Bond in Beams

Bond stress in a beam may be defined either as the nominal "flexural
bond" [2] value calculated from the expression u = V/~ojd, or it may be
defined as the average bond stress from a point of m a x i m u m tension to the
end of the bar. The latter has been termed "anchorage" or "development
bond," and it is generally considered to be more meaningful than the
"flexural" bond value.
Bond strength is determined using either beam or pullout specimens.
Pullout tests are generally satisfactory for measuring the relative bond
values of bars with different deformations. However, experience has shown
that bond values derived from pullout tests cannot be applied in general
1Vice president, Computerized Structural Design, Inc., Brown Deer, Wis. 53223.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
LUTZ ON BOND WITH REINFORCING STEEL
321
to design of reinforced concrete beams and that beam tests are necessary
to develop design criteria.
The difference between bond values obtained from pullouts and beams
are due primarily to the different crack patterns which develop in the two
types of specimens. Although longitudinal cracks develop along the reinforcing bars (as the stresses increase and slip becomes appreciable) in
both beams and pullouts, there is a basic difference in the cracking phenomena. In pullout specimens the concrete is in compression, while in the
beam specimens both the concrete and the reinforcement are in tension.
As the steel stresses increase, transverse cracks develop, and each new
transverse crack tends to initiate a new longitudinal crack. Transverse
cracks are absent entirely in pullout specimens and the cracking is confined
to longitudinal splitting. The bonding in the portion of the beam between
the support and the nearest transverse crack is probably the only bonding
that the pullout specimen properly models according to work done by
Mains [3]. Mains observed in his study of distribution of tensile and bond
stresses along reinforcing bars that the tensile stress distribution curves
for the beams and pullouts were quite similar when plotted to the same
scale and with the free end of the beam bar coincident with the free end
of the pullout bar.
The Bond Stress Committee of the American Concrete Institute (ACI)
in its report of February 1945 [4] described a test procedure providing a
uniform basis for comparison of bond values of different reinforcing bars.
This test procedure included length of embedment and position of bar as
cast (top or bottom) as parameters. This procedure was extensively used in
bond studies [5, 6], which culminated in development of ASTM Specification
A 305 covering the geometric requirements for deformed bars. The test
procedure was later adopted by the American Concrete Institute as ACI
Test Procedure to Determine Relative Bond Value of Reinforcing Bars
(208-58) (discontinued). This standard did not include any minimum performance criteria since its primary purpose was to establish relative bond
values for different bars.
The establishment of minimum requirements for the deformation of
deformed reinforcing bars led to a substantial increase in working bond
stress for deformed bars conforming to requirements of ASTM Specification A 305. The ACI Building Code for reinforced concrete (ACI Standard
Code Requirements for Reinforced Concrete (318)) adopted in 1951 incorporated new provisions which increased the allowable bond for deformed
bars in beams and slabs from 0.05 to 0.10 of the 28-day compressive
strength, with a maximum of 2.4 MPa (350 psi). This code also recognized
for the first time the difference between top and bottom bars, as cast. The
allowable bond stress for top bars was reduced to 70 percent of the bond
for bottom bars.
Although the test procedure and the test beams described in ACI Standard
322
208 served their purpose well, the standard was rather restrictive since it
limited bars to one size, concrete to one strength, and the embedment
length to a m a x i m u m of 406 m m (16 in.). In order to study the effect of the
various parameters on the bond values of deformed bars in simply supported
beams subjected to concentrated loads, a beam specimen and a test procedure were developed [7], which represent a considerable departure from
the ACI Standard 208. This procedure, described in the report of ACI
Committee 408 on Bond and Development of Reinforcements [8], was intended to provide greater flexibility in the design of the recommended
specimen and test procedure in order to permit the use of bars of different
diameters, more than one strength of concrete, and the longer lengths of
embedment needed to develop the high-yield stresses of modern deformed
bars. This bond specimen also was intended to provide a research tool
yielding data that readily could be translated into design criteria. This test
beam still has drawbacks in that the beam size, stirrup reinforcement, and
test bar placement are all fixed. Although this standardized beam is fine
for providing comparison of test results, the beam test still is not truly
representative of the real-life beam.
The test beam illustrated in Fig. 1 was designed to permit the measurement of the average value of bond stress a n d t h e slip at both the loaded
and free ends of the portion of the bar between the supports and the load
points. The beam was provided with T-shaped ends in order to shift the
reactions to points where they would not contribute to the restraint of
longitudinal splitting. A strip of metal embedded in the concrete directly
opposite each load point assured formation of a crack at that plane, and
slip measurements were made at each load point plane.
Another type of beam specimen designed for investigating bond in the
areas of bar cutoffs and points of inflection was used in a bond study
which included bar sizes from Nos. 3 to 11 [9]; two additional bar sizes,
Nos. 14 and 18, were included in a subsequent study by the same author [10].
The test beam shown in Fig. 2 was designed to place the development
length of the test bar, designated as L, in a negative moment region between
the point of inflection and the point where the other bars were cut off. This
arrangement permitted the calculation of the maximum steel stress and the
average bond stress in a given development length of the bar. The dimensions
of the beam specimen were varied in order to study the effect of the bar
cover and beam width.
The two bond investigations described elsewhere [7,9] were concerned
with different critical sections of beams, and in this sense they may be considered as supplementing each other. The beam shown in Fig. 1 was designed
to investigate the bond at the support, while the beam in Fig. 2 was concerned with the bond at the point of inflection and the section where tension
bars terminate within a span. In spite of the differences in the specimens,
both investigations showed that the ultimate bond stress is proportional to
LUTZ O N B O N D WITH
~I"
REINFORCING
LOADED
PLAN
RUBBER
TEST
'I ,
'
rrJ
I
Ia.
--E-
,TT
STRI
r' = 2 5 . 4 , . , ~
323
r~"DIA HOLE~
,,
STEEL
'
,
TUBING
BAR
1
I
I
a...4
_ll
- ~ _ t.- i.~
~,
iI. /
!
~.~
ELEVATION
FIG. 1--Bond test beam recommended by ACI Committee 408 in 1964.
~ff~[and inversely proportional to the bar diameter when other factors were
constant. This relationship was recognized by the ACI in the formulation
of the allowable bond stresses in the ACI Standard 318 for bars with sizes
and deformations conforming to ASTM Specification A 305.
An upper limit of bond stress in ACI Standard 318 functioned as the
ultimate bond stress for the smaller reinforcing bars. The parameter
alone was used in an expression for the tension bars conforming to ASTM
Specification for Special Large Size Deformed Billet-Steel Bars for Concrete
Reinforcement (A 408) (discontinued, replaced by ASTM Specification
for Deformed and Plain Billet-Steel Bars for Concrete Reinforcement
(A 615)) (Nos. 14S and 18S), w h e r e f ' c is the 28-day compressive strength
of concrete. Position of the bar in the beam was also a factor influencing
the bond stress; the bond stress for top-cast bars was limited to 70 percent
of the bond stress used for bottom and vertically cast bars. An ultimate
bond stress for reinforcing bars in compression was also in ACI Standard
318.
324
"t
z
......
'
PLAN
= " ~-iS'T.-4" ~.. -~- -~I
~--J J .
J.
. 1 - ~ - -L. . . . . . . . . . . . . . . . . . . . . . .
t
ELEVATION
FIG. 2--University of Texas beam.
Two basic changes were reflected by the bond provisions which were in
ACI Standard 318-71 [11]. Anchorage or development bond was considered
of such importance relative to flexural bond that all bond provisions were
presented in terms of development length rather than bond stress. Secondly,
the development lengths were 20 percent longer than those calculated using
the provisions of ACI Standard 318-63; this was done due to the observation
that closely spaced bars, which are quite typical, constituted a real problem
at the bond stress levels of ACI Standard 318-63.
The development length expressions are based on the bond stress expressions of ACI Standard 318-63 and consequently involve the same variables.
As a result the development length of small bars is proportional to the
bar diameter and the yield stress. For the larger bars up to No. 11, the
development length is proportional to bar area and yield stress and inversely
proportional to x/f'c. For Nos. 14 and 18 bars, separate expressions for
development length which are proportional to yield strength and inversely
proportional to ~ c are used.
The development lengths calculated are modified for various situations,
for example, top-casting of the reinforcing bar, use of lightweight concrete,
large spacing of the bars, and use of reinforcing bars with yield strengths
greater than 414 MPa (60 ksi). These situations which modify the development length are an indication of the areas of current thinking or research
in bond, or both. Three such areas are as follows.
1. The development length for top bars still represents a 70 percent
325
effectiveness in bond relative to bottom bars. However, there is evidence

that for the low-slump concretes now being used, the bond reduction is not
as large as indicated initially [5, 6].
2. The development length of reinforcing bars has been found not to be
directly proportional to the yield strength. The bond effectiveness decreases
as the yield strength increases in a bar [12].
3. The development length also appears to be appreciably influenced by
the spacing of reinforcing bars, the concrete cover of a bar, and the confinement of a reinforcing bar by lateral ties [12,13].
Recent bond testing has made use of two different specimens which are
shown in Figs. 3 and 4. The Stub-cantilever beam which is shown in Fig. 3
II
I
JACK
I
F I G . 3--Stub-cantilever beam used at West Virginia University.
has been used in several research projects [14-16]. Besides being cheaper
than the Texas beam (Fig. 2), it has flexibility of load application as the
relationships of bond wiih shear, moment, and dowel forces can be varied
easily. The symmetrical beams shown in Fig. 4 allow a number of different
tests to be run on the same type of beam. A splice test can be conducted
using the beam in Fig. 4(a) while a pair of top bar embedment tests can be
conducted with the beam in Fig. 4(b). Both beams could be used to test
bond development in the overhanging portions at each end.
In each case, the beam size and the number and arrangement of reinforcing bars are being varied. Much is being learned about Item 3 above, so
that in the near future the effects of bar spacing, concrete cover and conCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:32 EDT 2014
326
LAP
-I
,]
t
(A)
METAL CRACK
r"r"r "rr +
I i i i i I
L..L_I. 1 L
STARTER7
t
(B)
FIG. 4--Symmetrical beams.
finement by lateral ties may be explicit variables in development length

expressions as are now being suggested by recent work [12,13].
Bond
of Lapped
Splices
One of the areas of application in which bond plays a critical role is in

the tensile lapped splices. Longitudinal splitting is produced in the splice
length with increase in bar stress and is the basic cause for failure of lapped
splices.
Test data on tensile lapped splices are somewhat limited relative to data
available on development. The research work [17-19] which has led to the
splice provisions in ACI Standards 318-63 and 318-71 have emphasized the
following factors which influence the lap splice length required: (a) the increase of spacing between lapped splices decreases the splice length required, (b) an increase in concrete cover over the splice decreases the required splice length, (c) confinement of the splice by some form of lateral
327
reinforcement decreases the required splice length, and (d) the length of
splice required increases as the tensile force requirement increases.
The code provisions reflect these points by increasing the required splice
length as the percentage of spliced bars is increased and the level of stress
in the bars is increased. The splice length is increased if the spacing is less
than a certain value, unless lateral reinforcement of a certain amount is
used.
Tests of splices have been conducted primarily in test setups which place
the spliced bars in a constant moment region as illustrated in Fig. 4(a). An
exception is the work by Ferguson and Briceno [20] and Ferguson and
Krishnaswamy [21] where the splice is in varying moment region as produced by a single concentrated load on a simple beam span.
The code provisions for splices have always differed considerably from
the development length provisions. However, the recent work referred to
previously [12] which is attempting to express explicitly the effects of concrete cover, bar spacing, and confinement by lateral ties has indicated that
the expression for development length works equally well for splice length.
Bond in Prestressed Concrete
The bond between tensioned tendons and concrete plays a dual role in
prestressed concrete beams. One type of bond known as "prestress transfer
bond" is developed at the ends of concrete members prestressed by the pretensioning method. In this case, tension in the steel is transferred to the
surrounding concrete primarily by friction which is made possible by the
normal force produced by the Poisson effect, or lateral swelling of the steel
in the transfer zone.
The second type of bond stresses in prestressed concrete is brought about
by flexural action. When a prestressed beam is subjected to flexure, the
tendons function to some extent as ordinary reinforcing bars and bond
stresses develop in accordance with shears and moments in the beam.
However, prior to cracking of concrete the stresses in the prestressing steel
change quite slowly and the bond stresses are low. As cracking occurs,
the steel stresses at cracked sections increase rapidly and bond stresses increase accordingly. These bond stresses which are associated with flexural
cracking are commonly referred to as "flexural bond stresses."
In a study of the nature of bond in pretensioned prestressed concrete,
Janney [22] investigated the variation of the prestress transfer bond with
various parameters, as well as the interaction between prestress transfer
bond and flexural bond. Considerable variation in the anchorage length
and the shape of bond stress distribution curves were noted for wires of
different diameters and different surface conditions of the wire ranging
from rusted to lubricated. Tests of beams prestressed by pretensioning
indicated that as cracks occurred the flexural bond stresses increased
328
sharply and the bond stress "wave" progressed from the cracked midsection toward the ends of the beam. As the diameter is reduced and radial
pressure is relieved by the sharp increase in tension in the tendon after
cracking, the flexural bond capability is reduced considerably near the
ends of the beam. As a result, a bond failure may be expected when the
flexural bond stress wave reaches the anchorage zone, even prior to reaching the free end of the tendon. Janney [22] demonstrated this phenomenon
by strain measurements in the posttensioned beams.
An expression for development length, ld in millimetres, of three-wire
and seven-wire pretensioning strand based primarily on the work of Hanson and Kaar [23] appears in the ACI Standards 318-63 and 318-71. The
length is proportional to strand diameter, db, and composed of two parts,
a transfer bond length and a flexural bond length respectively as follows:
ld -~- 3.685 lf-,fe db _if._(fp, -- fse) db1
where
f,e = effective stress in prestressing steel after losses, MPa, and
fp, = calculated stress in prestressing steel at design load, MPa.
The equation considers failure to occur at the onset of general bond slippage.
Additional work ~related to development length and bond performance of
pretensioning strand appears in other references [24-27]. Martin and
Scott [25], for example, suggested a set of equations that would limit fp, if
the full development length could not be achieved in a particular situation.
Most bond testing of prestressed strand is done by concentrated loading
on simple span prestressed members. Little is known at present how the
size of member, position of the strand, and spacing of the strand may influence the bond integrity.
Influence of Bond in Control of Cracking and Deflections
The efficient and economical use of high-yield strength reinforcement is

limited by the necessity of avoiding excessively wide cracks which may be
objectionable from the standpoint of appearance or lead to corrosion of
reinforcement. In recent years considerable research was carried out with
new types of high-strength deformed reinforcing bars to study the mechanism of cracking and to determine the effect of such parameters as diameter
of bars, ratio of reinforcement, and cross-sectional dimensions of beams on
the width and spacing of cracks. Clark [28] investigated these variables in
a comprehensive series of tests of beams and slabs, and presented a formula
relating the average width of cracks to the computed stress in the reinforcement, the diameter of the bars, and the ratio of reinforcement. Chi and
329
Kirstein [29] made use of Clark's data in addition to their own to develop
a simplified formula for the width of crack, taking into account the distance
from the bar surface to the nearest beam surface as a parameter. Similar
conclusions were reached by Broms [30,31].
The earlier studies by Watstein and Parsons [32], and Watstein and
Seese [33] have demonstrated, by means of tension tests of axially reinforced
cylindrical specimens, the direct relationship between the bond strength
and both the spacing and the width of tensile cracks. A tension specimen
devised to measure the bonding efficiency of a bar [33] gave values which
correlated well with the spacing and width of cracks observed in an independent set of long axially reinforced tension specimens.
A statistical study conducted by Gergely and Lutz [34] of beam crack
width data from several investigations revealed that the bar stress, concrete
cover, and area of concrete surrounding a reinforcing bar were the primary
variables affecting crack width. The crack width expression recommended
by Gergely and Lutz formed the basis of the parameter used in ACI Standard 318 to control the distribution of reinforcement (and crack width) in
beams and one-way slabs.
The bonding efficiency of reinforcing bars also affects the rigidity of
flexural members because the quality of bond affects the contribution of
the concrete between tensile cracks which in turn improves the stiffness of
the beam. In other words, reduction of the slip between the concrete and
the reinforcement along the entire length of the b e a m does increase the
beam stiffness.
The resistance of the concrete between tensile cracks to flexure was considered at length in the 1957 RILEM Symposium on Bond and Crack
Formation in Reinforced Concrete [35]. The papers which discuss the
various parameters affecting the distribution of strain between cracks and
their effect on the flexural stiffness are by Soretz [36], Baker et al [37],
and Murashev [38].
Acknowledgments
Acknowledgment is given to David Watstein of the National Bureau of
Standards whose article on "Bond With Reinforcing Steel" formed the
basis of this updated article on the same topic.
References
[1] Abrams, D., "Tests of Bond Between Concrete and Steel," Bulletin No. 71, University
of Illinois, Urbana, Ill., Dec. 1913.
[2] "ACI Standard Building Code Requirements for Reinforced Concrete (ACI 318-63),"
Journal, American Concrete Institute, Vol. 60, June 1963.
[3] Mains, R. M., "Measurement of the Distribution of Tensile and Bond Stresses Along
Reinforcing Bars," Journal, American Concrete Institute, Vol. 48, Nov. 1951, p. 225.
330
[4] "Report of ACI Committee 208, Proposed Test Procedure to Determine Relative Bond
Value of Reinforcing Bars," Journal, American Concrete Institute, Vol. 41, Feb. 1945,
p. 273.
[5] Clark, A. P., "Comparative Bond Efficiency of Deformed Concrete Reinforcing Bars,
RP 1755," Journal of Research, National Bureau of Standards, Vol. 37, Dec. 1946.
[6] Clark, A. P., "Bond of Concrete Reinforcing Bars, RP 2050," Journal of Research,
National Bureau of Standards, Vol. 43, Dec. 1949.
[7] Mathey, R. G. and Watstein, D., "Investigation of Bond in Beam and Pullout Specimens with High-Yield-Strength Deformed Bars," Journal, American Concrete Institute,
Vol. 57, March 1961, p. 1071.
[8] "Report of ACI Committee 408, A Guide for Determination of Bond Strength in Beam
Specimens," Journal, American Concrete Institute, Vol. 61, Feb. 1964, p. 129.
[9] Ferguson, P. M., and Thompson, J. N., "Development Length of High Strength Reinforcing Bars in Bond," Journal, American Concrete Institute, Vol: 59, July 1962, p. 887.
[10] Ferguson, P. M. and Thompson, J. N., "Development Length of Large High-Strength
Reinforcing Bars," Journal, American Concrete Institute; Proceedings, Vol. 62, No. 1,
Jan. 1965, pp. 71-94.
[11] ACI Committee 318, "Building Code Requirements for Reinforced Concrete (ACI
318-71)," American Concrete Institute, Detroit, 1971, (plus cumulative supplement
issued annually).
[12] Orangun, C. O., Iirsa, J. O., and Breen, J. E., "The Strength of Anchored Bars: A
Reevaluation of Test Data on Development Length and Splices," Preliminary Review
Copy, Research Report 154-3F. Center for Highway Research, The University of Texas
at Austin, Jan. 1975.
[13] Kemp, E. L., and Wilhelm, W. J., "An Investigation of the Parameters Influencing
Bond Cracking," presented at the Interaction Between Steel and Concrete Symposium,
American Concrete Institute Convention, San Diego, Calif., 17 March 1977.
[14] Wilhelm, W. J., Kemp, E. L., and Lee, Y. T., "Influence of Deformation Height and
Spacing on the Bond Characteristics of Steel Reinforcing Bars," Civil Engineering Studies
Report No. 2013, Department of Civil Engineering, West Virginia University, 1971,
Federal Clearinghouse No. PB-200635.
[15] Gergely, P., "Splitting Cracks Along the Main Reinforcement in Concrete Members,"
Report to the Bureau of Public Roads of the U.S. Department of Transportation, Department of Structural Engineering, Cornell University, April 1969.
[16] Mirza, M. S. and Hsu, C. T., "Investigation of Bond in Reinforced Concrete Models-Eccentric Pullout Test," Proceedings, International Conference on Shear, Torsion, and
Bond in Reinforced and Prestressed Concrete, Coimbatore, India, Jan. 1969.
[17] Chamberlin, S. J., "Spacing of Spliced Bars in Beams," Journal, American Concrete
Institute, Vol. 54, Feb. 1958, p. 689.
[18] Chinn, J., Ferguson, P. M., and Thompson, J. N., "Lapped Splices in Reinforced Concrete Beams," Journal, American Concrete Institute, Vol. 52, Oct. 1955, p. 201.
[19] Ferguson, P. M. and Breen, J. E., "Lapped Splices for High Strength Reinforcing Bars,"
Journal, American Concrete Institute, Proceedings, Vol. 62, Sept. 1965.
[20] Ferguson, P. M. and Briceno, A., "Tensile Lap Splices--Part 1: Retaining Wall Type,
Varying Moment Zone," Research Report 113-2, Center for Highway Research, The
University of Texas at Austin, July 1959.
[21] Ferguson, P. M. and Krishnaswamy, C. N., "Tensile Lap Splices--Part 2: Design
Recommendations for Retaining Wall Splices and Large Bar Splices," Research Report
113-3, Center for Highway Research, The University of Texas at Austin, April 1971.
[22] Janney, J. R., "Nature of Bond in Pretensioned Prestressed Concrete," Journal, American
Concrete Institute, Vol. 50, May 1954, p. 717.
[23] Hanson, N. W. and Kaar, P. H., "Flexural Bond Tests of Pretensioned Prestressed
Beams," Journal, American Concrete Institute, Proceedings. Vol. 55, No. 7, Jan. 1959,
pp. 783-803. Also, Development Department Bulletin D28, Portland Cement Association.
[24] Kaar, P. H., LaFraugh, R. W., and Mass, M. A., "Influence of Concrete Strength on
Strand Transfer Length," Journal, Prestressed Concrete Institute, Vo]. 8, No. S, Oct.
1963, pp. 47-67. Also, Development Department Bulletin DT1, Portland Cement Association.
331
[25] Martin, L. D. and Scott, N. L., "Development of Prestressing Strand in Pretensioned

Members," Journal, American Concrete Institute, Vol. 73, Aug. 1976, p. 453.
[26] Anderson, A. R. and Anderson, R. G., "An Assurance Criterion for Flexural Bond in
Pretensioned Hollow Core Units," Vol. 73, Aug. 1976, p. 457.
[27] Salmons, J. R. and McCrate, T. E., "Bond Characteristics of Untensioned Prestressing
Strand," Journal, Prestressed Concrete Institute, Vol. 22, No. 1, Jan.-Feb. 1977.
[28] Clark, A. P., "Cracking in Reinforced Concrete Flexural Members," Journal American
Concrete Institute, Vol. 52, April 1956, p. 851.
[29] Chi, M. and Kirstein, A. F., "Flexural Cracks in Reinforced Concrete Beams," Journal,
American Concrete Institute, Vol. 54, April 1958, p. 865.
[30] Broms, B., "Crack Width and Crack Spacing in Reinforced Concrete Members," Journal,
American Concrete Institute, Vol. 62, No. 10, Oct. 1965.
[31] Broms, B. and Lutz, L. A., "Effects of Arrangement of Reinforcement on Crack Width
and Spacing of Reintorced Concrete Members," Jourm~l, American Concrete Institute
No. 11, Nov. 1965.
[32] Watstein, D. and Parsons, D. E., "Width and Spacing of Tensile Cracks in Axially
Reinforced Concrete Cylinders, RP 1545," Journal oJ" Research, National Bureau of
Standards, Vol. 31, July 1943.
[33] Watstein, D. and Seese, N. A., Jr., "Effect of Type of Bar on Width of Cracks in Reinforced Concrete Subjected to Tension," Journal, American Concrete Institute, Vol. 41,
Feb: 1945, p. 293.
[34] Gergely, P. and Lutz, L. A., "Maximum Crack Width in Reinforced Concrete Flexural
Members," Causes. Mechanism, and Control q/" Cracking in Concrete. SP-20, American
Concrete Institute, Detroit, 1968, pp. 1-17.
[3.5] Symposium on Bond and Crack Formation in Reinforced Concrete, RILEM, Stockholm,
1957.
[36] Soretz, S., "Factors Affecting the Stiffness of Reinforced Concrete Beams," Symposium on Bond and Crack Formation, RILEM, Stockholm, 1957.
[37} Baker, A. L. L., Ashdown, A. J., and Wildt, R. H., "'The Deformation of Cracked
Rectangular Reinforced Concrete Beams," Symposium on Bond and Crack Formation,
RILEM, Stockholm, 1957.
[38] Murashev, V. I., "Calculation of Crack Width in, and Flexural Rigidity of, Reinforced
Concrete Members in Flexure," Symposium on Bond and Crack Formation, RILEM,
Stockholm, 1957.
R. O. Lane1
Chapter 22
Abrasion Resistance
Introduction
This paper reviews the effects of important factors such as strength of

concrete, mixture proportioning, selection of concrete-making materials,
finishing procedures, and surface treatment on the wear resistance of concrete surfaces as determined by recent research. Newly revised or developed
ASTM standards, Test for Abrasion Resistance of Concrete by Sandblasting (C 418) and Test for Abrasion Resistance of Horizontal Concrete Surfaces (C 779), provide test procedures for various types of wear simulating
erosive or abrasive actions.
Definitions
The complex nature of the wear process is demonstrated by the multiplicity of definitions found in technical literature that attempt to describe the
mechanical action that takes place under various categories of inadvertent
particle removal. ASTM Definitions of Terms Relating to Erosion by Cavitation and Impingement (G 40) defines abrasion as wear by displacement
of materials from a solid surface due to hard particles or hard protuberances
sliding along the surface. Abrasion resistance, on the other hand, according
to American Concrete Institute Standard on Nomenclature (ACI 116), is
the ability of a surface to resist being worn away by rubbing and friction.
Erosive wear is loss of material from a solid surface due to contact with
a fluid which contains solid particles. When the relative motion of the solid
particles is nearly normal to the solid surface, the wear is called impact
erosion or impingement erosion. When the relative motion of the solid
particles is nearly parallel to the solid surface, the wear is called abrasive
erosion.
Cavitation erosion is the wear of a solid body moving relative to a liquid
in a region of collapsing vapor bubbles resulting in high impact. Repeated
collapse of cavities near a concrete surface causes pitting. This process
1Head, Materials Engineering Laboratory, Tennessee Valley Authority, Knoxville, Tenn.
37902.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
LANE ON ABRASION RESISTANCE
333
should not be confused with cavitation referring only to the formation and
collapse of cavities within a fluid.
The ACI has published a committee report on the subject of concrete
resistance to erosion as it relates to hydraulic structures which is an excellent source of information on this subject [1].2 Studies on cavitation [2]
disclosed that while good quality concrete withstands the flow of clear water
at high velocities, it will not resist severe forces of cavitation or the wearing
and grinding action of sand, gravel, or rock carried by the water.
Factors Affecting Abrasion Resistance

Materials related factors which may reduce the resistance of concrete to
abrasive action should be considered in the design and construction of
concrete surfaces which are to withstand wear due to rubbing, scouring,
sliding, impact, scraping, attrition, percussion, gouging, or cutting from
mechanical or hydraulic forces.
Frequently the failure of concrete to resist abrasion can be traced to
cumulative effects such as inadequate compressive strength, excessive air
content or slump, soft or resilient aggregate, inadequate bond strength of
the aggregate, improper curing or finishing, or overmanipulation of the
plastic concrete surface. Reducing the wear resistance of concrete involves
the surficial hardened cement paste, which initially resists abrasive forces.
As the paste is worn away, the coarse aggregate ultimately is exposed. Depending on the strength of the paste and subsequently on the bond and
hardness of the aggregate, surface deterioration may take place. Such wear
may be imposed by abrasive materials carried by wind or water flowing at
various velocities, cavitation, or may result from foot or vehicular traffic.
A brief discussion of significant factors relative to concrete resistance to
wear may illustrate the importance of the proper selection, composition,
and application of concrete based on the specific type of service condition.
Witte and Backstrom [3], as researchers before and after them, consider
the compressive strength the most important factor responsible for the
abrasion resistance of concrete. Studies indicated 34.5 MPa (5000 psi)
concrete develops twice as much resistance to abrasion or erosion as 20.7
MPa (3000 psi) concrete [4]. A minimum requirement of six bags of cement
per 335 kg/m 3 (yd 3) has been commonly accepted for concrete subject to
abrasion. However, experience shows that the wear resistance of concrete
of low to medium compressive strength can be improved appreciably by
providing a hard troweled finish. Conversely, concrete of higher strength
may show reduced wear resistance if given an ordinary commercial finish.
334
The significance of the void-cement system on the durability of concrete

was pointed out in extensive studies by Mielenz et al [5]. Generally speaking, abrasion resistance and compressive strength vary inversely with the
void-cement ratio.
Aggregate Selection
Next in importance is the quality of the fine and coarse aggregates. Smith
[6] suggests that although no conclusive correlation exists between abrasion
resistance and the quality of the coarse aggregate as determined by either
the ASTM Test for Soundness of Aggregates by Use of Sodium Sulfate or
Magnesium Sulfate (C 88), or ASTM Test for Resistance to Abrasion of
Small Size Coarse Aggregate by Use of the Los Angeles Machine (C 131),
the resistance can be increased appreciably by the use of dense, hard coarse
aggregates such as traprock, granite, silica, or metallic aggregate near the
surface. Schuman and Tucker [7] point out that the shape of aggregate particles, rounded or angular, regulates the water requirements for placing and
finishing and has a direct influence on the wear resistance of concrete.
Angular- to subangular-shaped aggregate is known to improve bond, usually
resulting in reduced wear. Other studies have shown that while concrete
of compressive strength below 55.2 MPa (8000 psi) is greatly benefited by
the use of hard aggregate, above this strength level, the effect of aggregate
on wear resistance is reduced greatly. Several researchers, among them
Scripture et al [8], Abrams [9], Jackson and Pauls [10], have investigated
specifically abrasion characteristics of stone. Pogany [11] states the rate of
wear of concrete generally is not affected by the aggregate provided the
coarse aggregate is equal or superior to the mortar matrix. As the exposure
to abrasion becomes more severe, it is advisable to reduce the amount of
material passing the 300/zm (No. 50) and 150/zm (No. 100) sieves in the
concrete mix. While mortar is more resistant to the forces of cavitation
than concrete of the same strength, it is less resistant to the forces of abrasion. ACI Standard on Construction of Concrete Floors (302) recommends
the use of the largest size aggregate permissable, usually 19 to 39 mm
(3/4 to 1~/2 in.), in concrete carrying light to medium traffic whereas for
heavy dtity concrete surfaces, the maximum-size aggregate should be limited
to 10 or 13 mm (3/8 or tA in.).
For concrete subject to light to medium abrasion, good quality aggregates meeting the requirements of ASTM Specification for Concrete Aggregates (C 33) are generally acceptable. Heavy duty floors and slabs exposed
to more severe abrasive action demand well-graded hard mineral aggregates.
Admixtures
The effects of various admixtures, including air-entraining and waterCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:33 EDT 2014
LANE ON ABRASION RESISTANCE 335
reducing admixtures and retarders, on the wear resistance of concrete

could not be established conclusively on the basis of the review literature.
The general trend in most studies reflects an increase in abrasion loss
proportional to increases in air content for any given water/cement (w/c)
ratio [12]. As the compressive strength is adversely affected by increased
air contents, so will the resistance of concrete to abrasion.
Water-reducing agents, by virtue of their positive effect on the w/c ratio
and consequently on increases in strength, tend to produce a concrete of
improved wear resistance.
Concrete not severely exposed to freezing and thawing or ice removal
agents should be placed at air contents not exceeding 3 percent to avert
any reduction in wear resistance [13].
Finishing Procedures
Wear resistance of a floor is also affected by the construction techniques
applied. Step-by-step procedures for 7 classes of floors including consolidation and finishing are comprehensively described in a committee report
of ACI 302 [14]. This report requires floor classes 4 through 7 to be given
special consideration when exposed to abrasive conditions and are classified
as (a) light duty, industrial, commercial; (b) industrial; and (c) heavy duty
industrial. General illustration of the effect of finishing techniques on
abrasion resistance is given in the appendix of ACI 302. One plot shows
greater wear loss due to a float finish as compared with a steel trowel finish.
The same effect is shown when comparing premature finishing on abrasion,
that is, floating or troweling the surface while bleeding water is present,
with delayed finishing. Finishing techniques are also compared with wear
resistance of concrete in a study by Fentress [15]. In this investigation,
abrasion forces were imparted on wood float, magnesium float, steel trowel,
and hard-steel trowel finishes. The wood float tends to tear the surface
and to displace the aggregate. The magnesium float, despite ease of finishing, causes a rough-textured surface and a lowering in wear resistance.
Both the steel trowel and hard-steel trowel produce smooth surfaces, closing
any existing imperfections and providing excellent resistance to abrasion.
The key to durable, cleanable concrete slabs according to Ytterberg
[16] can be found in proper finishing procedures and in a reduction in the
w/c ratio. Combining these demands in one application, he demonstrates
that a deferred topping finish from which surface water is removed by a
vibratory absorption process yields highest abrasion resistance. The technique requires that after the water of workability has been removed, the
surface is blade-floated, and upon stiffening, troweled. In this comparative study the deferred topping exceeded by far a monolithic application
in wear resistance.
Abrasion tests carried out by the U.S. Bureau of Reclamation confirm
336
that vacuum-treated concrete surfaces are in fact considerably more resistant to abrasive forces than surfaces with either nontreated lean or richer
concrete [17].
Wearing surfaces can be improved by the use of dust coats or "dry
shakes" consisting of dry cement or cement-aggregate mixtures sprinkled
on the surface just prior to initial set. The dry coating is power-floated
and then steel troweled into the surface without the use of additional water
to produce a hard, dense topping layer. Pockets of bleed water should be
absorbed before adding the dust coat and the troweling operation carefully supervised.
The use of steel fiber-reinforced concrete and topping mixes have recently
been used with some success for airport runway pavements. Relatively few
wear tests have been conducted on this type of surface, but the limited
data indicates abrasion resistance is increased somewhat.
Curing
In constructing concrete floors and slabs the quality of the concrete
operation determines the properties of the surface. Well-cured concrete
can produce a surface up to twice as wear-resistant as a surface cured for
only 3 days. ACI Durability of Concrete (201) suggests that for maximum
durability and abrasion resistance, curing should start as soon as possible
after the concrete has been finished and shall be continued preferably for
a minimum of 14 days. A correlation of curing time and wear reported by
Sawyer [18] involved a series of tests comprising a wide range of cement
contents, w/c ratios, and incremental curing. From this study it is apparent
that marked improvement in abrasion resistance can be expected with extended curing time, especially for surfaces composed of leaner concrete.
A laboratory program conducted by the California Division of Highways
[19] considers the effect on abrasion of such variables as slump, finishing,
curing, and surface treatments using linseed oil. Test data indicate among
these variables that the greatest abrasion losses encountered were associated
with less than adequate curing procedures.
For poorly-cured or porous concrete surfaces, beneficial effects from
treatment with linseed oil are demonstrated by an increase in abrasion
resistance of as much as 30 percent. Mayberry's studies also confirm appreciably higher wear resistance of concrete for emulsion cured bridge
decks [20]. The application of a curing compound immediately after finishing creates abrasion resistance approximately equivalent to that attained
after 7 days of curing by water spray and continuous dampness.
For concrete floor and slab construction using Type I portland cement,
at least 5 days of curing at concrete temperatures of 21~ (70~ or higher
are required and a minimum of 7 days at temperatures of 10 to 21~ (50
to 70~ to assure adequate abrasion resistance [14].
Floor Hardeners and Surface Treatment
Certain chemicals (magnesium and zinc fluorosilicates, sodium silicate,

gums and waxes) serve to prolong the life of older floors and are considered
an emergency measure for treatment of deficiencies in relatively pervious
and soft surfaces which wear and dust rapidly [21]. Treatment with fluorosilicates densities and hardens surfaces and improves abrasion resistance
but the improvement by the two types of fluorosilicates is not of equal
magnitude [22]. While the magnesium fluorosilicate is more resistant to
rubbing action simulated best by a revolving disk-type abrasion machine,
the zinc fluorosilicate shows greater resistance to impact type wear as imposed by the dressing wheel type machine. Liquid treatment of floor topping surfaces is less successful and improvements have proved of little
significance. Paints and coatings used to seal concrete surfaces and to protect the concrete from the attack of the environment or chemicals [23]
possess only limited wear resistance and any test simulation to evaluate
their effectiveness is difficult. However, some coatings, among them vinyls
and heavy rubber, if properly bonded remain fairly resilient and effective
in protecting concrete from erosive and abrasive actions.
Polymer impregnated concrete for highway bridges was investigated by
the University of Texas [24]. Concrete of 34.5 to 44.8-MPa (5000 to 6500
psi) compressive strength, nonair-entrained, treated with polymer showed
considerable increase in wear resistance.
Experimental studies by Krukar and Cook [25] and by Dikeon [26] also
demonstrated remarkable improvements in concrete properties, particularly
in strength and abrasion resistance due to impregnation and in-place polymerization of monomers in overlays and in mature concrete.
Abrasion Test Procedures
The techniques and test methods that have been employed for over a
century to evaluate the resistance of concrete to abrasion, erosion, or cavitation have attempted with varied success to reproduce the typical forces
detrimental to concrete surfaces. Preceding any standardization of laboratory methods or of testing machines, reliance was placed initially on the
use of rattlers which provide abrasive action based on the tumbling of
steel balls impacting a test surface [27]. Numerous modified versions of
the rattler principle continued to be the dominant test procedure along
with the revolving metal disk pressed against a small concrete specimen
[28]. The drill press-type abrasion machine [29] in modified form still is
used by highway departments and enjoys some popularity because of its
simple design. The Corps of Engineers Method of Test for Resistance of
Concrete or Mortar Surfaces to Abrasion (CRD C 52) describes a drill
press-type rotating cutter device.
ASTM Test for Abrasion Resistance of Concrete by Sandblasting (C 418)
338
dates back to 1958 and is based on the principle of producing abrasion

by sandblasting. This procedure was modified by ASTM Subcommittee
C09.03.13 on Methods of Testing Concrete for Resistance to Abrasion
from the original Ruemelin blast cabinet equipped with an injector-type
blast gun with high velocity air jet (Fig. 1). The control over such vari-
L[.~I
QQ
===LOADING DOOR=:=
m
~=1
i
I
L.,-
.COMP AIR CONN.
. . . . . .
,
q
J
:J
\CONTROL VALVE
FIG. 1--Sand blast cabinet.
ables as gradation of the silica sand, air pressure, rate of feed of the abrasive, distance of the nozzle from the surface, and the area of the shielded
surface is critical. The 1958 standard specified carborundum or silica sand.
This procedure was revised in 1964 calling for the explicit use of silica sand
which allows for closer control of the abrasive action. The 1976 revision
specifies the use of oil-base modeling clay of known specific gravity for
the filling of voids in the abraded concrete surface. This permits accurate
measurement of abrasion by volume displacement and replaces the old
weight-loss determination. The modified clay filler method is similar to a
procedure that evaluates floor and patio brick developed for ASTM Specifications for Required Safe Crushing Strengths for Sewer Pipe to Carry
Loads from Ditch Fillings (C 15, discontinued). Rushing [30] compared
the 1958 and 1964 procedures of ASTM Test C 418 in a 1968 laboratory
study and found the later revision to permit a more accurate determination
of abrasion loss. The sandblast method closely reproduces wear conditions
existing in hydraulic structures subject to abrasive materials carried by
water flowing at low to medium velocities; structures thus affected include

piers, spillways, tunnels, concrete stilling basins, etc. Other applications
of this test are relevant to the determination of the effectiveness of surface
treatments, among them curing compounds, hardeners, and coatings where
the performance of the protective "skin" is critical.
For almost two decades ASTM Test C 418 remained the only specified
standard for testing concrete resistance to abrasion. From 1964 through
1971, Subcommittee C09.03.13 initiated a number of major test programs
conducted by independent laboratories. The results of these comparative
programs established such parameters as severity and reproducibility,
leading to the selection of three abrasion machines and their inclusion in
one ASTM standard. ASTM Test for Abrasion Resistance of Horizontal
Concrete Surfaces (C 779) covers three testing procedures: (a) the revolving
disk machine, (b) the dressing wheel, and (c) the ball bearing machine.
All three machines are portable and adaptable for laboratory and in-place
field abrasion testing. The characteristics and applications of each machine
are discussed briefly. The revolving disk machine (Fig. 2) introduces frictional forces by rubbing and grinding. Sliding and scuffing is accomplished
by rotating steel disks in conjunction with abrasive grit. The disk machine
was redesigned by Master Builders Company, Cleveland, Ohio, and is
essentially patterned after a machine developed by the National Bureau of
Standards. Among the three standard methods, this machine provides the
most reproducible results.
A supply of 60 mesh silica carbide abrasive is fed to the disks at a rate
of 4 to 6 g/min. After every 15 min, specimens are brushed clean. Wear
is measured after 30 and 60 min, respectively, with a depth micrometer.
The smooth, plane surface produced by disk abrasion is explained by the
movement of the disks across the test surface causing the entrapped irregularly-shaped particles beneath to slide rather than to roll thus cutting the
surface by friction (Fig. 3). No high compressive stresses, fracture, or impact are imposed by the disk's action.
Davis and Troxell [31] report the depth of wear for periods of 30 to 60
rain to be about the same for ordinary slabs as for heavy duty concrete
surfaces. A plane surface resulted in either case regardless of the variation
in the quality of the hardened cement paste or the toughness of the aggregate which became exposed as the depth of wear increased. The abrasive
mode of this procedure best simulates wear by light to moderate foot traffic and light to medium tire wheeled traffic or moving of light steel racks,
etc.
In contrast to the disks the dressing wheel machine (Fig. 4) imparts
high concentrated compressive forces with high impact stresses. In principle
the dressing wheel follows the design of a machine described in the Corps
of Engineers Specification CRD C 52 and was modified by the Kalman
Floor Company, White Plains, N.Y. The dressing wheel machine is not
340
FIG. 2--Revolving d&k machine.
unlike the steel disk machine in appearance except for three sets of seven
dressing wheels mounted on horizontal shafts which take the place of the
three rotating steel disks. However, no abrasive material is employed with
this machine. Initial and intermediate measurements of the test path are
taken with a depth micrometer. Tracking of the dressing wheel normally
leaves a grooved path with the test surface being irregular and fairly rough
as harder aggregate particles stand out from the softer stone and mortar
which are abraded more quickly (Fig. 5). On ordinary finished concrete
slabs the dressing wheels produce a depth of wear more than double that
obtained with the disk machine for the same test duration. Yet approximately equal depths of wear are obtained from both machines for the
same test period when hard troweled finished floors are tested. The coeffiCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:33 EDT 2014
FIG. 3--Rotating disk after 120 rain of abrasion.

cient of variation established by the Berkeley study for slabs abraded with
the dressing wheel machine is several times as great as that for the revolving disk machine. These results are essentially concurrent with a parallel
study conducted in the PCA laboratories [32]. Abrasion of concrete induced
by the dressing wheel machine closely simulates the rolling, pounding, and
cutting action of steel wheels or the effect of resilient wheels with embedded
nails or stones. Repeated "dynamic" loading through strong impact, compressive forces, rolling, and high speed constitutes the abrasive action of
the ball bearing machine (Fig. 6). Designed by Baker [3], the machine
operates on the principle of a series of eight ball bearings rotating under
load at a speed of 1000 rpm and using water as a cutting aid. During the
test, wear readings are taken every 50 s with a dial micrometer mounted
directly to the supporting shaft allowing readings "on the fly." Readings
are continued for a total of 1200 s or until a maximum depth of 3.2 mm
342
FIG. 4--Dressing wheel machine.

(0.12 in.) or 120 mils is reached. In its abrasive severity the ball bearing
machine exceeds both the revolving disk and the dressing wheel, also producing the highest standard deviation among the three procedures. The
effects of the ball bearing machine on concrete surfaces indicate abrasive
action becomes progressively more severe as the test continues due to soft
and hard spots causing the core barrel to bounce to high angular speed
(Fig. 7). This test procedure has merit in that the ball bearings are similar
to rolling wheels and are more typical of actual loadings by steel wheel
traffic to which a concrete floor may be exposed.
Comparison of ASTM Test C 779 Abrasion Test Machines

The relative performance of the three abrasion testing machines on four
different floor types can be illustrated by the results of a program including
abrasion tests on 305 by 305-mm (12 by 12-in.) samples cut from ibur 1.2 m
by 3.0 m by 92-mm (4 ft by 10 ft by s/a-in.) slabs of the same age, each of a
different type of floor construction (see Table 1).
FIG. S--Dressing wheel machine wear path.
TABLE 1--Composition of various concrete floor slabs and finishing applications.

Slab
Type
Finish
1
2
monolithic
monolithic with cast-on traprock, 9.5 mm (3/8
in.) maximum and final topping of grit size
traprock worked into paste
monolithic with cast-on iron filings and carbon
black worked into paste
76-ram (3-in.) base slab with 15.9 mm (5/8 in.)
thick low water ratio
ordinary commercial
hard steel trowel
3
4
hard steel trowel

hard steel trowel
343
344
FIG. 6--Ball bearing machine.
For a comparison of the effect of the three test methods on the four slabs,
results are plotted (Fig. 8) with the standard deviation as ordinate versus
depth of abrasion as abscissa. It is apparent from this graph that the more
parallel the plotted graph line is to the abscissa the more reproducible is
the test method. As a general trend, the highest deviation is attributed to
the ball bearing machine and the lowest deviation--with exception of slab
4--is noted for the revolving disk machine.
Typical relationships between depth of abrasion versus time, for any test
procedure, are shown in Fig. 9. Characteristically, the first third of any
curve reflects the influence of paste strength, curing, or sealing compounds.
The quality of workmanship is disclosed by the shape of the middle third
of the curve and the final third is indicative of the quality of the aggregate.
Serving as a general guide, Fig. 10 illustrates possible applications of
the four standard procedures for various categories of wear. This tabulation attempts to correlate severity of in-place wear with the abrasive action
345
FIG. 7--Ball bearing machine cutting head and wear path.
particular with each of the test methods. However, it is necessary to bear

in mind that there is overlapping in the type of abrasion imparted by these
test applications and that a specific service condition may be reproduced
by more than one procedure.
Reproducibility
Repeatability of test results will depend largely on the relative uniformity
and consistent quality of surface texture and hardness of the concrete.
Experience with testing of abrasion demonstrates relatively good reproducibility over a fairly wide range of concrete surface types and conditions
when the abrasive action is moderate, for instance, through attrition. More
severe abrasion due to percussion, impact, cutting, or rolling creates erratic
action and is less conducive to reproducibility. While all types of wear can
be simulated by abrasion machines with a fair degree of realism, the rate
of wear differs for various types of in-place abrasion and simulated testing
is accelerated, of necessity. In addition, loss in test accuracy may develop
due to factors in the techniques: (1) rapid wear of dressing wheels, revolving disks, and steel balls and failure to replace them regularly as specified,
(2) failure to remove dust and loose-abraded material regularly from the
346
TESTS
AND
PROPERTIES
SLAB
NO.
OF CONCRETE
SLAB
IO00
NO.
I000
co
.J
IO0
z
/
/
F-
/
J
r~
rY
rr"
;z
IO0
0
I--
1.0
2
<~
I-..co
Ico
1.0
/ t
O I
20
DEPTH
40
OF
60
80
I00
20
A B R A S I O N , MILS
SLAB
NO.
DEPTH
40
OF
NO.
/
/
)00
I00
0
l==
I00
co. I 0 O 0
O9
J
ILl
80
SLAB
I000
60
)0
F-
/
,'-,
/..--.-~
t0
/
a
nr"
o
Z
r~
(Z
/J
//
1.0
I--
r~
<~
I-co
0.1
20
DEPTH
40
OF
60
80
1.0
0,1
IO0
20
ABRASION, MILS
DEPTH
40
OF
60
80
IO0
LEGEND :
----
BALL
BEARING
DRESSING
....
REVOLVING
WHEELS
DISKS
FIG. 8--Abrasive effects of the three test methods in A S T M Test C 799 on various floor
finishes (see Table 1).
347
TIME, M[N,
50
60
90
_J
25
Z
0
II1
50
u_
0
I
75
n
w
O
[00
FIG. 9--Typical time, depth (abrasion) curves.
test surface, (3) improper selection of a representative test area or specimen, (4) differences in age levels of concrete at which test results are compared, (5) arbitrary increase in the length of test or flow of sand, and (6)
insufficient number of wear readings.
Excluding the above variables, the standard deviations shown in Fig. 8
indicate the lowest coefficient of variation is obtained with the revolving
disk method, increasing in order with dressing wheels and ball bearings.
Conclusion
Concrete materials are susceptible to deterioration due to abrasive action

of environmental and man-made factors. In addition, intrinsic conditions
related to properties of concrete and construction techniques contribute to
the reduction in wear resistance. Proper mixture design and selection of
concrete-making materials, curing, and finishing procedures are the basic
requirements applying to any concrete subjected to abrasion from wind
or water-borne particles or vehicular traffic on floors and pavements.
Effects of variables on reduced wear resistance of concrete have proved to
be cumulative which may explain why only certain portions of a concrete
surface may fail in resisting abrasion while the remainder gives satisfactory
performance.
Abrasion testing of concrete in-place or in the laboratory according to
ASTM standard procedures will serve to (1) evaluate, predict, or accept
the quality of concrete surfaces; (2) evaluate specific effects of variables
such as concrete-making materials, curing or finishing procedures, surface
hardeners, coatings or dust-coating materials; (3) compare various types of
348
ASTH C 779
PROCEDURE
~fPE OF D~AR
ABRASIVE EROSION
IMPACT EROSION
CAVITATION
(HYDRAULIC & WIND BLOWN)
LIGHT & Dff[DILrM ]~AFFIC
HEAVY TIRE AND STEEL

~qiEEL TRAFFIC
MOVING OF >FO. RACKS,
\\</- ~ /.-'"
ETC.
/
IiEAk:Y STEEL AND TRAC!I

VEIIICLES
TIRE CtU\INS
HEi\V~" RACKS, ]ETC.
-
FIG. lO--Applications of test procedures.
concrete surfaces under simulated abrasion conditions; and (4) verify

products or systems to meet specifications.
Until recently the problem of evaluating the wear resistance of concrete
was compounded by the absence of standard testing procedures and
machines. ASTM Tests C 418 and C 779 offer four distinct procedures
which simulate various degrees of severity and types of abrasive or erosive
forces to determine depth of wear or the loss of volume of the abraded
material. The lack of sufficiently large populations of test data based on
the proper methodology in the past has hampered the preparation of precision statements for these procedures. Future concern of research dealing
with procedures for testing abrasion resistance should be with the fact that
these allow only an evaluation of a measure of relative quality without any
defined acceptable or unacceptable limits. At present, the absence of standard criteria for wear of concrete surfaces prevents specifications of quality
in terms of abrasion resistance.
349
References
[l] ACI Committee 210, Erosion Resistance of Concrete in Hydraulic Structures (ACI
210-55), Vol. 27, No. 3, Proceedings, Vol. 52, American Concrete Institute, Detroit, 1955.
[2] U.S. Bureau of Reclamation, "Erosion of Concrete by Cavitation and Solids in Flowing
Water," Laboratory Report No. C-342, 1947.
[3] Witte and Backstrom, "Some Properties Affecting the Abrasion Resistance of Air-Entrained Concrete," Proceedings, Vol. 51, 1951, p. 1141.
[4] Kennedy, H. L. and Prior, M. E., Abrasion Resistance, Significance of Tests and Properties of Concrete and Concrete-Making Materials, ASTM STP 169, American Society
tbr Testing and Materials, 1956, pp. 163-174.
[5] Mielenz, R. C., et al., "Origin, Evolution, and Effects of the Air Void System in Concrete," Journal, American Concrete Institute; Proceedings, Vol. 55, No. 1-4, 1958.
[6] Smith, F. L., "Effect of Aggregate Quality on Resistance of Concrete to Abrasion,"
Cement and Concrete, ASTM STP 205, American Society for Testing and Materials,
1956.
[7] Schuman, L. and Tucker, J., Jr., "A Portable Apparatus for Determining the Relative
Wear Resistance of Concrete Floors," Research Paper No. RP 1252, National Bureau
of Standards.
[8] Scripture, E. W., Benedict, S. W., and Bryant, D. E., "Floor Aggregates," Proceedings,
Title 50-18, American Concrete Institute, 1954.
[9] Abrams, D. A., "Wear Test of Concrete," Proceedings, Vol. 21, 1921, p. 1013.
[10] Jackson, F. H. and Pauls, J. T., "Accelerated Wear Tests of Concrete Pavements,"
Proceedings, American Society for Testing and Materials, Vol. 24, Part II, 1924, p. 864.
[11] Pogany, A., "Determining the Abrasion Resistance of Concrete From That of the Mortar
and Aggregates," Zement, Vol. 24, 1935, p. 522.
[12] Kennedy, T. B. and Mather, K., "Correlation Between Laboratory Accelerated Freezing and Thawing and Weathering at Treat Island, Maine," Proceedings, American
Concrete Institute, Vol. 50, 1954, pp. 141-172.
[13] ACI Committee 201, "Durability of Concrete in Service," Chapter 4, Abrasion of Concrete, Journal, American Concrete Institute; Proceedings, Vol. 59, No. 12, Dec. 1962.
[14] ACI Committee 302, Recommended Practice for Concrete Floor and Slab Construction,
(ACI 302-69), American Concrete Institute, Detroit, 1969.
[15] Fentress, Blake, "Slab Construction Practices Compared by Wear Tests," Journal,
American Concrete Institute, July 1973.
[16] Ytterberg, R. F., "Wear Resistance of Industrial Floors of Portland Cement Concrete,"
Journal. American Society of Civil Engineers, Civil Engineering. Jan. 1971.
[17] U.S. Bureau of Reclamation, "Erosion Resistance Tests of Concrete and Protective
Coatings," Concrete Laboratory Report, No. C 445, Feb. 1952.
[18] Sawyer, J. L., "Wear Test on Concrete Using the German Standard Method of Test and
Machine," Proceedings. American Society tor Testing and Materials, Vol. 57, 1957,
pp. 1143-1153.
[19] Spellman, D. L. and Ames, W. H., "Factors Affecting Durability of Concrete Surfaces,"
Materials and Research Department, California Division of Highways, Report No. M&R
250908-1, BRPD-3-7, 1967.
[20] Mayberry, Dean, U.S. Department of Agriculture, "Linseed Oil Emulsion for Bridge
Decks," Agriculture Research Service, Peoria, I11., Sept. 1971.
[21] Portland Cement Association, "Surface Treatments tor Concrete Floors," Concrete Intbrmation ST-37, p. 2.
[22] U.S. Bureau of Reclamation, "The Effect of Various Surface Treatments Using Magnesium and Zinc Fluorosilicate Crystals on Abrasion Resistance of Concrete Surfaces,"
Concrete Laboratory Report No. C 819.
[23] ACI "Guide for the Protection of Concrete Against Chemical Attack by Means of Coatings and Other Corrosion-Resistant Materials," ACI 515, Journal, American Concrete
Institute, Proceedings. Vol. 63, No. 12, Dec. 1966, pp. 1305-1392.
[24] Fowler, D. W., Houston, J. T., and Paul, D. R., "Polymer Impregnated Concrete for
Highway Application," Research Report No. 114-1, Center for Highway Research,
University of Texas, Austin, Feb. 1973.
350
{25] Krukar, M. and Cook, J. C., "The Effect of Studded Tires on Different Pavement and
Surface Texture," Washington State University, Feb. 1973.
[26} Dikeon, J. T., "Radiation Polymerization of Monomers," U.S. Bureau of Reclamation,
Jan. 1970.
[27] Scofield, H. H., "Significance of Talbot Jones Rattler as Test for Concrete in Road
Slabs," Proceedings, Highway Research Board, 1925, p. 127.
[28] Guttman, A., "Abrasion Tests on Concrete," Chem&alAbstracts, Vol. 30, 1936, p. 5750.
[29] Haris, D. H., "Apparatus for Testing the Hardness of Materials," Chemical Abstracts,
Vol. 38, 1944, p. 1338.
[30] Rushing, H. B., "Concrete Wear Study," Report PB 183410, Louisiana Department of
Highways, June 1968.
[31] Davis, R. E. and Troxell, G. E., "Methods of Testing Concrete for Resistance to Abrasion-Cooperative Testing Program," University of California, Berkeley, Oct. 1964,
unpublished report.
[32] Klieger, P. and Brinkerhoff, C. H., "Cooperative Study on Methods of Testing for
Abrasion of Concrete Floor Surfaces," Research and Development Laboratories, Portland Cement Association, Skokie, I11., March 1970, unpublished report.
[33] Baker, R. C., "Description and Development of a Bearing Abrasion Test Machine,"
Chin and Hensolt, Inc., Engineers, San Francisco, July 1970, unpublished report.
H o w a r d Newlon, Jr. t
Chapter 23 Resistance to
Weathering
Introduction
The purpose of this paper is to discuss the significance of currently standardized testing procedures for evaluating the resistance of concrete to
weathering under service conditions. The published literature on the subject
is extensive and only a somewhat superficial treatment can be given in a brief
paper. The reader should consult the similar papers in earlier editions of this
special technical publication by Scholer [1,2], 2 Powers [3], and Arni [4]. A
monograph by Woods [5] presents a comprehensive picture of durability,
and one by Cordon [6] discusses freezing and thawing in detail. Perhaps the
most comprehensive treatment of the influence of aggregates on resistance to
freezing and thawing is that of Larson et al [7]. An extensive annotated
bibliography on concrete durability published by the Highway Research
Board (HRB) in 1957 [8] with a supplement in 1966 [9] contains 534 citations. Unfortunately this bibliography has not been updated. The report of
ACI Committee 201 on Durability of Concrete [10] contains valuable recommendations for production of concrete to provide resistance to the various
destructive processes encountered in field exposures. The ACI Symposium
[11] on durability of concrete contains 17 papers on various theoretical and
operational aspects of producing durable concrete.
In the absence of contact with aggressive fluids or incorporation of aggregates susceptible to detrimental expansion by reaction with alkalies in cement, the resistance of concrete to weathering is- determined by its ability to
withstand the effects of freezing and thawing in the presence of moisture. 3
The testing of concrete for resistance to freezing and thawing had its
genesis in similar tests on building stones reported as early as 1837 by Vicat
1Associate head, Virginia Highway and Transportation Research 'Council, Charlottesville,
Va. 22903.
3Attack by aggressive fluids and cement aggregate reactions are treated elsewhere in this
publication.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
352
[12]. It was only natural that the advent of "artificial stone" in the form of
portland cement concrete would result in corresponding evaluations. Few
published tests from the nineteenth century survive, but when freezing equipment became commercially available for food processing in the 1930s the
reporting of such testing increased. With the recognition that tests conducted
in a variety of freezers designed for other purposes naturally gave variable
results, several agencies such as the U.S. Bureau of Reclamation, Corps of
Engineers, Bureau of Public Roads, Portland Cement Association, and the
National Sand and Gravel Association constructed specialized equipment for
testing concrete by freezing and thawing. Descriptions of some of this equipment will be found in reports by the agencies [3-8]. These efforts marked the
beginning of a systematic quest for standardization and an understanding of
factors influencing the resistance of concrete to freezing and thawing.
There are currently three standard test methods and one recommended
practice under the jurisdiction of ASTM Committee C-9 on Concrete and
Concrete Aggregates intended to aid in evaluating the resistance of concrete
to freezing and thawing. In addition, several petrographic procedures are
standardized that provide invaluable information for predicting the
resistance of concrete to freezing and thawing and for interpreting the results
of exposure in either the laboratory or under field conditions. Two older
freezing and thawing methods have been discontinued. The presently standardized methods as well as those discontinued have evolved over a period of
about 50 years and reflect the inevitable compromise between the need for
rapid assessment of resistance to weathering and the difficulty of translating
the results from accelerated laboratory testing to the varied conditions encountered in field exposures.
Of the currently standardized methods, ASTM Test for Resistance of Concrete to Rapid Freezing and Thawing (C 666) covers the exposure of
specimens to cyclic freezing and thawing. This test, the most widely used of
those available, is a consolidation of two earlier methods (ASTM C 290 and C
291) and provides for two procedures: Procedure A in which both freezing
and thawing occurs with the specimens surrounded by water, and Procedure
B in which the specimens freeze in air and thaw in water. ASTM Test for
Critical Dilation of Concrete Specimens Subjected to Freezing (C 671)
employs a single cycle of cooling through the freezing point with specimens
that are continuously wet. This test was developed to accommodate certain
theoretical criticisms of the cyclic method. ASTM Recommended Practice
for Evaluation of Frost Resistance of Coarse Aggregates in Air-Entrained
Concrete by Critical Dilation Procedures (C 682) describes the use of ASTM
Test C 671 to evaluate the influence of coarse aggregates on the resistance of
concrete to freezing and thawing. ASTM Test for Scaling Resistance of Concrete Surfaces Exposed to Deicing Chemicals (C 672) provides a procedure
for evaluating the effect of deicing chemicals on concrete and the effectiveness of modifications of the concrete or of surface coatings in mitigating
NEWLON ON RESISTANCE TO WEATHERING
353
the detrimental influence of such chemicals. Two of the initially approved

standards were discontinued in 1971. These methods, (ASTM Test for
Resistance of Concrete Specimens to Slow Freezing and Thawing in Water or
Brine (C 292), and ASTM Test for Resistance of Concrete Specimens to Slow
Freezing in Air and Thawing in Water (C 310)), which provided for one cycle
every 24 h, were dropped because of lack of use.
While much is still to be learned, an extensive body of knowledge has been
developed that permits evaluations of concrete to aid in minimizing
premature deteriorations from environmental factors.
While an extensive discussion of the theory and historical development of
freezing and thawing tests is beyond the scope of this summary, a brief treatment is necessary. The reader should consult the cited references for additional information on these subjects.
Historical Evolution
As early as 1837, Vicat, in his famous "Treatise on Calcareous Mortars
and Cements," reported the results from experiments by Brard [19] to
distinguish the building stones that were injured by frost from those that
were not [12]. These tests were conducted "by substituting for the expansive
force of the congealing water, that of an easily crystallizable salt, the sulfate
of soda." In 1856 Joseph Henry reported testing by 50 cycles of freezing and
thawing, samples of marble used in the extension of the U.S. Capitol [20].
During the two decades before and those after the turn of the twentieth
century, various tests of stone, concrete, brick, and other porous materials
were reported. In 1928 Grun [21] in Europe and Scholer [22] in the United
States reported results from accelerated freezing and thawing of concrete in
the laboratory. Following Scholer's initial paper, accelerated laboratory
testing greatly increased. In 1936 [23], 1944 [24], and again in 1959 [25] the
HRB Committee on Durability of Concrete reported results of cooperative
freezing and thawing tests designed to identify factors that influence
resistance of concrete to freezing and thawing. The first two cooperative test
series were conducted betbre any methods were standardized. The 1936
series concentrated upon the influence of cement using mortar prisms incorporating ten commercial cements. The 1944 series of tests used concrete
specimens. Both of these series emphasized the necessity for carefully
regulating the methods of making and curing the specimens, the air content
of the specimens, the degree of saturation of the aggregates at the time of
mixing the concrete, and the degree of saturation of the concrete at the time
of freezing.
In 1951 ASTM Committee C-9 on Concrete and Concrete Aggregates
formed Subcommittee III-O on Resistance to Weathering (presently
C09.03.15) with the major responsibility for proposing methods for
evaluating the resistance to freezing and thawing of aggregates in concrete
354
and standardizing the methods of tests. Drawing heavily upon the results of
the HRB cooperative test series and the experience of several laboratories
that had developed specialized equipment for conducting freezing and thawing tests, in 1952 and 1953 ASTM approved four tentative test methods that
later became standards. These methods were largely representative of the
methods and procedures then in use by the membership of Subcommittee
III-O. The general characteristics of the methods are reflected in their titles:
ASTM Test for Resistance of Concrete Specimens to Rapid Freezing and
Thawing in Water (C 290), ASTM Test for Resistance of Concrete Specimens
to Rapid Freezing in Air and Thawing in Water (C 291), ASTM Test for
Resistance of Concrete to Slow Freezing and Thawing in Water or Brine (C
292), and ASTM Test for Resistance of Concrete Specimens to Slow Freezing
in Air and Thawing in Water .(C 310). The two slow tests were adopted to
cover tests usually conducted in conventional freezers with manual transfer
of specimens between the freezing chamber and thawing tank. These
methods were dropped in 1971 since neither was in general use nor required
by any other ASTM specification. In 1971 the two rapid tests (ASTM C 290
and C 291) were combined as two procedures (A and B) in a single test
(ASTM C 666). In these tests deterioration of specimens is evaluated by the
resonant frequency method (ASTM Test for Fundamental Transverse
Longitudinal and Torsional Frequencies of Concrete Specimens (C 215)).
In the third cooperative test series, initiated in 1954, reported in 1959
[25], soon after standardization of the four methods, 13 laboratories participated using three conerete formulations representing a concrete with
good quality coarse aggregate and adequate air entrainment, one with good
quality coarse aggregate and deficient air entrainment, and one with poor
quality coarse aggregate and adequate air entrainment.
Arni [4] summarized the general conclusions from the 1944 and 1954 test
series as follows:
"1. Methods involving freezing in water were more severe, that is, produced failure in fewer cycles, than were those involving freezing in air.
2. Rapid freezing was more severe than slow freezing when done in air but
not when done in water.
3. Rapid freezing and thawing in water (ASTM Test C 290) appeared to
do the best job of detecting a difference between concretes both of which had
high durability.
4. Only the slow freezing in air and thawing in water method (ASTM Test
310) was able to discriminate adequately between concretes of low durability.
5. In general, the four methods tended to rate different concretes in the
same order of durability when there was a significant difference."
One of the difficulties with the four methods which was emphasized
especially by the HRB cooperative programs was poor repeatability and
reproducibility of results within and between laboratories. Good
355
reproducibility generally was obtained only for concretes which were very
high or very low in durability. For concretes in the middle range, a wide
spread in results was obtained. While the variabilities are large in the middle
range they are amenable to establishing a precision statement which has been
incorporated in ASTM Test C 666 as discussed later.
As noted earlier, Subcommittee III-O was formed initially in 1951 wih the
primary mission of developing freezing and thawing tests to be applied to the
evaluation of coarse aggregates. In 1961 the subcommittee reported t h a t
standardization of such tests was not warranted because of the high levels of
variability associated with the existing methods. However, in an unpublished
report the subcommittee did outline procedures for using ASTM Test C 290
if evaluations of coarse aggregates were required [26].
Lack of a standardized test method for evaluating resistance to freezing
and thawing of coarse aggregates in concrete coupled with theoretically
based criticisms of the cyclic methods, particularly by Powers [27], as
discussed later, led to the development of methods designed to determine the
length of time required for an aggregate to become critically saturated in
concrete. Critical saturation was defined when specimens exposed to continuous soaking and subjected to a cycle of cooling through the freezing point
exhibited dilation greater than a specified value. This method was first used
by the California Department of Highways [28] and later was refined and extensively evaluated by Larson and his co-workers [29]. The literature survey
prepared by Larson et al as part of the research is a particularly valuable
reference on all aspects of freezing and thawing studies related to aggregates
[7]. As an outgrowth of this work, in 1971 ASTM Test C 671 was approved
along with ASTM Recommended Practice C 682, which provides guidelines
for applying ASTM Test C 671 to the evaluation of coarse aggregates.
More recently, agencies studying "D-cracking" of concrete have generally
found a modification of the older procedure (ASTM Test C 666) to be useful
in identifying coarse aggregates susceptible to this type of deterioration. According to Stark [30] aggregates susceptible to causing distress in concrete
pavements can be identified by freezing and thawing of concrete in water at a
rate of two cycles per day.
Concern with surface mortar deterioration or deicer scaling, particularly
on highway and bridge deck pavements, led in 1971 to standardization of
ASTM Test C 672, which combined features of methods that had been
developed and used by various agencies for a number of years. The method
uses blocks fabricated to permit ponding of water on surfaces that are subjected to freezing and thawing in the presence of various deicing agents.
Of the methods, ASTM Test C 666 continues to be the most widely used.
Specialized equipment is commercially available for conducting the tests
under controlled conditions, but the essential elements of the method are
those that have been used for more than a century.
356
Theoretical Considerations
Cyclic freezing and thawing tests were developed on a pragmatic rather
than a theoretical basis. It was assumed that the destruction resulted from
the 9 percent volume expansion accompanying the conversion of water to ice,
and that this process was reproduced artificially in the laboratory environment. Powers and his co-workers, from their comprehensive study of the
structure of cement paste in 1945 [31], advanced the hypothesis that the
destructive stress is produced by the flow of displaced water away from the
region of freezing, the pressure being due to the viscous resistance of such
flow through the permeable structure of the concrete. According to this
theory, when the flow path exceeds a critical length, the pressure exceeds
the strength of the paste. Such flow would occur when the water content
exceeds the critical saturation point. This concept is called the hydraulic
pressure theory. The theory was amplified in 1948 [32] to explain the beneficial influence of entrained air. Since the resistance to flow at a given rate
is proportional to the length of the flow path, the air bubbles were conceived
as spaces into which the excess water produced by freezing could be expelled
without generating destructive pressures. Powers calculated a critical
thickness of the order of 0.25 mm (0.01 in.), a value that was approximately
the same as that suggested by Mielenz and his co-workers from experimental
studies [33]. Powers' initial studies suggested that the hypothesis advanced
by Taber [34] to explain frost heaving of soils did not apply to mature concrete. By this hypothesis, the stress is produced not by hydraulic pressure,
but by the segregation of ice into layers which enlarge as unfrozen water is
drawn toward the region of freezing rather than forced away from it.
Studies by Verbeck and Landgren [35] as well as those of Powers [36]
make clear that the paste and aggregate should be considered separately
when explaining the resistance of concrete to freezing and thawing. This is
because the paste not only may become critically saturated by moisture from
external sources but also must withstand pressure generated by water expelled from the aggregate particles during freezing.
The temperature at which water freezes in various pores within the paste
decreases with the size of the pore so that even if the concrete is at a uniform
temperature throughout, the water will be at various stages of conversion to
ice. As the water freezes, the solution in the pore becomes more concentrated. The existence of solutions of various concentrations in the pores of the
paste causes unfrozen water to move to the site of freezing in order to lower
the concentration made higher at the freezing site than the more dilute solution of the unfrozen water. This flow generates stress somewhat like osmotic
pressures, hence the designation "osmotic pressure hypothesis."
Powers, in a recent summary of his and other research [37], concludes that
in general all three of the theories, with some modification, are required to
account for the behavior of concrete subjected to freezing and thawing.
357
When freezing and thawing takes place in the presence of deicing

chemica]s, localized failures of the exposed surfaces occur that is called
"scaling," or surface mortar deterioration. Except where the concentration
of deicing chemicals is high enough to cause chemical attack, scaling results
from freezing and thawing. Various theories have been advanced to explain
the increased severity of the damage as compared with freezing and thawing
in water. Browne and Cady [38] have recently summarized these theories and
their own experiments. Although definitive answers have not been obtained,
the mechanism is probably most influenced by concentration gradients. In
addition to the lowering of the freezing temperature that accompanies
increased deicer concentration, flow of water from areas of lower to those of
higher concentrations generate stresses such as were described earlier.
Verbeck and Klieger [39], in work confirmed by others, found that deterioration was greater for intermediate concentrations of deicing chemical (3 percent to 4 percent) than for lower or higher concentrations (to 16 percent).
Such findings can be explained by generation of osmotic-like pressures accompanying the movement of water between areas of varying concentrations.
Rapid Freezing and Thawing Tests

In 1971 ASTM Tests C 290 and C 291 were combined as Procedures A and
B in ASTM Test C 666. The method is designated as "rapid" because it permits alternately lowering the temperature of specimens from 4.4 to --17.8~
(40 to 0~ and raising it from --17.8 to 4.4~ (0 to 40~ in not less than 2
nor more than 4 h. Thus a minimum of 6 and a maximum of 12 complete
cycles may be achieved during a 24-h period. The conventionally accepted
term of testing is 300 cycles, which can be obtained in 25 to 50 days. For Procedure A both freezing and thawing occur with the specimens surrounded by
water, while in Procedure B the specimens freeze in air and thaw in water.
For Procedure A, the thawing portion is not less than 25 percent of the total
cycles, while for Procedure B not less than 20 percent of the time is used for
thawing. The time required for the temperature at the center of any single
specimen to be reduced from 2.8 to --16.1~ (37 to 3~ shall be not less
than one half of the length of the cooling period, and the time required for
the temperature at the center of any single specimen to be raised from --16.1
to 2.8~ (3 to 37~ shall be not less than one half of the length of the heating
period. For Procedure A each specimen is surrounded by approximately 3
mm (1/8 in.) of water during the freezing and thawing cycles, while in Procedure B the specimen is surrounded completely by air during the freezing
phase of the cycle and by water during the thawing phase. The requirements
for Procedure A are met by confining the specimen and surrounding water in
a suitable container. The specimens are normally prisms not less than 76 mm
(3 in.) nor more than 137 mm (5 in.) in width and depth, and between 356
and 406 mm (14 to 16 in.) long.
358
During the early years of testing by freezing and thawing, laboratories constructed specialized equipment. In recent years, several types of equipment
have become available commercially that meet the requirements for ASTM
Test C 666. These handle from 20 to 60 specimens. Custom-built freezers can
hold 100 specimens. While some aspects of these machines have been subjected to criticism, they meet the needs for rapid testing within practical
limits.
Deterioration of specimens is determined by the resonant frequency
method, ASTM Test C 215. The fundamental transverse frequencies are
determined at intervals not exceeding 36 cycles of exposure and used to
calculate the relative dynamic modulus of elasticity
P~
= (hi2)
\--~--/ 100
where
P,. -----relative dynamic modulus of elasticity, after c cycles of freezing and
thawing, in percent,
n = fundamental transverse frequency at 0 cycles of freezing and thawing, and
nt = fundamental transverse frequency after c cycles of freezing and
thawing.
The fundamental transverse frequency is determined with the specimens at
a temperature of 5.6 +__ 2.8~ (42 +_ 5~
Calculation of Pc assumes that the
weight and dimensions of the specimens remain constant throughout the
test. While this assumption is not true in many cases because of disintegration, the test is usually used to make comparisons between the relative
dynamic moduli of specimens and Pc is adequate for the purpose.
The durability factor is calculated as
DF = PN/M
where
D F -----durability factor of the specimen,
P = relative dynamic modulus of elasticity at N cycles, percent,

N = number of cycles at which P reaches the specified minimum value
for discontinuing the test or the specified number of cycles at which
the exposure is to be terminated, whichever is less, and
M = specified number of cycles at which the exposure is to be terminated.
359
Because of the danger of damage to specimen containers and other parts of

the equipment, testing usually is terminated when the "relative dynamic
modulus of elasticity falls below 50 percent.
The scope of ASTM Test C 666 states that "both procedures are intended
for use in determining the effects of variations in the properties of concrete
on the resistance of the concrete to the freezing and thawing cycles specified
in the particular procedures. Neither procedure is intended to provide a
quantitative measure of the length of service that may be expected from a
specific type of concrete."
Procedure A is currently required in three ASTM specifications; namely,
ASTM Specification for Air-Entraining Admixtures for Concrete (C 260),
ASTM Specification for Chemical Admixtures for Concrete (C 494), and
ASTM Specification for Lightweight Aggregates for Structural Concrete (C
330). In both admixture specifications, the performance requirement is
stated in terms of a "relative durability factor" calculated as
DF
(or D F O = P N / 3 0 0
and
RDF = (DF/DFO
100
where
durability factor of the concrete containing the admixture under
test,
= durability factor of the concrete containing a reference admixture
(or in the case of ASTM Specification C 494, only an approved airentraining admixture),
= relative dynamic modulus of elasticity in percent of the dynamic
modulus of elasticity at zero cycles (values of P will be 60 or
greater since the test is to be terminated when P falls below 60
percent),
-----number of cycles at which P reaches 60 percent, or 300 if P does
not reach 60 percent prior to the end of the test (300 cycles), and
= relative durability factor.
DF =
DF1
N
RDF
Both ASTM Specifications C 260 and C 494 require that the relative
durability factor of the concrete containing the admixture under test be at
least 80 when compared with the reference concrete. The value of 80 is not intended to permit poorer performance than the reference concrete, but rather
to assure the same level of performance, with appropriate recognition of the
variability of the test method. The value of 80 was established before levels of
precision were established for ASTM Test C 666, but is consistent with
recently published precision values.
One criticism of rapid freezing and thawing tests has been variability of the
360
results both within and between laboratories. In response to the requirement

by ASTM for precision statements in all test methods, ASTM Subcommittee
C09.03.15 reviewed the data from major published studies. Statistical
analyses of these data, as described by Arni [40], showed that the variability
was primarily a function of the level of durability factor of the concrete for
the ranges of N and P normally used. In 1976 a precision statement was
added to ASTM Test C 666 for expected within batch precision for both Pro,
cedures A and B. Values for standard deviation (1S) and acceptable range
(D2S) as defined by ASTM Recommended Practice for Preparing Precision
Statements for Test Methods for Construction Materials (C 670) are given in
Tables 2 and 3 of ASTM Test C 666 for ranges of average durability factor in
ten increments and numbers of specimens averaged from 2 through 6. Values
for 6 specimens (as required by ASTM Specifications C 260 and C 494) are
given in Table 1.
TABLE 1-- Within-laboratory precisions for averages of

6 beams tested in accordance with A S T M Standard C 666.
Procedure A
Range of
Average
Durability
Factor
Standard
Deviation
(1S)
0 to 5
5 to 10
10 to 20
20 to 30
30 to 50
50 to 70
70 to 80
80 to 90
90 to 95
Above 95
0.3
0.6
2.4
3.4
5.2
6.4
4.7
2.3
0.9
0.4
Acceptable
Range
(D2S)
0.9
1.8
6.8
9.7
14.7
18.1
13.4
6.5
2.4
1.3
Procedure B
Standard
Deviation
(1S)
0.4
1.7
3.3
4,3
6,3
8.2
7.0
3.6
1.6
0.8
Acceptable
Range
(D2S)
1.2
4.7
7.4
12.2
17.8
23.2
19.7
6.8
4.5
2.3
These values confirm the long recognized facts that the variability is less
for very good or very poor concrete than for concretes of intermediate
durability and that results from Procedure A are somewhat less variable than
those obtained from Procedure B. While Procedure A of ASTM Test C 666 is
specified in ASTM Specification C 330, no values are given. The specification states that "in the absence of a proven record of satisfactory durability in
structural concrete, lightweight aggregates may be required to pass a concrete freezing and thawing test satisfactory to the purchaser." ASTM Test C
666 is listed as a method of sampling and testing in ASTM Specification for
Concrete Aggregates (C 33), but is not mentioned elsewhere in the document.
As noted, the primary measure of deterioration is the relative dynamic
361
modulus calculated from determinations of resonant transverse frequency.

Two other characteristics that are commonly used are weight loss and length
change. Determination of reductions in tensile or compressive strength have
been used infrequently because such testing is destructive.
The question of which measure of deterioration is best to use is complicated by the facts that the different available tests measure different
things, and the manner and extent to which the properties measured are
related to freezing-and-thawing damage, especially under natural conditions, are matters of disagreement. Thus, the particular measure used often
depends on the philosophy of the laboratory using it and on the particular
purpose for which the tests are being made. Weight loss measures loss of
material or sloughing from the surface of the specimens. Length change is
based upon the fact that internal damage is accompanied by expansion
rather than contraction during cooling or by a permanent dilation after a
freezing and thawing cycle, as discussed by Powers [27]. Either weight loss
due to sloughing or reduction in resonant frequency may occur without accompanying length change, and in Procedure A weight loss often occurs
without a reduction in resonant frequency or expansion. Resonant frequency
and expansion reflect internal disruptions that are caused by unsound aggregates or deficient air void characteristics, while weight loss reflects
primarily surface mortar deterioration. ASTM Subcommittee C09.03.15 is
considering adding requirements for length change measurements to ASTM
Test C 666, which presently requires weight and resonant frequency determinations.
One important influence that is not covered in ASTM Test C 666 is the effect of the container used to hold the specimen during testing with Procedure A. A note states that "for Procedure A, this requirement contemplates
that the specimen and the surrounding water will be kept in a suitable container while in the freezing-and-thawing apparatus. Experience has indicated that ice or water pressure, during freezing tests, particularly in equipment that uses air rather than a liquid as the heat transfer medium, can
cause excessive damage to rigid metal containers, and possibly to the specimens therein. Results of tests during which bulging or other distortion of
containers occurs should be interpreted with caution." This situation is
particularly noticeable when the specimens are exposed with the long
dimension vertical. Ice forms quickly on the open top rather than uniformly along the long dimension of the container. Pressures from continued conversion of water between the specimen and container cannot
be relived with consequent bulging of the container. Various adjustments
have been made to mitigate this problem, including exposure of the specimens in containers with the long dimensions horizontal to provide a larger
open area; inclusion of flexible windows, or corners, in containers; or
inclusion of a rubber ball in the bottom of containers to absorb the pres-
362
sures. All of these have been reported to overcome the problem by some
and not by others. That the degree of restraint offered by the container
influences the number of cycles required to reach a specified level of relative dynamic modulus is clear as shown by Cook [41], who found that
the number of cycles necessary to reach a given level of durability factor
was increased dramatically when the specimens were tested in rubber
containers as compared with containers made of steel.
Another important influence on the results is the degree of saturation of
the concrete and the aggregates both at the time of mixing and throughout
the course of the testing. ASTM Test C 666 and the specifications citing its
use require that testing begin after 14 days of moist curing. The mixing procedures referenced in ASTM Method for Making and Curing Concrete Test
Specimens in the Laboratory (C 192) require that coarse aggregates be
immersed for 24 h prior to mixing. Thus the comparatively high degree of internal saturation and the early age at which testing begins result in a relatively severe test when compared with field exposures in which a period of
drying normally occurs before exposure to freezing and thawing. Even a brief
period of drying greatly improves resistance to freezing and thawing, since
it is difficult to resaturate concrete that has undergone some drying. But
accommodating the many possible variations of saturation that might be
encountered is impractical so that the most consistently reproducible condition is that of continued moist curing.
Because the testing is initiated at a fixed age, considerable variation of
strength at the time of exposure to freezing and thawing may be encountered
with cements of different strength gain characteristics for concrete with
sound aggregates and satisfactory air void characteristics. There is not a
great body of data on the influence on resistance to freezing and thawing of
strength at the time that exposure begins. Buck et al [42] reported tests
indicating that because of its relationship with the amount of freezable
water, a given level of maturity (strength) was necessary to provide an acceptable degree of frost resistance as indicated by a durability factor of 50 for
concrete containing satisfactory aggregates and entrained air. They cited
earlier work by Klieger [43], who reported similar findings in his studies of
salt scaling. Consideration of strength at the time of initial freezing is particularly important in testing concrete made with blended cements and
pozzolanic admixtures that gain strength more slowly than concrete without
such admixtures. This influence is minimized when the evaluations are
made by comparing concretes made with similar materials as required in
ASTM Specifications C 260 and C 494.
Cyclic freezing and thawing methods were developed and applied only for
laboratory mixed concrete until 1975, when they were extended to cores or
prisms cut from hardened concrete. Experience with testing of specimens
from hardened concrete is limited.
363
Dilation Methods
In 1955 Powers published a critical review of existing cyclic methods for

freezing and thawing tests [27]. He was particularly critical of what he considered unrealistically high freezing rates of from 6 to 60~ (10 to 100~
as compared with natural cooling rates that seldom exceed 3~ (5~
He
also noted the significance of moisture conditioning of aggregates and concretes in test methods which generally provided for only saturated concrete as
compared with natural exposures where some seasonal drying is possible.
He stated that durability was not a measureable property but that expansion occurring during a slow cooling cycle when the concrete or its aggregates
became critically saturated was measurable and would provide an indication
of potential resistance to damage by freezing and thawing. He proposed that
specimens be prepared and conditioned so as to simulate field conditions and
then be subjected to periodic slow rate freezing and storage in water at low
temperature between freezing exposures. Concrete subject to frost damage
should reach some critical saturation level, after which it would expand on
freezing. The length of time required to become critically saturated would be
compared with the field exposures to be encountered. If the period during
which freezing would be expected was less than the time to reach critical
saturation, then no damage would be expected. As opposed to a single
durability factor, the time to critical saturation could be more readily interpreted for various field exposures.
The California Division of Highways was first to report, in 1961, a practical application of Power's proposed method [28]. They developed specimen
preparation and conditioning methods, testing and measuring techniques,
and performance criteria. The method was used to evaluate several
aggregates for a major highway construction project. The aggregates judged
acceptable by the California procedure would have been rejected by other
conventionally accepted criteria. The performance of this concrete has been
reported as satisfactory after 15 years of service.
The procedures proposed by Powers and the method developed in California were extensively studied and refined by Larson and others [29]. In 1971
ASTM Test C 671 was standardized. This method provides for cylindrical
specimens 75 mm (3 in.) in diameter and 150 mm (6 in.) long that are stored
in water at 1.7 ___ 9~ (35 ___ 2~
At two-week intervals the specimens are
cooled in water-saturated kerosene at a rate of 2.8 _ 0.5~
(5 ___ 1 ~
During the cooling cycle the specimen is placed in a strain frame to permit
measurement of length change. A typical plot of length change versus
temperature is shown in Fig. 1. Prior to critical saturation the length change
will proceed along the dashed curve without dilation. Critical dilation is
defined as a sharp increase (by a factor of 2 or more) between dilations on
successive cycles. Highly frost-resistant concrete may never exhibit critical
364
200
-400
w
t.9
15z
-600
T
O
-800
Z
kd
-J
E"
25
(.9
Z
bJ
.J
G
o ~
35
:E
I
o
TIME
(H)
FIG. 1--Typical length change and temperature charts.
dilation. ASTM Recommended Practice for Evaluation of Frost Resistance

of Coarse Aggregates in Air-Entrained Concrete by Critical Dilation Procedures (C 682) was also standardized in 1971. This practice is based largely
upon the work of Larson and his co-workers. Procedures essentially in accordance with ASTM Test C 671 have been used by Buck [44] and he found that
a specimen that is frost resistant will now show increasing dilation with continuously decreasing temperature; it will show some limited initial dilation as
all moist specimens do, but dilation will not continue throughout the cooling.
In addition to ASTM Test C 671 criterion for critical dilation in terms of
increase from one cycle to another, he suggested that a criterion for critical
dilation applicable to results of a single test be as follows:
(a) If the dilation is 0.005 percent ( = 50 millionths) or less, the specimen
may be regarded as frost resistant, that is, the dilation is not critical.
(b) If the dilation is 0.020 percent ( = 200 millionths) or more, the
specimen may be regarded as not frost resistant, that is, critical dilation has
been exceeded.
(c) If the dilation is in the range between 0.005 percent and 0.020 percent,
an additional cycle or more should be run.
ASTM Test C 671 and Recommended Practice C 682 have not been used
extensively. Although the apparatus is comparatively inexpensive and larger
numbers of specimens can be processed than with ASTM Test C 666 equipment, significant storage capacity is required. The procedures of ASTM
Recommended Practice C 682 are extensive and complex, largely because of
365
requirements designed to bracket a broad range of potential exposure conditions. It is suggested that aggregates and concrete be "maintained or brought
to the moisture condition representative of that which might be expected in
the field." However, it is noted that "aggregate moisture states other than
dry or saturated are very difficult to maintain during preparation of
specimens. Reproducibility of overall test results is likely to be affected
adversely by variability in aggregate moisture." While the complexity of the
evaluation procedure limits its general applicability, where large projects or
economic consequences of detailed aggregate evaluation are warranted, the
procedures may be justified.
Scaling Resistance
In the early 1960s it became apparent that the increasing use of deicing
chemicals as part of a "bare pavement" policy adopted for the nation's
highways was being reflected in widespread surface scaling of pavements and
bridge decks. It has long been known that dense, high quality concrete, with
adequate entrained air and with adequate curing and a period of drying
before the first application of deicing agents are essential in preventing
damage [I0]. Widespread scaling demonstrated that all of these requirements were not being met consistently and brought forth a plethora of
remedial or preventive products including admixtures, surface treatments,
and curing agents. In 1971 ASTM Test C 672 was standardized. The test is
based upon experience of a number of agencies who used blocks fabricated to
permit ponding of water on one surface which could be exposed to freezing
and thawing in the presence of deicing agents. The test "covers determination of the resistance of scaling of a horizontal concrete surface subject to
freezing and thawing cycles in the presence of deicing chemicals. It is intended for use in evaluating the surface resistance qualitatively by visual
examination. The test can be used to evaluate the effect of mix design, surface
treatment, curing or other variables on resistance to scaling. It is not intended as a test method for durability of aggregates or other ingredients of
concrete."
The specimens must have a surface area of at least 0.046 m 2 (72 in. 2) and
be at least 75 mm (3 in.) deep. The specimens are placed in a freezing space
after moist curing for 14 days and air storage for 14 days. Provisions are
included for applications of protective coatings if desired at the age of 21 days.
The method calls for covering the surface with approximately 6 mm (1A in.)
of a solution of calcium chloride and water having a concentration such that
each 100 ml of solution contains 4 g of anhydrous calcium chloride.
Modifications of the deicer and application procedures, including freezing of
water and addition of the solid deicer, are allowed where there is need to
evaluate the specific effect.
The specimens are placed in a freezing environment for 16 to 18 h and in
366
laboratory air for 6 to 8 h. While the method describes laboratory procedures, it has been used for outdoor exposures as well.
The specimens are rated visually according to a scale from 0 (no scaling)
through 5 (severe scaling) after 5, 10, 15, 25, and every 25 cycles thereafter.
Some investigators have weighed the detritus but this is not required by the
method.
As noted in the method the ratings are ranks, and as such may not be
subjected appropriately to analyses based on the calculation of averages and
standard deviations nor other techniques that assure continuous distributions. If groups of similar specimens are to be reported or compared with
other groups, such nonparametric quantities as the median and range may
be used.
Experience with ASTM Test C 672 has been generally satisfactory for
evaluating the variables for which it was developed.
Other Weathering Processes

Other weathering processes that have been suspected of causing deterioration include heating and cooling and wetting and drying. As noted, chemical
attack and alkali aggregate reactions are treated elsewhere in this publication. Although cases have been reported where deterioration was attributed
to aggregate with an abnormally low coefficient of thermal expansion [45] or
where the freezing and thawing resistance was influenced by aggregates with
different coefficients of thermal expansion [46,47], it is now generally believed that "thermal incompatibility" has a minor, if any, effect on concrete
durability within the normal temperature range. Elevated temperatures such
as are encountered in certain parts of nuclear construction are beyond the
scope of this paper.
While wetting and drying induce variation in moisture content that influences the resistance to freezing and thawing, aggregates cracking from excessive drying shrinkage, and an increased concentration of dissolved salts,
all of which reduce resistance to weathering, the writer is not aware of cases
where alternate wetting and drying per se has caused deterioration.
Summary
The resistance of concrete to weathering in the absence of chemical attack
or detrimental cement-aggregate reactions depends upon its ability to resist
freezing and thawing. Dry concrete will withstand freezing and thawing indefinitely whereas highly saturated concrete, even if air entrained, may be
severely damaged by a few cycles. Research and experience have shown that
resistance to freezing and thawing requires a low water/cement (w/c) ratio,
an adequate volume of entrained air with the proper void distribution and
characteristics, and exposures which reduce the opportunity for critical
367
saturation. While direct translation of results from laboratory freezing and

thawing tests is difficult at best because of the variety of exposures encountered, the currently approved standards are very useful when properly
conducted and interpreted and have undoubtedly resulted in a significant
improvement in the resistance of concrete to weathering from natural forces.
References
[1] Scholer, C. H., "Durability of Concrete," Report on Significance of Tests of Concrete and
Concrete Aggregates, A S T M STP 22,4, American Society for Testing and Materials, 1943.
[2] Scboler, C. H., "Hardened Concrete, Resistance to Weathering--General Aspects,"
Significance of Tests and Properties of Concrete and Concrete Aggregates. A S T M STP
169, American Society for Testing and Materials, 1955.
[3] Powers, T. C., "Resistance to Weathering--Freezing and Thawing," Significance of
Tests and Properties of Concrete and Concrete Aggregates, A S T M STP 169, American
[4] Arni, H. T., "Resistance to Weathering," Significance of Tests and Properties of Concrete
attd Concrete-Making Materials. A S T M STP 169A, American Society for Testing and
Materials, 1966.
[5] Woods, Hubert, "Durability of Concrete Construction," ACI Monograph No. 4, American
Concrete Institute, 1968.
[6] Cordon, W. A., "Freezing and Thawing of Concrete-Mechanisms and Control," ACI
Monograph No. 3, American Concrete Institute, 1966.
[7] Larson, T., Cady, P., Franzen M., and Reed, J., "A Critical Review of Literature Treating
Methods of Identifying Aggregates Subject to Destructive Volume Change When Frozen in
Concrete and a Proposed Program of Research," Special Report 80, Highway Research
Board, 1964.
[8] "Durability of Concrete: Physical Aspects," Bibliography 20, Highway Research Board,
NAS-NRC Publication 493, 1957.
[9] "Durability of Concrete: Physical Aspects: Supplement to Bibliography No. 20/' Bibliography 38, Highway Research Board, NAS-NRC Publication 1333, 1966.
[10] ACI Committee 201, "Guide to Durable Concrete," Journal, American Concrete Institute,
Dec. 1977.
[11] "Durability of Concrete," Publication SP-47, American Concrete Institute, 1975.
[12] Vicat, L. J., "Treatise on Calcareous Mortars and Cements," (translated from the French
by J. T. Smith), J. Weale, London, 1837.
{t3] "Symposium on Freezing-and-Thawing Tests of Concrete," Proceedings, American Society
for Testing and Materials, Vol. 46, 1946.
114] Wuerpel, C. E., and Cook, H. K., "Automatic Accelerated Freezing-and-Thawing Apparatus for Concrete," Proceedings, American Society for Testing and Materials, Vol. 45,
1945.
[15] Walker, S., and Bloem, D. L., "Performance of Automatic Freezing-and-Thawing Apparatus for Testing Concrete," Proceedings, American Society for Testing and Materials,
Vol. 51, 1951.
[16] Arni, H. T., Foster, B. E., and Clevenger, R. A., "Automatic Equipment and Comparative Test Results for the Four ASTM Freezing-and-Thawing Methods for Concrete," Proceedings, American Society for Testing and Materials, Vol. 56, 1956.
[17] Cordon, W. A., "Automatic Freezing-and-Thawing Equipment for a Small Laboratory,"
Bulletin No. 259. Highway Research Board, NAS-NRC Publication 768, 1960.
[18] Cook, H. K., "Automatic Equipment for Rapid Freezing-and-Thawing of Concrete in
Water," Bulletin No. 259, Highway Research Board, NAS-NRC Publication 768, 1960.
[19] Brard, On the Method Proposed by Mr. Brard for the Immediate Detection of Stones
Unable to Resist the Action of Frost, Hericart de Thury, Annales de Chimie et de Physique,
Vol. 38, 1828, pp. 160-192.
[20] Henry, Joseph, "On the Mode of Testing Building Materials and an Account of the Marble
Used in the Extension of the U.S. Capitol," American Journal of Science, Vol. 72, 1856.
368
[21] Grun, R., "Investigations of Concrete in Freezing Chambers," Zement, Vol. 17, 1928.
[22] Scholer, C. H., "Some Accelerated Freezing and Thawing Tests," Proceedings, American
Society for Testing and Materials, Vol. 28, 1928.
[23] Mattimore, H. S., "Durability Tests of Certain Portland Cements," Proceedings, Highway
Research Board, Vol. 16, 1936.
[24] Withey, M. O., chairman, "Progress Report, Committee on Durability of Concrete," Proceedings, Highway Research Board, Vol. 24, 1944.
[25] Foster, B. E., chairman, "Report on Cooperative Freezing-and-Thawing Tests of Concrete," Special Report No. 47. Highway Research Board, 1959.
[26] ASTM Committee C-9, Minutes of December 1961 Meeting.
[27] Powers, T. C., "Basic Considerations Pertaining to Freezing-and-Thawing Tests,"
Proceedings, American Society for Testing and Materials, Vol. 55, 1955.
[27] Tremper, B. and Spellman, D. L., "Tests for Freeze-Thaw Durability of Concrete Aggregates," Bulletin 305, Highway Research Board, 1961.
[29] Larson, T. D. and Cady, P. D., "Identification of Frost-Susceptible Particles in Concrete
Aggregates," NCHRP Report No. 66, Highway Research Board, 1969.
[30] Stark, David, "Characteristics and Utilization of Coarse Aggregates Associated with
D-Cracking," Living with Marginal Aggregates, ASTM STP 597, American Society for
Testing and Materials, 1976, p. 45.
[31] Powers, T. C., "A Working Hypothesis for Further Studies of Frost Resistance of Concrete," Journal, American Concrete Institute, Feb. 1945.
[32] Powers, T. C., "The Air-Requirement of Frost Resistant Concrete," Proceedings. Highway Research Board, Vol. 29, 1949.
[33] Mielenz, R. C., Wolkodoff, V. E., Burrows, R. W., Backstrom, J. L., and Flack, H. E.,
"Origin, Evolution and Effects of the Air Void System in Concrete," Journal, American
Concrete Institute, July 1958. Aug. 1958, Sept. and Oct. 1958.
[34] Taber, Stephen, "The Mechanics of Frost Heaving," The Journal of Geology. Vol. 38,
No. 4, 1930.
[35] Verbeck, George and Landgren, Robert, "Influence of Physical Characteristics of Aggregate on the Frost Resistance of Concrete," Proceedings. American Society for Testing
and Materials, Vol. 60, 1960.
[36] Powers, T. C., "The Mechanism of Frost Action in Concrete," Stanton Walker Lecture
No. 3. National Sand and Gravel Association, 1965.
[37] Powers, T. C., "Freezing Effects in Concrete," Durability of Concrete. Publication SP-47,
American Concrete Institute, 1975.
[38] Browne, F. P., and Cady, P. D., "Deicer Scaling Mechanisms in Concrete," Durability
of Concrete, Publication SP-47, American Concrete Institute, 1975.
[39] Verbeck, G. J. and Klieger, P., "Studies of 'Salt' Scaling of Concrete," Bulletin 150.
[40] Arni, H. T., "Precision Statements without an Interlaboratory Test Program," Cement,
Concrete, and Aggregates, to be published Jan. 1979.
[41] Cook, H. K., "Effects of Fluid Circulation and Specimen Containers on the Severity of
Freezing and Thawing Tests," informal presentation, Session 13, Annual Meeting of
Highway Research Board, Jan. 1963.
[42] Buck, A. D., Mather, B., and Thornton, H. T., Jr., "Investigation of Concrete in Eisenhower and Snell Locks St. Lawrence Seaway," Technical Report 6-784, Waterways Experiment Station, July 1967.
[43] Klieger, Paul, "Curing Requirements for Scale Resistance of Concrete," Bulletin 150.
[44] Buck, A. D., "'Investigation of Frost Resistance of Mortar and Concrete," Technical
Report C-76-4, U.S. Army Engineer Waterways Experiment Station, Vicksburg, Miss.
[45] Pearson, J. C., "A Concrete Failure Attributed to Aggregate of Low Thermal Coefficient,"
Proceedings, American Concrete Institute, Vol. 48, 1952.
[46] Callan, E. J., "Thermal Expansion of Aggregates and Concrete Durability," Proceedings.
American Concrete Institute, Vol. 48, 1952.
[47] Higginson, E. C. and Kretsinger, D. G., "Prediction of Concrete Durability from Thermal
Tests of Aggregate," Proceedings, American Society for Testing and Materials, Vol.
53, 1953.
L. H. T u t h i l l ~
Chapter 24--Resistance to Chemical

Attack
Introduction
Fortunately most concrete in service is not subjected to chemical attack.
Concrete is generally less able to resist successfully corrosion of this kind
than disintegration by other forces. By chemical attack is meant leaching
and acid or sulfate attack.
Although concrete is often disfigured by deposits of efflorescence [1] 2 as
lime water leach reaches outside surfaces, very little good-quality concrete
has actually been destroyed or made unserviceable by leaching of lime from
internal or external surfaces. In addition to good workmanship, probably
the best initial correctives are either an ample cement content with a lowwater content and water/cement (w/c) ratio or a cementitious mixture of
portland cement and a good, active pozzolanic material to minimize permeability.
Acid attack is commonly encountered in sewers above the flow line, on
floors of food processing plants, and elsewhere. In the case of sewers, attack
on concrete is by sulfuric acid; little can be done about it directly but much
can be done to prevent it by proper sewer design. The lactic and acetic acid
attacks are milder and can be minimized by good workmanship, good concrete, and a good pozzolanic material added to the cement to make the
lime less subject to attack by acids.
Sulfate attack can completely disintegrate average concrete made with
ordinary cement in a very few years (Figs. 1, 2, and 3). Many field structures
have become unserviceable from disintegration by sulfate salts, usually of
sodium, magnesium, or possibly calcium, which exist naturally at varying
intensity in many soils, shales, and ground waters. This chemical attack
can be controlled for many years (Fig. 4) by using at least six sacks of Type
V sulfate-resistant cement per cubic yard of concrete at the lowest practil Concrete engineering consultant, formerly concrete engineer, California Department of
Water Resources, Sacramento, Calif. 95822.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
370
FIG. 1--This World War H air base barrack floor in a desert location showed distress
in a few years.from sulfate attack.
FIG. 2--Corrosion by suljhtes has completely disintegrated lower parts of this drop structure in a western canal.
TUTHILL ON RESISTANCE TO CHEMICAL ATTACK
371
FIG. 3--Concrete under ground was intact, but rising sulfate solutions caused exposed
surfaces to spall away as salt co,stals developed in concrete pores when moisture evaporated.
EXPANSIONOF3x3xl6-iNCHCONCRETEBARS DURING7-YEARS IN SOAPLAKE WATER (4.6%-Nae S04)
COLUMBIA BASIN PRO~EGT, WASHINGTON
\
\
\
\
z\ L
,
Unk~s Type ~[ cement co~tains 5.5% C3A, an

txtro sock pen cu yotd wifl not produce sulfote
r~sista~ce equQ{ to ~'hot developed by petter
thon borderline Type v sulfote resistin~ cement~
It should be noted thor Q Type I! cement c o ~ i n in9 this omount of CIA is opprooching the
moximum orr~unt Of 1his compound ollowed ton
TylDeV. An extt'o sock of TypeVcement will
inceeo~ s~Hote resistonce more l'hon 5 0 %
..'/~5 $ACKS/ CU. Y D.
_o
e:
4.5 SACKS/CU. YD.
TYPE V
3
TYPE I I
5
TRIGALGIUM
ALUMINATE
(C3A)
rO
II
12
tN P E R C E N T
FIG. 4--The lower the percentage o[" tricalcium aluminate in the cement, and the richer
the mix the better will be the resistance o['concrete to sulfate attack.
372
cable w/c ratio. Much greater resistance can be obtained by substituting

an effective pozzolan for 15 to 30 percent of the cement by weight.
With this summary of the general problem of chemical attack and the
means for producing concrete to resist it, this report will proceed with a
more detailed discussion of the significant aspects of how these chemical
actions damage concrete, what can be done to combat them, and in some
cases what tests can be used to measure the degree of attack or the success
of corrective measures.
Leaching
Hydrated lime is one of the compounds formed when cement and water
combine in concrete. This lime is readily dissolved by water, particularly if
the water is lime-free and contains dissolved carbon dioxide. Snow water in
mountain streams and reservoirs is often particularly aggressive because it
is unusually cold (calcium hydroxide is more soluble in cold than in warm
water), pure, and consequently lime-hungry, and, like most surface water,
contains carbon dioxide. This produces a mild carbonic acid solution that
has a higher capacity for dissolving lime than does pure water without it
[1-81.
As a result, inside surfaces of concrete conduits, flumes, and canal linings
develop a sandy appearance from having the cement matrix leached and
weakened by contact with these lime-hungry waters. Scandinavians report
[9,10] serious attack and decay of this kind in dams and conduits for relatively lime-free water; however, the later report [9] indicates that its severity
was due to the poor quality of concrete at these locations.
Attack by snow water has been noted in power flumes and lined canals
carrying water from the California Sierras (see Fig. 5). An outstanding California example is the 24 km (15-mile) Tiger Creek flume of the
Pacific Gas and Electric Co. A few years after construction in 1930, the
surface was found to be considerably roughened, and it was feared furthec
deterioration would seriously reduce its capacity. After a trial length that
had been coated with well-cured, troweled gunite using ordinary portland
cement showed that it too was being attacked, the entire flume interior
was painted with a coal-tar-base paint which had performed best in tests
of many paints and coatings previously tried in the flume. Even with this
treatment, repainting was required after 8 years' service.
Objectionable results of leaching are not confined to surfaces in contact
with pure mountain water. Exposed surfaces of tunnel linings, retaining
walls, abutments, and other structures, where ground water has access to
the opposite side, are often disfigured by lime deposits. These are formed
by water that has come through the concrete, either along cracks or joints
or through porous areas, taking lime into solution and becoming saturated
with it. At the surface the solution absorbs carbon dioxide which reacts
373
FIG. 5 - - L i m e - h u n g ~ or slightly acid water has leached sur/uce mortar and exposed small
aggregate on surfaces of this siphon inlet.
with the calcium hydroxide and causes precipitation of a white deposit of

calcium carbonate. Although unsightly, this kind of leaching rarely has
detracted significantly from the basic serviceability of such structures made
of good concrete.
Investigators have attempted to measure the influence of cement composition and the use of pozzolanic material on the degree of leaching that
may occur. An unpublished report by Stenzel for the Metropolitan Water
District of Southern California in 1936 found for typical Types I, II, and V
cements that "the leaching effect is similar in nature and magnitude for all
the cements studied. The effect [for a portland-pozzolan cement] is, however, definitely less than for the portland cement group, even when the reduced actual cement content (80 percent) is taken into consideration." In
120 days of leach tests, more than 20 percent less solids were dissolved
from the mortar made with the 80-20 portland-pozzolan cement. Stenzel
further concluded: "Since the laboratory experiments indicate that the
actual quantity of cement dissolved is of a low order of magnitude, it may
be said that failures from this source are not likely to occur before those
due to weathering and other causes." He added: "As the lime is exhausted
from the rich surface layer of the concrete, it is to be expected that the
leaching process will greatly subside" and be further inhibited by slower
diffusion.
From these observations and other studies of the problem, the following
ways to minimize leaching are recommended:
1. Use of: (a) aluminous cements, (b) portland-blast furnace slag ceCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:35 EDT 2014
374
ments, (c) portland-pozzolan cement with 20 to 30 percent of a good pozzolanic material that is strongly active in combining with lime to form insoluble lime silicates, or (d) "low-lime" portland cement with less tricalcium
silicate than dicalcium silicate.
2. Use of sufficient cement and a low w/c ratio to ensure good dense
concrete together with air entrainment, to minimize permeability and
capillarity, inhibit diffusion, and retard dissolving of cement increasingly
after the surface has been leached, particularly in water channels.
3. Design and use of contraction and construction joints that are watertight and frequent enough to prevent intermediate cracking.
4. Proper control of concrete mixes, placing procedures, and curing so
that hardened concrete free from permeable imperfections is ensured.
5. When practical, provision for drainage facilities necessary to prevent
water from standing behind structure walls.
6. Consideration of a provision for a protective coating of durable and
effective surface sealing material, if such can be found and reliably applied.
(See Portland Cement Association Bulletin "Effect of Various Substances
on Concrete and Protective Treatments, Where Required," and Ref [11].
Acid Attack in Sewers
Problems of acid attack above water levels in sewers usually are assumed
to be the result of sulfuric acid formed from hydrogen sulfide gas, generated by bacteria from sulfur compounds in the sewage, which rises and
combines with oxygen and with moisture condensed on upper surfaces of
the sewer conduit [12]. This is not a direct attack of acid in the sewage.
Records indicate that the usual run of sewage and waste water averages
very close to neutral, fluctuating only slightly above and below a pH of 7.0.
Predominance of certain industrial wastes in certain locations may change
this significantly and present a special problem.
For most sewers, however, since the flow itself is not aggressive, problems of acid corrosion in the crown portions are best solved by design or
operating modifications that result in the sewers running full or with ventilation at higher velocities. Lower temperatures also tend to reduce the
production of acid-forming hydrogen sulfide gas [12]. When sewage is
stagnant or moves slowly, bacterial creation of the sulfide is often too rapid
to be oxidized by air dissolved in the sewage.
Unfortunately many existing systems are not amenable to such changes,
and serious problems of acid corrosion exist in their crown portions (see
Fig. 6). The fact should be faced that, in such aggressively corrosive situations, portland-cement concrete is not the proper material to use since
none of the various compositions of portland cement are resistant to acid
corrosion. Other materials or protective coverings or coatings should be
employed.
375
FIG. 6--Disintegration of concrete in the crown qf'a sewer is caused by suljuric acid jormed
as hydrogen su(tide gas rising from the sewage and combining with oxygen and moisture
on the concrete.
It is reported [13,14] that calcareous aggregate tends to neutralize acids

in sewage and so reduces the rate of deterioration of concrete even though
it will still be attacked. Hence, use of limestone or dolomite aggregate may
be a desirable measure where concrete must be used in sewerage for reasons
of design or otherwise.
Smith and Backstrom, in an unpublished report for the U.S. Bureau of
Reclamation in 1952, found that in 5 percent solutions of sulfuric acid
with a pH of 0.2, both plain and air-entrained concrete lost strength rapidly,
and lost weight from progressive surface disintegration, amounting to approximately 50 percent in 12 weeks' immersion. Actually, 5 percent is
much stronger than the acid solutions attacking sewers, but it is considered
that this concentration provides a reliably indicative accelerated test of
relative resistance to acid attack. The tests were made on 3 by 6-in. cylinders
of concrete having a w/c ratio of 0.52, 9.1 m m ( 90
max aggregate,
and a slump of about 68.6 m m (2.7 in.).
Investigators are continuing the search for materials or treatments that
will make concrete a suitably resistant material to chemical attack. ASTM
Test for Chemical Resistance of Mortars (C 267) describes a test method
which is essentially a description of methods used by most investigators.
376
Results of these many research tests have identified materials and mixes
that have not been resistant to various forms of chemical attack in service.
Such methods are considered to be significant because they have pointed
the way to other materials and practices which are giving far better service.
Many investigators have used immersed-storage and flow-over-the-surface
tests with specimens of various materials, mixes, sizes, and shapes being
exposed to various acid solutions for intervals between which surfaces were
brushed to remove loosened or softened material [15,16]. Such tests, when
suitably representative of the nature of attack, gave a significant indication
of the degree of corrosion and the degree of resistance that may be expected.
It is best that results be judged and compared from depth of scaling rather
than from loss in weight, since loss varies with volume of the specimens,
whereas depth does not.
A German investigator has reported initial encouraging results with a
process called "Ocrat-Concrete" [17]. The treatment [18] consists of subjecting the concrete in a vacuum to the action of silicon tetrafluoride gas
(SiF4): 2Ca(OH2 + SiF4 = 2CaF2 + Si(OH)4. When given proper time in
the treatment, the fluoride gas penetrates to some depth and provides more
than a thin shell of resistance to acid attack.
So far this treatment is applicable only to precast units, and it has the
drawback of being so toxic that exceptional care would have to be exercised
to protect workmen. However, the results of tests are encouraging (Table 1).
Other Acid Attack
Common among other acid attacks are lactic and acetic acids in dairy
and fruit products spilled on concrete floors of food processing plants. Although attack from these products is comparatively mild, it is persistent
and can result in softening a working floor so that it wears rapidly and
becomes uneven and unsatisfactory for smooth operation of wheeled trucks,
dollies, and carts. Another common area of acetic acid attack (pH 3.4 to
3.9) is due to silage in concrete silos [15,16]
Swedish moor water with pH values as low as 4.5 has been destructive to
somewhat porous concrete pipe. The water contained considerable aggressive
carbonic acid [9]. One English moor water was made weakly sulfuric acidic
by contamination from a polluted industrial atmosphere and from pyrites
in the soil [19]. Another soft English moor water was quite acidic, averaging
a pH of 4.4, probably due mainly to organic acids washed from peat,
sometimes collectively described as humic acid. Exposed to a flow of this
water for 4 years, 101.6 mm (4-in.) concrete cubes lost considerably more
strength than would be indicated by their 4 percent loss in weight [20].
These observations suggest that need for precautions against acids may
exist and must be recognized. Not all aggressive surface waters are sulfates,
although they are the most common.
1--Comparison
Conversion factors -l in. : 25.4 ram, and

1 psi = 6.89 kPa.
Portland cement and Rhein sand 0 to I/8-in.;

mix 1:3.5 by weight, plastic; w/c
ratio = 0.68; cured 3 days, then
air dried at 6 weeks, psi
Portland-blast furnace slag cement, 30 percent max slag; w/c ratio -- 0.36;
mix 1:5 by weight, earth moist: strength
at 6 weeks, psi
Dycherhoff white cement; mix 1:3 by weight,
same materials:
Strength at 6 weeks, psi
Strength at 8 weeks, psi
Strength after 6 weeks immersion in acids and
sulfate solutions after ocrating at 28 days,
psi:
sodium sulfate solution:
portland cement
portland-blast furnace slag
10 percent acetic acid solution:
portland cement
portland-blast furnace slag cement
dycherhoff white cement
lactic acid:
portland cement
dycherhoff white cement
5 percent hydrochloric acid:
Removed in abrasion test, in.3/in.e; Aged 6 Weeks
TABLE
1252
3300
3357
4025
4395
1920
1721
2091
1465
2176
2247
384
569
626
1152
1195
411
484
455
427
612
583
destroyed after 5 h
2930
Compression
Specimen
1707
1408
1408
1394
1565
1422
1380
1735
1579
1166
1180
754
1408
Tension
Specimen
0,040
0.061
711
Tension
Specimen
Ocrated
Untreated
o f u n t r e a t e d concrete a n d O c r a t - C o n c r e t e
7069
6472
4964
4592
5362
4893
4992
6429
5206
7368
8335
3001
6045
Compression
Specimen
..4
--I
-t
r-
(3
"I"
m
Z
O
ITI
--I
O
6O
2O
m
O
Z
C
..q
-1I-"
I-"
378
The weakening effect of these acids on portland-cement mortar is noticeable in the strengths reported in the German data shown in Table 1 [16].
Among these same tests it may be noted that the portland-blast furnace
slag cement mortar was improved in strength during 6 weeks of storage in
these acid solutions. This cement contained not over 30 percent slag. In
1927 Schlyter [21] recommended slag or aluminous cement for better acid
resistance.
The American Concrete Pipe Association [22] reports results from tests
in which pipe concrete with Type I cement was immersed for 9 months in
acetic acid (pH = 2.5) after 6 months of water curing. Clearly superior
resistance was found in specimens in which 25 percent of the cement had
been replaced with a calcined opaline shale pozzolan. However, this report
suggests that reliance should not be placed on serviceability of such portland-pozzolan cement concrete if pH values of attacking acids are less than
5.0.
Super-sulfated cement is a mixture of granulated blast furnace slag with
some anhydrite (calcium sulfate) and a small addition of lime or portland
cement. This cement is being investigated as a means for creating a concrete resistant to acids, especially in sewers, but tests by contemporary
investigators give no indication that super-sulfated cement is resistant to
acid attack [20]. Most investigators report best results with it in resistance
to sulfates and sea water [12,23] and imply, by omission of the subject,
that it contributes nothing to acid resistance.
Basically, no portland-cement concrete is acid-resistant. However, there
is great difference in the degree of resistance obtained in the wide range of
cementing materials and workmanship used. When the most resistant
cements are used in good, well-placed, impermeable concrete in welldesigned and well-drained structures, much better serviceability may be
expected. In detail, the same recommendations made for resistance to
leaching are suggested for best resistance to mild acids. However, where
strong acid attack is expected, other materials should be used, or at least
a protective surface treatment of proved effectiveness should be provided.
Sulfate Corrosion
Some of the most spectacular chemical attacks encountered in concrete
have been a result of sulfates in adjacent soil and ground water (see Figs.
1-3). The degree and rate of this attack vary. These attacks increase as the
concentration of sulfates in surrounding water increases, and decreases
as the amount of dry exposure increases. Dry concrete in dry sulfate-bearing
soils will not be attacked. Where saturation is continuous in strongly sulfatebearing ground water, the attack will be rapid and severe. It is even more
severe where saturation and drying are frequently alternated. Average concrete with Type I cement has been completely disintegrated in only a very
379
few years when severely exposed in drainage pipe, floors, and lower parts
of canal structures.
The most common and aggressive sulfates are those of sodium and magnesium, but some structure and canal locations are in soil containing
appreciable amounts of calcium sulfate (gypsum). Because gypsum is relatively insoluble, sulfate attack on concrete may be negligible. However, if
ground water movements are sufficiently sluggish, it is possible for it to
become a saturated solution of gypsum which contains 1100 ppm SO4.
This is a concentration that is capable of a considerable degree of sulfate
attack, although it is likely to be at a slower rate than with sodium sulfate
because it has been determined that calcium sulfate reacts with hydrated
tricaleium aluminate but not with hydrated lime or the hydrated silicates.
Both the manner of sulfate corrosion and one means of retarding it are
mentioned in the following statement of Bogue [23]: "Since calcium sulfoaluminate forms most readily in concentrated solutions of Ca(OH)2, and
since the disintegration of cements in sulfate solutions is due chiefly to the
formation of that salt, the presence of active silica will retard the formation
of the sulfoaluminate and thus delay the disintegration of the structure."
Presumably this implies that active silica or pozzolan is helpful in reducing
the severity of sulfate attack by combining with Ca(OH)2 and making much
of it insoluble.
A less severe effect has been noted in some exposures where underground
concrete was unaffected, but the surface of the concrete immediately above
the ground was pitted and scaled away (see Fig. 3). By a wicking action,
sulfate solution rises in the concrete. In arid climates, as it evaporates from
concrete in surfaces just above the ground, crystals of sulfoaluminate and
sulfate develop in pores under the surface with sufficient force to flake off
the concrete [24].
The widespread nature of the problem and the vigor of sulfate attack
stimulated a search for means to combat it, and as early as 1920 Miller
commenced his lifelong study of this problem with Manson. This is summarized in their University of Minnesota Technical Bulletin No. 194, "LongTime Tests of Concretes and Mortars Exposed to Sulfate Waters." Other
investigators, including the Portland Cement Association, the U.S. Bureau
of Reclamation, and the Metropolitan Water District of Southern California [25], similarly arrived at the same basic conclusion: that the lower
the percentage of tricalcium aluminate in the cement, the better would be
the resistance of concrete to sulfate attack. Typical results of tests by the
Bureau of Reclamation are plotted in Fig. 4.
Essentially, these research tests involved exposure of concrete specimens
immersed in sulfate waters. Specimens ranged from half briquets of standard 20-30 Ottawa sand mortar to approximately 150 by 300 mm (6 by
12-in.) concrete cylinders and lengths of commercial precast concrete pipe.
These specimens contained a wide variety of cements, pozzolanic materials,
380
and admixtures in mortars and concretes of varying richness and curing.

Effects of curing, temperature, and exposure to drying were determined.
Concentrations of sulfate were varied, and effects of both sodium and magnesium sulfates were noted. Lack of resistance was evaluated: from expansion, 0.5 percent being considered as complete failure; from loss of
strength compared to that of specimens stored in plain water; from loss in
dynamic modulus, complete failure being a loss of 40 percent; from significant loss of weight; and from obvious visual failure.
In these tests, made under normal exposure conditions, it did not take
long to identify unresisting concretes, but it took many years to prove the
relative merit of comparatively resistant materials and treatments. Consequently, the Bureau of Reclamation laboratories in Denver, Colo., are
using currently an accelerated sulfate test for evaluating sulfate resistance
of concretes. This test produces results comparable to those of the 21.1~
(70~ continuous soaking test in approximately one-sixth the time. In this
test approximately 75 by 150 mm (3 by 6-in.) specimens are alternately
soaked in a 2.1 percent solution of sodium sulfate (12 600 ppm SO4) at
about 22.8~ (73~ for 16 h and then dried in air at 54.4~ (130~ for
8 h. The criterion of failure in this test is established as being a loss of 15
percent in dynamic modulus or an expansion of 0.2 percent. At this point,
internal disruption has definitely started, and complete failure of the concrete, indicated by a 40 percent reduction in dynamic modulus or an expansion of 0.5 percent, is imminent. Indications from this accelerated test
have been consistent with long-time results well established during the past
25 years.
Whenever sulfate resistance of a comprehensive range of cement compositions has been obtained and tabulated against the percent of tricalcium
aluminate in each composition, such as in Lerch's Table 1, "Summary of
Sulfate Resisting Performance of Cements in the Long Time Study of the
Portland Cement Association," reported in Ref 26, and in Miller's Table 6,
"Summary of 122 Commercial Cements," p. 33 of Ref 27, and in many
other such summaries, the conclusion is inescapable that low tricalcium
aluminate content correlates positively and most reliably with greatest sulfate resistance.
Actually, in the long-time tests, except for differences in the effects of
pozzolanic materials and admixtures, it has been found that the time required to develop an expansion of 0.10 percent correlates quite accurately
with ultimate resistance performance. Information in Fig. 4 is presented on
this basis. Results of the long-time exposure tests have proved quite significant in that concrete in service, made with the cements, mixes, and
curing found most resistant in these tests, has resisted sulfate attack far
better than concrete made with most other cements.
Although the low tricalcium aluminate in Type V cement is without
doubt the most important factor in improving resistance of concrete to sulCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:35 EDT 2014
381
fates (Fig. 4), there are several other provisions and precautions that will
contribute materially to sulfate resistance:
1. The concrete should have an ample cement content. A w/c ratio of
not more than 0.50 should be used. An additional hundredweight of Type
V sulfate-resistant cement will increase resistance at least 50 percent as
shown in Fig. 4. It is also evident that sulfate resistance, equal to that provided by Type V cement with 4.5 percent tricalcium aluminate, is not obtained by using an additional hundredweight of Type II cement per cubic
yard unless the Type lI cement contains less than 5.5 percent of tricalcium
aluminate.
2. An active pozzolanic material may be substituted for 15 to 30 percent
by weight of the cement, provided the pozzolan, as most are, is one proved
to be effective in increasing sulfate resistance of Type II and Type V cement.
All otherwise acceptable flyashes tested have improved significantly (as
much as tbur to ten fold [28, Fig. 5] the sulfate resistance of Type II and
Type V cements [29], with the exception of sub-bituminous and lignite
flyashes [30]. Among the N-type pozzolans, those least effective were usually
(but not always) those causing appreciably higher mixing water requirement. Accordingly, with N-type pozzolans best results are more likely to
result with only the 15 percent replacement [28]. Pozzolan effectiveness is
substantially greater the longer curing can be provided prior to sulfate exposure, since its combination with the lime is slow and will stop without
moisture.
3. A low-lime and low-aluminate cement, having less than 50 percent
tricalcium silicate and less than 12 percent tricalcium aluminate plus tetracalcium aluminoferrite, in which less than 4 percent is tricalcium aluminate,
makes the most sulfate-resistant portland cement. Specifications for sulfateresistant cement having these compound limits were written in 1934 for
work in southern California [25].
It was felt that when long-range durability of concrete is at stake in an
aggressive environment it is not as important to require composition limits
that most manufacturers can meet, as it is to specify limits that will provide all the sulfate resistance that can be put into the cement. This is
advisable even at a somewhat higher premium, perhaps to only one mill, if
there is only one mill that can make the cement to those safer limits. As
noted in item 2 above, however, later research investigation has shown
that if arrangements can be made to use a suitable pozzolan, greater or
more prolonged resistance can be obtained with a somewhat less resistant
composition of cement.
4. Air entrainment is generally recognized as being measurably beneficial
to sulfate resistance. Inasmuch as a chemical explanation for this is unlikely, it is considered to be due to the improvement in impermeability
usually resulting from air entrainment of the concrete.
5. Precast units such as concrete pipe and block can be made appreciably
382
more resistant by a period of several weeks of drying following good curing,

as indicated by the data in Fig. 7. P r e s u m a b l y c a r b o n a t i o n occurring d u r i n g
this period contributes to improved sulfate resistance, as it also improved
resistance of leaching, by reducing permeability.
35
3c
Cement V-.
I
/'~emen't
!zo
io
/i
R--q.t/
....
"f
"~"~'~
Cement
Z0
DAYS
30
DRYING
40
AFTER
28
Hell briquettes m I% solufion of

No2504
Type I cemenf-MWD Reporf~702
""Ce e.t C
.-
io
"/
50
C0
70
OAYS M O I S T CURING
80
~0
FIG. 7--Resistance to sulfate corrosion is materially increased by do, ing concrete several
weeks alter good curing.
TABLE 2--Attack on concrete by waters containing sulfate (as S04).
Relative Degree of
Sulfate Attack
Minor
Positive
Severe
Very severe
Parts per million Sulfate

(as SO4) in Water
Least Resistant
Cement Advisable
0 to 150
150 to 1 500
I 500 to t0 000
Type I or II
Type I1 or V
Type V or V + extra
cement or a pozzolana
Type V + extra cement
+ a pozzolan
10 000 or more
" A pozzolan proved to increase effectivelysulfate resistance of Type V cement.
6. As indicated by Miller [27], steam curing at 1 7 6 . 7 ~ (350~

or
above greatly improves the resistance of specimens, i n c l u d i n g those m a d e
with ordinary p o r t l a n d cements.
7. C a l c i u m chloride reduces sulfate resistance regardless of the type of
cement used.
It m u s t be recognized that the intensity of the corrosive sulfate attack
will vary with the parts per million of sulfate (SO4) in water that wilt be in
383
contact with the subject concrete. The U.S. Bureau of Reclamation Concrete
Manual, 8th ed., Table 2, p. 11, recommends somewhat the following
schedule of cement types for appropriate resistance to the indicated degree
of sulfate exposure (see Table 2).
In this modified table an alternative cement of superior resistance is
shown for each degree of sulfate attack, because the factor of time has not
been taken otherwise clearly into consideration, and because of the possibility that the service exposure may become more aggressive. As time goes on,
an exposure of lesser degree may well have a cumulative greater effect in
the long run. At today's (1978) prices for concrete in place, the small fraction
of extra cost for this more assured protection for each degree of severity of
prospective attack could well be a wise investment.
For purposes of acceptance testing and for comparing sulfate resistance
of different cements, Subcommittee C01.29 on Sulfate Resistance of ASTM
Committee C-1 on Cement has developed an expansion test which in 14
days reflects with fair reliability the resistance to sulfate attack which may
be expected from the cement used to make up the test bars. The method is
described in ASTM Test for Potential Expansion of Portland Cement
Mortars Exposed to Sulfate (C 452). Prompt sulfate action is generated by
mixing sufficient powdered calcium sulfate with the cement for it to contain 7.0 percent of SO3.
After a decade of experience with this test it was noted that the results
of such tests agree fairly well with the long-term behavior of cements in
concrete in aggressive sulfate exposure, but no better than could have been
predicted at the start from the calculated CaA contents of the cements.
Accordingly, some feel that the results of this test can be used best only as
a supplemental indication to compound composition requirements, since
on occasion cements having 6 to 9 percent tricalcium aluminate have been
indicated by this test to be comparatively sulfate resisting.
Because such high values of C3A would raise serious doubts as to actual
sulfate resistance, the test would serve no reliable purpose as an acceptance
test that is not served by simply specifying the compound composition requirements, preferably those mentioned in item 3 above. The work on this
test has recently been reviewed by Mather [31].
Attack by Other Chemicals

Most widespread among attacks by other chemicals is the scaling of
concrete pavements and sidewalks caused by the use of sodium or calcium
chloride to melt ice and snow in freezing weather. Actually this scaling is
not primarily the result of chemical attack as evidenced by the considerable
freedom from scaling obtained when concrete is amply air entrained. In
such a rapid physical attack, these salts have the effect of increasing the
severity-of freezing-and-thawing cycles and, unless concrete is unusually
384
resistant to such weathering, surface disintegration results. The physical

actions involved in such weathering may be supplemented by leaching because of the greater solubility of calcium hydroxide in calcium or sodium
chloride solutions, or by the formation of chloro-aluminates under wetting
and drying conditions. Presumably by such chemical actions calcium
chloride has caused disintegration of concrete where brine has dripped
from refrigerating equipment and where no freezing was involved.
Since a pavement is disintegrated mainly by cycles of freezing and thawing, experience has shown that this result is largely prevented by using 5 to
7 percent of entrained air in the concrete, by keeping slumps low, and by
avoiding overmanipulation in finishing which brings water to the surface
and works entrained air out of the concrete.
Sea water is mentioned mainly to point out that most concrete failures
in this connection have not been due to chemical attack of the sea water so
much as to porosity which enhanced leaching and expansive corrosion of
reinforcing bars, and to freezing and thawing at tide and wave levels [2, 4-7,
12,32]. In one notable example in southern California, alkali-aggregate reactivity contributed to failure. However, sulfate constitutes a little over
2500 ppm of sea water, a concentration that is regarded as sufficiently
severe when found in ground water to warrant use of Type V sulfate-resistant
cement and other precautions previously mentioned to secure sulfate resistance. The additional precautions of good workmanship, air entrainment, low w/c ratios, and ample cover over reinforcing steel should largely
offset the other recognized causes of failure of concrete in sea water. European literature speaks favorably of the performance of aluminous and slag
cement and of portland-pozzolan combinations in sea water exposure [18,
33,34].
With increasing attention to conversion of saline water, plants of larger
capacity may utilize as much concrete as possible structurally. This possibility has raised the question of the effect of hot sea water and of hot distilled water on concrete durability.
The U.S. Bureau of Reclamation has conducted pertinent tests [35] and
reports [36] that conventional portland cement concrete, uncoated and untreated, is not suitable for conducting mineral-free distilled water or, at a
temperature of 143~ (290~
a brine as strong as synthetic sea water.
However, such concrete is suitable at temperatures not exceeding 121~
(250~
where only in contact with distilled water vapor or with a sea
water brine, provided exposed surfaces for either, have from 25.4 to 50.8
mm (1 to 2.5 in.) of additional sacrificial concrete. At temperatures of the
brine below 93~ (200~ no sacrificial concrete need be used. Concrete
for the large plant to be built shortly at Yuma, Ariz., for desalination of
return-flow Colorado River treaty water deliveries to Mexico, will be used
with the above limitations.
385
Sulfur water is an occasional source of concrete deterioration according

to reported tests [3]. Evidently, aggressive sulfate salts exist in this water
and are responsible for the attack. Arrangements were made to submerge
approximately 150 by 300 m m (6 by 12-in.) cylinders of various cements
and mixes in continuously flowing water at the job site for comparison
with similar cylinders stored in fresh water. Results indicated that "the
special cements--Lumnite, Trass, and Super," or mixes with flyash or
pumicite substituted for part of the cement were even better than rich
mixes and low w/c ratios. Appropriate warning is given that heat generated
by Lumnite cement in confined massive placements may be formidable
and objectionable unless precautions are taken to care for it.
Relation of Aggregates to Resistance to Chemical Attack
Except in its relation to resistance to chemical attack, it is not the purpose
of this paper to discuss the influence of aggregate on concrete durability.
This is covered in greater detail in other papers in this symposium. Indications from available records of experience are that aggregate is secondary
in responsibility for lack of resistance to chemical attack. Although some
aggregates may be somewhat vulnerable, as limestone, for example, is
vulnerable to acid, failure of concrete to resist chemical attack is primarily
a failure of the cement paste. If the paste can be made resistant, the concrete will be resistant and serviceable, if the aggregate otherwise makes
serviceable concrete.
Conclusions
This brief survey indicates that there are many resources in materials
and benefits from careful workmanship which can be marshalled quite
effectively against many kinds of chemical attack on concrete in service.
However, there are certain exposures to which concrete is inherently unresistant, and no means are now known by which it can be made satisfactorily resistant. These cases should be recognized in advance, and other
materials of proved resistance should be used.
Acknowledgment
The author is indebted to A. B. Crosby of the staff of the U.S. Bureau of
Reclamation Concrete Laboratory for his considerable assistance in reviewing pertinent technical literature. Recognition is made also of valuable
work reported by contemporary investigators mentioned in the text and in
the references.
386
References
[1] Anderegg, F. O., "'Efflorescence,"ASTM Bulletin No. 185, American Society for Testing
and Materials, Oct. 1952, pp. 35-45.
[21 Mather, K., "Leaching of Lime from Concrete," Journal, American Concrete Institute,
Feb. 1950; Proceedings, Vol. 46, pp. 473-475.
[3] Nelles, J. S., "Concrete Exposed to Sulfur Water," Journal, American Concrete Institute,
Vol. 37, Feb. 1941, p. 441.
[4] Terzaghi, R. D., "Concrete Deterioration in a Shipway," Journal, American Concrete
Institute, Vol. 44, June 1948, p. 977.
[5] Terzaghi, R. D., "Concrete Deterioration Due to Carbonic Acid," Journal, Boston
Society of CiviJ Engineers, Vot. 36, No. 2, April 1949, pp. 136-160.
[6} Terzaghi, R. D., "Leaching of Lime from Concrete," Journal, American Concrete Institute, Feb. 1950; Proceedings, Vol. 46, p. 475.
]7] Tremper, B., "The Effect of Acid Waters on Concrete," Journal, American Concrete
Institute, Sept. 1931; Proceedings, Vol. 28, p. 1.
[8] Wolley, C. T., "Leaching of Lime from Concrete," Journal, American Concrete Institute,
May 1950; Proceedings, Vol. 46, pp. 753-755.
[9} Granholm, H., "Long-Time Pipe Study," Transactions, Chalmers University of Technology, Gothenburg, Sweden, 1944.
[10] Hellstrom, B., "Decay and Repair of Concrete and Masonry Dams," Journal Institution
of Structural Engineers, May 1933, p. 210.
[11] ACI Committee 201, "Guide to Durable Concrete," Journal, American Concrete Institute, Dec. 1977.
[12] Hammerton, C., "The Corrosion of Cement and Concrete," The Surveyor. 1 Dec. 1944,
p. 587.
[131 Pomroy, R., "Protection of Concrete Sewers in Presence of H~S," Water and Sewage
Works, Oct. 1960.
[14] Swab, B. H., "Effects of H2S on Concrete Structures," Proceedings, Journal of Sanitary
Engineering Div., American Society of Chemical Engineers, Sept. 1961.
[15] Hughes, C. A., "Permeability, Acid and Absorption Tests of Mortars Used in Dry
Tamped Silo Staves," Journal, American Concrete Institute, June 1940; Proceedings,
Vol. 36, p. 553.
[16] Miller, D. G., Manson, P. W., and Rogers, C. F., "Laboratory Tests of Concretes and
Mortars Exposed to Weak Acids," Agriculture Engineering. Vot. 20, No. I1, Nov. 1939.
[17] Wittekind, W., Zement, Kalk-Gips, July 1952.
{18] German Patent No. 849, 225, 5 Dec. 1951.
[19] Lea, F. M., "Deterioration of Concrete Owing to Chemical Attack," Cement and Cement Manufacture, June 1936, p. 130.
[20] Halstead, P. E., "An Investigation of the Erosive Effect on Concrete of Soft Water of
Low pH Value," Magazine of Concrete Research, Vol. 6, No. 17, Sept. 1954, p. 93.
[21] Schlyter, R., "Damage to Cement Structures," Bulletin. Government Testing Station,
Stockholm, Sweden, 1927.
[22} Davis, R. E., "Pozzolanic-Material--with Special Reference to Their Use in Concrete
Pipe," Technical Memorandum. American Concrete Pipe Association, 1 Sept. 1954.
[23] Bogue, R. H., The Chemistry of Portland Cement, 2nd ed., Reinhold Publishing Corp.,
New York, N.Y., 1955.
[24] Stutterheim, N., "The Deterioration of Concrete and Reinforced Concrete Due to
Chemical and Natural Agencies," Bulletin No. 12, National Building Research Institute,
South African Council for Scientific and Industrial Research, June 1954.
[25] Tuthill, L. H., "Resistance of Cement to the Corrosive Action of Sodium Sulfate Solutions," Journal, American Concrete Institute, Vol. 33, Nov.-Dec. 1936, p. 83.
]26] Lerch, W., "Significance of Tests for Sulfate Resistance," Proceedings. American Society
for Testing and Materials, Vol. 61, 1961, p. 1043.
[27] Miller, D. G. and Manson, P. W., "Long Time Tests of Concretes and Mortars Exposed to Sulfate Waters," Bulletin No. 194, University of Minnesota, May 1951.
[28] Kalousek, G. L., Porter, L. C., and Benton, E. J., "Concrete for Long-time Service
in Sulfate Environment," Cement and Concrete Research. Vol. 2. 1972, pp. 79-89.
387
[29] "Fly Ash Increases Resistance of Concrete to Sulfate Attack," Research Report No. 23,
U.S. Bureau of Reclamation, 1975.
[30] "Performance of Lignite and Subbituminous Fly Ash in Concrete," REC-ERC-76-1,

U.S. Bureau of Reclamation, Jan. 1976.
[31] Mather, K., "Tests and Evaluation of Portland and Blended Cements 1or Resistance to
Sulfate Attack," Cement Standards--Evolution and Trends, A S T M STP 663, American
[32] Wakeman, C. M., Dockweiler, E. V., Stover, H. E., and Whiteneck, L. L., "Use of
Concrete in Marine Environments," Journal, American Concrete Institute, Vol. 54, April
1958, p. 841.
[33] Lea, F. M. and Desch, C. H., Chemistry of Cement and Concrete, Edward Arnold
and Co., London, 1937.
[34] Miller, D. G., Manson, P. W., and Chen, R. T. H., Bibliography on Sulfate Resistance
of Portland Cements, Concretes and Mortars, with Abstracts, American Society for Testing and Materials, April 1952.
[35] Graham, J. R., Backstrom, J. E., Redmond, M. C., Backstrom, T. E., and Rubenstein,
S. R., "Evaluation of Concrete for Desalination Plants," Report Rec-ERC-71-15, U.S.
Bureau of Reclamation, March 1971.
[36] Backstrom, J. E. and Graham, J. R., "Evaluation of Concrete and Related Materials
for Desalination Plants," Executive Summary (35), U.S. Bureau of Reclamation, June
1971.
Peter Smith 1
Chapter 25 Resistance to High

Temperatures
Introduction
One of the main reasons why portland cement concrete is so widely used
in building construction is that it can help satisfy a cardinal need for public
safety in face of the hazards of fire better than most of its competitors.
Concrete is incombustible and a reasonable insulator against the transmission of heat. These qualities alone help confine the fire and limit the
extent of the damage, and though the surface may crumble or spall, the
essential engineering properties of the body of the concrete remain intact. A temperature of 649 ~ (1200~ may be reached before 25 percent
of the original compressive strength of a calcareous aggregate concrete is
lost (the corresponding temperature for a siliceous aggregate concrete is
about 427~ (800~
However, the main role of concrete in a fire is to
protect any embedded steel for as long as possible against a rise in temperature to the point where its physical properties are reduced significantly,
causing excessive structural deflections that might lead ultimately to collapse. Fifty percent of the yield strength of mild steel is lost by about
593 ~ (1100 ~ (the corresponding temperature in the case of the tensile
strength of prestressing wire is about 427 ~ (800 ~
Every building code contains minimum fire protection requirements
based on (a) a combination of knowledge of the physical properties and
past experience of the behavior of various building materials when exposed
to fire, and (b) upon fire endurance ratings specified by the survival times
of specific structural assemblies or components in standard laboratory
fire tests. Usually only the ambient temperature regime is controlled in
these tests, to a maximum of 1260 ~ (2300 ~ after 8-h survival in ASTM
Fire Tests of Building Construction and Materials (E 119), for example.
Therefore, such tests do not provide much information about the effect
~Director, Engineering Research and Development Branch, Ministry of Transportation
and Communications, Toronto, Ont., Canada.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
SMITH ON RESISTANCE TO HIGH TEMPERATURES
389
of specific high temperatures on the properties of concrete, its constituent

materials, or reinforcing steel except in a very general way. To improve
fire resistance when designing and building a new concrete structure or
for assessing the condition and possibilities of repair of one damaged by
fire, then much more than mere survival needs to be known about the
thermal and mechanical properties of steel and concrete at elevated temperatures, or those residual properties after slow or quench cooling. It is
relatively easy to determine the residual properties by standard test methods
and the results provide much of the information needed to determine
what can be saved after a fire. However, experiments to determine the
same properties at sustained high temperature require special equipment
and measuring techniques. Yet this kind of information has become increasingly in demand in recent years for two main reasons. Firstly, it is
needed for computer-based modeling techniques of heat flow and structural behavior that offer great promise for extending the results of fire
tests on discrete assemblies to the more general case and of demonstrating
the beneficial effects of restraint and support that from unaffected cooler
parts of the structure have on overall fire resistance. Secondly, advanced
industrial and military applications require a greater knowledge of the
physical properties of various types of concrete under sustained or repetitive moderately high temperature regimes.
Concretes made with normal cements and aggregates have been used
successfully over many years in locations such as soaking pits, foundry
floors, furnace and compressor supports at moderately high, sustained
or repetitive temperatures, in some cases up to 371~ (700~
without
special forethought having been given to the properties of the concrete
used. However, some of the newer industrial applications do demand a
detailed knowledge of the structural behavior of concrete at moderately
high temperatures. One such application is in prestressed nuclear reactor
primary containment vessels where complex regimes of mechanical and
thermal stress may occur in comparatively massive concrete sections. An
objective in recent years has been to increase the operating temperatures
within the concrete without the need for artificial cooling. While some
investigations have pursued the subject up to 1000~ (1832~
in present
practice design bulk concrete temperatures at failure (in order to still retain the reactor core) are in the range from 204 to 316~ (400 to 600~
though normal operating bulk concrete temperatures are considerably
lower, in the order of 66~ (150~ with local hot spots reaching 121 ~
(250 ~
For the ultimate doomsday accident where the reactor core melts,
the resistance of the concrete is dependent on those properties that impart
fire resistance with the added interest of the ease with which the concrete
may vaporize.
Concretes made from normal constituent materials usually cannot provide for sustained performance at temperatures above 538~ (1000~
390
special concretes are necessary for use in furnace, flue, and kiln linings,
as thermal insulation; and in reactor thermal and radiation shields. Reactor shields require heavy weight aggregate concretes, while light weight
aggregate concretes may be useful for thermal insulation. Refractory concretes made with high alumina cements and heat resistant aggregates are
capable of resisting sustained temperatures up to 1800 ~ (3272 ~
At temperatures approaching 1200~ (2192~
above the point where
normal concretes have decomposed to mush, ceramic bonds develop which
might reestablish concrete as a useful engineering material. As yet this
potential has received little attention beyond producing ceramic glazes
by heating the faces of concrete units with an oxyacetylene flame. The
extreme case of high temperature use for normal concrete is probably
scarificial exposure in some missile launching sites where temperatures
of 2760~ (5000~ may be reached for a few moments.
Interest in the high tempeature behavior of structural concretes for
most applications starts at a lower bound temperature of 100~ (212~
and immediately above as free water starts to be driven off. Generally
speaking, the engineering properties and behavior of concrete at these
temperatures vary by only a few percentage points from those measured
at room temperatures. However, above about 149~ (300OF) the progressive continuum of cement dehydration reactions, thermal incompatibilities between paste and aggregate, and eventual physiochemical deterioration of the aggregate leads to high thermal stresses, microcracking,
and a rapid worsening in most mechanical properties of structural value.
However, the end of structural usefulness of a particular concrete does
not occur suddenly at a specific temperature. Many influencing factors,
some not fully understood, enter the picture.
Fortunately for most needs, three of the most important factors are
within control prior to placing concrete: (a) the structural system and
design details chosen, (b) the depth of cover provided to any embedded
steel, and (c) the type of coarse aggregate specified. By selection of the
best combination of these for a particular application, significant improvements in overall fire resistance can be achieved. Specialized concrete applications such as those in the nuclear, missile, or refractory field require
many other considerations to be taken into account. While not all of the
information necessary for such uses of concrete is in the public domain,
an appreciation and references sufficient for a general understanding are
given. Concrete manufacture by autoclaving or other processes involving
high temperatures is excluded from the material presented.
Fire Testing, Endurance Standards, and Modeling

In North America the fire resistance of building components is usually
measured, and fire endurance is specified, on the basis of ASTM Test
391
E 119. A standard fire exposure is defined by a time-temperature curve

in which 538~ (1000~ is attained in 5 min, 927~ (1700~ in 1 h,
and 1260~ (2300~
in 8 h (assuming an end point has not yet been
reached). Exposure to the furnace temperature is on one side only for
slabs and walls, beams are exposed from beneath and from the sides,
and columns are exposed on all faces. Minimum specimen sizes (areas
or lengths exposed) and a moisture condition requirement prior to testing
are specified and superimposed loads may be added to simulate in-service
conditions.
End point criteria in the test are intended to simulate modes of failure
in an actual fire. They include development of a passage for flames or
gases through to the unheated side, increase in temperature of steel or the
unexposed surface due to heat transmission, and failure to support load
after hosing down. Fire endurance is then stated in terms of time to reach
whichever end point occurs first (such as 1, 2, 3, or 4-h rating). The appropriate fire endurance requirements, necessary to ensure saftety in a
particular set of circumstances, are called forward in the applicable building
code along with related design criteria. It must be emphasized that the
prime purpose of fire tests and related criteria is to ensure safety of the
building and occupants in case of fire and not to ensure structural survival
or ease of repair after the fire (though obviously the more fire resistant
the construction is, the more likely that it will be salvageable).
Both before, and ever since, the first version of the standard fire tests
was established in 1917, the requirements and significance of such test
procedures have been debated, with improvements upon them being introduced as the development of knowledge and experience dictated. In
their literature survey Uddin and Culver [1] 2 cite 149 references covering
the period from 1884 to 1961, when an American Concrete Institute Symposium was held. The proceedings of that symposium provide a good opening
to present day thinking on fire resistance through four comprehensive papers. Carlson [2] looked at the effect of moisture, specimen size, restraint,
surface temperature criteria, and radiant heat. Benjamin [3] reported on
thermal factors arising from moisture content and type of aggregate, on
structural factors such as concrete strength, cover to reinforcing steel, spalling, and lateral restraint as affecting fire endurance ratings. Sheridan [4]
discussed concrete as a protection for structural steel. Troxell [5] addressed
the special problems of prestressed concrete sections. In the last decade, fire
testing laboratories around the world have worked actively to evaluate
new types of construction and improve fundamental knowledge. One particular problem, the moisture conditioning of specimens prior to exposure
to high temperature, is discussed later when the matter of moisture content in concrete is considered.
392
The establishment of fire endurance ratings by means of standard fire

tests is both expensive and time consuming. The results are specific and
particular to the component or assembly tested and cannot be extended
with confidence to include even comparatively minor variations in the configuration tested. Lastly, it is difficult during the test to simulate the
beneficial effects of restraint by cooler parts of a concrete structure that
are clear of the actual fire.
Accordingly, with the advent of computers, mathematical models are
being developed for the heat flow process in a variety of structural concrete units to establish temperature distributions and, taking material
and structural parameters into account, to then predict fire endurance and
structural behavior at high temperatures [6]. The regimes to be modeled
are not simple; moisture content, thermal, and mechanical properties
are interactive in the concrete mass and the structural geometry may be
complex. However, reasonable calibration with actual fire tests is being
achieved and the modeling approach is greatly increasing understanding
of the interaction of material properties and the design and performance
of concrete structures at elevated temperatures.
Behavior of Concrete
In their review, Uddin and Culver [1], list six material properties as
having particular influence on the high temperature performance of structural concrete. They are: (1) type of aggregate, (2) free moisture in the
concrete, (3) stress levels in both steel and concrete, (4) cover over reinforcing steel, (5) modulus of elasticity, and (6) thermal conductivity, diffusivity, specific heat, etc.
Most of these factors interrelate and the format has been adopted of
first considering the influences of component materials and then considering
the properties or particular problems of concrete at high temperature.
Throughout the discussion experimental techniques to measure various
properties will be identified or, where this is not essential, left to the details provided by the referenced sources. Two references [7,8], not later
discussed, are specific to experimental techniques and support a general
observation that a property that can easily be determined at normal temperatures by standard tests often can only be determined at elevated temperatures with difficulty, that is, by modifying the normal procedure or developing a new or indirect method through considerable ingenuity. As a
result, many of the differences reported in measured values of various
properties may be ascribed to differences in test conditions. However,
while the actual numbers may differ, there is a usually good agreement on
the general trends reported.
Two general tests are useful in obtaining an understanding of the ob-
393
served behavior of concrete at high temperatures. Thermogravimetry tests

provide insight into chemical reactions involving loss of weight while dilatometry tests that measure change of length may detect other reactions not
necessarily accompanied by loss of weight.
Harmathy and Allen [9], among other investigators, used such analysis
to determine the overall stability of a variety of concretes and to pinpoint
specific chemical and physical changes that occur with increasing temperature. Cement dehydration or the decomposition of carbonate aggregate
and the chemical stability of lightweight or siliceous aggregate concretes
are very evident in their loss-of-weight test data. One important result
from the change-of-length test was to show that the a-~ quartz transformation at 573 ~ (1063 ~
while certainly causing an increase in volume in
siliceous aggregates, did not cause a corresponding change in the concrete.
This Harmathy and Allen attributed to the accommodating plasticity of
the cement paste at high temperatures. The overall pattern of their length
change data confirms the complex superposition of dehydration shrinkage
in the cement paste and expansion of the aggregate as being a major factor
in high temperature performance of concrete. The corresponding cooling
curves demonstrate that most of the reactions and events that occur during
heating are irreversible.
Influence of Cement Paste
Choosing one portland cement over another will do little to improve

the properties of concrete at high temperature. While a case may be made
for using pozzolanic or blast furnace slag cements mainly because these
produce lesser amounts of calcium oxide as a result of dehydration and
also for using microfillers (fly ash, silica flour, and calcined shaledust)
[10], in practice these are little used for such benefit alone. High alumina
cements are used in refractory concretes but may not be appropriate for
structural purposes because of the conversion phenomena which may occur under normal service conditions. The development of distress and
the change for the worse in the thermal and mechanical properties of
portland cement pastes that occurs with increasing temperature are the
result of an uninterrupted series of physical-chemical reactions that are
accompanied by shrinkage and microcracking. Desorption of what is commonly called evaporable water first takes place as the temperature increases. Then, and overlapping, chemically bound water (nonevaporable)
is released progressively from the complex system of low crystalline order,
calcium silicate hydrates, and other hydrates in the cement paste and from
the calcium hydroxide crystals formed when the cement originally hydrated.
The dehydration of calcium hydroxide to form the oxide is a relatively
simple process occurring generally between 400 and 600 ~ (752 and 1112 ~
394
Subsequently the calcium oxide thus formed, after cooling and in the
presence of moisture, will rehydrate to calcium hydroxide with a disruptive
14 percent increase in volume. On the other hand, the dehydration of
the silicate hydrates and other compounds present, through any intermediate stages, is more complicated and gradual and not fully understood. Total dehydration is only complete at 800~ (1472~
or above.
The stage reached at any given temperature is dependent also upon the
rate of heating since all the reactions are not instantaneous. Most of these
decompositions are irreversible so damage to concrete from this cause at
sustained high temperatures is essentially permanent; thus for dehydrated
cement paste to serve a useful present and future purpose, its temperature
endpoint is probably about 538~ (1000~
After exposure to sustained
temperatures of 649 to 816 ~ (1200 to 1500 ~
concrete is friable, porous,
and, after cooling, usually can be taken apart with the fingers. Many of
the above events are evident in thermogravimetric, dilatometric, and differential thermal analysis [9, 11].
Zoldners' review of experimental thermal data [12] shows that the variation of thermal properties of the cement paste with temperature, in particular thermal volume change, is a very important component to the overall behavior of concrete because of incompatibility with the same properties
in the aggregates. There appear to be two components to thermal volume
change of paste prior to decomposition: (a) a true thermal expansion that
is essentially constant and reversible, and (b) an apparent thermal expansion that is a hygrothermal contraction dependent on the internal transportation of moisture between various states in the capillaries and gel
pores, resulting in shrinkage that is irreversible. Thermal expansion is
additive to shrinkage due to dehydration, with the result that many investigators have confirmed that cement paste expands up to 100 ~ (212 ~
and then with further heating and loss of moisture, it contracts rapidly
to a degree that more than equals the initial expansion. Above 500~
(932~
shrinkage again changes to expansion. Actual values of the apparent coefficient of contraction or expansion are time dependent as is
dehydration, one of the causative mechanisms. Thermal conductivity of
cement paste is also dehydration dependent; the lower the moisture content
the lower the thermal conductivity. At 750~ (1382~ where the paste
is much dryer and more porous, thermal conductivity is only half its value
at room temperature.
In addition to the effects of thermal incompatibilities, the decrease in
evaporable and chemically bound water with progressive dehydration of the
paste modifies the physical bonds existing initially in the concrete and promotes microcracking. The resulting influence for the worse in the mechanical
properties of concrete with increasing temperature is clear, though separate
data related to the paste component alone are sparse.
395
Influence of Aggregates
Unlike the situation with cements, the selection of one aggregate over
another can have a great influence on the resistance of concrete to high
temperature. Though not entirely justified by fact, many building codes
and insurance rating handbooks group concretes into two or four classes
of fire endurance rating depending on the aggregate used.
Crushed firebrick and fused aluminum oxide (corundum) and other
special aggregates used in refractory concretes are at the top of the line.
Then in descending order of fire endurance come expanded slags, shales,
slates, and clays; air-cooled slag; basic, finely grained igneous rocks such
as basalt; calcareous; and lastly siliceous aggregates. The advantages of
lightweight aggregates, man-made or natural, are discussed in connection
with lightweight concrete and attention naturally focuses on the reasons for
the significantly different behavior of calcareous and siliceous aggregates
which are the most widely used in normal weight concretes.
The first point is that carbonate aggregates are decomposed chemically
by heat whereas siliceous aggregates in general are not. Between 660 to
979 ~ (1220 to 1795 ~ calcium carbonate breaks down into calcium oxide
with the release of carbon dioxide. Magnesium carbonate is likewise decomposed between 741 and 838~ (1365 and 1540~
Both reactions
are endothermic thus absorbing heat and delaying temperature rise in the
concrete. The calcined material is also less dense and hence a better insulator. Furthermore, the carbon dioxide, escaping from the surface of the
concrete in considerable volume, forms an inert insulative layer thus further
retarding temperature rise in the concrete.
On the other hand, though siliceous aggregates may be chemically stable,
they have serious deficiencies in their physical properties at high temperatures. Most striking is the transformation of the quartz crystal from the
a to [3 polymorph at 573 ~ (1063 ~ with an increase in volume. At higher
temperatures other internal volume changes occur and the crystal form
may be metastable. Furthermore, quartz has a much greater coefficient of
thermal expansion than most other rocks up to 600~ ( l l 1 2 ~
The
consequence in concrete is much greater thermal incompatibility between
the cement paste and aggregate and hence greater internal thermal stresses.
Many people have ascribed a greater tendency to spalling for these reasons,
in particular the crystal form transformation, and this is discussed later
when the spalling phenomena are considered.
To summarize the pros and cons of carbonate and siliceous aggregates
--on the whole, the benefit of carbonate aggregates is the positive one
of being a better insulator.
Since the aggregate fraction is about 70 to 80 percent of the concrete
and it dominates high temperature performance, it is wise to look at some
396
of the thermal properties of rocks in more detail before turning to concrete. The desirable properties in case of fire are low thermal conductivity
that delays temperature rise, a thermal expansion as close to that of the
paste as possible in order to minimize development of parasitic thermal
stresses, and a high specific heat to absorb heat.
At normal temperature the highest values of thermal conductivity are
associated with silica-rich rocks such as quartzites and sandstones. Igneous
rocks such as granites, gneisses, and rhyolites have intermediate values
as do limestones and dolomites. Low values are found in basalt, anorthosite,
dolerite, and barite rocks. Upon heating, thermal conductivities decline
(with the exception of amorphous and monoerystalline rocks), in some
cases by as much as 50 percent between 100 and 573 ~ (212 and 1063 ~
Mirkovich [13] provides experimental values over the range from 100 to
800~ (212 to 1472~
He found the greatest decrease in thermal conductivity occurs with those rock types that show the greatest thermal expansions, that is, quartz, sandstone, and granite, while anorthosite, basalt,
and limestone show relatively little change. This connection between thermal
conductivity and expansion is ascribed by Zoldners [12] to the fact that
higher thermal expansion stresses cause more microcracking of crystals
and loosening of grains resulting in increased rock porosity and hence
decreased thermal conductivity. The "ideal" aggregate from a thermal
conductivity point of view would seem to require an amorphous microstructure and porous macrostructure, a specification best met by lightweight aggregates.
Thermal expansion of aggregates at high temperatures depends, as it
does at more normal temperatures, on mineral composition and structure.
A rock of only one mineral type usually has the average coefficient of expansion of the individual crystals. Some minerals are anisotropic, expanding differently in different directions, and through the orientation
of the crystals this is reflected in the rocks formed from them. An amorphous
mineral can have a much lower coefficient of expansion than the same
mineral when crystallized and again this is reflected in rocks. Most rocks
contain different minerals and crystal forms and, therefore, have composite
complex coefficient of thermal expansion that are nonlinear with temperature and, where the rock is porous, are also moisture content dependent. However, mineral composition is the dominant factor and therefore it is reasonably possible to compute the coefficient of thermal expansion of composite rocks from the weighted average of that of the individual
minerals.
Studies by Mines Branch, Canada, reviewed by Zoldners [I2] probably
provide the most recent complete sets of data available on rock thermal
expansions up to 1000~ (1832~
For most rocks, the rate of thermal
elongation is many times greater at the 500~ (932~
mark and above
than at room temperatures, and a strong influence of the presence or
397
absence of quartz is evident. An abrupt leveling of the expansion curve

at about 600 ~ (1112 ~ indicates the presence of quartz (after the 573 ~
(1063~ inversion). Further changes involving sudden volume increases
may occur at higher temperatures and contribute to the generally less
desirable performance of siliceous aggregates. Calcium carbonate rocks
show contraction after about 900~ (1652~
and dolomites after about
800~ (1472~ (on decomposition to the oxides). A high expansion rate
after 650~ (1202~
is stated to possibly indicate a high biotite mica
content. Of all the rocks examined in this test series, anorthosite showed
the best thermal stability up to 1000~ (1832~
Although basalt shows
a similar low expansion up to 600~ ( l l 1 2 ~
this increases at higher
temperatures and rapid expansion occurs above 900~ (1652~
One
general point to note is that fine grained rocks generally seem more durable
than coarse grained ones. A few rock types such as phonolite, tinguaite,
and zeolite inclusions in basalt are unstable at temperatures as low as
300~ (572~
because they contain hydrous minerals that dissociate
readily.
One surprisingly simple test for the overall durability of an aggregate
was developed for use in determining the resistance of aggregates to high
temperatures during the drying process prior to the proportioning and
mixing of bituminous concrete. However, the test does not appear to have
been applied to portland cement concrete aggregates. Unheated and heated
aggregates (after 1 day at 250~ (482~ or 15 min at 500~ (932~ or
2 min at 1000~ (1832~
were tested by Shergold [14] in a modified
aggregate crushing and impact test (British Standard Sampling and Testing
of Mineral Aggregates, Sands, and Fillers (812-51)). From the measured
values of strength reduction an ordering of various types of aggregate by
merit was developed. This ordering appears to be in general agreement
with that of the other investigations referenced previously.
Embedded Steel and the Influence of Structural Systems

A prime objective in the design of fire-resistant structural concrete is
to prevent collapse for as long as possible. This requires keeping any embedded steel as coot as possible for as long as possible, and employing
statically indeterminate structural systems, continuity over supports or
providing restraint of members. Such systems include continuous slabs
or beams, frames, arches, and plates that are less prone to collapse than
statically determinate structures such as simply supported beams.
The ultimate moment capacity of a normally reinforced concrete section
at a particular temperature is a direct function of the steel yield strength,
among other factors. Hot-rolled steel reinforcing bars show a linear decrease in yield strength with increase in temperature up to about 370~
(700~ amounting to about 15 percent. After that the decrease is more
398
rapid and reaches 50 percent by about 593~ (1100~

Less than 20
percent of the original yield strength remains at 760 ~ (1400 ~
The length of time before steel reaches a critical temperature in a fire
depends on the amount of cover to the steel and on the insulative properties
of the concrete. The influence of cover is striking. For the same normal
weight concrete exposed in a standard fire test for 1 h steel with only 25-mm
(1-in.) cover would reach over 427~ (800~ whereas with 50-mm cover
(2-in.) the steel temperature would be about 204 ~ (400~
With 7S-ram
(3-in.) cover the steel would probably be hardly warm to the touch. Similar
orders of limitation on steel temperature with increased cover are applicable to other periods of fire exposure.
Both the coefficient of thermal expansion and the thermal conductivity
of steel influence the performance of the surrounding concrete in a fire.
The coefficient of thermal expansion of steel varies with temperature in a
dissimilar manner to that of concrete at very high temperatures, but is
generally close to that of concrete where compatibility is critical at lower
temperatures. Changes in the coefficient of thermal expansion of steel at
high temperatures are related to changes in the phase and crystal composition of the steel. The coefficient increases with temperature up to 649~
(1200~
decreases to zero in the range up to 816~ (1S00~
and then
increases again. The thermal conductivity of steel is much greater than
that of concrete and sometimes after a fire a thin, whitish, porous layer
is evident surrounding reinforcing steel embedded in otherwise sound concrete. Usually this happens if part of the bar has become exposed to the
heat source (as by spalling), and the steel temperature then rapidly becomes high enough to dehydrate the cement paste. This effect and the difference in coefficients of thermal expansion may adversely affect the local
bond of concrete to steel at higher temperatures. Up to about 316~
(600~ the bond strength between steel and concrete is the same as, or
slightly higher than, at room temperature. Thereafter it declines. However, steel that has remained encased in concrete will recover much of its
lost strength on cooling and only exposed reinforcement that has become
distorted and twisted will usually need replacing or supplementing. Fire
protection requirements relating to structural steel encased in concrete
are covered in Ref 4.
When prestressed concrete first came to the fore, concern was expressed
that the thinner concrete sections would offer less protection to the steel
tendons, resulting in direct loss of prestress and relaxation of steel stress
at high temperatures [5]. It was thought that this could result in excessive
deflections and earlier failure than with normal reinforced concrete, especially in simply supported members. High tensile prestressing wire and
strand is cold drawn from hot-rolled, high-carbon steel and then stress
relieved. Above about 260 ~ (500~ the loss of tensile strength is greater
than that for normal reinforcing steel. A S0-percent loss in yield strength
399
of prestressing steel occurs at about 427~ (800~ whereas this loss is

only reached with hot-rolled reinforcing bars at a higher temperature of
about 593~ ( l l 0 0 ~
High strength alloy steel bars used in some prestressing systems reach the 50-percent loss level at about 565 ~ (1050 ~
On cooling, much of the direct tensile strength loss is recovered. However,
when a prestressed tendon is heated, stress relaxation also reduces the
prestressing force and substantial relaxation losses may remain after cooling. Treated "stabilized" wire or strand now coming into wider use has
improved relaxation properties, and shows some-modest advantages in
high temperature performance.
Actual fire experience and numerous studies of prestressed concrete
structural assemblies in standard fire tests result in a well placed confidence in the high-temperature performance of prestressed concrete provided its singular features in this respect are taken into account during
design and construction.
It should be emphasized that there is no one critical temperature or
condition at which embedded steel (normal reinforcement, prestressed,
or structural steel) suddenly fails. The dead and live load intensity, structural indeterminacy, continuity, and member restraint all influence the
outcome in terms of the distortion, deflection, and ultimate collapse of a
member. A combination of established engineering principles, computerbased modeling of temperature-structural regimes, and applicable material
properties offers promise that the rational design of concrete structures
for a particular high temperature resistance may soon become as routine
as any other design task. Standard references in design handbooks cover
the properties of structural and reinforcing steel at high temperatures.
Further discussion of design philosophy by Gibbons is found in Ref
15. Uddin and Culver [1] published a state-of-the-art review in 1975 on the
effects of elevated temperatures on structural members. While the 246
references cited are only up to 1971, discussion of their paper in 1976
added a further 21 references.
Strength--Compressive and Flexurai

The observed strength of the same concrete when heated to a particular
temperature, or after heating and cooling, depends on the circumstances
of the superimposed conditions. Variables include: (a) whether the concrete is restrained or not during heating, (b) whether free moisture is contained or not during heating, (c) whether the concrete is subject to thermal
cycling, (d) whether it is tested while hot or after cooling, and (e) whether
it is quenched or allowed to cool slowly. In addition, each concrete, depending on the behavior of the cement paste and aggregates, will have
different strength values at different temperatures. While little data directly
related to tensile, shear, or bond strength are available, compressive and
400
flexural concrete strengths at elevated temperatures have been studied

widely. For purposes of a general understanding only the more recent and
significant studies have been referenced. Access to earlier work is by Ref 1.
Because compressive strength is one of the prime qualities by which concrete is judged, it is probably the most investigated property at high
temperature. In simply supported members, differences in the strength of
concrete in the compressive zone have relatively minor consequences and
do not affect fire endurance. However in the design of complex frames and
continuous or restraine~t members, the compressive strength is important
and specific data are needed if fire endurance is to be improved by these
means.
Various techniques have been used to modify the compressive strength
testing procedure used at normal temperatures (ASTM Test for Compressive Strength of Cylindrical Concrete Specimens (C 39)) to permit
testing at high temperatures in the usual type of universal testing machine.
Usually the specimen is heated by means of an electrically or gas heated
furnace which jackets the concrete cylinder. To maintain uniform temperature conditions, load is applied from the testing machine platens through
preloaded cement-asbestos disks above and beneath steel cylinders. Concrete specimens (cy]inders) have varied in size from 50 mm (2 in.) diameter
and 100 mm (4 in.) length to 100 mm (4 in.) by 200 mm (8 in.), the ends
of which usually are machined flat. Where length changes as well as
strength are to be measured, then extension rods or optical devices may be
used to bring the gage points outside the furnace. Temperatures are
measured and controlled by surface or embedded thermocouples.
Using such apparatus and 75 by 150-mm (3 6 in.) cylinders, Abrams
[16] determined the compressive strength of concrete between 93 and 871 ~
(200 to 1600~ for concretes in the strength range of 23 to 45 MPa (3300
to 6500 psi) that contain carbonate, siliceous, and light-weight (expanded
shale) aggregates. The testing regimes included heating without load and
testing while hot, heating at three stress levels and testing while hot, and
testing after slowly cooling.
In summary, his results showed: (1) carbonate and sanded lightweight
aggregate concrete retained more than 75 percent of original strength up to
649~ (1200~ when heated unstressed and tested hot; for siliceous aggregate concrete the corresponding temperature was 427~ (800~
(2) when loaded during heating, compressive strengths were 5 to 25 percent
higher but were not affected by the applied stress level, (3) residual
strengths after cooling were a little lower than the corresponding hot
compressive strength, and (4) the original strength of the concrete had
little effect on the percent reductions in strength observed.
Turning to Malhotra [17], who tested 50 by 100-mm (2 by 4 in.) concrete
(10-ram (3/a-in.) siliceous aggregate) cylinders with maturities of from 42 to
227 days conditioned at 24 ~ (75~ and 55 to 60 percent relative humidity,
401
provides the following observations on some additional factors: (1) aggregate/cement ratio had little bearing on changes in coriapressive strength
up to 204~ (400~
however, above this temperature lean mixes showed a
proportionally smaller reduction in compressive strength than occurred
with richer mixtures, and (2) water/cement (w/c) ratios in the range from
0.37 to 0.65 had little influence on reduction of compressive strength of
concrete (up to 593~ (1100~ in his tests).
His tests also confirmed those of Abrams with respect to the benefit of
the concrete being under compressive stress (in the order of design stress)
when heated. He also recorded an additional reduction of compressive
strength of about 20 percent in specimens tested after cooling.
If concrete specimens are quenched, as may happen to actual concrete
hosed down with water during a fire, then as Zoldners [18] showed, up to
about 500~ (932~ there is a greater reduction in residual compressive
strength than when slowly cooled. This he attributed to thermal shock. He
also noted a harder outer shell in cylinders quenched due to rehydration of
the cement, and that at higher temperatures this largely eliminated the
difference in strength between quenched and slowly cooled specimens.
Zoldner's [18] results include the effect of four different aggregate types
on compressive strength at elevated temperatures. They were a gravel
comprised mainly of a mixture of metamorphic and granitic igneous rock,
a relatively pure sandstone, a high-calcium, fine-grained limestone, and an
expanded slag.
His conclusions provide three additional pieces of information in addition to generally agreeing with Abrams [16]: (1) the crystalline igneous
and metamorphic rock gravel aggregate deteriorated more rapidly than
did the limestone aggregate concrete; at 400~ (752~ only 85 percent of
the initial strength remained as compared with over 95 percent for the
limestone, (2) the sandstone aggregate concrete showed a significant
strength gain up to about 300~ (572~ after which deterioration was
rapid, and (3) above 525~ (977~ the expanded slag aggregate concrete
outperformed the others whereas below that temperature it was equal or
inferior in compressive strength.
The observation about an increase in strength in the lower range of
temperature is an interesting one that other investigators [10,16] bare also
noted in certain conditions of test.
The commonest procedure for determining flexural strength appears to
be a modification of three-point loading of small beams along the lines of
ASTM Test for Flexural Strength of Concrete (Using Simple Beam with
Third-Point Loading)(C 78) [18,19]. The decreases observed in flexural
strength are not necessarily of the same order and magnitude as those for
compressive strength of concrete of the same mix proportions and containing the same constituents [18]. Speaking generally, the decline of
flexural strength is much greater than that of compressive strength. For
402
example, Sullivan and Poucher [19] found the flexural strength of a

siliceous aggregate concrete at 400~ (752~
varied from 25 percent to
0 percent of the original strength. Zoldners [18] obtained similar results,
but more hopefully found only a 50 percent flexural strength reduction
with limestone aggregate concrete at 400~ (752~
(a reduction level
which he found to occur at over 600~ ( l i 1 2 ~
in the case of compression). Again quenching resulted in greater strength losses than occurred
when specimens were tested hot.
Modulus of Elasticity, Poisson's Ratio, and Bulk Modulus

The ratio of stress to strain is one of the most important of concrete
structural parameters. At high temperatures the value of E, the modulus of
elasticity, may be obtained dynamically as by Philleo [20], or by displacement measurements as by Sullivan and Poucher [19] on concrete specimens
heated within a furnace. In general terms, there is a decrease in E in the
order of 10 to lS percent between initial values at room temperature and
those at 100~ (212~
E then remains relatively constant up to 300~
(572~ after which a progressive decline sets in so that by 538~ (1000~
it is only half or less than half of that before heating. Sullivan and Poucher
[19] noted in their tests, made up to 400~ (752~ that for a given mortar
or concrete (with siliceous aggregates), the value of E after cooling was
essentially the same as that prevailing at the high test temperature, thus
determining that the reduction in E with increasing temperature is a
permanent one. Philleo [20] noted that the value of E depended on initial
w/c ratio, curing regime, and age. Between unheated values for calcareous
gravel aggregate concrete ranging from 3.20 to 7.34 X 10 -6 and final
values at 760~ (1400~ of from 0.91 to 2.30 X 10 -6 some benefit was
found in low w/c ratios and air drying.
Marechal [21] looked at the variation of E with temperature for expanded clay, crushed rock aggregate (mainly porphyry and quartzite),
and gravel aggregate (siliceous and siliceous limestone) at one cement
factor and found the reduction in this modulus to be aggregate dependent,
in the above order, up to 400~ (752~
Philleo [20] was able to compute Poisson's ratio since he measured both
flexural and torsional frequencies of vibration. Because of sensitivity complications in the test procedure the results were erratic but a general
tendency to decline with increasing temperature was apparent. Marechal [21]
also calculated Poisson's ratio, but did so directly from the longitudinal
axial load and measured transverse strain data and a similar decline was
found after the evaporable water had been driven off. He made further
calculations for the bulk modulus K which declined with temperature in a
curve parallelling that for E.
403
Creep
Data on inelastic behavior of concrete at high temperature is essential
for nuclear applications, since long term behavior under complex mechanical thermal loadings is involved, and for fire resistance design, if the
effects of restraint are to be taken into accout. If inelastic behavior is not
taken into account and if only data secured on small specimens are used,
overestimations of the affected parameters will occur. A first step in overcoming the continuing elusiveness of creep at high temperature was to
devise apparatus for measuring it as reported by Cruz [22].
A study by Wang [23] of high strength concretes 38 and 46 MPa (5500
and 6700 psi) containing quartz gravel aggregate over the temperature
range from 93 to 427~ (200 to 800~
showed that the shape of creeptime curves was the same as at room temperature, but the creep rate was
higher when the concrete was subject to both high temperature and a high
stress-to-strength ratio. He also found that (a) the latter affects the creep
rate more than the temperature, (b) there was a nonlinear relationship
between creep and stress level, and (c) with a lower w/c ratio, creep was
less. Sullivan and Poucher [19] attributed the fact that creep rates for concrete above 200~ (392~ were higher than those for mortar, while the
reverse was true below 200~ (392~
to the development of microcracking between aggregate and paste. At 300~ (572~ they measured creep
rates in concrete at three times those at 125~ (257~
and above 400~
(752~ they observed that creep increased very rapidly at low stress levels
as a plastic stage was reached. Dependence on creep on moisture content
is stressed by Marechal [24]. Higher creep rates occur in saturated or
sealed specimens for which a fully accepted explanation does not exist.
In addition to creep tests, Cruz [22] also carried out exploratory stress
relaxation tests under constant strain over a 5-h period. Stress was reduced
by 2, 32, and 74 percent at 24, 316, and 649~ (75, 600, and 1200~ respectively.
Recent investigations of concrete under transient high temperature conditions, including time-dependent effects such as creep and relaxation,
confirm that a very complex pattern of behavior probably exists [25].
Obviously, there is scope for more refined understanding of the viscoelastic properties of concrete at elevated temperatures if the practical effects are to be predicted by theoretical analysis supported by test data.
Thermal Properties
The basic thermal properties of concern to the behavior of concrete at
elevated temperatures are: thermal conductivity (the ability of the material
to conduct heat, a ratio of the flux of the heat to the temperature gradient),
404
thermal diffusivity (the rate at which temperature changes can take place),
and the two more widely understood parameters of specific heat and the
coefficient of thermal expansion. These properties and test methods are
more fully discussed in the chapter, "Thermal Properties" by J. A. Rhodes,
which also contains tables summarizing experimental values (also given in
Ref 12). However, it is essential to stress a number of very significant differences between the thermal behavior of concrete, mortars, cement pastes,
and aggregates at elevated temperatures as compared with those at more
normal temperatures.
Thermal Conductivity, Diffusivity, and Specific Heat

Thermal conductivity and diffusivity of a mixture such as concrete are
not additives of the properties of the constituents and there are difficulties
in experimental determination at a given elevated temperature. These are
due principally to accompanying moisture migration-desorption, though
other physiochemical changes also contribute. Thompson [26] discusses the
shortcomings of steady state methods such as the hot plate (ASTM Test
for Steady-State Thermal Transmission Properties by Means of the Guarded
Hot Plate (C 177)) or hot box (ASTM Test for Thermal Conductance and
Transmittance of Built-up Sections by Means of the Guarded Hot Box
(C 236)) standards and indicates that meaningful determination would
only be possible after moisture equilibrium is reached, a self-defeating
state of affairs. Transient methods, such as the hot wire method, also
present problems. Experimental errors are greater, since the hot wire does
not present a surface of constant flux and the initial temperature rise in the
wire in a region of high conductivity will be less than in a region of low
conductivity, both of which exist side by side in a nonhomogeneous material such as concrete. However, Harmathy [27] offers two variable state
methods which appear to overcome the various problems associated with
other methods. One is based on fitting a temperature distribution curve
used (assuming linear heat flow) to determine all thermal properties, and
the other on the time-temperature rise response to heat pulses to determine
difussivity. Both are accompanied by the necessary mathematical interpretations.
Difficulties also plague direct measurement of true specific heat at
elevated temperatures. Equilibrium at the test temperature is disturbed in
making the determination by standard physical methods and chemical
dehydration reactions take place in the cement phase which may add or
subtract heat. Such latent heat effects, while overridden to some extent by
the greater contribution of the aggregate fraction to the overall apparent
specific heat of concrete, must be taken into account if reasonably accurate results are to be attained in heat flow model studies.
Data from Harmathy (and Allen) [11,9] and other investigators show
405
considerable dispersion in values for the specific heat of concrete. This is

not surprising in view of the experimental difficulties encountered but, in
general, specific heat appears to be insensitive to the aggregate used or to
the mix proportions of the concrete. Values reported in Ref 9 for a range
of concretes over a temperature range from normal ambient to 650~
(1202~ show a slight trend to higher values with higher temperature. A
high specific heat, including the contribution from moisture present, has
the advantage of taking up more heat and delaying temperature rise in a
fire.
Fortunately, thermal conductivity (k), thermal diffusivity (h), and
specific heat (c) are simply related by the expression
k ---- hcp
where p is the density.
Conductivity and diffusivity hence vary in step and, with the relatively
small variation in specific heat with temperature, density of the concrete is
an important overall parameter when low density is due to air voids. This is
most apparent between a dry lightweight concrete and a normal weight
concrete. Moisture content has a significant role in heat conduction in concrete until dehydration is complete after which, and with decomposition of
the constituents, the now dry, porous, normal weight concrete becomes a
much better heat insulator than at lower temperatures. The greater the
moisture content the greater the thermal conductivity according to data
summarized by Zoldners [12]. For normal weight concretes, both thermal
conductivity and diffusivity show considerable decrease at high temperature.
Harada et ai [28] cite a 50-percent decrease at 800~ (1472~ as compared with thermal conductivity at room temperature. For a lean concrete
as compared with a rich one, a 20-percent increase in thermal conductivity
was noted assuming both were consolidated fully. From an examination of
]6 concretes made with Type 1 cement and a range of aggregates and
processes used in the manufacture of concrete masonry units, Harmathy
and Allen [9] confirm this order of reduction and find that all thermal
diffusivities became approximately equal at about 600~ (1112 ~ whereas
at lower temperatures there was considerable difference between Iightweight
aggregate and normal weight aggregate concretes. The lightweight aggregates showed lower diffusivities and less reduction with increase in
temperature than did concretes of normal weight. Their data on conductivity, while showing a similar trend, did not merge the aggregate
groups at the upper temperature.
The three thermal properties so far discussed (conductivity, diffusivity,
and specific heat) are important to overall behavior of concrete at elevated
temperature because they essentially determine the inward progression of
406
temperature rise into concrete and hence the extent of damage that may
occur to steel or concrete in a fire. They are also important in heat flow
calculations necessary in the design of structures for performance at elevated temperatures as, for example, in nuclear reactors where significant
temperature gradients are a feature of normal operations. On the microscale they influence the thermal incompatibilities that occur between the
cement paste and aggregates due to differential thermal volume changes.
Thermal Volume Change and Thermal Incompatibilities

The observed thermal expansion in concrete is an additive of those of the
components. As already discussed, the thermal expansion of the paste is
highly dependent on moisture content while that of the aggregate is highly
dependent on mineral composition.
There are three components of volume change in concrete at elevated
temperatures to consider. As well as true thermal volume changes, volume
changes due to chemical reactions, for example, decomposition of calcium
hydroxide or rchydration of the lime, and changes in physical form of the
component materials, for example, the transition in quartz crystals, also
occur. The term "thermal incompatibility" is used widely to describe those
differential volume changes which result in differential strains and hence
stresses between aggregate particles and paste and between the constituent
minerals of each. Unless relieved by creep or plasticity in the paste, these
parasitic thermal stresses may be very high.
Looking at Philleo's data [20] for a calcareous aggregate, concrete shows
thermal expansion in linear up to about 260~ (500~
then after a
transition period there is a further iinearity about 427~ (800~
but at a
higher coefficient, after completion of dehydration. Above 538~ (1000~
the coefficient is again lower, probably due to a permanent dilation as a
result of decomposition. Philleo's experimental technique seems fairly
typical of those used to secure such data. The horizontal displacement
between two wires hung vertically from gage points on a horizontal concrete beam is measured at the test temperatures, clear of the furnace containing the specimen.
Zoldners [12] summarizes data from other investigations which show
that the thermal coefficient of expansion of concrete is in approximate
proportion to the weighted value of those of the aggregates once initial
shrinkage of the cement paste due to dehydration has occurred. The coefficient of thermal expansion of concrete with increasing temperature thus
tends to approach asymptotically that of the rock type in the aggregate [28].
A point of particular interest in this data is that the expansion curve for
limestone aggregate concrete is very close to that of mild steel over the
temperature range up to 700~ (1292~
Between the "best" natural
407
rock aggregate concrete investigated, andesite and limestone or sandstone

concrete, the coefficient of thermal expansion was halved [28].
Zoldners' own data [18] show a very rapid expansion above about 400~
(752~ for a predominately metamorphic and igneous gravel aggregate
concrete and for a sandstone aggregate concrete, as compared with limestone aggregate concrete and, in the case of an expanded slag or shale
lightweight concrete, a slight overall contraction with an almost constant
coefficient up to 900~ (1652~
In all cases an initial shrinkage up to
about 300~ (572~ is noted. These data are for only one cement factor,
one w/c ratio, and one aggregate/cement ratio. However, Philleo's work
[20] did include three cement factors and corresponding w/c ratios. Above
427~ (800~ these variables and initial curing (moist or air dried) had
little effect; however, below 260~ (500~ all these variables did affect the
coefficient values to a greater extent. Specimens having a higher w/c ratio
and lower cement factor tended to have the higher coefficients.
Crispino [32] pursued an investigation to find the best combination of
thermally compatible materials that would hence give good performance as
a structuraI concrete for nuclear reactor pressure vessels at high temperature under thermal cycling. As compared with concretes containing limestone aggregate and a variety of portland and high alumina cements, he
found barytes aggregate with portland cement far superior. In addition to
documenting the thermal properties, he provided full data on most mechanical properties up to 500 ~ (932 ~
The importance of thermal incompatibilities and the resulting thermal
strains and stresses cannot be overemphasized. While the cement paste is
shrinking as dehydration proceeds, the aggregate is expanding. After dehydration is complete, both are expanding, though at different rates. As a
result, there is a clear, if not as yet fully explained, effect on the mechanical properties of concrete in which, as will be discussed, moisture effects
may play a substantial role. Considerable hypothesizing continues toward
a full explanation of the behavior of concrete at all temperatures with
respect to thermal properties. Further reading is provided by Dougill [30].
No doubt, thermodynamic based theories that fully account for the role
water and a variety of physico-chemica! phenomena play will be forthcoming one day in a graduate thesis.
Thermal Cycling
Since most investigations in the past have addressed fire resistance, it
is not surprising that the effect of more than one thermal cycle has only
really been addressed after interest arose in using concrete in nuclear reactor vessels.
The effect of thermal cycling on strength has been investigated by Camp-
408
bell-Allen and Desai [31] and Campbell-Allen et al [32] who compared the
compressive strengths of linestone, expanded shale, and firebrick aggregate concretes up to 302~ (STS~ and noted a progressive deterioration with increasing number of thermal cycles though most of the loss
occurred in the first few cycles. Similar effects are observed with flexural
strengths and modulus of elasticity attributable to incompatible dimensional changes between cement paste and aggregate and to observed microcracking [19,31,32].
The interrelationship of thermal cycling, moisture content, and thermal
volume change is also important in nuclear applications of concrete. Polivka
et al [33] examined concrete that allowed moisture a controlled escape or
sealed it in, over durations up to 14 days of thermal exposure cycling up to
143~ (300~
They showed that:
1. For a constant moisture condition (sealed) of the concrete its thermal
coefficient of expansion decreased during thermal cycling to 143~ (300~
A similar decrease was observed for concretes which successively dried out
during thermal exposure.
2. After the first thermal cycle (21 to 143 to 21~ (70 to 300 to 70~ of
the concrete in a moist condition, a significant residual expansion was observed, and this expansion successively increased with the number of cycles.
This was also true for concrete specimens that were permitted to lose moisture during thermal exposure. A larger permanent expansion was observed for concretes of high moisture contents than for those of lower
moisture contents.
In examining the significance of data from thermal cycling and sustained
high temperature tests, the nature of the thermal regime must be considered since many of the reactions influencing the measured thermal and
mechanical properties including the internal transport or the explusion of
water, are dependent on time or the rate of temperature increase.
Behavior Mechanisms and Influence of Moisture Content
The stresses from thermal loadings due to overall expansion or contraction, when added to those arising from mechanical loadings, determine the
imposed total stress levels in the concrete. However, thermal stresses on a
smaller scale between particles determine the resisting strength, compressive
or tensile, of the concrete and the effect on other mechanical properties
such as the various moduli. Thermal strain incompatibilities and the resulting parasitic thermal stresses are thus critical to performance of concrete at high temperature, expecially where thermal cycling occurs.
Thus investigators have explained their observations on compressive
strength changes with temperature in general terms of the changes in
physical properties of the cement and aggregate and the thermal incompatibilities between these components with temperature. From petroCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:36 EDT 2014
409
graphic examination, Abrams [16] draws attention to microcracking and

deterioration in aggregate to paste bond. The advantage in compressive
strength where the concrete is preloaded in compression is ascribed to the
suppression of thermal microcracking in the cement paste and the increase
in compressive strength at moderately high temperatures has been attributed by Zoldners [18] to "a heat stimulated cement hydration and
densiflcation of the cement gel due to the loss of absorbed water."
From his pulse velocity measurements, Zoldners [18] attributed the
greater reductions in flexural strength as compared with compressive
strength at the same temperature to microcracking. Whereas small defects would have little effect on compressive strength, they would immediately show up as a reduction in flexural strength.
Since the reductions in modulus of elasticity, Poisson's ratio, and bulk
modulus K with increase in temperature follow the general pattern for
reduction in compressive strength, these have usually likewise been put
down to thermal incompatibilities, microcracking, and relaxation of bonds.
Lankard et al [34] reviewed existing data secured on specimens at elevated temperatures from which moisture could escape, made their own
tests on both sealed and unsealed specimens up to 260~ (500~ and
provided this summary.
1. First, and probably most important, the influence of heat exposure
per se on the structural properties of concrete is critically dependent on the
moisture content of the concrete at the time of heating and is quite different for concrete in which the moisture is free to evaporate as compared
to concrete sealed against moisture loss.
2. For concrete that is heated slowly at atmospheric pressure, the primary factor influencing changes in the structural properties is the loss of
free water. Partial loss of chemically combined water (dehydration of
hydrated cement phases) occurs above 121~ (250~ and primarily affects
the flexural strength.
3. For concrete slowly heated at saturated steam pressures, the primary
factor influencing changes in the structural properties is the hydrothermal
reactions that can occur in the cement phase and that convert the original
highly cementitious calcium silicate hydrates into crystalline, lime-rich
calcium silicate hydrates producing lower strength.
4. Deterioration of structural properties under both moisture conditions (sealed and unsealed) can be expected from any test variable that
can produce large temperature differentials in the concrete on heating
(for example, quench heating and cooling, use of very large specimens,
etc.)
While their data strongly support the role of moisture content in influencing mechanical properties of concrete at moderately high temperatures, they do provide somewhat of a challenge to the views expressed by
others on the role of microcracking.
410
For unsealed specimens they noted the microcrack system was no worse
after careful heating, possibly not as bad as that in the original concrete.
This observation helps to explain that compressive strength decline is not
significant below 260~ (500~ (and in some cases may even increase)
provided all the free water is removed on heating. It also adds to the usual
explanation of the development of flaws and stress concentrations around
flaws to account for the reduction in tensile strength with temperature
being greater than that for compressive strength. In this case they considered that the benefit of free water removal is offset by the loss of
chemically bound water which reduces the critical stress requirements for
tensile failure.
The removal of free, nonload bearing water also appears to account for
the observed behavior of the independence of the reduction in modulus of
elasticity with temperature at moderately high temperatures. However,
once dehydration proceeds to a significant loss of other forms of water
and to decomposition of the cement hydration products, and once aggregate deterioration and thermal incompatibilities increase with increasing temperature, then these latter mechanisms appear to account
fully for the observed behavior. But it does appear that the presence or
lack of free water is the controlling parameter at lower temperatures.
Polivka et al [33] established that at 143~ (300~ a maximum reduction in strength and stiffness, in the order of 75 and 55 percent respectively,
occurred at an intermediate moisture content corresponding to about a 50percent loss of the free water in the concrete.
Most of the data so far discussed on the thermal and mechanical properties of concrete have been secured by testing small unsealed specimens.
The justifications for this are that most tests at normal temperature are
made on small specimens and, in a fire, moisture is usually free to be
progressively driven off from the concrete, sometimes accompanied by
spalling in which moisture expulsion plays a significant role. However,
nuclear reactor shields and primary containment vessels present a different set of conditions in which the concrete essentially is sealed against
moisture loss. The prestressed concrete sections tend to be massive and
concrete may be hot and sealed by a steel liner on one side, even where it
is cool and open on the other. In hot sealed areas the exposure conditions
may be quite akin to autoclaving mature concrete and, with chemical and
mineralogical changes in the component phases of the concrete and restriction on the mobility of moisture, there is a greater reduction with
temperature in the physical properties of concrete than occurs when the
moisture is freely driven off.
In "saturated steam" conditions, Lankard et al [34] found a continual
and much greater decline in compressive strength and modulus of elasticity
up to 260~ (500~
Compressive strength at the upper temperature was
only 50 percent of original while the modulus of elasticity fell to only 31
411
percent of the original value. In observations of the composition of the

hydrated cement phases under these sealed heated conditions, they determined from X-ray analysis the presence of new, more highly crystalline,
probably lime-rich calcium silicate hydrates some of which, if they occur
in autoclaved products, are known to lead to lower strengths.
The amount of free water in concrete also has a major influence on
thermal properties. The contraction of cement paste is due to the expulsion
of free water and while water is a relatively poor direct conductor of heat
as compared with aggregate, it is a many times better conductor of heat
than is the air which would be occupying voids, capillaries, or pores if the
water was not present.
Because it influences almost every parameter of concrete behavior, including spalling, control, and uniformity of moisture content in the concrete prior to fire tests has been of long standing concern. This led to investigations and proposals to modify ASTM Test E 119 by preconditioning
the specimens [35]. Natural drying has been replaced with a specified conditioning regime of relative humidity and temperature and moisture content at the start of the test. While it is still not possible to cover every circumstance of possible moisture condition in a standard test, the present
requirements are an improvement and save time and money.
No doubt the active debate and investigation into the role of moisture in
concrete at elevated temperatures under both actual exposure conditions
and in laboratory test conditions will continue the drive to further refine
fire tests and develop better models to predict heat flow and structural
behavior. Many of the disparities in the results of physical tests for strength
and other properties in the investigations referenced are probably explained by lack of control on the moisture parameter and the temperature
gradient in the concrete. However, this apparent lack of control is the real
situation when mature or young concrete is exposed to fire and sometimes it spalls and sometimes it does not.
Spalling and Cracking

Spalling of concrete surfaces, particularly at arris, is an evident feature
of most fire damaged concrete. The worst consequence of this process is
that it opens up fresh surfaces to the heat source, thus making for more
rapid inward progress of the heat into the concrete and increasing the
probability of embedded steel being adversely affected, especially if it
becomes exposed. There are two possible causes of such spalling: thermal
incompatibilities and moisture entrapment. For many years direct thermal
causes, in particular such things as the 2.4 percent increase in volume of
quartz at 573~ (1063~
were considered a main contributing factor.
However, the quartz volume change may not be as significant as earlier
thought. Experienced investigators such as Harmathy and Allen state
412
they have never identified spalling due to this cause [9]. It seems that by
the temperature necessary for quartz expansion to occur the plasticity of
the paste is great enough to accommodate the expansion, through sudden
pop-outs of siliceous aggregate particles at or near the surface will likely
occur. Moisture entrapment now is identified [36,37] as predominate at
least in the earlier stages of spalling deterioration.
A simple view is that the disruptive pressure causing spalling is the result
of rapid heating with sufficient moisture present to produce saturated or
superheated steam. However, this does not appear to be an adequate explanation and it is worthwhile to examine a more plausible explanation in
some detail.
Shorter and Harmathy [37] offer an explanation called "moisture clog
spalling" along the following lines. As moisture is desorped in a layer adjacent to the heated surface, the vapor migrates toward the colder region
and is readsorbed. As the thickness of the dry layer increases a completely
saturated layer builds up at some distance in from the exposed face and
later a sharply defined front forms between the two layers. As the temperature of the exposed surface increases, a very steep temperature gradient
develops across the dry layer resulting in a high rate of heat flow, intensified desorption at the frontal plane and vapor pressure build-up. Because the concrete to the rear is saturated, the vapor can only leave through
the dry layer, expanding and meeting increasing flow resistance as it goes.
If the permeability of the material is low the pressure build up at the front
continues until a tensile spall occurs separating the dry layers from the
rest.
In establishing an overall picture of the effects of moisture on the fire
resistance of concrete, Harmathy [37] stressed two very important practical
points. High moisture contents are conducive to spalling but if spalling
did not take place (because, if resistance of the pores to moisture flow was
not too high, the existing large vapor pressure differential would move the
"moisture clog region" towards the colder region and the pressure buildup
would level off), then the moisture in the concrete is most beneficial. The
absorption of heat associated with the desorption of moisture checks the
rise in temperature in the concrete.
Data on the thermal properties of cement paste, aggregates, and concrete emphasize the thermal contraction and expansion strain incompatibilities that develop with increasing temperature, within and between
the constituents, and between the hotter and cooler parts of the concrete itself. The resulting thermal stresses may be sufficient to induce cracking.
Thermal cracking may range from crazing on the surface and microcracking to deeper and substantial spalling or cracks of structural significance at locations where overall structural deformations or thermal
stresses exceed the tensile capacity of the concrete. Planes of weakness
413
may also develop at the level of the reinforcing steel, and because of
thermal shock, if the surface of the concrete is quenched with water.
Selvaggio and Carlson [38] in fire tests of prestressed concrete beams
observed some cases of longitudinal tensile splitting through the full
depth of the flanges and webs when the temperature at the level of the steel
tendons was as low as 149~ (300~
By analysis they established that the
differentia] thermal expansion between concrete and embedded steel can
be a source of tangential tensile (bursting) stresses of considerable magnitude. Two concrete factors were considered to contribute to lack of
resistance to these stresses, a high modulus of elasticity, and reduced
tensile and compressive strength of concrete when simultaneously exposed
to tensile and compressive stresses (from the longitudinal prestress), together with factors originating from the steel arrangement and moisture expulsion.
Sometimes it appears as though additional spalling occurs as the concrete cools down. While the normal contractions are opposed by a volume
increase if moisture is present to rehydrate the calcium hydroxide, these
processes are not rapid and most likely the apparent spalling is simply the
falling away of already loosened pieces of concrete.
Investigation of Fire Damage

Once the structure is safe to enter after the fire, a first step is to try to
obtain an estimate of the temperature reached and its duration and hence
a first estimate of the likely remaining properties of the concrete. Many
circumstantial clues may be found by sifting through the debris and identifying the remains of common objects that have known melting points. The
appearance of the concrete can also provide important evidence [39,40].
Permanent color changes may accompany dehydration of the cement paste.
As temperature increases the color of siliceous or limestone aggregate concrete changes through pink or red to gray at about 600~ (ll12~
to
buff at about 900~ (1652~
Harmathy [41] proposes thermogravimetric
and dilatometric testing of samples secured within a day or two of the fire
as a means of determining a temperature history inwardly into the concrete. The approach is based on the fact that the continuum of reactions in
the cement paste remains frozen at the temperature level reached (for at
least a few days) and by comparison with a reference sample of unaffected
concrete or standard data this temperature marker can be identified.
Abrams and Erlin [42] provides a method based on examining the
microstructure and microhardness of prestressing steel for estimating the
temperature reached in a fire and hence its remaining tensile strength.
Though the effect of the temperature on normal reinforcing steel is not so
critical, similar metallographic methods may be useful.
414
Once the likely temperature regime has been established, the available
data on the engineering properties of the concrete and steel at either the
elevated temperature reached or after cooling can be incorporated in the
appropriate structural analysis to determine respectively the mode of
failure or if the structure should be demolished or can be repaired.
Confirmation of the residual engineering properties of the concrete and
steel after cooling can be determined by securing samples for standard
laboratory testing. Standard methods for in situ nondestructive testing
such as ultrasonic pulse velocity measurements can be used to help delineate
the damaged areas and compare the properties in situ of the affected
concrete with that untouched by the fire. Where doubt still exists after the
structural analysis or repairs have been completed, confirmatory fullscale load tests can be made.
Smith [43] described the investigative, material, and structural evaluation techniques and restoration methods available to examine young
concrete damage in construction fires. Numerous accounts of investigations
and repair of fire damage to mature concrete have been published (Ref 40
provides an excellent general review). The account by Fruchtbaum [44] in
1941 remains as a classic and that by Eisiner, Moreno, and Parma [45] of
the 1973 Avianca fire is typical of the advances incorporated in such investigations in recent years to match the greater sophistication of concrete
constructions.
Lightweight Aggregate Concrete

The thermogravimetric, dilatometric, and thermal studies made by
Harmathy and Allen [9] on concrete used in masonry units confirm the
reasons why concretes made from lightweight aggregates have superior heat
resistance to normal weight concretes. Lightweight aggregates (man-made
expanded shales, clays, slates, slags, perlites, or micas and natural pumice
or scoria) aid in the endurance of concrete to high temperatures. The
merits they impart appear to be due to their significantly lower thermal
diffusiveness (up to 600~ (1112~
and thermal conductivities (at all
temperatures), the capacity of the aggregate to retain moisture, and its
chemically stable vesicular and glassy (amorphous) composition. Abrams
[I6], among others, points out that the compatibility in relative stiffnesses
between the cement paste and lightweight aggregates is also beneficial. He
further noted a good bond strength between paste and aggregate.
Zoldners (and Wilson) [I0,18] among others, have looked at the residual
engineering properties of lightweight concretes after exposure to temperatures of 300, 500, 700, and 1000~ (572, 932, 1292, and 1832~ and after
cycling 5 or 25 times from 100 to 300~ (212 to 572~
They concluded
from ultrasonic pulse velocity, compressive, and flexural strength determinations made after cooling that all lightweight (expanded shale and
415
slag) concrete stood up to the effects of high temperature better than did
semi-lightweight concrete. They also found that although there was no
advantage in using Type 1S (blended portland blast furnace slag) cement
as compared with a normal Type 1 cement, the addition of microfillers
(fly ash, silica flour, and calcined shale dust) as 20 or 40-percent cement
replacements did improve the heat resistance of the concretes up to 500~
(932~
In all cases where microfillers were used, the improvement in
residual flexural strength was in the range of 10 to 20 percent though
similar improvements were not always noted in compressive strength. The
pulse velocity measurements, taken to be an indicator of porosity and
microcracking in the matrix, generally supported the conclusions drawn
from the strength data. Residual strengths for all lightweight concretes
were reduced to about 80 percent by heating to 300~ (572~ and to
about 45 percent by heating to 700~ (1292~ as compared with the unheated reference concretes. After heating to 1000~ (1832~
residual
strengths were negligible. Since failure is largely because of the deterioration of the cement paste matrix, by using high alumina cement rather than
portland or blended cements, insulating lightweight refractory concretes
good up to about 950~ (1742~ can be made.
Heavyweight Aggregate Concrete

Heavyweight concrete is used for radiation shielding in nuclear reactors.
This shielding is usually in the shape of a hollow cylinder or sphere. Under
moderately high operating temperatures, thermal gradients induce moisture migrations, and complex triaxial stresses can occur. In summarizing
the results of compressive strength of limonite, serpentine, magnetite and
steel punchings, portland cement concretes, and combination thereof,
Desov et al [46] reported reductions of up to 40 percent in hot tested specimens after heating to 350~ (662~ for between 3 and 7 h. As with
normal weight concrete, these effects were ascribed to differences in
thermal expansion between the cement paste and aggregates and to the
dehydration of the paste.
Refractory Concretes
Concretes used to withstand prolonged high temperatures, for example,
in monolithic linings for kilns, can be made with high alumina cements
and common refractory materials, such as crushed firebrick, as aggregate
[47]. This type of concrete will be stable and retain considerable strength
at dry heat temperatures of up to 1300~ (2372~ because ceramic bonds
have replaced the hydraulic bonds which were lost during desiccation. By
using aggregates such as fused aluminum oxide with a cement that is
essentially calcium aluminate, service temperatures for a refractory conCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:36 EDT 2014
416
crete can be as high as 1800~ (3272~

A 1978 report by ACI Committee
347 [48] provides a definitive discussion of refractory concrete properties
for those having deeper interest.
Elevated Temperature Coupled with Blast

At the time of the introduction of jet aircraft into service, there was
considerable concern that the hot blast might be damaging to concrete test
cells, warm-up aprons, and runway ends. Immediately at the jet exhaust,
temperatures of 677~ (1250~ occur and this may be as high at 1927~
(3500~ where afterburners are used. The velocity of the attendant hot
blast has been estimated at 1067 m/s (3500 ft/s) and the impingement of
this on the concrete was particularly exaggerated by the downward inclination and low slung engines on many early jet aircraft. Tail and higher wing
mountings in more recent planes have largely removed the problem from the
practical sphere and a 1953 paper by Bishop [49] remains definitive.
The launching of rockets from pads or underground silos for space exploration or military purposes presents another aspect of the hot blast
problem and is one for which all the answers have not yet reached the
public domain. However, it is known that the solutions tried include sacrificial concrete and protective shields including ceramic ablative shields.
The conditions though extreme, involving temperatures of about 2760~
(5000~
and velocities over 2438 m/s (8000 ft/s), last for only a few
seconds, however, and the choice between sacrificing the concrete or trying
to shield it is one of cost and convenience.
Concluding Remarks
Concrete has a longstanding and justified reputation as a fire resistant
material, howbeit with long recognized room for improvement in testing
methods [50]. Development of temperature regime-structural behavior
models and greater understanding of material properties are leading to
significant improvements in the specification and design for fire endurance
based on updated standard fire tests [51]. Deficiencies in concrete performance at elevated temperatures arising from moisture effects and
thermal incompatibilities long have been recognized also [50]. References
cited to studies in connection with both fire resistance and nuclear reactors
show that in the last decade striking advances have been made in our
understanding of the mechanism and consequences of these phenomena on
the thermal and mechanical properties of concrete at high temperatures.
However, full physico-chemical, thermodynamic explanations that tie
together all the aspects of the very complex viscoelastic, moisture-dependent
behavior pattern in concrete at elevated temperatures still are awaited.
417
References
[1] Uddin, T. and Culver, C. G., "Effects of Elevated Temperatures on Structural
Members," Journal of the Structural Division; Proceedings, American Society of Civil
Engineers, Vol. 101, No. ST7, July 1975, pp. 1531-1549 and Errata STll, November
1975, pp. 2463-2464 and Discussion ST3, Vol. 102, March 1976, p. 685; ST6, Vol. 102,
June 1976, p. 1268; STI2, Vol. 102, Dec. 1976, p. 2376.
[2] Carlson, C. C., "Fire Endurance Testing Procedures," Publication SP-5, American Concrete Institute, 1962, pp. 3-22.
[3] Benjamin, I. A., "Fire Resistance of Reinforced Concrete," Publication SP-5, American
Concrete Institute, 1962, pp. 25-39.
[4] Sheridan, R. R., "Fire Resistance with Concrete as Protection," Publication SP-5,
American Concrete Institute, 1962, pp. 43-55.
[5] Troxell, G. E., "Fire Resistance of Prestressed Concrete," Publication SP-5, American
[6] Salse, E. and Lin, T. C., "Structural Fire Resistance of Concrete," Journal of the
Structural Division: Proceedings, American Society of Civil Engineers, January 1976,
pp. 51-63.
[7] Bertero, V. V., Bresler, B., and Polivka, M., "Instrumentation and Techniques for
Study of Concrete Properties at Elevated Temperatures," Publication SP-34, American
[8] Purkiss, J. A. and Dougill, J. W., "Apparatus for Compression Tests on Concrete at
High Temperatures," Magazine of Concrete Research, Vol. 25, No. 83, June 1973,
pp. 102-108.
[9] Harmathy, T. Z. and Allen, L. W., "Thermal Properties of Selected Masonry Concrete
Units," Journal VTO, American Concrete Institute, Feb. 1973, pp. 132-142.
[10] Zoldners, N. G. and Wilson, H. S., "Effect of Sustained and Cyclic Temperature Exposure on Lightweight Concrete," Publication SP-39-9, American Concrete Institute,
1973, pp. 149-177.
[11] Harmathy, T. Z., "Thermal Properties of Concrete at Elevated Temperatures," Journal
of Materials, Vol. 5, No. 1, March 1970, pp. 47-74.
[12] Zoldners, N. G., "Thermal Properties of Concrete under Sustained Elevated Temperatures," Publication SP-25, American Concrete Institute, 1971, pp. 1-32.
[13] Mirkovich, V. V., "Experimental Study Relating Thermal Conductivity to Thermal
Piercing of Rocks," International Journal of Rock Mechanics in Mining Sciences
(England), Vol. 5, 1968, pp. 205-218.
[14] Shergold, F. A., "The Effect of High Temperatures on the Strength of Road Making
Aggregates," Roads and Road Construction, Vol. 31, June 1953, pp. 161-163.
[15] Gibbons, A. T., "Some Aspects of Structural Fire Resistance of Concrete," Fire Technology, Vol. 7, No. 1, 1971, pp. 61-68.
[16] Abrams, M. S., "Compressive Strength of Concrete at Temperatures of 1600~ '' Publication SP-25, American Concrete Institute, 1971, pp. 33-58.
[17] Malhotra, H. L., "The Effect of Temperature on the Compressive Strength of Concrete,"
Magazine of Concrete Research, Vol. 8, No. 23, 1956, pp. 85-94.
[18] Zoldners, N. G., "The Effect of High Temperatures on Concrete Incorporating Different Aggregates," Proceedings, American Society for Testing and Materials, Vol. 60,
1960, pp. 1087-1108.
[19] Sullivan, P. J. and Poacher, M. P., "The Influence of Temperature on the Physical
Properties of Concrete and Mortar in the Range 20~ to 400~ '' Publication SP-25,
[20] Philleo, R., "Some Physical Properties of Concrete at High Temperatures," Journal,
American Concrete Institute; Proceedings, Vol. 54, No. 10, April 1958, pp. 857-864.
[21] Marechal, J. C., "Variations in the Modulus of Elasticity and Poisson's Ratio with
Temperature," Publication SP-34, American Concrete Institute, 1972, pp. 495-503,
[22] Cruz, C. R., "Apparatus for Measuring Creep of Concrete at High Temperature,"
Journal, Portland Cement Association, Research and Development Laboratory, Vol. 10,
No. 3, Sept. 1968, pp. 36-42.
418
[23] Wang, C. H., "Creep of Concrete at Elevated Temperatures," Publication SP-27,

[24] Marechal, J. C., "Creep of Concrete as a Function of Temperature," Publication SP-34,

[25] Schneider, U., "Physical Properties of Concrete under Nonsteady State Condition,"
Proceedings, VII, Fe'd6ration Internationale de la Precontrainte, Congress, New York,
1974.
[26] Thompson, N. E., "A Note on Difficulties of Measuring the Thermal Conductivity of
Concrete," Magazine of Concrete Research, Vol. 20, No. 62, March 1968, pp. 45-49.
[27] Harmathy, T. Z., "Variable-State Methods of Measuring the Thermal Properties of
Solids," Journal of Applied Physics, Vol. 35, No. 4, April 1964.
[28] Harada, T., Takeda, J., Yamane, S., and Furumura, F., "'Strength Elasticity and
Thermal Properties of Concrete Subjected to Elevated Temperatures," Publication SP-34,
[29] Crispino, E., "Studies of the Technology of Concretes Under Thermal Conditions,"
Publication SP-34, American Concrete Institute, 1972, pp. 443-479.
[30] Dougill, J. W., "Some Effects of Thermal Volume Changes on the Properties and Behaviour of Concrete," Proceedings, International Conference on the Structure of
Concrete, London, 1968.
[31] Campbell-Allen, D. and Desai, P. M., "The Influence of Aggregate on the Behaviour of
Concrete at Elevated Temperatures," Nuclear Engineering and Design (Amsterdam),
Vol. 6, No. 1, 1967, pp. 65-77.
[32] Campbell-Allen, D., Low, E. W. E., and Roger, H., "An Investigation of the Effect
of Elevated Temperatures on Concrete for Reactor Vessels," Nuclear Structural Engineering (Amsterdam), Vol. 1, No. 2, 1965, pp. 382-388.
[33] Polivka, M., Bertero, V. V., and Gjorv, O. E., "The Effect of Moisture Content on the
Mechanical Behaviour of Concrete Exposed to Elevated Temperatures," Proceedings,
International Conference on Mechanical Behaviour of Material, Vol. IV, 1972, pp.
203-213.
[34] Lankard, D. R., Birkimer, D. L., Fondriest, F. F., and Snyder, M. J., "Effects of
Moisture Content on Structural Properties of Portland Cement Concrete Exposed to
Temperatures up to 500~ '' Publication SP-25, American Concrete Institute, 1971,
pp. 59-102.
[35] Abrams, M. S. and Orels, D. L., "Concrete Drying Methods and Their Effect on Fire
Resistance," ASTM STP 385, American Society for Testing and Materials," 1964,
pp. 52-73.
[36] Meyer, Ottens C., "Explosive Spalling Occurring in Reinforced and Prestressed Concrete
Structural Members of Normal Dense Concrete Under Fire Attack," (in German),
Deutscher Ausschuss f'tir Stahlbeton, H248, 1975.
[37] Harmathy, T. Z,, "Effect of Moisture on the Fire Endurance of Building Elements,"
ASTM STP 385, American Society for Testing and Materials, 1965, pp. 74-94.
[38] Selvaggio, S. L. and Carlson, C. C., "Fire Resistance of Prestressed Concrete Beams,
Study B, Influence of Aggregate and Load Intensity," Journal, Portland Cement Association, Research and Development Laboratory, Vol. 6, No. 1, Jan. 1964, pp. 41-64.
[39] Bessey, G. E., "Investigation on Building Fires, Part 2--The Visible Changes in Concrete or Mortar Exposed to High Temperature," National Building Studies, Technical
Paper 4, HMSO, London, 1950.
[40] "Concrete Structures after Fires, Evaluating Fire Damage to Concrete Structures and
Reinstating Fire Damaged Structures," Three part series on "Repair of Fire Damage,"
Concrete Construction, Vol. 17, No. 4, S, and 6, April, May, June 1972.
[41] Harmathy, T. Z., "Determining the Temperature History of Concrete Constructions Following Fire Exposure," Journal, American Concrete Institute, Vol. 65, No. 11, Nov.
1968, pp. 959-1064.
[42] Abrams, M. S. and Erlin, B., "Estimating Post Fire Strength and Exposure Temperature
of Prestressing Steel by a Metallographic Method," Journal, Portland Cement Association, Research and Development Laboratory, Vol. 9, No. 3, Sept. 1967, pp. 23-33.
[43] Smith, P., "Investigation and Repair of Damage to Concrete Caused by Formwork and
[44]
[45]
[46]
[47]
[48]
[49]
[50]
I51]
419
Falsework Fire," Journal, American Concrete Institute, Proceedings, Vol. 60, No. 11,
Nov. 1963, pp. 1535-1566.
Fruchtbaum, J., "Fire Damage to General Mills Building and Its Repair," Journal,
American Concrete Institute, Proceedings, Vol. 37, No. 3, Jan. 1941, pp. 201-252.
Eisiner, R. C., Moreno, J., and Parma, D., "Lessons from the Avianca Fire in Bogota,
Columbia," Proceedings, VII Congress, Fe'de'ration Internationale de la Precontrainte,
New York, 1974.
Desov, A. E., Nekrasov, K. D., and Milovaniov, A. G., "Cube and Prism Strength of
Concrete at Elevated Temperatures," Publication SP-34, American Concrete Institute,
pp. 423-434.
Zoldners, N. G., Malhotra, V. M., and Wilson, H. S., "High-Temperature Behavior
of Aluminous Cement Concrete Containing Different Aggregate," Proceedings, American
Society for Testing and Materials, Vol. 63, 1963, pp. 966-995.
American Concrete Institute, 1978 Refractory Concrete Report by Committee 347, to be
published.
Bishop, J. A., "The Effect of Jet Aircraft on Airforce Pavements; Investigation Conducted by the Bureau of Yards and Docks," Proceedings, American Society of Civil
Engineers, Separate 317, Oct. 1953.
Menzel, C. A., "Tests of the Fire Resistance and Thermal Properties of Solid Concrete
Slabs and Their Significance," Proceedings, American Society for Testing and Materials,
Vol. 43, 1943, pp. 1099-1153.
"Fire Performance Ratings 1977, Supplement No. 2, National Building Code of Canada," National Research Council of Canada, Ottawa, 1977.
M i l o s P o E v k a 1 a n d H. S. Davis ~
Chapter 26--Radiation Effects and

Shielding
Introduction
Portland cement concrete is an ideal material for use in construction of
radiation shields. Although there are other materials that could be employed for radiation shielding purposes, concrete is not only economical,
but it also has the advantage of being a material that can be cast into any
desired homogeneous structural shape. Concrete is now commonly used for
shielding of atomic research facilities, nuclear power plants, and for radiation medical and research units or equipment.
Conventional concrete of sufficient thickness can be and is being used
for such purposes. However, where usable space is a major consideration,
the reduction in the thickness of the shield is accomplished by the use of
high-density concretes. Such high-density concretes, produced by the use
of heavy aggregates, will usually have a unit weight ranging from 3350 to
3850 k g / m 3 (210 to 240 lb/ft3), which is about 50 percent higher than the
unit weight of conventional concrete. However, they can be produced with
densities up to about 5300 k g / m 3 (330 lb/ft 3) using iron as both fine and
coarse aggregate. For certain types of biological shields there is a need to
include materials of low atomic weight in the concrete.
At the present time there are two standards, ASTM Specification for
Aggregates for Radiation-Shielding Concrete (C 637) and ASTM Descriptive
Nomenclature of Constituents of Aggregates for Radiation-Shielding Concrete (C 638), that pertain to concrete-making materials for biological
shielding. In addition, several of the existing ASTM standards pertaining
to concrete and concrete-making materials are applicable to shielding concretes. Specific references to some of these standards and to their applicability are covered later herein.
1Professor of civil engineering, Universityof California, Berkeley, Calif. 94720.

2Consulting engineer, United Nuclear Industries, Inc., Richland. Wash. 99352.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
POLIVKA AND DAVIS ON RADIATION EFFECTS AND SHIELDING
421
Extensive literature is available on the subject of concrete for radiation

shielding. Pertinent papers were reviewed and are cited in the list of references. Recognition by ASTM Committee C-9 on Concrete and Concrete
Aggregates of the significance of concrete for radiation shielding as an
engineering material is evidenced by the activities of its Subcommittee
C9.02.08 on Concrete for Radiation Shielding.
This research subcommittee was established for the purpose of carrying
out a study and research on the general problem of concrete for radiation
shielding. After a 3-year study, the committee acquired information which
was made available to the profession at a symposium held at Gatlinburg,
Tenn. in 1965. This symposium was sponsored jointly with the Shielding
Division of the American Nuclear Society. The subcommittee prepared
ASTM Standards C 637 and C 638 and is responsible for revision of these
standards as required. Currently, the subcommittee is preparing a specification covering test methods for materials and procedures associated with
the preplaced-aggregate method for constructing radiation-shielding structures. It is also investigating boron-containing aggregates and their potential
use in shielding concrete.
This paper consists of a review and discussion of radiation shielding
and of properties of concrete and concrete-making materials for shielding
purposes and of their effects and significance. The work of other writers
on the subject, when used, is appropriately referenced.
Types of Radiation
There are two general classes of radiation that are considered in the
design of biological shields: (a) electromagnetic waves and (b) nuclear
particles [1].3
Electromagnetic Waves
Of the electromagnetic waves, the high-energy, high-frequency waves
known as X- and gamma rays are the only types that require shielding for
protection of personnel. Gamma rays are identical with X-rays except for
the source. Gamma rays are given off spontaneously from naturally occurring heavy elements or from lighter elements produced in atomic piles
or by accelerators [2]. Both X-rays and gamma rays have a high power of
penetration but can be adequately absorbed by an appropriate thickness of
concrete shield.
Nuclear Particles
Nuclear particles consisting of nuclei of atoms or fragments thereof,
422
include neutrons, protons, and alpha and beta particles. Of these, all but
the neutrons possess an electrical charge.
Neutrons, the heavy particles of atomic nuclei, do not carry an electrical
charge and thus are unaffected by electrical fields surrounding atoms.
They are slowed down only on collision with atomic nuclei. Neutrons have
high powers of penetration but do not follow a defined path through a given
material. Neutrons are classified as fast, intermediate (epithermal), and
slow (thermal) neutrons. The energy range of fast neutrons is between
10 000 and 10 000 000 eV, of intermediate neutrons between 0.S and 10 000
eV and of thermal neutrons less than 0.5 eV.
Protons and alpha and beta particles carry electrical charges which interact with the electrical field surrounding the atoms of the shielding material,
thus decreasing the energy of the particle. They generally do not constitute
a separate shielding problem, although accelerated protons at high-energy
levels may require heavy shielding comparable to that required for neutrons
of equal energies.
Tolerance Levels
Concrete shields are installed to protect personnel from radiation emitted
by nuclear reactors, accelerators, nuclear explosions, and some medical
and research units or equipment. The magnitudes of radiation must be
determined, using theoretical calculations and experimental data, to design
properly a biological shield that will ensure that the level of radiation on
the outside of the shield is maintained below the maximum permissible
intensity. Information on the tolerance levels for gamma rays and neutrons
was published by the National Bureau of Standards [3, 4].
Low Background Concrete

In addition to the use of concrete shields as protection from radiation,
there is also the problem of designing shields which will minimize the socalled "background" radioactivity. As reported by Wollenberg and Smith [5],
concrete was used successfully as a shield against radioactivity in structures
enclosing areas where low levels of gamma radioactivity are measured,
that is, low-background counting facilities. Such concretes must not only
provide a shield from extraneous radioactivity, but the concrete itself must
be low in radioactivity.
Shielding Problems
As discussed above, the principal types of radiation requiring consideration in the design of a shield are gamma rays and neutrons. A shield
adequate to reduce gamma radiation is normally also sufficient for protons
423
and alpha and beta particles. Thus, the general shielding problem resolves
itself into attenuation of neutrons and gammas. Almost any material of
sufficient thickness can serve satisfactorily as a shield, but neutrons and
gammas are not equally attenuated by the same materials.
Neutron and Gamma Ray Shielding

Neutron Shielding--For an efficient neutron shield it is often desirable
to incorporate three classes of materials. The shield should contain some
heavy material, such as iron, whose atomic mass is 56, or elements of higher
atomic number. These heavy elements slow down the fast neutrons by inelastic collisions. It is desirable to have light elements, preferably hydrogen,
to slow down further the moderately fast neutrons through elastic collisions.
Hydrogen is particularly effective because it weighs about the same as a
neutron. Finally, it is necessary to remove the slow, thermal neutrons by
absorption. Hydrogen is effective in this action, but in the process emits a
2.2 million volt gamma ray, which requires considerable shielding itself.
Boron, on the other hand, has not only a very high cross section (absorbing
power) for neutrons, but emits gamma rays of only 0.0478 million volts.
For these reasons boron-containing materials are often used in neutron
shields. The gamma rays given off during absorption of neutrons are a
factor in the required shield thickness.
Gamma Ray Shielding--Gamma radiation may originate inside the
shield through neutron capture as well as from the primary source behind
the shield. Attenuation of only the high-energy gamma radiation (1 to
10 000 000 V) is discussed here but not that of the lower energy X-rays.
Most materials attenuate gamma radiation primarily by the Compton
scattering process [2] with the attenuation being approximately proportional
to the mass of the material in the radiation path, but very little influenced
by the type of material, because different materials of the same mass have
approximately equal protective properties against gamma radiation [6].
Shielding Requirements
Concrete shields are usually of two distinct types: (a) those required to
attenuate gamma radiation only and (b) those required to attenuate fast
neutrons and associated gamma radiation [7]. For the first case, high
density and low cost are of prime importance. For the second case, it is
desirable to have materials of high or medium atomic mass (to slow down
fast neutrons); materials of low atomic weight, especially hydrogen (for
reducing neutrons of intermediate range to the thermal range); and finally
materials for absorbing thermal neutrons. These requirements for a neutron
shield are summarized in more detail by Henrie [8].
424
Shielding Ability of Concrete

Concrete is an excellent shielding material that possesses the needed
characteristics for both neutron and g a m m a ray attenuation, has satisfactory
mechanical properties, and has a relatively low initial as well as maintenance cost. Also the ease of construction makes concrete an especially
suitable material for radiation shielding.
Since concrete is a mixture of hydrogen and other light nuclei, and nuclei
of higher atomic number, and can be produced within a relatively wide
range of density, it is efficient in absorbing gammas, slowing down fast
neutrons, and absorbing resonance and slow neutrons [9]. The hydrogen
and oxygen, contained in chemically combined form in the hydrated cement, moderate the neutron flux satisfactorily. The oxygen may also be
present in the concrete in another form in addition to water; concrete with
silica aggregate is about one half oxygen.
Intensity of g a m m a radiation rather than of neutrons often determines
the thickness of the shield, in which case the required thickness will vary
almost directly with the density. The efficiency of a concrete shield of given
thickness depends mainly on its density and uniformity. The density of
concrete can be controlled within wide limits by selection of appropriate
heavy aggregates. The desired uniformity of the shielding quality requires
that the concrete form a homogeneous mass with uniform distribution of
the aggregate particles. Excessive settling, segregation, or formation of
pockets during placing may cause excessive radiation to penetrate such
faulty sections.
Materials used in a shield should not be subject to decomposition or
weakening under the influence of radiation. Mather [9] reports that data
reviewed indicate that the radiation stability of concrete is adequate for
most applications. In some cases, such as removable shielding plugs, materials that become highly reactive when exposed to intense thermal neutron
radiation should be avoided. Troublesome elements include vanadium,
chromium, manganese, cobalt, and others [7]. Also concrete containing
potentially deleteriously reactive cement-aggregate combinations should be
avoided.
Aggregates for Shielding Concrete

Concrete shields may be small structures housing a research nuclear
unit or may be massive monolithic structures or structures consisting of
precast assembled elements of various sizes. The thickness of the concrete
shield will vary depending on the type of nuclear unit but may range up to
1.8 to 2.4 m (6 to 8 it). Conventional concrete has been used to provide
radiation shielding. If there are no space limitations, conventional concrete
will provide a most economical shield.
425
Frequently, however, it is desired to produce concrete of improved properties with respect to shielding ability, especially when space limitations require that the thickness of the shield be reduced. In such cases the shielding
ability of the concrete is increased by increasing its density and in the case
of neutron shields by including a sufficient amount of light elements. The
increase in the density of the concrete is accomplished by the use of heavy
aggregates, such as naturally occurring iron ores, manufactured iron and
steel, ferrophosphorus, and barite. Aggregates containing bound water are
sometimes added to the mix to shield against fast neutrons. They include
hydrous-iron ores, serpentine, or bauxite. Also, to shield against neutrons,
boron-containing admixtures are sometimes added to the concrete.
Information on conventional concrete and concrete-making materials is
covered thoroughly by the various papers of this publication. The remainder
of this paper will refer to concretes whose properties were enhanced with
respect to their shielding ability.
Types
Naturally occurring iron ores, manufactured iron and steel, ferrophosphorus, and barite are the major heavy aggregates that have been used to
increase the density of concrete. A listing of these aggregates as prepared
by Davis [7,10] is given in Table 1. Included in this table are boron-containing additives used in neutron shields.
Natural iron ore aggregates can produce concrete densities up to 3850
kg/m z (240 lb/ft3). The most commonly used have been barite, magnetite,
and ilmenite. Along with hematite, these are used chiefly for gamma attenuation. To improve neutron attenuation, hydrous ores including goethite,
limonite, serpentine, and bauxite, and borates such as colemanite, boron
frit, and borocalcite have been used. The manufactured heavy aggregates
listed in Table 1 (except for the boron additives) are only used when the
required concrete density is greater than about 4000 kg/m 3 (250 lb/ft 3) [10].
Composition
In general, the exact chemical composition of heavy aggregates is not
critical as long as the required density is met. As a control of the uniformity
of the high density of some of these aggregates, chemical limitations may
be applied. For example, barite ore should have a barium sulfate content
greater than 90 percent, while the iron content of magnetite and limonite
aggregate should be in excess of 60 and 55 percent, respectively. However,
such limits will greatly depend on the economy of using easily available
materials which may justify using lower values. For example, magnetite
ore from Wyoming has an iron content of about 50 percent, while ilmenite
ore from Quebec has an iron content of only 38 percent. The specific gravity
426
TABLE 1--Aggregates used in shielding concrete.

Natural Mineral
Local sand and gravel
calcareous
siliceous
basaltic
Hydrous ore
bauxite
serpentine
goethite
limonite
Heavy ore
barite
magnetite
ilmenite
hematite
Boron additives
calcium borates
borocalcite
colemanite
gerstley borate
(2.5 to 2.7)~
(2.7 to 3.1)
(1.8 to 2.3)
[15 to 25%] b
(2.4 to 2.6)
[10 to 13%]
(3.4 to 3.8)
[8 to 12%]
(4.0 to 4.4)
Manufactured
Crushed aggregates
heavy slags
ferrophosphorus
ferrosilicon
Metallic iron products
sheared bars
steel punching
iron shot
Boron additives
boron frit
ferroboron
borated diatomaceous
earth
boron carbide
(-5.0)
(5.8 to 6.3)
(6.5 to 7.0)
(7.7 to 7.8)
(7.5 to 7.6)
(2.4 to 2.6)
(5.0)
( - 1.0)
(2.5 to 2.6)
(4.2 to 4.8)
(2.3 to 2.4)
(2.0)
aSpecific gravity is shown in ( ) .

bWater of hydration is indicated in [ ].
of both of these a g g r e g a t e s is in excess of 4.5. To require the iron content

to be 55 p e r c e n t or g r e a t e r would e l i m i n a t e these two reliable sources o f
heavy aggregate [10].
C o m p o s i t i o n of aggregates b e c o m e s very i m p o r t a n t when they are to be
used in l o w - b a c k g r o u n d counting facilities r e q u i r i n g aggregates having
low levels of n a t u r a l g a m m a activity. O f p r i m a r y c o n c e r n are the concentrations of p o t a s s i u m , u r a n i u m , a n d t h o r i u m . Based on extensive studies,
W o l l e n b e r g a n d S m i t h [5] indicate t h a t serpentinized u l t r a m a f i c rock is
the most suitable source of l o w - b a c k g r o u n d aggregate.
Radiation A t t e n u a t i o n a n d A b s o r p t i o n
The c a p a c i t y of the various heavy aggregates to a b s o r b g a m m a rays is
almost directly p r o p o r t i o n a l to their density. Also the heavier elements are
more effective in a b s o r b i n g fast n e u t r o n s by inelastic collisions t h a n are
the lighter ones. Davis [7,10] gives representative values of a b s o r p t i o n
factors for various aggregates a n d o t h e r c o n c r e t e - m a k i n g materials.
However, density is not the only factor to consider in the selection of
an aggregate for a n e u t r o n concrete shield. T h e d e s i r e d increase in hydroCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:36 EDT 2014
427
gen content, required to slow down fast neutrons, can be accomplished by

the use of hydrous ores. These materials contain a high percentage of
water of hydration (fixed water content) as shown in Table 1. However, on
heating of the concrete some of this fixed water in the aggregate may be
lost [10]. Limonite and goethite are reliable sources of hydrogen as long
as shield temperatures do not exceed 200~ (390 ~F), whereas serpentine is
good up to at least 400~ (750~
Measurement and Significance of Properties

Members of Subcommittee C9.02.08 of ASTM Committee C-9 have
made an evaluation of ASTM standards on aggregates with regard to their
applicability to heavy aggregates. Most of the pertinent existing standards
are applicable in their present form. A discussion follows of some of the
important properties and their significance.
Nomenclature and Mineralogy--ASTM Descriptive Nomenclature of
Constituents of Natural Mineral Aggregates (C 294) does not include a
number of the more probable constituents of heavy aggregates, and treats
several of the important constituents of heavy aggregates, such as iron
oxides, only in their capacity as accessory constituents of rocks in use as
normal aggregates. Members of Subcommittee ASTM C9.02.08 chose to
prepare a new document (ASTM Method C 638) rather than request revision of ASTM Method C 294.
ASTM Recommended Practice for Petrographic Examination of Aggregates for Concrete (C 295) applies to aggregates for shielding concrete as
well as it does to other aggregates. Mather [11] states: "The high-density
aggregates, and aggregates containing water that are used in shielding
concrete, present special problems, but they are still problems in which
petrographic identification of reactive substances, or substances likely to
produce problems in concrete production and placement, can be useful."
Chemical Properties--Raphael [12] has reported expansions of mortar
bars, tested according to ASTM Test for Potential Alkali Reactivity of
Cement-Aggregate Combinations (Mortar-Bar Method) (C 227), containing
barite aggregate and high-alkali Type I cement, which were large enough
to cause the combination to be regarded as alkali reactive. These expansions
might have been caused by reactive silica or by calcium sulfate in the
aggregate.
ASTM Test C 227 is applicable to heavy aggregates. Based on results of
a limited study made by Dolch, for the information of Subcommittee
C9.02.08, it appears that ASTM Test for Potential Reactivity of Aggregates
(Chemical Method) (C 289) is also applicable to heavy aggregates.
The origin of the majority of deposits of heavy aggregates makes the
presence of organic impurities very unlikely. However, presence of such
organic impurities should be checked for by the use of ASTM Test for
428
Organic Impurities in Sands for Concrete (C 40). If such a test is included

in a specification for heavy aggregate it should contain an explanation
similar to that given for lightweight aggregates, ASTM Specification for
Lightweight Aggregates for Structural Concrete (C 330), as follows: "Heavy
aggregates that, upon being subjected to the test for organic impurities,
produce a color darker than the standard shall be rejected, unless it can
be demonstrated that the discoloration is due to small quantities of materials not harmful to the concrete."
Deleterious Constituents--ASTM standards for deleterit, us constituents
of aggregates include tests for clay lumps (ASTM Test for Clay Lumps and
Friable Particles in Aggregates (C 142)) and lightweight pieces (ASTM Test
for Lightweight Pieces in Aggregates (C 123)) are applicable to, and can be
used in specifications for, heavy aggregates. However, the test for hardness
(ASTM Test for Scratch Hardness of Coarse Aggregate Particles (C 235)),
when applied to heavy aggregates, needs some interpretation.
The hardness of the minerals magnetite, hematitie, and ilmenite, which
produce ores of high specific gravity (4.2 to 4.8) is about 5 to 6 on Mohs'
scale. Barite ore, which has a specific gravity of 4.0 to 4.5, has a Mohs'
hardness (of the mineral) of only 2.5 to 3.5. The Mohs' hardness of calcite,
by definition, is 3. Thus if ASTM Test C 235 is intended to classify calcite
as just barely harder than "soft," it would be expected to classify barite
sometimes as "hard" and sometimes as "soft." Therefore, it would not be
reasonable to ask for a barite aggregate under a specification that includes
a limit of soft particles, as is the case of ASTM Specification for Concrete
Aggregates (C 33), for particles scratched by the ASTM Test C 235. For
these reasons, it is necessary to use careful judgment in the application of
ASTM Test C 235 to heavy aggregates.
Gradation--The gradation requirements of heavy aggregates can be
within the limits given in ASTM Specification C 33. The sample of heavy
aggregate used in the sieve analysis should be greater in weight than the
samples specified in ASTM Test for Sieve or Screen Analysis of Fine and
Coarse Aggregates (C 136). This increase in weight should be obtained by
multiplying the specified weight (ASTM Test C 136) by the ratio of the
specific gravity of the heavy aggregate to that of ordinary aggregate (for
which a value of 2.65 may be assumed).
Unit WeightwThe unit weight of heavy aggregates can be determined
by the procedures of ASTM Test for Unit Weight of Aggregate (C 29).
However, it might be desirable to employ larger capacity measures than
those specified in ASTM Test for Unit Weight of Aggregate (C 29). For
example, a 0.03 m 3 (1 ft 3) capacity measure is much too small for 100-mm
(4-in.) aggregate. Also, the use of a vibrating table in lieu of rodding (ASTM
Test C 29) is suggested as a better procedure for heavy aggregates.
Specific Gravity--In the design of a concrete shield the unit weight
429
selected for the concrete is of primary importance. The specific gravity

of the aggregate is the major factor influencing unit weight and is, therefore, of considerable significance in specifications for heavy aggregates.
The specific gravity, for both fine and coarse aggregate, is determined on
a saturated surface-dry basis in accordance with appropriate sections of
ASTM Test for Specific Gravity and Absorption of Coarse Aggregates
(C 127) and ASTM Test for Specific Gravity and Absorption (C 128).
Again, the sample weights should be increased according to the procedure
described earlier under Gradation. To produce concrete weighing 3700 to
3850 kg/m ~ (230 to 240 lb/ft 3) the aggregate must have a specific gravity of
4.5 or greater. To produce concrete weighing around 4800 kg/m ~, (300
lb/ftJ), the aggregate must have a specific gravity of 6.0 or greater. It is
necessary in specifications for heavy aggregate to give the minimum values of
specific gravity. Davis [10] gives the following minimum values: limonite or
goethite 3.25, magnetite 4.50, ferrophosphorus 6.10 for fine, and 5.90 for
coarse aggregate. The range of specific gravities of the various aggregates
used in shielding concrete is given in Table 1.
Resistance to Abrasion--The resistance of heavy aggregates to abrasion
and their tendency to fracture during handling and mixing varies considerably, and allowances must be made for the natural characteristics of
the heavy ores being used [10]. Magnetite and ilmenite, which can be purchased in satisfactory gradings, are harder than barite and the hydrous
ores, goethite and limonite, and are less prone to produce excess fines
during handling and mixing [9]. Since concrete shields are generally not
subjected to severe abrasion, aggregates which do not pass the abrasion
resistance requirement of ASTM Specification C 33 (maximum abrasion
loss of 50 percent) may be used, provided these aggregates produce concrete
of otherwise satisfactory properties.
Concrete for Radiation Shielding

The special concretes used for the construction of shielding structures
are distinguished from conventional concretes primarily by their higher
density, obtained through the use of heavy aggregates, and in some eases
by containing boron additives to improve attenuation of neutrons [13].
The practices to be employed in the proportioning, mixing, and curing of
the shielding concretes are basically the same as those used for conventional
concretes. Test methods used for the evaluation of the properties of fresh
and hardened conventional concretes are discussed in this publication by
various authors. In this paper only the special properties of shielding concretes, some of which are of less consequence in conventional concrete
structures, are discussed.
430
Special Properties and Their Significance

In the selection of mix proportions for a concrete to be used for shielding
purposes several properties need special consideration. They include the
workability and setting characteristics of the fresh concrete and the density
and strength of the hardened concrete. The effects of high-temperature
exposure and thermal cycling on the properties of hardened concrete also
need to be considered.
Workability--To facilitate the proper placing of concrete in forms (frequently of complex shape and containing many embedded items) and to
ensure the integrity of the concrete in the shield (avoiding segregation and
cold joints), the concrete mix must be extremely workable. Because heavy
aggregate mixes tend to be harsh, more fine aggregate is usually used than
in conventional concrete, and the sand frequently contains a high percentage
(10 to 16 percent) of material passing the No. 100 sieve [7,10]. Cement
contents of the concrete are usually high 350 to 450 kg/m a (560 to 750
lb/yd 3) and even higher, because of desired workability or fixed water content. Frequently, when a shield includes closely spaced reinforcement or a
high concentration of embedded items, it becomes necessary to employ other
than conventional methods of placing, such as the preplaced-aggregate
method.
Setting Characteristics--The setting characteristics of a concrete need to
be considered when boron additivies are used in neutron shields, since
some of these borates may adversely affect the setting and strength properties of the concrete. As discussed earlier, boron is an efficient element in
shielding against neutrons. However, most naturally occurring borates like
colemanite and borocalcite are somewhat soluble, and borate solutions are
powerful retarders of the set of concrete and also of the rate of strength
gain [8, 9]. In addition, borate ores frequently contain impurities that either
tend to reduce the ultimate strength of concrete or cause erratic setting
behavior or both. Henri [8] discussed the influence of boron-containing
additivies, such as colemanite, borocalcite, and boron frit, on the setting
characteristics of concrete. Colemanite is generally considered the poorest
of the three in this regard. However, it has the advantage of being more
economical in some geographical areas. It contains about 10 percent boron.
Borocalcite is a calcium borate similar to colemanite and contains about
13 percent boron. It also causes erratic setting characteristics of the concrete. To reduce the effect of these naturally occurring borates, all the fines
passing a No. 30 sieve are usually eliminated. This lessens the exposed
surface area of a given weight of the material and thereby reduces its solubility. These materials were used successfully in concretes, frequently in
conjunction with calcium chloride (about 1 percent by weight of cement) to
overcome the retardation.
Low-sodium boron frits are an excellent source of boron (about 18 perCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:36 EDT 2014
431
cent) and have little effect on the setting characteristics of concretes. These
materials, being refined products produced by fusion of different compounds, may be more costly than the naturally occurring borates.
It is highly desirable that prior to the use of borate additives a thorough
evaluation of their effects on the properties of concrete, especially setting
characteristics, be made.
Density--Though the density or unit weight of the concrete used in a
shield is a major consideration, the designer also must consider all of the
shielding characteristics of the concrete, as well as cost and construction
problems. As stated earlier, conventional concrete (weighing say, 2400
kg/m 3 (150 lb/ft 3) provides the most economical and is the most commonly
used shielding material. Only due to limited space or other conditions of
construction will concretes of higher densities be employed.
The density of the concrete is governed principally by the specific gravity
of its aggregate, as previously discussed. The density is also influenced by
the cement, water, and air content of the concrete mix. ASTM Test for
Unit Weight, Yield, and Air Content (Gravimetric) of Concrete (C 138) is
applicable to the determination of the unit weight of shielding concrete.
Typical ranges of densities commonly specified for concretes made with
some of the heavy aggregates are listed in Table 2.
TABLE 2--Densities of shielding concretes commonly specifiedfor given heavy aggregates.
Type of Aggregate
Density, Ib/ft 3
Barite
Magnetite
Ilmenite
Ferrophosphorus
Barite and boron additive
Magnetite and boron additive
215
220
220
285
200
210
to
to
to
to
to
to
225
235
240
300
215
225
Density, kg/m 3
3450
3500
3500
4550
3200
3350
to
to
to
to
to
to
3600
3750
3850
4800
3450
3600
Densities of about 5300 kg/m 3 (330 lb/ft 3) can be produced when using
iron as both fine and coarse aggregate. Densities other than those shown
in Table 2 can be obtained by the use of various combinations of natural or
manufactured aggregates. References included with this paper give the
properties of many typical concrete mixes containing heavy aggregates.
Strength--Although structural strength may be a requirement for highdensity shielding concrete, this property usually causes no problem because
sections are thick and the mixes have high cement content [14]. For structural shielding concrete, 21 to 35 MPa (3000 to 5000 psi) compressive
strength is usually sufficient; for massive walls, 14 MPa (2000 psi) is adequate. However, concrete for radiation shielding must have adequate
432
strength to resist cracking during its service life. Some of the strengthdetermining properties of hardened concrete are discussed by Lott and
Kesler [15].
In a study by the Corps of Engineers [9,16], it is shown that highstrength concrete can be made using materials that provide high density.
The high-density, high-strength concrete produced for this investigation
was made with a Type I portland cement, using magnetite or ilmenite
aggregate of 12.5 mm (or 11/2 in.) 37.5 mm maximum size. The concretes
had cement contents of 430 to 575 kg/m 3 (720 to 970 lb/ft3), water/
cement (w/c) ratios of 0.30 and 0.35 by weight, and slumps of about
12.5 mm (1/2 in.). The unit weight of these concretes was about 3700
kg/m 3 (230 lb/ft3). The compressive strengths at 7 days were in the range
52 to 65 MPa (7600 to 9400 psi), at 28 days, 62 to 76 MPa (9000 to 11 000
psi), and at 90 days, 69 to 83 MPa (10 000 to 12 000 psi). It is concluded
in this study [9,16], that although no new technology is needed to produce
uniform high-density concrete of high strength, adherence to the best
practices in production and control of such concretes is absolutely essential.
Effects of High-Temperature Exposure--The influence of high temperature and thermal cycling on properties of concretes are of major importance in the design of concrete shielding structures [17]. These effects of
temperature were studied by Davis [ 7,13.17,18], and the following evaluation
of effects of temperature is taken from a recent paper [18].
Fortunately, from a strength standpoint, concrete in a reactor shield is
usually over 6 or 9 months old before it is subjected to nuclear heating and
is, therefore, mature in hydration and strength.
The loss in compressive strength that occurs when mature concrete is
subjected to high temperatures ranges from 20 to 50 percent, even when
the concrete is heated as high as 320 or 430~ (600 or 800~ for long
periods. However, for important structural members, concrete should not
be exposed continuously to temperatures much above 260 to 320~ (500 to
600~
Concrete is almost completely dehydrated when heated above
430~ (800~
If made with heat-resistant aggregates, it may still retain
25 to 50 percent of its strength at temperatures as high as 700 or 760~
(1300 or 1400~
On the other hand, if unstable aggregates or high w/c
ratios (over 0.60 by weight) are used in the mix, deterioration at temperatures above 320 to 430~ (600 to 800~ may be appreciable.
If the concrete is subjected to wide and frequent fluctuations in temperature, the loss of strength of the concrete may be several times as great as
that observed for concrete exposed to a constant high temperature.
If concrete contains special aggregates, such as limonite or serpentine,
to provide hydrogen for neutron attentuation, the temperatures will be
limited to a safe value below that which dehydrates the aggregate: about
180 to 200~ (350 to 400~ in the case of limonite or as high as 370~
(700~ for serpentine. On the other hand, the maximum temperature in a
433
concrete shielding structure may be limited to less than 90~ (200~

to
minimize moisture loss, temperature gradients, or surface cracking.
Davis [17] shows that at high temperatures, tensile, flexural, and bond
strength are affected more than compressive strength. In most shielding
structures, thermal effects are of prime consideration in design.
References
[1] Callan, E. J., "Concrete for Radiation Shielding," Proceedings, American Concrete
Institute, Vol. 50, 1954, pp. 17-44.
[2] Foster, B. E., "Absorption by Concrete of X rays and Gamma Rays," Proceedings.
American Concrete Institute, Vol. 50, 1954, pp. 45-63.
[3] "Permissible Dose from External Sources of Ionizing Radiation," NBS Handbook 59,
National Bureau of Standards, 24 Sept. 1954, and insert dated 8 Jan. 1957.
t4] "Protection Against Neutron Radiation Up to 30 Million Electronvolts," NBS Handbook 63. National Bureau of Standards, 22 Nov. 1957.
[5] Wollenberg, H. A. and Smith, A. R., "Low-Background Concrete," Transactions,
American Nuclear Society, Vol. 8, No. 1, May, 1965, p. 183 (summary only). Paper presented at the 1965 annual meeting of the American Nuclear Society, Gatlinburg, Tenn.,
1965. Also in report form: UCRL-11674, Lawrence Radiation Laboratory, University of
California, Berkeley, Calif., 18 Sept. 1964.
[6] Komarovskii, A. N., "Shielding Materials for Nuclear Reactors," translated from Russian
by V. M. Newton, Pergamon Press, London, 1961, p. 145.
[7] Davis, H. S., "Concrete for Shielding, Nuclear Radiations," Progress Report, Committee on Nuclear Structures and Materials, Proceedings. American Society of Civil
Engineers, Vol. 88, No. ST 1, Feb. 1962, Part I, Paper 3062, 1962.
[8] Henri, J. O., "Properties of Nuclear Shielding Concrete," Proceedings, American Concrete Institute, Vol. 56, 1960, pp. 37-46.
[9] "High-Strength, High-Density Concrete," Technical Report No. 6-635, Corps of Engineers, Waterways Experiment Station, Nov. 1963, p. 65.
[10] Davis, H. S., "Aggregates for Radiation-Shielding Concrete," Transactions, American
Nuclear Society, Vol. 8, No. 1, May 1965, p. 178 (summary only), paper presented at
the 1965 annual meeting of the American Nuclear Society, Gatlinburg, Tenn., 1965.
Full text published by ASTM in Materials Research and Standards. Vol. 7, No. 11, Nov.
1967.
[11] Mather, K., "Application of Petrography to Radiation-Shielding Concrete," Transactions.
American Nuclear Society, Vol. 8, No. 1, May 1965, pp. 177-178, paper presented at
the 1965 annual meeting of the American Nuclear Society, Gatlinburg, Tenn., 1965.
[12] Raphael, J. M., "Structural Properties of Magnetite Concrete," Proceedings, American
Society of Civil Engineers, Vol. 84, No. ST 1, Jan. 1958, Paper 1511, 1958.
[13] Davis, H. S., "High-Density Concrete for Shielding Atomic Energy Plants," Proceedings,
[14] Washington, W. O., "Design Requirements of Shielding Concrete," Transactions,
American Nuclear Society, Vol. 8, No. 1, May 1965, pp. 176-177 (summary only), paper
presented at the 1965 annual meeting of the American Nuclear Society, Gatlinburg,
Tenn., 1965.
[15] Lott, J. L. and Kesler, C. E., "Problems of Hardened Concrete," Transactions, American Nuclear Society, Vol. 8, No. 1, May 1965, pp. 178-179 (summary only), paper presented at the 1965 annual meeting of the American Nuclear Society, Gatlinburg, Tenn.,
1965.
[16] Mather, Katharine, "High Strength, High Density Concrete," Proceedings, American
[t 7] Davis, H. S., "Thermal Considerations in the Design of Concrete Shields," Proceedings,
American Society of Civil Engineers, Vol. 84, No. ST 5, Sept. 1958, Paper 1755. 1958.
434
[18] Davis, H. S., "Effects of High-Temperature Exposure of Concrete," Transactions.

American Nuclear Society, Vol. 8, No. 1, May 1965, p. 182 (summary only), paper presented at the 1965 annual meeting of the American Nuclear Society, Gatlinburg, Tenn.,
1965. Full text published by ASTM in Materials Research and Standards, Vol. 7, No. 10,
Oct. 1967.
Bibliography
"Concrete for Nuclear Reactors, Vol. I, II and llI," Publication SP-34, American Concrete
Institute, 1972. (Includes 75 papers presented at an international seminar on concrete for
nuclear reactors, Berlin, 5-9 Oct. 1970. Several of the papers were prepared and presented
by members of ASTM Committee C-9.
Davis, H. S., "Concrete Shields--Their Preparation and Use," General Electric Review,
Vol. 59, No. 6, June 1956, pp. 43-45.
Davis, H. S., "High-Density Concrete for Reactor Construction," Civil Engineering, Vol. 26,
No. 6, June 1956, pp. 52-56.
Davis, H. S., Brown, F. L., and Witter, H. C., "Properties of High-Density Concrete Made
with Iron Aggregate," Proceedings, American Concrete Institute, Vol. 52, 1956, pp. 705726.
Davis, H. S., "Cement and Aggregates for Shielding in Atomic Energy Plants," Mining Engineering, Vol. 9, No. S, May 1957, pp. 544-548.
Davis, H. S. and Borge, O. E., "High-Density Concrete Made with Hydrous-Iron Aggregates,"
Proceedings, American Concrete Institute, Vol. 55, 1959, pp. 1141-1147.
"Engineering Compendium on Radiation Shielding, Vol. II Shielding Materials," International Atomic Energy Agency, Springer-Verlag, Berlin, 1975.
Polivka, M., "Annotated Bibliography on Concrete for Radiation Shielding," Concrete for
Nuclear Reactors, Vol. III, Publication SP-34, American Concrete Institute, 1972, pp. 16391731.
S. B. H e l m s ~
Chapter 27
Weight
Air Content and Unit
Terminology
The density of concrete is the mass of the solid per unit of volume and,
considering weight synonymous to mass, the term unit weight can be used
to signify density expressed as kilograms per cubic metre (or pounds per
cubic foot).
In engineering and scientific work, density of substances is frequently
indicated as grams per cubic centimetre and when the numbers so expressed
are listed without stating the units they may be assumed equal to the specific
gravity for all practical purposes.
In the technology of concrete and aggregates, where the specific gravity
is known, 1000 is the practical multiplier used to determine the unit weight
in kilograms per cubic metre (or 62.4 in lb/ft3). As an example, "solid
unit weights" of aggregates derived from these multipliers are used in mix
design procedures involving the summation of solid volumes of the mix
components.
Air content is expressed as percent by volume irrespective of the size
distribution of the minute air voids beneficially incorporated in air-entrained
concrete by air-entraining agents.
Determining Test Values

The measurement of density of hardened concrete is based on procedures
that are simple and direct, using samples such as molded specimens, drilled
cores, or portions taken from hardened structures and trimmed to cylinders
or prisms of sufficient size to be representative.
It is reasonable to require unit weight by displacement measurement to
be reported to the nearest 1.6 kg/m 3 (0.1 lb/ft 3) and within an accuracy of
0.1 percent. Accurate weighings are used in determinations by displacement techniques such as suspended-immersed methods, or by weight-inI Materials engineer, Bechtel Power Corp., Gaithersburg, Md. 20760.
Universidade
do Gois pursuant
Agreement.
No further reproductions authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
436
air of the sample divided by its volume as calculated from measured dimensions. This latter procedure is especially adaptable for determinations on
dry specimens using precise measurements of linear dimensions within an
accuracy of 0.2 percent. The apparatus and instruments used for these
purposes are familiar and it is unnecessary to describe the detailed procedures commonly used or to comment on their significance.
In contrast, the accepted techniques for determination of air content,
after concrete has set, require specialized equipment and sophisticated
techniques if exact information is desired. Where an estimated air content
will suffice to differentiate, or to be indicative of the general level of air
content, less discriminating methods may be used which are predicated on
precise density measurement, specific gravities of the concrete constituents,
and known or determined percentages of the ingredients. While methods
for unit weight are traditional, those for air content have developed since
the time the advantages of entrained air were recognized, as the means for
studying the effects of air content on other properties and especially the
resistance to frost action.
Representative Densities and Ranges of Porosity

The unit weight block diagram was developed to describe the principal
types of concrete which are of commercial significance. Figure 1 shows
these in one illustration which serves to indicate the scope of this paper,
insofar as air content and unit weight should be considered, whereas the
other characteristics of concrete in certain other major categories (such as
cellular, lightweight, radiation shielding) are covered separately as topics of
special interest in this publication.
Figure 1 does not delineate exact limits because such variables as moisture content and richness of mix have an appreciable effect on the density.
Hardened unit weight exceeds the respective fresh unit weight due to the
influences of absorption, bleeding (water gain), and settlement. Voids in
vesicular aggregates contribute to reduced density but are not considered
to be part of the air content charted on the figure. The arbitrary dividing
lines between zones were drawn to correspond to available data in published
references.
For purposes of this discussion, concretes are classified on the "block
diagram" according to the air content as: "non-AE" 0 to 2 percent, "air
entrained" 2 to 10 percent, and "aerated" from 10 to 30 percent air. Other
subclasses, according to unit weight, are: "lightweight" 320 to 1922 kg/m ~
(20 to 120 lb), "normal weight" 1922 to 2643 k g / m 3 (120 to 165 lb), and
"heavyweight" 2643 to 5286 kg/m 3 (165 to 330 lb/ft3). Influences of moisture content and variations of aggregate physical properties affecting concrete density were not considered critical in constructing the block diagram
which serves a descriptive purpose.
437
HELMS ON AIR CONTENT AND UNIT WEIGHT
] ~Stee]
~ --Steel
500~
Shot. Punchings & Magnetite Aggregate
coarse & Magnetite fine
--Steel coarse & Limonite fine
l
~
~Eerro
Alleys
5286
--Iron & Limonice

--lron OIe & S t e e I
400
Punchings
--Magnetite coarse and fine
f~
/ B a r i t e Aggregate - USBR Design

--H,,dreus Iron Aggregate - NPDL Design
--llmenice Aggregate - HEPC Design
300
e~
Natural Heavy Iron Ores

~ M i c h i g a n Limonite Aggregate
CO3~ACTED PIPE
centrifugated,
~ S S
CONCRETE and
ASBESTOS CEMENT PIPE
2000
~SONRY
~ORTAR CUBES
1922
PASTE-FOAM-ORAVEL
10
STRUCTURAL
LI~I~T
.~'LIGHT
1000
~-643
E~A~ED AGGRE~TE
FILL CONCRETE
WT. BLOCK
EXPANDED AGGRECATE
"low fines" Concrete
Pumice
H ~
L ~ N LIGHT ~ .
]
MIXTURES including[
Perlite-Sand Blends
INSULTING P ~ S T E R S
Wermiculite and
Perlfte MORTARS
[- . . . . . . .
--_~ .~
320
non
air entraineo
AE , l ~ . - - - - c o n c r e t r
~
~ir entrained
aerated mixtures
mortars
~0
AIR C O ~ E ~ ,
,
2O
per cen~
&
30
FIG. 1--Unitwe~htblock d~gram.
Review of Principles of Density

Feret [1] 2 represented the "density" of mortar as the total volume of the
solid particles per unit of volume. This was similar to the current application
of the method of absolute volumes involved both in proportioning methods
and in the analysis of the physical constants of concrete. Detailed discussions of the validity of the relationships established by Feret are included
in Refs 2 and 3.
After the relationship between water content and richness of mix was
established by Abrams' water/cement (w/c) ratio, Talbot and Richart [4]
utilized the concept of the cement/space ratio which now may be applied
in evaluating data of air-entrained and lightweight concrete. Their Fig. 12
[4] applicable to their Series 2-G mortars and concretes, made with Wabash
438
River aggregates, shows a linear relationship between the unit weight of

concrete and the voids as represented by the line
unit weight, k g / m 3 = 16.02[167(1-v-c) + 193c + 55v]
or
unit weight, lb/ft a :
167(1-v-c) + 193c + 55v
This line closely fitted all experimental points plotted for these materials;
v and c are absolute volumes of voids and cement, v being the sum of water
and air voids. Talbot and Richart recognized that levels of water content
affected compressive strength at a given ratio and employed a strength
reduction curve to normalize their charts of compressive strength versus
cement/space ratio.
The effect of air voids on unit weight is more pronounced than that of
water voids, and changes of unit weight are therefore indicative of the air
content of air-entrained concrete. Feret [5] distinguished between "airentrained" in mortar, water voids, and cavities quite a long time ago.
Heretofore there has been no method to distinguish and discriminate the
different types and sizes of air voids in freshly mixed concrete; in 1969 a
technique was developed to facilitate such distinction. The device called
void spacing indicator [6, 7] was introduced for studying the quality of the
entrained air system in plastic concrete and was shown to be a valuable indicator of the frost resistance of the hardened concrete.
The cement/space ratio concept advanced by Talbot and Richart was
utilized by Morris [8] who developed a mortar voids methods for designing
concretes. Illustrations in Morris' paper correlate strength with the cement/
space ratio as shown in the original Bulletin No. 137 [4]. Such concepts
were extended to the design of air entrained concrete mixtures by Thornburn [9]. Charts of the type used by Morris have unique utility in establishing strength versus voids correlations in presentation of design data for
the lightweight insulating or aerated classes described by Fig. 1.
Effect of Aggregate Density

Condensed data from Bureau of Reclamation reports [10] presented in
Table 1 show the variation of concrete unit weight, in round numbers, as
attributable to specific gravity of the aggregate used in the mix. For structural concrete the values usually range from 2371 to 2435 kg/m 3 (148 to
152 lb/ft 3) with 2403 (150) being the reasonable assumption for dead load,
whereas it can range as high as 2499 k g / m 3 (156 lb/ft 3) in massive construction using larger maximum size aggregate.
Effect of Paste Content and Degree of Compaction

Cement pastes in the practical range of 0.019 to 0.034 m 3 (5 to 9 gal)
,=
<
[..
E~E~
439
440
of water per sack of cement, or w/c ratio of 0.44 to 0.80, have air-free unit
weights from 1602 to 1922 kg/m 3 (100 to 120 lb/ft 3). Since normal aggregates usually range from about 2563 to 2803 kg/m 3 (160 to 175 lb/ft3),
it is clear that a reduction of paste volume by substitution of aggregate
will increase the unit weight. Table 2, which applies to mass concrete, is
included as an example of increased unit weight due to reduction of the
amount of paste or mortar in the mixture. In mixtures of stiffer consistency,
of 50 mm (2 in.) slump or less, the degree of compaction as effected by
vibration would have a more pronounced effect.
Important factors in the consolidation of concrete are reviewed in the
report of ACI Committee 309 [11]. Some applicable test results given in
the paper by Grieb et al [12] are shown in condensed form in Table 3. By
using vibration to facilitate compaction it is practical to lower the sand
proportion and reduce the slump required for proper placing. Each of these
changes permits lower water content and thereby increases unit weight.
With improved densification and adjustment of mix proportions, as described, the mortar or paste will be of better quality and show higher
strength for a given cement content. The advantage of proper compaction
of concrete in the field is as obvious as the adverse effects of incomplete
or improper consolidation.
The ACI 309 report cites the benefits of vibration and revibration in
elimination of entrapped water under aggregate particles, and under steel
reinforcement, and the elimination of extraneous air voids, to improve
bond. Published results [11] on effects of overvibration have shown that
it can reduce entrained air to low levels; however, it was found that extra
vibration to eliminate undesired large voids "will not damage the parameters of the air void system . . . essential to best durability if the concrete
initially contains the amount of air recommended in ACI 211-73. While
half of the initial air could be lost due to overvibration, the requirements
for frost resistance apparently would still be met unless the initial air content was lower than that prescribed in ACI 211 Recommended Practice."
Recognition of Advantages of Entrained Air

A study launched about 1935 [13] showed that cores from certain experimental pavements having high resistance to freezing and thawing contained portland cement that had been blended with either of two natural
cements which happened to contain about 0.07 percent of fat or grease.
Lowered unit weight was the clue leading to understanding that resistant
slabs contained larger amounts of entrained air than did nonresistant slabs,
and studies of air entraining materials were started. It has been stated [14]
that the first work correlating freezing-and-thawing tests with air content
was reported by the Portland Cement Association in June 1938. But the
term "air-entraining cement" was not introduced until after their Study of
h2.98
1:2.80:1.05
1:2.79:2.65
1:2.46:4,26
1:2.34:6.07
No. 4 (4.8 mm)

3/8 in. (9 mm)
90 in. (19 mm)
11/2 in. (38 mm)
3 in. (76 mm)
3.6
3.8
3
3.4
3,2
(91)
(96)
(76)
(86)
(81)
Slump,
in. (ram)
42.9
36.7
30.1
26.5
22.4
Volume %
of Paste b
Vibrated
137.2
141.8
148.0
152.9
154.6
137.7
140.6
149.2
155.1
158.1
2198
2272
2371
2449
2477
2206
2252
2390
2485
2533
lb/ft 3
kg/m 3
Rodded Vibrated Rodded Vibrated
2360
2480
2450
2440
2280
2400
2590
2480
2400
2140
psi
Rodded Vibrated
16.27
17.10
16.89
16.82
15.72
16.55
17.86
17.10
16.55
14.75
MPa
Rodded Vibrated
7 Day Compressive Strength
aCondensed from Table 61 Bureau of Reclamation, Bulletin 4, Cement and Concrete Investigations, Boulder Canyon Project.
b Water/cement ratio 0.56 by weight.
Weight
Proportions
Maximum
Sieve Size
of Aggregate
Unit Weight
TABLE 2--Unit weight and strength of hand rodded and vibrated specimens data from mass concrete investigations, a
I
ill
0
--r
c
z
-q
>
z
c~
m
z
-H
O
O
6O
O
Z
>
m
t-
"Mix No. 2 results.

bFor six cited tests of representative retarders.
150.3
150.1
147.0
145.3
151.1
149.9
149.0
153.5
152.5
152.2
150.8
153.8
152.1
152.5
5
8
11
12
19
21
Average b
Nonvibrated
lb/ft 3
Vibrated
No.
Retarder
2459
2443
2438
2416
2464
2436
2443
Vibrated
Unit Weight of Hardened Concrete
2407
2404
2355
2327
2420
2401
2386
kg/m 3
Nonvibrated
6500
6050
6240
5690
6170
6210
. . .
psi
44.8
41.7
43.0
39.2
42.5
42.8
MPa/m 2
Vibrated Cylinders
1
1
2
16
2
5
4.5
Increase Due
To Vibration, %
T A B L E 3--Data representative o f Grieb's Table 10, R e f 12. Effect o f delayed vibration on compressive strength o f concrete at 28 days. a
-I
I'1"1
0
0
Z
0
O
.-fl
(,9
O
-o
m
33
"O
2~
m
o~
--4
co
4~
Ix)
J~
443
Cement Performance was instituted in 1941, when these were known as

"treated" cements. Improved scaling resistance of pavements made with
"treated" cement was attributed to entrained air by Moore [15] who stated:
"Both Vinsol resin and tallow reduce the unit weight of concrete by introducing air, uniformly distributed throughout the mass in microscopic
voids . . . . The reduction in compressive strength and unit weight of concrete
which accompanies the increase in resistance to freezing and thawing may
appear disturbing at first thought for it has generally been accepted that
the denser and stronger the concrete, the greater its durability [16]. The
belief that durability increases with strength is undoubtedly true up to the
point where requirements for good concrete are satisfied, but beyond this
point the theory that strength is an index of durability may not be true."
Vollmer [17] utilized fresh unit weight data to show the effect of calcium
chloride admixture on the calculated air content of concretes made with
Vinsol resin cement and concluded that use of calcium chloride reduced
air. However, more refined tests by Pigman [18], Gonnerman [19], and
others [20] have shown that entrained air is increased slightly by the incorporation of calcium chloride.
The density of saturated hardened concrete is consistently found to exceed that of the respective fresh concrete. This is probably attributable
to two effects combined; a small reduction in the air content is usually
experienced and higher water content can result from prolonged moist
curing or resaturation. Therefore, the fresh unit weight should not be used
as the basis of absolute volume calculations for air content of the cured
concrete unless approximate values for air will suffice.
Data selected from Tables 5 and 6 of Bugg's paper [21], while seemingly contradictory to the foregoing statement, are of interest as listed in
Table 4. The mixtures had 76 to 102 mm (3 to 4 in.) slump and contained
335 kg/m 3 (6 sacks) of Type 1 cement. The unit weight relationships could
have been affected by variations of coarse aggregate in samples, bleeding
tendency, degree of compaction, curing, and moisture condition of beams
at time of measurement. The fact that the sets of unit weight data for
several mixes agree Closely could be misleading, since unit weight of moistcured specimens should be higher than that of fresh mix. In a few cases a
surprising difference in unit weight is shown, corresponding to higher
computed air in hardened mix, because of variations such as loss of moisture. See Table 11 of Ref 12.
Axon [22] listed the changes in weight of hardened concretes containing varying amounts of entrained air during the progress of moist curing,
water curing, and air curing; beams gained approximately 0.5 percent in
weight in 28 days of water curing and were shown to lose from 2 to 2.7
percent in weight during 2 weeks of air drying. Since sonic E is a function
of the true density of the specimen, E values in the last column of Table 4
show a reasonable relationship to hardened unit weight.
Moisture Condition,
Prior to Batching
vacuum saturated
vacuum saturated
immersed 24 h
immersed 24 h
vacuum saturated
vacuum saturated
immersed 24 h
immersed 24 h
vacuum saturated
vacuum saturated
Type, and
Characteristics
Limestone 67-25
0.7 % absorption
gravel 79-IG
1.6% absorption
gravel 79-IG
1.8% absorption
gravel 82-IG
1.8% absorption
gravel 82-1G
2.5 % absorption
Coarse Aggregate
3.3
0.6
3.6
0.3
3.6
0.3
4.0
0.9
4.8
0.5
Air
Content,
Percent
148.8
152.8
150.1
153.9
150.1
154.6
146.1
150.9
144.4
151.0
2383
2448
2404
2465
2404
2476
2340
2417
2313
2419
Fresh Concrete
Unit Weight
lb/ft 3
kg/m 3
149.7
154.1
148.9
153.6
145.1
149.9
143.8
149.8
152.8
148.8
2383
2448
2398
2468
2385
2460
2324
2401
2303
2399
Cured
Unit Weight
lb/ft 3
kg/m 3
825
865
800
830
755
870
680
735
645
725
5.69
5.96
5.52
5.72
5.21
6.00
4.69
5.07
4.45
5.00
Flexural
Strength
psi
MPa
At Age 28 Days
5.52
5.99
5.77
6.04
5.68
6.11
5.59
6.00
5.25
6.00
38.1
41.3
39.8
41.6
39.2
42.1
38.5
41.4
36.2
41.4
S o n i c E X 10 - 6
psi
kPa
TABLE 4--Unit weights of fresh concrete and (76 by 102 by 406 ram) (3 by 4 by 16 in. ) size hardened beams, after Bugg [21].
m
.-1
m
o
0
z
0
O
11
O
"o
ill
z13
-13
-n
--I
o')
.Ix
445
It is also of interest that Bugg studied similar concretes with air contents
ranging up to 7.5 percent and concluded that no significant increases in
freezing-and-thawing durability occurred with air contents in excess of 3
percent except when very porous aggregates were placed in concrete in a
highly saturated condition.
Measurement of Air Content of Hardened Concrete
The simplest estimate of total amount of air in concrete can be derived

from unit weight, provided the specimen is representative of known mix
proportions and is in a saturated state. Several papers in the 1947 ASTM
Symposium on Measurement of Entrained Air [23] listed data for volumetric
method versus the gravimetric (unit weight). Most published gravimetric
tests apply to fresh concrete. Figure 3 of Swanberg's paper [23, p. 871],
and Table X in Gregg's paper [24, p. 74] used unit weight test data to
substantiate results by the pressure method.
In checking hardened concrete for the presence of entrained air, such
as in correlation studies of pavement construction, it is not always essential
to determine the size distribution of air voids when one is primarily interested in knowing the amount of air contained, on the assumption the
void sizes are normal.
Development of the high pressure air meter was recounted by Lindsay
[25] who reported the successful application of a high pressure meter of
special design, operating at a test pressure of 34.5 MPa (5000 psi) to determine air content of hardened concrete. The equipment accommodates
drilled cores up to 127 mm (5 in.) in diameter and 254 mm (10 in.) in
length. Seventy-two hours of oven drying with subsequent 48-h resaturation
of specimens is essential to conduct tests according to the procedure deve.loped at Illinois Division of Highways laboratory under Larsen and Blandin.
This meter is an excellent means of differentiating between concretes of
different air contents, and comparisons of cylindrical specimens have
shown reasonable agreement with air contents determined by microscopical
means. Kennedy's discussion [25] indicated his appreciation of this method.
Other tests at Waterways Experiment Station [26] have shown the high
pressure meter to be a satisfactory tool for such tests.
Additional Illinois Division of Highways data obtained under conditions
of laboratory control are given in Table 5. Some data for 51-mm (2 in.)
diameter cores were omitted for brevity. However, 102-mm (4 in.) diameter
cores were the size usually tested in the high pressure meter. Results for
the original fresh mix are compared in Table 5 to cores drilled from hardened 152 by 152 by 760-mm (6 by 6 by 30-in.) beam specimens molded
from the test concretes.
Erlin [27] modified this apparatus to eliminate using a testing machine
to apply high pressure.
446
TABLE 5--Air meter test comparisons, values for laboratory concrete in plastic state and the
same concrete in the hardened state, after Dykins and Blandin [28]. (Selected data from
Ref 28 Table 2; 4-in (102-mm) diameter cores only.)
Air Content in Hardened Concrete, %
Subgroup
Individual Core Determinations
Average Air Concent, %
Hardened Concrete
Plastic
Concrete
GravelCoarseAggrega~
A
C
D
E
0.5
2.0
3.6
5.7
0.3
1.8
3.5
5.3
F
H
I
J
0.7
3.0
4.8
7.1
0.7
2.9
4.5
7.0
0.1
t.8
3.2
4.9
0.1
1.6
2.9
4.9
0.2
1.8
3.3
5.2
0.2
2.3
3.6
5.0
0.6
2.8
4.6
6.9
0.4
2.7
4.4
6.5
~ushedStone ~ a ~ e A ~ r e g a t e
0.4
2.6
4.5
6.8
0.4
2.5
4.5
6.7
Loss of Entrained Air as Shown by Tests on Hardened Concrete
Whereas tests under closely controlled conditions result in good agreement between measured air contents in fresh and hardened states, it is
generally accepted that a loss of entrained air will be experienced between
the mixing of concrete and finally placing it in the forms. Compilations
of the State of Illinois have indicated this difference can range from a few
tenths of a percent up to nearly 1.0 percent air.
Sweet [29] pointed out that the air volume in hardened material may
differ from the volume of air entrained in fresh mix, due to changes in
water content, concrete volume, and specific volume of ingredients. In his
ASTM paper [30], data were presented on changes in weight (water absorbed) and unit weight in curing prior to freezing-and-thawing tests, and
it was shown [29] that the average air content fell from an initial value of
3.0 percent to 2.0 percent at 28 days.
Searching for an explanation for about 20 percent less air in hardened
concrete, computed from absolute volumes, Bloem and Walker [31] studied
the apparent specific gravity of cement in water versus time and, on this
basis, declared that use of such specific gravity as indicated for 72 h, the
time concrete displacement data were obtained, would account for the
differences found, after allowance was made for water lost in molding.
In U.S. Bureau of Reclamation tests made at about the same time,
changes in theoretical unit weight of concrete with time, due to hydration,
were also recognized, as shown in Table 6.
When the foregoing values were applied in calculations, the agreement
with the pressure test method provided evidence of accuracy. In a series of
447
TABLE 6--Changes in theoretical unit weight of concrete with time.
Time After Mixing, h

0
1
2
3
4
Theoretical Unit Weight ~

lb/ft a
kg/m 3
149.5
150.4
150.9
151.2
151.4
2395
2409
2417
2422
2425
a See chart on Fig. 2 of Ref 32.
ACI articles by Mielenz et al [33], interesting facts were presented on the

evolution of the air void system and the possible changes which take place
in freshly mixed concrete. Part I of the series covered the action of air bubbles in freshly mixed concrete, comprising a treatise on factors governing
size, distribution, and stability of entrained air during hardening of the
mixture. Their Table 4, Part I [33], includes data for concrete specimens
containing two superior air entraining agents and shows the air content
of hardened mixture 0.5 to 1.1 percent lower than the 5.1 percent measured
by pressure meter on the fresh mix.
Bruere [34], referencing the papers by Mielenz et al, studied air-entrained pastes during setting. In one of his experiments, using a neutralized
Vinsol resin air-entraining agent, the bubble systems of pastes of normal
and accelerated setting characteristics were compared. While the case for
paste-aggregate combinations was not studied, the conclusion based on
Bruere's paste data was that rearrangement of bubble sizes and solution
of air from small bubbles were negligible in the period during which airentrained pastes are fluid prior to setting. By this work Bruere has provided a means whereby the size distribution of air voids in fresh concrete
may be studied by arresting the fresh void system (introducing an agent
which induces quick set) then using microscopical methods developed for
samples of hardened concrete. Otherwise, it is practically impossible to
determine the size distribution of air voids in tests of fresh concrete. Powers
[35] provides a practical insight regarding the cited papers and other related references for those who need to apply the principles of studies such
as these in further investigations.
Quantitative Estimate of Air Content by ASTM Recommended Practice

Several techniques have been developed for hardened material which
are based on inspection of polished sections. Verbeck [36] applied the
camera-lucida method successfully by tracing the voids on an enlargement
of a microscopic view at a magnification of about 100 diameters to determine air content at desired locations within a specimen, making compariCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:38 EDT 2014
448
sons with other methods. An interesting feature of Verbeck's paper was

the tabulation of size and volume of voids. (The nature and size range of
air voids in Types I and IA cement concretes were shown clearly in cameralucida drawings in Klieger's [37] paper.) The published discussion of
Verbeck's paper and subsequent developments have shown that skilled
technicians at the laboratories of the Corps of Engineers, the Missouri Highway Department, and elsewhere were meeting the need for specialized
microscopical examination of hardened air-entrained concrete; this is discussed by other authors in this publication.3 With the additional evidence
provided by new techniques [38] for determining the nature of air voids,
Powers was able to advance the theory of permissible spacing to show that,
for high resistance to freezing and thawing in cement pastes, the average
void-to-void distance should be 0.5 mm (0.02 in.) or less.
The results reported by Warren [39] at the University of Wisconsin include photomicrographs of carefully polished interior sections of air-entrained concrete magnified to 40 diameters. Voids filled with fluorescent
dye in Canada balsam were photographed in ultraviolet light. Void properties, calculated from measurements by the plane intercept method, are
interesting because those introduced by two accepted air-entraining agents
had essentially the same specific surface, as is seen in Table 7 herein.
ASTM Recommended Practice for Microscopical Determination of AirVoid Content and Parameters of the Air-Void System in Hardened Concrete (C 457) has provisions for both linear traverse and modified pointcount methods. These procedures require skill and experience to achieve
desired accuracy. The recommended practice includes specific instructions
for proper conduct of the microscopical examination with numerous notes
explaining special details, and is extensive. Stated briefly, the technique is
based on preparation of representative plane interior surfaces by sawing
and fine grinding on suitable laps, then traversing the prepared surface
under a microscope to determine the relative amount, size, and spacing of
air voids present. The ability to distinguish air voids is dependent on the
qualities of the ground surface, locating a typical field for scanning, effect
of magnification, and the operator's skill using stereoscopic binocular with
the surface illuminated from one direction at a low angle. In this way the
air voids are shadowed and recognizable down to 15 #m size or less; 4 low
incident light (around 30 deg from horizontal) also aids in detecting air
voids, shallow because the cut surface may be close to the bottom of the
void. In determinations of void systems of high specific surface, better
perception at higher magnification will substantially increase the microscopically determined air. Although this procedure requires precision and
capable technicians, it is conducted regularly in research studies and has
proven to be a practical method for study on job concrete. The ASTM
3See p. 132 and p. 787.
4For details see ASTM Standards C 457, Section 5.2; C 823, Section 11.3; and C 586, Table 5.
Cement Factor:
( k g / m 3)
28.0
5.43
0.0604
0.118
5.5
140.3
(2248)
4 sacks
(223)
39.4
5.39
0.0669
0.100
5.5
143.1
(2292)
7 sacks
(390)
29.1
5.18
0.0536
0.113
5.2
141.1
(2260)
4 sacks
(223)
39.6
5.33
0.0567
0.094
5.3
142.8
(99~,~
-7 sacks
(390)
Neutralized Vinsol Resin

Admixture
~Defined as surface area of voids per cubic millimetre air in concrete.

bOne half average m a x i m u m distance from void to void within paste, about 0.004 in. At a given cement content, the listed data show little
difference between concretes made with the two agents. Photographs showed, at air content of 7 percent, the air voids were more numerous and
smaller in the concrete with higher cement content. Curves showed the average diameter of air voids was only slightly affected by cement content
and angularity of coarse aggregate.
Specific surface of voids, a m m 2

Air content, %
Average diameter of voids, m m
Void spacing factor, b m m
Net air content, pressure meter, %
Fresh unit weight, lb/ft 3
( k g / m 3)
Type I Cement Concrete Mixes
Darex Air Entraining

Admixture
TABLE 7--Void properties in 76-mm (3-in.)slump air-entrained gravel concretes (after Warren [39]).
.Ix
r
-r
.-4
c
z
z
E2
z
-4
m
z
3o
(3
O
-1m
rE
6o
O
Z
450
Recommended Practice C 457 procedures are the available methods for

the determination of A, the air-void content; a, the specific surface of air
voids; and Z,, the computed spacing factor.
Contrast of Purpose for Higher Levels of Air Incorporated in Concrete

Usually the amount of entrained air in hardened concrete is intended
to be within the range of 3 to 7 percent in the form of minute air voids of
high specific surface, ranging from 275 t o 550 cm 2/cm3 (700 to 1400 in.2/
in.a), in order tO minimize the sacrifice of mechanical properties and ensure
satisfactory durability. On the other hand, the inclusion of air in excess
of 8 percent will markedly reduce compressive strength. Figure 2 is a photo-
FIG. 2--Concrete containing excessive entrained air (X15). Air determined microscopically
to be approximately 30 percent.
micrograph of a concrete containing air voids at the high level of 30 percent which would be prohibitive from the structural standpoint but commonplace if cellular concrete were desired. This figure forms an interesting
contrast with the illstrations presented by Mielenz et al [33, Part 2, Aug.
1958, p. 262] to show normal void systems of concrete with 3.5 percent
air.
Polished sections of concrete with high air content, or weak for other
reasons, present problems in surface preparation. According to Mielenz
and Wolkodoff [33], " . . . at very high air content the bubbles tend to
coalesce, and the thin septa and ridges between the voids are easily lost
451
during grinding," adding the thought that these factors would impair the
accuracy of the linear traverse technique.
Thus there are two main purposes for incorporating air in concrete.
1. Air entrainment for frost resistance, and improved workability. If
properly controlled, the voids will be of high specific surface and provide
the required spacing factor of 0.2 mm (0.008 in.) or less.
2. Large amounts of air, introduced by aeration or generation of gas, in
the range of 10 to 30 percent or more air content, primarily to reduce the
density of the concrete product. For purposes of insulation and weight reduction the void size is not as critical as in structural applications.
Referring to the "non-AE" subclass mentioned earlier, in the explanation of Fig. 1, the smaller amount of air present in non-air-entrained concretes is known as incidental air, or as entrapped air, or accidental air.
In normal weight concrete this "incidental" entrained air is less than 2
percent, although higher values may result from inadequate compaction.
The amount of incidentally entrained air is dependent on mix proportions,
aggregate particle shape, aggregate absorption, amount of fines, organic
impurities, and other minor factors.
Determination of Unit Weight Using Standardized Procedures

Calibrated standard cylindrical measures are used under ASTM Test
for Unit Weight, Yield, and Air Content (Gravimetric) of Concrete (C 138)
to determine the unit weight of freshly mixed concrete. Theoretical weight
per cubic metre (cubic foot) was mentioned previously. When batch proportions and weights of ingredients are accurately known, the air content
is determinable by the method of absolute volumes--deducting the volume
of component ingredients from the calculated volume of the batch.
Where composition is known, some predictions of the air dry unit weight
can be made or calculated by making certain assumptions. Such predictions are dependent on the accumulation of data pertaining to equilibrium
air dry weight at a given relative humidity. To a certain extent air dry unit
weight values, while fictitious in nature, are useful in dead-load computations.
ASTM Test for Unit Weight of Structural Lightweight Concrete (C 567)
includes the testing of hardened concrete by determining the air dry weight
of standard 152 by 305-mm (6 by 12-in.) cylinders moist cured 7 days and
air dried at 50 percent relative humidity for 21 days. As mentioned in an
introductory paragraph of this paper, specimen volume is determined in
ASTM Test C 567 by making suspended-immersed weighings. Notably the
procedure given in Section 4 of this test is applicable to normal weight
concrete specimens, too. Similar displacement procedures are available for
tests of concrete masonry units under ASTM Method of Sampling and
Testing Concrete Masonry Units (C 140); however, ASTM Method C 140
452
specifies a draining period of I rain which is required to properly calculate

the effective specimen volume. The short period required for completing
the "saturated surface dry" and "immersed" weighings minimizes the
possibility of error in determined volume due to further reabsorption, unless the specimen is not immersed for the required 24-h period. On the
other hand, the 1-min drain time tends to increase the calculated volume
of a porous concrete, as in the case of a lightweight concrete masonry unit,
where increased drainage time of 5 to 10 min could have the effect of reducing the apparent volume of the specimen by 1 or 2 percent.
The same type displacement procedures may be used to determine volume
and unit weight of specimens of various sizes obtained from hardened concrete according to ASTM Method of Obtaining and Testing Drilled Cores
and Sawed Beams of Concrete (C 42). Having obtained physical data on a
given concrete sample, an analytical approach can be made according to
the method of absolute volumes by establishing the proportions and specific gravity of ingredients. In our experience, this principle has been used
to indicate the absence of entrained air. By measuring the volume and
weight of a suitable resaturated specimen, the concrete may be disassembled
after dehydration at 600~ in the manner described by Blackman [40],
and since the proportions of dry solids are determinable, unit air and water
contents may be calculated with reasonable accuracy. Axon [41] developed
and published a procedure for determination of air content and estimation
of original unit water content based on physical and microscopical methods.
Entrained Air and Related Properties

Though affected by coarse aggregate character, the absorption of concrete is primarily dependent on the quantity and quality of paste or mortar
component and, for given materials, absorption is a useful criterion of
quality. Hansen [42] published data showing the effects of overrodding on
63-mm (2i/2 in.) maximum crushed stone concrete with different sands
and cements, including air-entraining cement. Table V of Ref 2 included
interesting saturated surface dry unit weight and absorption data of resaturated air dry concrete, as abridged from Hansen's paper.
In discussing the reabsorption of water, Klieger's data [43], shown in
Table 8 representing a wider variety of proportions, are of interest because
the sand was reduced to compensate for the increased amounts of air. The
selected results show reduction of absorption in lean mixes where there was
a marked reduction of mixing water when air was entrained.
Following the line of paradoxes expressed in Gilkey's speech [44] which,
incidentally, traced the relationships of voids, density, and air content-it is appropriate to emphasize that air-entraining cement will increase consolidation of concrete block mixtures of dry tamp consistency. Swayze [45]
mentioned this in an article in Civil Engineering when "treated" cement
Gravel
Gravel
Stone
Stone
Type I
Type IA
Type I
Type 1A
6 (152)
6 (152)
3 (76)
3 (76)
4.4
3.6
3.4
3.1
90 in.
(19 mm)
max
2% in.
(63 ram)
max
2.7
2.5
...
...
1% in.
(38 ram)
max
2.7
2.5
3.3
2.9
2.6
2.2
1.6
1.7
2.2
2.0
1.5
1.4
1.5 in.
(38 ram)
max
...
...
1.6
1.4
2.5 in.
(63 ram)
max
51/2-Sack a Mixes
307 k g / m 3
90 in.
(19 mm)
max
a l sack/yd 3 : 55.77 k g / m 3. Values for kg cement/m 3 shown in parentheses.
Coarse
Aggregate
Cement
Slump,
in.
(mm)
4-Sack a Mixes
223 k g / m 3
1.7
1.7
1.2
1.2
90 in.
(19 ram)
max
1.6
1.6
1.1
1.1
1.5 in.
(38 ram)
max
7-Sack a Mixes
390 k g / m 3
Absorption, %, by 72-h Resaturation of Air-Dry Concrete
TABLE 8--Absorption of air-dry concrete (after Klieger [43]).
1.2
1.1
2.5 in.
(63 ram)
max
O'1
Ga
1"
>
z
o
z
-H
z
-H
O
O
O
z
"1"
m
v-
454
was being introduced to the block industry, and Helms and Bowman [46]
presented quantitative data. Results supporting this principle were given
in Table 6 of Ref 47 for laboratory specimens containing expanded slag
aggregate, including a double dosage of dry powdered air-entraining agent
which, in plastic concrete, could cause an inordinately high air content.
Whereas, in dry tamp concrete, increased compaction was experienced.
Varied Amounts of Voids and Special Purpose Concretes
Methods used in commercial practice for imparting dense structure to
concrete by manipulative means include: (1) production of hard floor finishes from dry-tamped concrete using power floats [48], (2) densification
of initially overwet concrete by vacuum techniques to achieve high early
strength, (3) making of concrete pipe by spinning and packer methods,
and (4) manufacture of lightweight masonry units [49].
Various reports have traced the development of the use of gas-forming
agents, such as aluminum powder and hydrogen peroxide, to make lightweight compositions. Aluminum powder has been used for many years in
other countries such as Sweden, and in Canada, to manufacture cellular
products, but these products have never proved feasible for production in
our country. Manufacturing problems and economic factors are reasons for
their not gaining acceptance here. Erratic behavior of aluminum powder
dependent on the sensitivity to chemical properties of cement, and difficulty
in distributing the powder to get uniform results, are among the factors to
be considered in controlling manufacture. Foamed concretes that use large
amounts of air-entraining agent as well as gas concretes are likely to show
high shrinkage, and stabilization by autoclaving is desirable. Lightweight
cellular concrete [50] is discussed more fully in the paper by Lewis 5 in this
publication.
In the range of lowest density concretes, usually made with lightweight
aggregates, an intended use for thermal insulation often makes strength a
secondary consideration. The criterion of strength/weight ratio may be
used, however, to rate the efficiency of compositions of similar unit weight.
The principle of "no fines" concrete [51] could even be modified to the
use of lightweight aggregate and foamed paste to produce a variety of densities down to the low limit. Low specific weight is dependent on the presence of voids, whether formed by air or water, as holes in honeycomb or
as pores in the aggregate. The insulating characteristics of low density
concrete may be correlated closely with the unit weight of the oven dry
specimens and the compressive strength with air dry density. Although
slight advantages can be claimed for certain aggregates in weaker, leaner
mixtures, the plotted thermal conductivity in Kluge's [52] Fig. 14 permits
5See p. 503.
455
citation of K values which could be assumed applicable to any oven-dry

lightweight concrete mix with low unit weight, as shown in 'Table 9.
TABLE 9--K values for oven-dry lightweight concrete.
Oven-Dry" Unit Weight of Concrete
lb/ft 3
20
42
55
70
(kg/m 3)
(320)
(673)
(881)
(1121)
Conservation Maximum
Thermal K Valueh"~
btu, in./h-ft 2- ~
0.7
1.5
2.0
2.5
W/m.K
0.10
0.22
0.29
0.36
a Slight advantages for certain combinations could be overshadowed by the effect of small
amounts of moisture retained by air dry specimens.
bLess conservativevalues may be found elsewhere [53].
c Results tabulated by Brewer [54] tend to confirm the above portrayal of the thermal conductivity of low density concrete.
The plotted data of expanded perlite concretes in Fig. 3 illustrate the
relations of voids to weight and total solid volume to compressive strength.
Results for 40 test combinations are shown which included three aggregate
samples, three volume proportions of aggregate, and usually either an airentraining agent or a fine plasticizing admixture. The curve of strength
versus solids has limitations, since cement content is reflected only as
weight, but it has unique applicability in the extreme range of high voids.
At the other extreme, the utilization of heavy aggregate because of its
weight is of direct engineering importance. Much ordinary concrete, used
in gravity dams, lock walls, counterweights, and some foundations, is
dependent on its relatively high mass for its usefulness.
Keyser [55] studied heavy aggregate concrete for making counterweights
as far back as 1932. Although not directed toward nuclear application, the
mixtures described contained iron ore and steel punchings which have
been used as aggregates for radiation shielding. In order to avoid segregation the quantity of mortar used was only slightly greater than that required
to fill the voids in the coarse aggregates. Revised compositional data selected
from the final mixtures, listed in Table 10, are of historical interest.
The special concretes used for radiation shielding require dense aggregate materials such as iron ores; for this purpose Callan [56] described
high unit weight concretes in the range of 3604 to 4406 k g / m 3 (225 to 275
lb/ft3). This topic is covered adequately in Polivka's paper 6 in this volume.
Concrete in the weight range above 4005 k g / m 3 (250 lb/ft 3) has been used
extensively for permanent ballast in ships [57].
In conclusion, it should be mentioned that unit weight tests are the time6See p. 420
456
65O
40
..4
(,,)
m
600
9.0
8,(
X%
7.(
"~"~
~ so~
o..
4.(
\oo.,.
600
40Q
_,
25
9 %,,.
9 .,,,<--
2.0
,Q_
200
-~.
400
\
\
9 ~
350
20
"70
"'
75
80
'""
Pet Ccnf Voids
FIG. 3--Relationship of unit weight and strength to percent voids, Type IIl cement insulating concrete made with expanded perlite aggregate.
honored basis for determination of yield,7 and sometimes it is expedient,

as for the data plotted in Fig. 3, to calculate this information from weight
and displacement or weight and dimensions of hardened specimens. The
determination of yield and interdependent unit contents of cement and
water is necessary in both laboratory and field for the proper control of
concrete mixtures.
7Measured by ASTM Test for Unit Weight, Yield, and Air Content (Gravimetric) of Concrete (C 138).
15.4
MPa
psi
2230
4378
4333
kg/m 3
lb/ft 3
273.3
270.5
334
188
227
520
1594
1515
4378
kg/m 3
563
317
383
876
2687
2553
7379
Ib/yd 3
11
2190
psi
273.0
272.1
15.1
MPa
4373
4359
kg/m 3
"46"8
1584
1821
4373
"789
2670
3069
7371
lb/ft 3
334
166
kg/m J
563
280
lb/yd 3
12
2750
psi
273.3
268.6
lb/ft 3
1"6"27
3204
1667
7378
563
317
lb/yd J
a 1 sack yd 3 = 55.77 k g / m 3. Value for kg c e m e n t / m 3 shown in parentheses.

bSized 1/2 to 1 in. (13 to 25 ram) in diameter and 1/4 to s/8 in. (6 to 16 m m ) thick; oil and rust free.
Compressive strength, 28 days
Design unit weight

Unit weight, 28 days age
Cement
Water
Sand
Fine iron ore
Coarse iron ore
Rivet punchings v
Total
Reference Mix No.
13
19.0
MPa
4378
4303
kg/m 3
"9"6"5
1901
989
4378
334
188
kg/m 3
kg/m 3
4260
psi
272.6
271.8
lb/ft 3
29.4
MPa
4367
4354
kg/m 3
563
334
224
133
744
441
. . . . . .
3487
2069
2342
1390
7360
4367
lb/yd 3
14
(38-mm )-max iron ore concretes, probably no-slump.
Quantities per Cubic Yard or per Cubic Metre
TABLE l O--A bridged average data from Keyser's Table 1 [55]. 6-sack a (335 kg/m 3) 1 89
t-
O'1
",4
4~
-1--I
c
z
),
z
u
m
z
o
0
z
:12
60
O
z
--ira
458
When air is incorporated for the purpose of reducing density, the size
of voids is of little consequence if they are uniformly distributed throughout
the mix. But the size of air voids entrained to improve durability is extremely important, and it has been realized for many years that if air is
efficiently dispersed into the finest attainable size a smaller amount will be
required for durability with subsequently less effect on strength.
REFERENCES
[1] Feret, R., "Sur la Compacite (Density) des Mortars Hydrauliques," Vol. 4, 1892; "Annales des Ponts et Chausses" and Bulletin de la Societe d'Encouragement pour l'Industrie Nationale, Vol. 2, 1897, p. 1604.
[2] Helms, S. B., "Air Content and Unit Weight," Significance of Test and Properties of
Concrete and Concrete Aggregates, ASTM STP 169, American Society for Testing and
Materials, 1955, p. 208.
[3] Woods, Hubert, discussion, "Water Cement Ratio Versus Strength etc," Journal, American Concrete Institute, Dec. 1961, Part 2; Proceedings, Vol. 58, p. 1871.
[4] Talbot, A. N. and Richart, F. E., "The Strength of Concrete--Its Relation to the Cement, Aggregates, and Water," Bulletin No. 137. Engineering Experiment Station,
University of Illinois, Urbana, I11., 1923.
15] Feret, R., discussion, "The Laws of Proportioning Concrete," Transactions, American
Society of Civil Engineers, Vol. 59, 1907, p. 154.
[6] Torrans, P. H. and Ivey, D. L., "The Void Spacing Indicator," TTI, Texas A & M University, Research Report No. 103-3, June 1969.
[7] Ivey, D. L. and Torrans, P. H., "Air Void Systems in Ready-Mixed Concrete," Research
ReportNo. 103-4F, June 1969; and Journal of Materials, Voh S, No. 2, June 1970, p. 492.
[8] Morris, M., "The Mortar Voids Method of Designing Concrete Mixtures," Journal,
American Concrete Institute, Sept. 1932; Proceedings, Vol. 29, p. 9.
[9] Thornburn, T. H., "The Design of Concrete Mixes Containing Entrained Air," Proceedings, American Society for Testing and Materials, Vol. 49, 1949, p. 921.
[10] Concrete Manual, U.S. Bureau of Reclamation, 7th ed., Table 4, p. 34.
[11] ACI Committee 309, "Recommended Practice for Consolidation of Concrete," Journal,
American Concrete Institute, Vol. 68, No. 12, Dec. 1971, p. 893.
[12] Grieb, W. E., Werner, G., and Woolf, D. O., "Water-Reducing Retarders for Concrete--Physical Tests," Public Roads, Feb. 1961, Vol. 31, No. 6, p. 136.
[13] Paul, Ira, "Chloride Salts Resistant Concrete in Pavements," Proceedings, American
Association of Highway Officials of the North Atlantic States, Feb. 1938, p. 144.
[14] Avery, W. M., "Portland Cement Association, 45 Years of Outstanding Service to Industry," Pit and Quarry, July 1947, p. 87.
[15] Moore, O. L., "Pavement Scaling Successfully Checked," Engineering News-Record,
10 Oct. 1940, p. 61.
[16] Jackson, F. H., "Concretes Containing Air-Entraining Agents," Journal, American Concrete Institute, June 1944; Proceedings, Vol. 40, p. 509. (An account of Kansas Highway
Department experience, 1934).
[17l Vollmer, H. C., "Effect of CaC12 on Water Requirement, Specific Weight and Compressive Strength of Concrete Made with Plain and Treated Cement," Proceedings,
[18] Pigman, G. L., "Vacuum Method of Measuring Air Content of Fresh Concrete," Journal,
American Concrete Institute, Nov. 1941; Proceedings, Vol. 38, p. 121.
[19] Gonnerman, H. F., Research Report, Portland Cement Association, 1 Feb. to 31 July
1943.
[20] McCall, J. T. and Claus, R. J., "Effect of Pellet and Flake Forms of Calcium Chloride
in Concrete," Bulletin 75, Highway Research Board, 1953.
[21] Bugg, S. L., "Effect of Air Entrainment on the Durability Characteristics of Concrete
Aggregates," Proceedings, Highway Research Board, Vol. 27, 1947, p. 156.
459
[22] Axon, E. O., Willis, T. F., and Reagel, F. V., "Effect of Air-Entrapping Portland Ce-
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
ment on the Resistance to Freezing and Thawing of Concrete Containing Inferior Coarse
Aggregate," Proceedings, American Society for Testing and Materials, Vol. 43, 1943,
p. 981.
Swanberg, J. H., Thomas, T. W., Klein, Alexander, Pirtz, David, Schweizer, C. B.,
Cordon, W. A., Brewer, H. W., and Pearson, J. C., Symposium on Measurement of
Entrained Air in Concrete, Proceedings, American Society for Testing and Materials,
Vol. 47, 1947, pp. 871,884, 894, 918.
Gregg, L. E., "Experiments with Air Entrained in Cement Concrete," Bulletin No. 5,
Engineering Experiment Station, University of Kentucky, Lexington, Ky., 1947.
Lindsay, J. D., "Illinois Develops High Pressure Air Meter for Determining the Air
Content of Hardened Concrete," Proceedings, Highway Research Board, Vol. 35, 1956,
p. 424, and discussion by T. B. Kennedy, p. 434; including: T. B. Kennedy, Concrete,
Vol. 60, No. 12, Dec. 1952, p. 18.
McCoy, E. E., Jr., "High Pressure Test for Determining Air Content of Hardened Concrete," WES Miscellaneous Paper No. 6-286, U.S. Army Corps of Engineers, Waterways Experiment Station, Oct. 1958.
Erlin, B., "Air Content of Concrete by a High-Pressure Method," Bulletin 149, Research
Department, Portland Cement Association, (Journal, Sept. 1962).
Dykins, F. A. and Blandin, F. H., "Field and Laboratory Air-Content Studies of SaltDamaged Concrete Structures," Research on Aggregate, Cement, Concrete, and Epoxy
Bonding, Highway Research Board, Highway Research Record 62, Publication 1246,
Dec. 1964.
Sweet, H. S., discussion, "Air-Entrained Concrete (Durability)," Proceedings, Highway
Sweet, H. S., "Research on Concrete Durability as Affected by Coarse Aggregates,"
Bloem, D. L. and Walker, S., "Studies of Concrete Containing Entrained Air," (Series
66) Journal, American Concrete Institute, June, 1946; Proceedings, Vol. 42, p. 629.
"Loss of Entrained Air in Freshly Mixed Concrete and Effect on Properties of Hardened
Concrete," Report No. C-431, Materials Laboratories, U.S. Bureau of Reclamation,
May 1949.
Mielenz, R. C., Wolkodoff, V. E., Backstrom, J. E., and Flack, H. L., "Origin, Evolution, and Effects of the Air Void System in Concrete, Part I, Entrained Air in Unhardened Concrete," Journal, American Concrete Institute, July 1958; Proceedings, Vol. 55,
p. 95.
Bruere, G. M., "Rearrangement of Bubble Sizes in Air-Entrained Cement Pastes During
Setting," Australian Journal of Applied Science, Vol. 13, No. 3, Sept. 1962, p. 222.
Powers, T. C., "Topics in Concrete Technology 4. Characteristics of Air Void Systems,"
Bulletin 174, Research Department, Portland Cement Association, Journal, Jan. 1965.
Verbeck, G. J., "The Camera-Lucida Method for Measuring Air Voids in Hardened
Concrete," Journal, American Concrete Institute, May 1947; Proceedings, Vol. 43, p.
1025, also discussion.
Klieger, P., "Effect of Entrained Air on Concrete with 'Sand-Gravel' Aggregates,"
Journal, American Concrete Institute, Oct. 1948; Proceedings, Vol. 45, p. 149.
Brown, L. S. and Pierson, C. U. "Linear Traverse Technique for Measurement of Air
in Hardened Concrete," Journal, American Concrete Institute, Oct. 1950; Proceedings,
Vol. 47, p. 117, also discussion.
Warren, C., "Determination of the Properties of Air Voids in Concrete." Bulletin No.
70, Highway Research Board, 1953.
Blackman, J. S., "Method of Estimating Water Content of Concrete at the Time of
Hardening," Journal, American Concrete Institute, March 1954; Proceedings, Vol. 50,
p. 533.
Axon, E. O., "A Method of Estimating the Original Mix Composition of Hardened Concrete Using Physical Tests," Proceedings, American Society for Testing and Materials,
Vol. 62, 1962, p. 1068.
Hansen, W. C., "Influence of Sands, Cements, and Manipulation Upon the Resistance
460
of Concrete to Freezing and Thawing," Journal, American Concrete Institute, Nov.

1942; Proceedings, Vol. 39, p. 105, also discussion.
[43] Klieger, P., "Effect of Entrained Air on Strength and Durability of Concrete Made with
Various Maximum Sizes of Aggregates," Proceedings, Highway Research Board, Vol.
31, 1952, p. 177; further studies, Vol. 34, 1955.
[44] Gilkey, H. J., "The Zig-Zag Course of Concrete Progress," Journal, American Concrete
Institute, April 1950; Proceedings, Vol. 46, p. 573.
[45] Swayze, M. A., "Less Sand and Water in Air-Entraining Batches Improves Concrete,"
CivilEngineering, Vol. 16, No. 7, July 1946, p. 302.
[46] Helms, S. B. and Bowman, A. L., "Method of Making Vibrated Dry-Tamp Concrete
Cylinders Applied to Tests of Lightweight Aggregate and Block Mixtures," Proceedings,
[47] Helms, S. B., "Air Content and Unit Weight," Significance of Test and Properties of
Concrete and Concrete Aggregates, ASTM STP 169A, 1966, p. 309.
[48] Klock, M. B., "Monolithic and Bonded Floor Finishes," Journal, American Concrete
Institute, June 1949; Proceedings, Vol. 45, p. 725.
[49] Wendt, K. F. and Woodworth, P. M., "Tests on Concrete Masonry Units Using Tamping and Vibration Molding Methods," Journal, American Concrete Institute, Nov. 1939;
Proceedings, Vol. 36, p. 121, also discussion.
[50] Valore, R. C., Jr., "Cellular Concretes," Journal, American Concrete Institute, May,
June 1954; Proceedings, Vol. S0, pp. 773,817.
[51] Petersen, P. H., Building Materials and Structures Report BMS 96, National Bureau
Standards, 1943. Valore, R. C. and Green, W. C., "Air Replaces Sand in 'No-Fines'
Concrete," Journal, American Concrete Institute, June 1951; Proceedings, Vol. 47,
p. 833.
[52] Kluge, R. W., Sparks, M. M., and Tuma, E. C., "Lightweight-Aggregate Concrete,"
Journal, American Concrete Institute, May 1949; Proceedings, Vol. 45, p. 625, also discussion.
[53] Neher, Journal, American Concrete Institute, Sept. 1951, Fig. 3, p. 68.
[54] Brewer, "Relation of Heat Flow Factors to the Unit Weight of Concrete," Journal, Portland Cement Association Research and Development Laboratories, Jan. 1967, p. 51.
[55] Keyser, C. C., "Designing Concrete for Weight of 271 Pounds Per Cubic Foot," Journal,
American Concrete Institute, April 1932; Proceedings, Vol. 28, p. 525.
[56] Callan, E. J., "Concrete for Radiation Shielding," Journal, American Concrete Institute,
Sept. 1953; Proceedings, Vol. 50, p. 17.
[57] Journal of Commerce, 19 Dec. 1942.
Bibliography
Beaudoin, J. J. and Mclnnis, C., "Mechanism of Frost Damage in Hardened Cement Paste,"
Cement and Concrete Research, Vol. 4, No. 2, March 1974, p. 139.
Forder, Colin, "Lightweight Concrete's Place in the Insulation Spectrum," Concrete (London),
Jan. 1975, p. 28.
Gaynor, R. D., "High Strength Air-Entrained Concrete," Joint Research Lab Publication No.
17, National Sand and Gravel Association and National Ready Mixed Concrete Association,
March 1968.
Gilkey, H. J., "Water Cement Ratio Versus Strength--Another Look," Journal, American
Concrete Institute, April 1961; Proceedings, Vol. 57, p. 1287.
Harland, D. G., "A Radio-active Method for Measuring Variation in Density in Concrete
Cores, Cubes and Beams," Magazine of Concrete Research, MCORA, Vol. L*, No. 55,
June 1966, p. 95.
Hoff, G. C., "Porosity-Strength Considerations for Cellular Concrete," Cement and Concrete
Research, Vol. 2, No. 1, Jan. 1972, p. 91.
Hogberg, E., "Effect of Vibration on Air Content of Cement Mortar and Concrete," (in
Swedish) Nordisk Betong, Vol. 13, No. 1, 1969, p. 27.
Kraft, L. M., "Compaction of Concrete Slabs by Vibration," Journal, American Concrete
Institute, June 1971, Proceedings, Voi. 68, No. 6, p. 462.
461
Larson, T. D. and Cady, P. D., "Systems Approach to Analysis of Hardened Concrete,"

Materials Research and Standards, MTRSA, Vol. 8, No. 10, Oct. 19.68, p. 8.
Larson, T. D., Cady, P. D., and Malloy, J. J., "Protected Paste Volume Concept Using New
Air-Void Measurement and Distribution Techniques," Journal of Materials, JMLSA, Vol.
2, No. 1, March 1967, p. 202.
Lewis, R. K., "Note on Density of Concrete as Affected by Variations in Properties and Proportions of Ingredients," Construction Review (Australia) Vol. 38, No. 9, Sept. 1965, p. 21.
Preiss, K., "Measuring Concrete Density by Gamma Ray Transmission," Materials Research
and Standards, MTRSA, Vot. 5, No. 6, June 1965, p. 285.
Riley, O. and Cooper, S. B., "Concrete Control on a Major Project," Journal, American Concrete Institute, Feb. 1971; Proceedings, Vol. 68, No. 2, p. 107.
Torrans, P. H. and Ivey, D. L., "Air Void Systems Affected by Chemical Admixtures and
Mixing Methods," Highway Research Record, Highway Research Board, National Research
Council, No. 226, 1968, p. 1.
U.S. Army Corps of Engineers, CRD-C83-58, "Method of Test for Air Content of Hardened
Concrete," 1958.
Vellines, R. P. and Ason, T., "A Method for Determining the Air Content of Fresh and
Hardened Concrete," Journal, American Concrete Institute, May 1949; Proceedings, Vol.
45, p. 665.
W. G. H i m e I
Chapter 28
Cement Content
Introduction
Analyses for cement content are requested frequently by those who have
encountered concrete that has shown low strength or other improper performance. Although such requests may often be poor substitutes for the
far better "Determine why this concrete performed badly," there are many
occasions when a knowledge of the cement factor is desirable.
The state of the art and science of the analysis of concrete for its components was well detailed by Minnick [i]2 in the previous edition of this
book (ASTM STP 169.4). T h a t review concluded that " . . . the chemical
procedure has not yielded a high degree of accuracy." Unfortunately,
except for one recently developed technique that at least requires only a
minimum of effort (see Maleic Acid Method section), the situation has
scarcely improved. However, the analyst who can report not only the data
but also estimates of the error and its direction is of inestimable value. To
such an end this chapter is written.
Determination of Cement Content
Sample Selection
A sampling program should be carefully designed to provide the desired
information. Concrete specimens should be secured to represent each area
in question. Where possible, specimens of acceptable concrete should also
be obtained so that comparisons of determined cement contents can be
made, hopefully obviating aggregate interference.
Although several 4-kg (10-1b) specimens are required if strict adherence
is made to ASTM Test for Cement Content of Hardened Portland Cement
Concrete (C 85), a smaller sample is often all that is available. It may
prove quite satisfactory if it is cut from a concrete core or cylinder. Conl Vice president, Erlin, Hime Associates, Northbrook, 111.60062.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
HIME ON CEMENT CONTENT
463
crete portions obtained by hammering or by selection from rubble are

undesirable as they are likely to reflect a loss of the mortar fraction.
If several 4-kg (10-1b) specimens are not available, the size of the aggregate should dictate the sample weight. Assuming a 4-kg (10-1b) minimum
sample requirement for a 25-mm (1-in.) top-sized aggregate, the sample
weight for other sizes of aggregate might, if necessary, be chosen in direct
proportion to the aggregate size.
Sample Preparation
The pulverization operation should not lead to significant loss of fine

material, nor gain from the grinding device itself. A magnet may be used
to remove any iron contaminant introduced by steel disks, crushers, or
disk pulverizers.
If soluble silica or lime is to be determined, fine grinding should be
avoided if aggregate correction procedures are not to be used. An analytical sample that just passes a 30-mesh screen will usually permit complete cement solubility while minimizing aggregate interference.
Cement Content Analysis Procedures

Methods Based on Soluble Silica or Lime
A S T M Test C 8 5 - 5 4 - - T h e original 1954 version Of ASTM Test C 85 was
based on determinations of soluble silica and calcium in the concrete, as
suggested as early as 1924 by Kriege [2]. That method, or a modification
of it, has been cited in specifications of many countries and has been
evaluated by dozens of investigators [3-6].
The method determines acid-soluble silica and calcium (generally soluble calcium is equivalent to total calcium) by straightforward, but quite
lengthy, "wet chemical" techniques of analysis, and assumes them to be
entirely due to the portland cement component. Unfortunately an error,
often major, occurs if the aggregate fraction contains calcareous and
siliceous components that are soluble under the conditions of the test. All
calcareous aggregates, such as limestone or dolomite, interfere with the
calcium procedure. Many siliceous aggregates, except quartz, are soluble
or partly soluble in acid, and interfere with the silica procedure. Excellent
tables detailing the acid solubility of aggregates were given in Minnick's [1]
review.
One phenomenon has been overlooked by other investigators: siliceous
aggregates that were originally insoluble in acid may have reacted, at least
in their surface regions, with alkalies from the portland cement fraction,
forming acid-soluble silicates. An unanticipated positive error may thus occur.
464
Corrections for aggregate interference in the ASTM Test C 85-54 procedure can be made if the original aggregate is available for analysis, a
generally infrequent situation. Such corrections may themselves be in
error due to the fact that their magnitude depends upon an assumption of
aggregate proportions. Additionally, the solubility of siliceous aggregates
may vary greatly with the time of treatment with the acid, the temperature
of the acid, and the particle size of the aggregate. It is nearly impossible
to duplicate these parameters, especially the latter. The hardness and
initial size of the concrete fragments being analyzed will not be the same as
those of the virgin aggregate particles, and thus identical grinding, or even
grinding to pass a given sieve, may produce quite unidentical particle sizes,
and thus unidentical solubilities.
Numerous investigations of the errors due to aggregate solubility have
been made. Ford [4] and many others have found that a cement content
determination will generally provide a high result without aggregate corrections, and a low result when corrections are applied.
Under the most favorable conditions, a 10 percent error should be assumed.
A S T M Test C 85-66--The current version of ASTM Test C 85 differs
from the original one (1954) in that it provides correction procedures for
aggregate interference, using the aggregate in the sample itself. This procedure, in the hands of an expert analyst, can provide excellent results for
most samples and is the method of choice for research purposes.
However, the correction procedure can lead to overcorrection for aggregate solubility, and consequently a low cement content. This is especially
true when the analysis is performed by an inexperienced analyst. Thus,
it is possible that doubt may be cast on good concrete, and guilt on an
innocent concrete producer.
Correction errors may arise for any of a number of reasons.
1. The coarse fraction of the concrete is treated with acid to dissolve
the cement adhering to the aggregate. It is difficult to stop this treatment
at the precise moment when all the cement, but none of the aggregate,
has dissolved. An exact timing is an impossibility with calcareous aggregates.
2. The fine fraction is divided into two parts to determine soluble silica
and lime in the cement-sand combination, and in the sand fraction. For
the cement-sand, material passing the No. 6 sieve is used; for the sand
fraction, only that portion retained on the No. 50 sieve is used. Composition may thus vary. Additionally, although both parts are further ground
to pass a No. 200 sieve, their final particle sizes, and thus their acidsolubilities, may differ significantly.
3. Both the coarse and fine aggregate may have reacted with the cement
alkalies to produce a layer now soluble in acid. This layer is removed by
465
the process that produces the aggregate fraction used for the correction
analysis.
Author's Procedure--"Negative" errors (underestimation of true cement
content value) can cause undue hardship to a concrete producer and may
lead to a false interpretation of the cause of improper concrete performance. Thus, a method that minimizes such errors has frequent application.
In the author's laboratory, a method has been developed to provide
determined cement contents at or above the actual level. Additionally, an
estimate of the magnitude of (positive) error is possible. The method is
essentially as follows.
1. The concrete specimen is ground to just pass a 30-mesh screen (by
limiting grinding time and size of charge).
2. The soluble silica procedure of ASTM Test C 85-66 is followed, except that the three portions are not each digested for 15 min, but for 10,
15, and 20 min, respectively. Siliceous aggregates that are partially acidsoluble are revealed by this treatment. Tables 1 and 2 present the data for
TABLE 1--Silica data f o r concrete containing partially soluble siliceous aggregate.
Determined
Acid Digestion Period
SiO2, g
SiO2, %
Portland Cement, %
10min
15min
20min
0.0689
0.0786
0.0825
3.44
3.93
4.12
16.38
18.71
19.62
TABLE 2--Silica data f o r concrete containing no soluble siliceous aggregate.

Determined
Acid Digestion Period
SiO2, g
SiO2, %
Portland Cement, %
10 min
IS rain
20 min
0.0764
0.0771
0.0767
3.82
3.85
3.38
18.19
18.33
18.24
such an aggregate, and for an aggregate containing no soluble siliceous

material (for example, calcareous aggregate and most quartz aggregates).
If the silica contents increase with time, the 10-min data usually provide
the most accurate results. Indeed, European practice is to employ a short
digestion procedure with ice-cold 80-percent hydrochloric acid ("Florentin"
acid) to minimize aggregate solubility.
466
3. The soluble calcium oxide procedure of ASTM Test C 85-66 is followed, or, alternatively, any rapid, accurate method for calcium may be
employed. An atomic absorption procedure employing a nitrous oxide
flame and cement standards has proven excellent for this determination,
providing results within 1 h.
4. The fraction of cement by weight is multiplied by the determined or
estimated unit weight to provide the cement factor.
Maleic Acid Method--A method based on the solubility of hydrated
cement in a methanolic maleic acid solution was developed by Tabikh et
al [7]. This method provides results within 24 h, and requires only about
2 man-hours.
Although the authors found that most aggregates were insoluble in
maleic acid solution, and thus did not interfere, Clemena [8], in an extensive test of the method, found that most calcareous and some siliceous
aggregates were partially soluble, and could cause very significant errors.
Unpublished work by Pistilli [9] at the Portland Cement Association found
that this interference could be minimized by reducing the maleic acid
concentration from 20 to 14.5 percent.
More extensive investigations have shown that fly ash, slag, and some
other aggregates can lead to error. Negative errors can arise if the aggregate (for example, a clay) contains water or other substances volatile at
600~ (for example, organics), or if the cement paste had carbonated
(masonry mortars are often extensively carbonated).
Data from the method must thus be used with care. However, its speed
and simplicity recommend it highly. The use of the 14.5 percent maleic
acid concentration, and a water-correction ignition temperature of only
520~ are suggested.
Sulfate Method--As presented by Kossivas [10] and several others, cement content can be determined through analyses for sulfate. Since few
aggregates contain sulfate, the method is remarkably free of interference,
a unique situation. Unfortunately, the sulfate content of the cement must
be known or estimated--an infrequent event in the former case, and
usually an unacceptable burden in the latter. Because sulfate contents of
cement, as SO3, vary from less than 2 percent to almost 4 percent, a 30
percent error can result from this factor alone.
The method can prove a helpful "backup" to confirm other data, or
suggest the need for additional work.
Microscopic Methods--Polivka et al [11] and Axon [12] have presented
methods for the determination of cement, air, aggregate, and water contents of hardened concrete. The methods have not found wide usage since
they require special equipment and expertise, and since they do not provide a direct determination of cement content, but rather of cement paste.
467
The proportion of cement in the paste fraction must be estimated, and the
accuracy of such estimates is open to question.
Since the methods do provide information on all the major concrete
components, they can have important value in concrete investigations.
Methods of detailed petrographic techniques for the analyses of concrete
and the wealth of information that can be obtained are aptly given by
Mather [13], Mielenz [14,15], and Erlin [16,17]. In the hands of petrographers well versed in concrete and concrete-making materials, the petrographic techniques and information thereby obtained are extremely useful
to the chemical analyst in his understanding of the potential error to which
his analytical results are susceptible.
Miscellaneous Methods--Many other methods for cement content have
been presented, but none has attained popularity nor shown promise except in limited applications.
For large-scale projects, nuclear devices that determine calcium or silicon automatically have been proposed by Covaut and Poovey [18], Iddings
et al [19], and others. These devices are extremely costly, and are not
only subject to the same interferences as the soluble silica and calcium
methods, but also to all forms of siliceous and calcareous aggregates.
Determination of Other Concrete Components

Although an extended discussion of methods for other components of
concrete is beyond the scope of this chapter, the analyst charged with a
determination of cement content should be aware that analyses can also be
made for additives, admixtures, air voids, aggregate, and water. Methods
that have been found to be especially applicable are as follows.
Additives and Admixtures

The chloride content of concrete can be determined by acid extraction
and potentiometric titration. A rugged, relatively inexpensive method
utilizing a silver electrode has been described by Hime [20]. A chloride
electrode method developed by Berman [21] is the basis of most governmental analyses in connection with bridge deck corrosion.
Methods for several organic additives and admixtures have. been developed. A general discussion was given by Hime et al [22]; procedures for
lignosulfonates were developed by Hime et al [22] and Kroome [23]; for
sugars by Shima and Nishi [24]; for polyhydroxycarboxylic acids by Frederick and Ellis [25]; for triethanolamine by Connolly and Hime [26]; and
for vinsol resin by Hime et al [22] and Kroome [23]. Methods for several
types of retarders were also given by Halstead and Chaiken [27].
468
Air Voids
The air void content of concrete is determined by point count or linear
traverse methods in accordance with ASTM Recommended Practice for
Microscopical Determination of Air-Void Content and Parameters of the
Air-Void System in Hardened Concrete (C 457).
Aggregate
The aggregate content of concrete can be determined by difference (100
percent less cement and water), as acid-insoluble residue for certain siliceous aggregates, or by microscopic methods such as those of Polivka et al
[11] and Axon [12].
Water
Generally, the determination of the water content of hardened concrete
is of academic interest only. In favorable cases, free water can be determined as loss at 110~ and combined water as loss from 110 to 1000~
correcting for carbon dioxide as determined by any usual technique. Differential thermal analysis and thermogravimetric analysis techniques
generally allow assignment of water to specific substances (for example,
calcium hydroxide).
The determination of original mix water is of much greater importance
but, unfortunately, often an approximation at best. For non-air-entrained
concrete, Blackman [28] and Brown [29] have presented an applicable
method determining original water as the sum of combined water and
water calculated as having led to development of voids. Water content also
has been estimated by the microscopic techniques of Axon [12].
General Reference Works

Methods of analysis of cements and concrete were presented by Hime
in a chapter of Encyclopedia of Industrial Chemical Analysis [30].
Methods of analyses of concrete and many other cement products have
been presented in books by the Society of Chemical Industry [31] and Figg
and Bowden [32]. The latter contains a bibliography of 193 references.
References
[1] Minnick, L. J., "Cement Content," Concrete and Concrete Making Materials, A S T M
STP i69A, American Society for Testing and Materials, 1966, pp. 326-339.
[21 Kriege, H. F., "Determining the Cement Content of Concrete," Engineering NewsRecord, Vol. 92, No. 21, 1924, p. 892.
[3] "Cement Content Determination and Reactive Aggregate Cooperative Tests," Proceedings, American Society for Testing and Materials, Vol. 52, 1952, pp. 282-285.
469
[4] Ford, C. L., "Study of Methods for the Determination of the Portland Cement Content
of Hardened Concrete," Bulletin, American Society for Testing ~ind Materials, April
1952, pp. 47-52.
[5] Kriege, H. F., "Cement Content of Hardened Concrete," Report on Significance of
Tests of Concrete and Concrete Aggregates, A S T M STP 169, American Society for Testing and Materials, 1955, p. 221.
[6] Forss, Bengt, "Analysis of Hardened Concrete," Nordish Betong, Vol. 14, No. 1, 1970,
pp. 21-26.
[7] Tabikh, A. A., Balchunas, M. L, and Schaefer, D. M. "A Method for the Determination of Cement Content in Concrete," Highway Research Record 370, Highway Research Board, Washington, D.C., 1971, pp. 1-7.
[8] Clemena, G. G., "Determination of the Cement Content of Hardened Concrete by
Selective Solution," Virginia Highway Research Council, 1 Oct. 1972. Available NTIS
PB-213-855.
[9] Pistilli, M. F., "Cement Content of Hardened Concrete," presented at Cement Chemist's
Seminar, Portland Cement Association, Skokie, Ill., 1976.
[10] Kossivas, K. G., "Cement Content Determination in Hardened Concrete," presented
at Rock Products Chemical Research Seminar, Chicago, W. R. Grace and Company,
1969.
[11] Polivka, M., Kelly, J. W., and Best, C. H., "A Physical Method for Determining the
Composition of Hardened Concrete," Papers on Cement and Concrete, A S T M STP 205,
American Society for Testing and Materials, 1956, pp. 135-152.
[12] Axon, E. O., "A Method of Estimating the Original Mix Composition of Hardened
Concrete Using Physical Tests," Proceedings, American Society for Testing and Materials, Vol. 62, 1962, pp. 1068-1080.
[13] Mather, Katharine, "Petrographic Examination (Hardened Concrete)," Significance
of Tests and Properties of Concrete and Concrete-Making Materials, A S T M STP 169A,
[14] Mielenz, R. C., "Petrography Applied to Portland Cement," Reviews in Engineering
Geology, Geological Society of America, Vol. 1, 1962, pp. 1-38.
[15] Mielenz, R. C., "Diagnosing Concrete Failures," Lecture No. 2, Stanton Walker Lecture Series on the Material Sciences, University of Maryland, 18 Nov. 1964.
[16] Erlin, Bernard, "Methods Used in Petrographic Studies of Concrete," Analytical Techniques for Hydraulic Cement and Concrete, A S T M STP 395, American Society for
Testing and Materials, 1966, pp. 3-17; also Research Department Bulletin 193, Portland Cement Association.
[17] Erlin, Bernard, "Analytical Techniques," Observations of the Performance of Concrete
in Service," Special Report 106, Highway Research Board, Washington, D.C., 1970,
pp. 29-37.
[18] Covault, D. O. and Poovey, C. E., "Use of Neutron Activation to Determine Cement
Content of Portland Cement Concrete," Publication 1022, Highway Research Board,
1962, pp. 1-29.
[19] Iddings, F. A., Arman, Ara, Perez, A. W., II, Kiessel, D. W., and Woods, W.,
"Nuclear Techniques for Cement Determination," HRB Bulletin 268, Highway Research
Board, 1969, pp. 118-130.
[20] Hime, W. G., "Portland Cement," in Standard Methods of Chemical Analysis, F. J.
Welcher, Ed., Vol. 3A, Van Nostrand Company, Inc., 1966, Chapter 40, pp. 962-974.
[21] Berman, H. A., "Determination of Chloride in Hardened Portland Cement Paste, Mortar and Concrete," Journal of Materials, Vol. 7, No. 3, September, 1972, pp. 330-335.
[22] Hime, W. G., Mivelaz, W. F., and Connolly, J. D., "Analytical Techniques for Hydraulic Cement and Concrete," A S T M STP 395, American Society for Testing and Materials, 1966, pp. 18-29.
[23] Kroome, B., "A Method of Detecting and Determining Lignin Compounds in Mortars and Concrete," Magazine of Concrete Research, Vol. 23, No. 76-76, June-Sept.,
1971, pp. 132-134.
[24] Shima, Ikuo and Nishi, Teiji, "Determination of Saccharose in Hydrated Cement,"
Semento Gijutsu Nempo, Vol. 17, 1963, pp. 106-109 (Chemical Abstract, Vol. 61,
1964, 5344b).
470
[25] Frederick, W. L. and Ellis, J, T., "Determination of a Polyhydroxy Carboxylic Acid

Retarder in Hardened Concrete," Materials Reseaeh and Standards, Vol. 8, March
1968, pp. 14-18.
[26] Connolly, J. D. and Hime, W. G., "Analysis of Concrete for Triethanolamine," Cement
and Concrete Research, Vol. 6, 1976, pp. 741-746.
[27] Halstead, W. J. and Chaiken, B., "Water-Reducing Retarders for Concrete--Chemical
and Spectral Analyses," Public Roads, Vol. 31, No. 6, 1961, pp. 126-135.
[28] Blackman, J. S., "Method for Estimating Water Content of Concrete at the Time of
Hardening," Proceedings, American Concrete Institute, Vol. 50, 1954, pp. 533-541.
[29] Brown, A. W., "A Tentative Method for the Determination of the Original Water/
Cement Ratio of Hardened Concrete," Journal of Applied Chemistry, Vol. 7, Oct. 1957,
pp. 565-572.
[30] Hime, W. G., "Cements, Mortars and Concrete," Encyclopedia of Industrial Chemical
Analysis, Vol. 9, Wiley, New York, 1970, pp. 94-155.
[31] "Analysis of Calcareous Materials," SCI Monograph No. 18, Society of Chemical Industry, 1964.
[32] Figg, J. W. and Bowden, S. R., The Analysis of Concretes, Her Majesty's Stationery
Office, 1971.
R. D. G a y n o r 1
Chapter 29--Ready-Mixed Concrete
Introduction
Ready-mixed concrete is concrete manufactured for delivery to a purchaser in a plastic and unhardened state [1,2]. 2 The American Society for
Testing and Materials (ASTM) has maintained ASTM Specification for
Ready Mixed Concrete (C 94) since 1933 and ASTM Specification for Concrete Made by Volumetric Batching and Continuous Mixing (C 685) since
1971. This paper will review requirements of these two specifications and
their relationship to other more general concrete specifications and codes.
This discussion will be supplemented by background information on the
industry, operating procedures, and special problems and circumstances.
The references are intended to identify additional information in particular
subject areas.
The discussion represents procedures and conditions in the United States
and Canada. The interested reader should recognize that within the past 10
or 15 years there has been a tremendous increase in the use of ready-mixed
concrete in the industrial nations in the rest of the world. Foreign literature
is an increasingly important source of information on ready-mixed concrete.
History of the Industry

The first concrete mixed off-site and delivered to the job may have been
furnished in about 1913 but, ready-mixed concrete was not a recognizable
industry until the late 1920s when the first revolving drum agitators and
mixers were developed [3]. In 1933 when the first ASTM specification appeared, the industry used only about 5 percent of the portland cement produced. That percentage doubled and tripled during World War II. By 1950
the industry used a third of the cement produced in the United States and
today it uses about two-thirds of that produced. The peak production oc1Vice president of engineering and research, Giant Portland and Masonry Cement Co.,
Harleyville, S.C. 29448.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
lntcrnational
www.astm.org
472
curred in 1973 when a total of 174 106 m 3 (228 106 yd3) were produced.
In the recent recession the volume dropped to 131.5 106 m 3 (172 X l0 s
yd3) in 1975. By the end of 1977 the volume should be about 15 percent
above 1975 levels but still substantially under the 1973 level.
The industry started as small individual family-owned businesses. It has
retained much of that character. A 1971 Bureau of Census survey of employment revealed that 40 percent of the companies had less than 7 employees,
75 percent less than 18 and 94 percent of the companies less than 50
employees. Industry surveys [4,5] have consistently shown that the 10 percent of largest companies produce well over half of the total volume of readymixed concrete.
Although it is difficult to identify all of the small companies it appears that
there are about 5000 companies in the industry and perhaps 8 to 10 thousand
plants. An estimate of total employment in the industry in 1975 was 80 000
[6]. Few companies operate in more than one or two geographical areas.
As yet there are no significant national chains. These facts and the U.S.
antitrust laws and regulations produce a diffusion of economic power in the
ready-mixed concrete industry that is unlike that in most other industrialized
nations.
Specifications
ASTM Specifications C 94 and C 685 are specifications for the material
freshly mixed concrete. They address the separate and joint responsibilities
of the various parties in a typical job--the owner, specifying agency, contractor, and ready-mixed concrete producer. Included are the basic elements
of both prescription and performance specifications. Appropriate test methods are defined. Included is the basis or unit of purchase--the cubic yard-and the appropriate test method.
Since strength test results are weighted heavily in the acceptance of the
material and since strength is not usually determined until the concrete has
been in the structure for 28 days, it is considered prudent to regulate further
the procedures used to proportion, batch, mix, and deliver the concrete.
Under certain prescriptive alternatives where strength is not specified,
procedural and compositional requirements are necessary. The requirements
govern measuring of materials, the batch plant, and the mixing and delivery
operation.
These two ASTM specifications are meant to be incorporated by reference
in the general job specifications. The responsibility for acceptable performance of concrete in place is divided inevitably between the concrete producer
who must deliver acceptable material, the contractor who must place it, and
the specifier who must structure appropriate specifications and plans.
American Concrete Institute (ACI) Specifications for Structural Concrete
for Buildings (ACI 301) is intended to cover all of the usual requirements for
cast-in-place concrete for buildings and should be used essentially by referCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:31:15 EDT 2014
GAYNOR ON READY-MIXED CONCRETE
473
ence. The batching, mixing, and delivery of ready-mixed concrete is by reference to ASTM Specification C 94. A section of ASTM Specification C 94
states that when requirements of the job specifications conflict with requirements of ASTM Specification C 94 the job specifications govern. Therefore,
the sections of ASTM Specification C 94 dealing with materials, proportioning, testing, and evaluation are not used in AC1301.
In preparing both ASTM Specifications C 94 and C 685 the ASTM Committee has produced a compromise of requirements somewhere between the
needs of the small project where no job specifications are prepared and for
large projects where specifiers provide detailed specifications requiring
highly specialized procedures tailored to specific job conditions.
The ACI Building Code Requirements for Reinforced Concrete (ACI 318)
is another example of the use of ASTM Specification C 94 by reference. In
this instance the Building Code is designed for inclusion in local or state
building codes, generally for the protection of the public interest. The requirements for mix design approval, testing, and evaluation of both ACI 301
and ACI 318 differ in important respects from those in the ASTM specifications for ready-mixed concrete.
An additional document that falls in the general classification of a specification is the National Ready Mixed Concrete Association (NRMCA) Check
List for Certification of Ready Mixed Concrete Production Facilities [7]. It is
a system for establishing the adequacy of the production facilities in a readymixed concrete operation. A company interested in certification must retain
the services of a registered professional engineer to perform the inspection
and to certify that the facilities met requirements at the time they were inspected. The check list covers only the physical facilities, not personnel,
procedures, or actual performance of the total operation. Conformance to
the check list is required by ACI 301 and several highway or transportation
departments and is often used in nuclear work.
In the past 10 or 15 years there has been a gradual change away from the
prescription to the performance or strength specification in large or important work. Although there will always be a need for prescriptive specifications for small work, the concrete industry is moving increasingly to the
performance approach. Two factors have been principally responsible for
this trend--the difficulty of obtaining competent testing services and the
almost complete reliance on strength tests made 28 days after the concrete
has been placed in the structure. One additional factor has been that the
mix design and proportioning expertise has been with the specifying agencies
or commercial testing laboratories employed by them.
Increasingly, ready-mixed concrete producers are assuming technical
responsibility. The acceptance of water-reducing admixtures which are now
used in about 50 percent of the concrete and the recent interest in the use of
fly ash, which is now used in about 12 percent of the concrete, are evidence of
this increased interest in performance specifications.
This writer believes that conditions will change rapidly in the future.
474
Ready-mixed concrete producers will become increasingly responsible for

self-inspection systems--quality control testing. They will rely on accelerated
strength testing procedures which will estimate 28-day strengths at 1 or 2
days and they will insist on accurate uniformity information from suppliers of
cement, admixtures, fly ash, and aggregates.
Owners will have to continue to perform acceptance testing but inevitably
the amount of such testing done will be reduced as the quality of the information from the concrete producers improves.
One additional change will be necessary if the system is to evolve. That is,
the prescriptive requirements must be reduced or eliminated and the concrete producer must be able to change materials and proportions rapidly in
response to varying materials, weather, job conditions, and acceptance testing procedures used for or by the owner.
Operating Procedures
Operating procedures are considered to include materials handling,
weighing, charging the mixer, mixing, delivery to the job, and final discharge
into the contractor's placing equipment. ASTM Specification C 94 defines
three basic types of ready-mixed concrete operations and ASTM Specification C 685 includes the fourth. Central mixed concrete is mixed completely
in a central mixer and transported in a revolving drum agitator (that is,
mixer), an agitator, or a nonagitating unit (that is, dump crete). Truckmixed concrete is mixed and delivered in a revolving drum truck mixer.
Shrink-mixed concrete is partially mixed in a central mixer with the final
mixing and transportation in a revolving drum truck mixer. The ASTM
Specification C 685 system is generally known by the name of the only equipment currently available for this process--the Concrete Mobile. The available equipment consists of truck mounted material bins, volumetric proportioning equipment, and continuous rather than batch-type mixing equipment.
The technology in the area of batch equipment and automation has been
developing at an exponential rate and "automation" and the associated
controls will be discussed in some detail.
Materials Handling
Aggregates--The objectives of a well-designed aggregate storage and
handling system are to provide an adequate supply and to prevent segregation, breakage, and contamination. ASTM Specification C 94 has no
specific requirements in this area. The requirements of the NRMCA check
list [7] outline the objectives and leave to the inspecting engineer's judgment
the question of the adequacy of the system. General precautions are described in ACI Recommended Practice for Measuring, Mixing, TransCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:31:15 EDT 2014
475
porting, and Placing Concrete (ACI 304) and the U.S. Bureau of Reclamation Concrete Manual [8].
Where the degradation and breakage produce detrimental clay or claylike fines rehandling and stockpiling must be minimized. Generally, the
plant will include some relatively large secondary storage at the site in stockpiles, large silos, barges, or even rail cars. Primary or overhead storage is
needed to supply the weigh hoppers or aggregate batchers. This is generally
above the batcher and the materials flow by gravity to the weigh hoppers.
Aggregates are moved from secondary to primary storage on belts, with
rubber tired loaders or combinations of these.
The principles of building and reclaiming materials from stockpiles is
discussed in Refs 9 and 10. A number of problems recur. When coarse aggregates are produced in a dry process and they either contain fine material or
degrade in handling, the fines tend to accumulate in the lower portions of the
pile and the loader operator must exercise care in reclaiming these portions
of the stockpile. A similar condition occurs with moisture in the lower portions of a stockpile. To avoid excessive variations in moisture content this
material will have to be elevated by the loader operator so that it can drain or
it must be blended with drier material as it is transferred to overhead storage.
In some instances, generally in the West, specifying agencies have required
"finish screening" of coarse aggregate with rinse water as it goes into the primary or overhead storage. This procedure should be necessary only when
excessive detrimental fines are generated in handling. Further, it is probably
more important in warm dry climates where slab concretes tend to develop
plastic or other early cracking.
Cernentitious Materials--Most cement is delivered by truck and is blown
into overhead storage with pneumatic pumps. Bag filters are used to handle the
excess air. Adequate pressure relief valves and bin level indicators are
necessary to protect the filters. Low pressure air pads are necessary at the
bottom of most bins to ensure flowability. Bin gates, rotary feeders, screws,
and air slides deserve close attention to ensure accurate and complete cutoff
during weighing operations.
Contamination of cement with fly ash has presented serious problems. Fly
ash flows through small openings and cracks in bin partitions and gates
much more easily then cement. Precise shutoff of gates is more difficult to
achieve and therefore cement should be weighed first and the fly ash last in a
cumulative butcher. Multicompartment cement silos must be inspected carefully for small openings in seams. Some operators will not store fly ash in
multicompartment bins adjacent to cement because of the possibility of conrumination. When fly ash is delivered by truck the fill pipe connections
should be reversed or of a different design so it will be impossible to blow fly
ash into a cement bin. Packing, plugging, and flowability has always presented problems in handling cement and fly ash is not immune to these
difficulties. With fly ash, poor flowability is likely to be associated with
476
increased carbon, increased moisture content, or an increase in coarse

fractions.
Admixtures--Most liquid chemical admixtures are furnished in bulk from
the supplier and stored in tanks ranging from about 2 to 8 m 3 (500 to 2000
gal). As with cement and aggregates, care must be exercised to ensure that
the material is placed in the proper tank. Many, if not most, admixtures
must be protected from freezing and tanks and supply lines suitably protected.
Measuring Materials--Batching
The various specification requirements for batching systems are designed
to accomplish one or more objectives. The primary concern is the quality of
the product. Important secondary objectives are to see that the customer
also obtains the quantity of concrete ordered and the quantities of cement
and admixtures where those are a part of the purchase agreement. Specifications for most scales or weigh batchers contain two kinds of tolerances-scale accuracy and batehing accuracy tolerances. The scale accuracy tolerances are applicable during calibration and are designed to ensure that
accurate equipment is used. Batching accuracy tolerances are much more
liberal and are intended to govern under operating conditions where it is not
practical to require the scale to come to static equilibrium. The batching
tolerances generally apply to the indication of the primary or master device.
Somewhat more liberal tolerances are permitted for secondary devices such
as slave dials or recorders. Materials such as cement and aggregates are
generally weighed, although in special circumstances, lightweight aggregates
may be measured by bulk volume. Under ASTM Standard C 685 cement
and aggregates are measured by bulk volume. Water is sometimes weighed
on a scale but most of the time it is measured by volume in a meter. Liquid
admixtures are measured by volume in a variety of meters and other positive
displacement devices, calibrated containers, and sometimes even with calibrated timers. Use of timers should be prohibited, except perhaps for calcium chloride solutions.
Scales or Weigh Batchers

A scale consists of a container generally suspended at the four corners and
transmitting reduced loads through a system of levels to a calibrated beam or
dial indicating device. In some equipment a load cell is used to replace the
dial or beam as the indicating device. The scale accuracy as a percentage of
the total capacity of the scale is 0.4 in ASTM Specification C 94, 0.2 in the
NRMCA Check List [7] and 0.1 in NBS Handbook 44 [i1] and CPMB Standards [12].
Scale calibration procedures vary widely between different agencies and
477
are not highly standardized. ASTM Specification C 94 requires that "adequate standard test weights" be available and the NRMCA Check List requires 227 kg (500 lb) of test weights. Generally, the scales are checked by
alternatively placing the weights on the scale and adding 227 kg (500 Ib) of
material to the hopper. Dial scales can be checked with reduced loads within
the lever system, checking the scale within each of the four quadrants. Occasionally very large individual test weights up to 2250 to 4500 kg (5 000 to
10 000 lb) each are available. When there is sufficient access and suitable
capacity hangars, large aggregate batchers can be checked quite rapidly with
such large individual test weights.
Some specifications require individual batchers for cement, fly ash, and
each aggregate size. These were found principally in large dam projects.
Often individual batchers were needed to increase batching speed. Currently
available automation weighs rapidly and there is little need for individual
batchers. ASTM Specification C 94 permits weighing cement and pozzolans
on the same scale, but not with aggregates. Aggregates may be weighed
either cumulatively or individually.
Table 1 is a summary of batching accuracy requirements of ASTM Specification C 94, NRMCA check list, and CPMB standards. Generally the
tolerances are ___1 percent for cement. ASTM Specification C 94 permits
cumulative weighing of cement and fly ash or pozzolans if the pozzolan is
weighed last, and prohibits weighing cement on the same scale as aggregates.
Both ASTM Specification C 94 and the CPMB standards add the clarifying
statement that the tolerance is _ 1 percent of the required cumulative
weight. Below 30 percent of scale capacity ASTM Specification C 94 requires
overweighing cement by up to 4 percent. The check list and CPMB standards
set a tolerance of ___0.3 percent of scale capacity which of course equals 1
percent of the amount being weighed at 30 percent of scale capacity. The
general rationale is that most automation is able to compensate for material
falling into the batcher to about 0.3 percent of capacity without unduly slowing down the operation. This is about 3 scale divisions on a typical dial scale.
Observation of batch operators on manual scales suggests that they generally are able to weigh to about this same 3 scale divsion. On aggregate batchers the tolerances are generally 2 percent for individual batchers and 1
percent of the cumulative weights for cumulative batchers. Again the 0.3 percent scale capacity restriction becomes equal to 1 percent of the amount
being weighed at 30 percent of capacity and 2 percent at 15 percent of scale
capacity.
The tolerances on measuring water by weight or volume are generally _ 1
percent except that the NRMCA check list permits _ 1.5 percent. Generally
these tolerances are stated as a percent of the total water in the batch, including ice, moisture on the aggregates, etc. The reason is that batches may require small additions of tempering water and it is unrealistic to require measurement to within a few millilitres. The CPMB standards which are for the
--+3 or +--0.3b or
dose per sack
(the larger)
_+3c or _+ dose per sack
-+1 c or + 0 . 3 b
(the larger)
_+1 c or 3.7 litre (1 gad
(the larger)
+0.3 b
+1 a
+0.3 b
+2 r
_0.3 b
-+1 c.d
CPMB a
Dec. 1977, %
aNote that tare compensated controls are required to meet tolerances for individual rather than cumulative batchers.
/'Percent of scale capacity.
Cpercent of amount weighed or measured.
aPercent of required cumulative weight.
ePercent of total mixing water required in the batch including water on aggregates, ice, and water batched from plant and mixers.
fContains requirements on aggregate moisture compensation and slump control.
b. by weight
-----3 or + dose
per sack
(the larger)
same as by volume
or + dose
per sack
(the larger)
same as by volume
...f
+ 0 . 3 b or 3 d
(the smaller)
-+3 e
+-- 1 d
+0.3 b
-+ 1 d
b. total mixing water in

4. Admixtures
a. by volume
-+ 2 c
+0.3 b
-+ 2 c
...
...
---+1.5e or --+ 3.7 litre (1 gal)

(the larger)
--0.3 b
--0, + 4 c
NRMCA
Check List, %
+1 e
-+ 1 c'd
-+ 1 c
ASTM Specification
C 94, %
by weight or volume
3. Water
a. added
by weight
by volume
1. Cement
above 30 b
below 30 b
2. Aggregates
a. individual batchers
above 15 b
below 15 b
b. cumulative batchers
above 30 b
below 30/'
Material
TABLE 1 - - B a t c h i n g accuracy r e q u i r e m e n t .
m
--.4
m
o
0
z
-n
"12
~o
0
13
m
oo
--4
m
,ix
",4
O0
479
equipment require somewhat different minimum tolerances for weigh batchers--___0.3 percent of scale capacity and + 3 . 8 mm 3 ( + 1 gad for volumetric
batchers.
Batching tolerances for admixtures are generally ___3 percent or an amount
necessary to dose a sack of cement. Some available admixture dispensers
have a discrete pulse rate of 0.3 mm 3 (4 oz) and when they are used for
small batches it may be necessary to dilute concentrated admixture solutions.
Volumetric Batching for Water or Admixtures

Although water weigh batchers are more reliable than water meters, they
are used in about 20 percent of the plants. Meters require more frequent
calibration and attention, but the initial cost is much less. Space limitations
are a consideration. The meters used are generally accurate to -t-3.8 mm 3
(___1 gal). At low flow rates meters are likely to be less accurate.
A great variety of volumetric dispersing systems have been used ranging
from simple pumps and timers to positive displacement devices which can be
equipped to record the amount batched. The early devices were graduated
tubes operated manually. Timers depended heavily on constant pressure and
freedom from build-up in lines or valves. Timer systems are permitted only
for calcium chloride in the NRMCA check list. Most systems include a sight
glass or other means of checking the amount batched which is visible to the
batchman. Some dispensers will handle several different admixture solutions
and are provided with a water flush between measurements.
The arrangement of the admixture piping system is of great importance
and deserves careful attention to ensure that variable amounts of material
are not trapped in the lines, that valves will not leak, and that the amount of
admixture measured actually is placed in each batch.
Volumetric Batching of Solids

Generally batching by bulk volume is prohibited under ASTM Specification C 94 and required under ASTM Specification C 685. A few producers
have found that control of structural lightweight aggregate concrete is improved if the coarse aggregate is batched by bulk volume instead of by
weight. With certain lightweight coarse aggregates specific gravities vary
widely and batching by weight produces significant variations in yield. A few
producers have built specially calibrated weigh hoppers and other devices to
be able to batch a constant bulk volume of the material. The principal purpose is to provide improved control of yield. ASTM Specification C 685 requires volumetric measurement of materials and continuous mixing. The
currently available equipment includes a vane feeder for cement with
vibrators and other devices to ensure constant calibration. The aggregate
feeder consists of adjustable gates near the bottom of the bin with a belt that
480
pushes aggregate up against the gates. The speed of the belt and rotation of
the cement feeder are coordinated and operate at constant speed. Lean mixes
are produced by increasing the aggregate gate openings.
Batching Systems and Controls

The CPMB standards provide convenient consistent terminology to define
batching systems and controls. A manual control is one that is actuated
manually with the accuracy dependent on the operator's visual observation.
A semiautomatic batcher is one that, when actuated, starts the weighing
operation and stops automatically when the required weight has been
reached. A "semiautomatic interlocked" control is similar but is furnished
with an interlock to prevent batcher discharge until the required material is
within tolerances. An automatic batcher starts and stops automatically and
includes interlocks to (a) prevent charging until the scale returns to zero, (b)
prevent charging if the discharge gate is open, (c) prevent discharge if the
charging gate is open, and (d) prevent discharge until the material is within
tolerances.
A relatively new type of control made possible by microprocessors and
computers is the tare compensated control which treats the start of each
weighing operation as zero. In this system cumulative batchers are required
to meet batching accuracy tolerances for individual material batchers in the
CPMB standards.
Batching systems consist of the required combination of individual batchers meeting the requirements for semiautomatic, semiautomatic interlocked,
or automatic controls. The number of combinations of these various types of
batchers and the interlocks that are provided are almost endless, but the
following have been selected in the CPMB standard:
1. A manual system consists of a combination of manual batchers except
that admixture or water batchers may have semiautomatic or automatic controls.
2. A partially automatic system is one with at least one control for cement or aggregate batchers which is semiautomatic or automatic and
the system does not meet requirements for semiautomatic or automatic systems. The degree of interlocking is optional.
3. A semiautomatic batching system is one that contains the necessary
individual weigh batchers and volumetric devices (if water or admixture is measured volumetrically) the controls of which are either semiautomatic interlocked or automatic or a combination of these. Interlocks other than those for the individual batchers is optional.
4. An automatic batching system consists of the required combination
of individual weigh batchers and volumetric devices the controls of
which are all automatic. Additionally:
(a) All must be actuated by a single starting mechanism. Water or
481
admixtures not batched at the same time as other ingredients

may have separate starting mechanisms.
(b) Each batcher must return to zero tolerance and volumetric devices must signal empty and reset to start.
(c) Discharge of any ingredient may not start until all individual
batchers have been cleared of the previous batch and returned to
zero within tolerance. Discharge of any weighed ingredient may
not start until all weighed ingredients have been batched.
A review of the plants currently certified under the NRMCA check list
in Table 2 shows a steady decline in manual batching since 1966. Based on
the estimated purchases of equipment in Table 3, manual batching has
remained fairly constant at about 20 percent in this period. Part of the difference is due to the fact that plants remain in service for a long time and
that in 1966 automated plants had been available for only a relatively limited
period.
TABLE 2--Equipment and operating procedures in NRMCA certified plants.
Plants, %
Mixing, % of plants
Truck mixing
Central mixing
Shrink mixing
2. Batching
Manual
Partially automatic
Semiautomatic
Automatic
3. Recording
Cement only
Cement and aggregate
Cement and aggregate
Cement, aggregate, and water
Total
1966
1970
1973
1977
83
15
2
80
19
1
75
23
3
71
27
2
54
45
26
10
17
47
1.
ii"
25
ii
34
36
10
21
33
...
"5 "
5
2
9
12
1
8
10
1
19
21
NOTE--The number of plants certified has varied from 400 to 620. Figures for 1977 may not be
representative of the industry since they include a reasonably large number of plants on
nuclear projects where sophisticated equipment is required.
Weight Setting--The question of how the required weight is to be designated is one that has undergone significant development in the past 10 years.
In a manual system a beam poise is positioned on a lever or the operator
simple observes a dial scale. In an early semiautomatic control a reed switch
was attached to a dial scale pointer and magnets were positioned around the
dial face. Other systems used various mechanical limit switches and other
devices to sense beam balance or position. Later precision potentiometers
482
TABLE 3--Types o f equipment purchased by ready-mixed concrete producers in 1966 and

1976. Values are for equipment purchased in the indicated year. Equipment in use at any given
time may be quite different.
% Estimated Purchases
1. Automation of batehing, %
a. manual
b. partial and semiautomatic a
c. automatic a
d. total
2. Water batching, %
a. by weight
b. by volume
c. total
3. Plant types, %
a(1) central and shrink mixing
(2) truck mixing
(3) total
b(1) permanent
(2) portable
(3) mobile
(4) total
4. Truck mixers
a. mixer size, under 5.3 m ~ (7 yd 3)
5.3 m 3 (7 yd 3)
6.1 m 3 (8yd ~)
6.9 m 3 (9yd 3)
7.6 m 3 (10yd 3)
over 7.6 m 3 (10 yd 3)
total
b. avg capacity, m 3 (yd 3)
5. Mixer drive type
a. chain drive, power take-off
b. chain drive, hydraulic
c. hydraulic direct drive
1966
1976
20
4
76
100
18
32
50
100
21
79
100
12
88
100
10
30
60
100
23
77
100
22
21
57
100
11
38
28
12
9
2
100
6.0 (7.8)
1
5
40
27
22
5
100
6.8 (8.9)
64
36
N
N
33
65
aChanges in the definitions of automatic systems make comparisons between 1966 and 1977
data uncertain, but the percent of manual batchers has been relatively constant.
were attached to the pointer shaft on a dial scale and various systems used to
measure or match the voltage produced. Control panels with provisions for
preset mixes were popular. These presets were potentiometers incorporated
in the panel to produce a reference voltage to be matched during batching.
The development of reliable low cost digital volt meters has greatly simplified
and reduced the cost of converting a voltage into a digital indication. This
has greatly facilitated the use of punched cards. More recently the development of low cost microprocessors and computers promises to permit an even
greater range of available options where mix formulations can be stored in
memory.
483
The use of the term "computer" controlled or operated in descriptions of

batching systems has a certain popular appeal, but there are practical limitations that must be recognized. Although it is theoretically possible to proportion mixes to produce the required strength, the required information is
seldom available in sufficient time to make such calculations worthwhile.
A few plants have been built where a computer has been used to control
and speed up the weighing operations. Here the computer can sense flow
rates of material into the scale, anticipate the correct amount, and shut off
flow with increased precision avoiding the usual system of gate chatter to
achieve final tolerance. A computer can also sense the rate of discharge of
material from a batcher and control that rate instead of relying on preset
combinations of limit switches or predetermined average settings. The significance of control of the sequence and rate of discharge of batchers is discussed in a later section.
Virtually all automated batching systems include provisions to switch to
manual control. Although such provisions are viewed with some skepticism
by specifying agencies, the provisions are necessary to permit weighing small
batches and permit operation in case of malfunction of the automatic system.
Another not so obvious reason is to correct out-of-specification batches or to
add materials to returned concrete. The importance of use or reuse of returned concrete and the precautions that should be observed will be discussed in a later section.
Remote Indicators
The development of punch card automated plants has done much to improve working conditions for the plant operator. The automation required
protection from dirt and dust and often the plant operator's station is air
conditioned with rugs on the floor and windows to observe the operations.
Correspondingly this has increased the need for remote indication, scale
followers, or slave dials. The digital volt meter and cathode ray tube (CRT)
display will further increase the use of remote indicators. The master scale is
required to be accurate to 0.2 percent of scale capacity in the NRMCA check
list. The tolerance is increased to 4-0.25 percent for digital readout. The
scale monitors or slave dials are required to repeat the indication of the master scale within ___0.2 percent of scale capacity. A related item is that digital
recorders are required to reproduce the (master) scale reading within 0.1
percent of capacity. Graphical recorders, because of more limited resolution,
must register within 2 percent of scale capacity.
Recorders
The use of recorders is increasing. Although Table 1 may overestimate the

use of recorders in the industry it shows a dramatic increase from 5 to 21
484
percent of the plants in the period from 1966 to 1977. Graphical recorders
were first used on large projects where large numbers of identical batches
were produced. Under these conditions, inspection for gross errors was easy
and rapid. More recently digital recorders have been most popular. Early
models produced a sequential record on a tape. This record was of relatively
little immediate value except that after a problem was detected the record
could be inspected to see if batching was correct. Early recorders were used
only for cement or aggregates. Later specifications required recordation of
presence or absence of admixtures.
More recently the amount of information recorded has greatly increased.
The minimum is generally (a) batch weights or volumes of materials required, (b) batch count or identification, (c) Time of day to the nearest minute and date, and (d) empty balance of scales and meters. Additional information may include (a) identity of each material, (b) size of batch, and (c)
sand moisture setting.
Increasingly digital recorders are being used to prepart delivery tickets,
keep inventory, and input data directly to accounting for billing.
Additional data on truck operation are being used to help the dispatching
and truck scheduling. Data on truck utilization can be made available to
customers to help increase their efficiency on the job.
Mixing Concrete
As outlined earlier there are three general types of conventional readymixed concrete operations: truck mixing, central mixing, and shrink mixing.
Table 2 shows that of the plants certified under the check list, central mixing
has increased between 1965 and 1977 from 15 to 27 percent. Shrink mixing
is reported at about 2 percent. Estimates of the sales of central and shrink
mixing plants in 1966 and 1976 are less--12 and 23 percent respectively. Because of the long life of such equipment the real percentages of central mix
plants in the industry are even lower. Significant growth in popularity of
central mix plants in the industry is apparent.
Within the technical community the advantage of central mixing over
truck mixing is improved control of quality. However, the CPMB studies of
economic factors [13,14] show that the decision to use central or truck mixing will depend upon a large number of other variables such as the market
area, demand, blade life, and truck utilization.
The low profile plant is another important development. The traditional
permanent plant has bins, batchers, central mixers, and holding hoppers
stacked one above the other so that materials flow vertically by gravity
throughout the batching and charging operation. In a low profile plant the
aggregate batcher is at ground level and materials are elevated on a belt for
charging either a truck or stationary central mixer. The cement batcher can
be located either near the aggregate batcher to discharge on to the belt or it
485
can be located at the elevated end of the belt to discharge cement into the
truck or central mixer. Strehlow [15] discusses equipment, batching techniques, and other factors which affect efficient production of concrete in
different plant configurations.
It is obvious that the primary function of any concrete mixer is to blend
uniformly the separate ingredients. The fact that may not have always been
obvious is that the method and sequence of charging both central and truck
mixers is of great importance in determining whether the concrete will be
mixed [16,17]. The other fact that may not always be appreciated is that
considerable additional mixing is accomplished in the process of discharging
and handling the concrete for placement in the work [18]. That improper
handling can produce segregation in the structure is common experience, but
may be overemphasized.
Truck Mixing
Revolving Drum Truck Mixers--Excluding, for the moment, volumetric
batching and continuous mixing under ASTM Specification C 685, there are
two types of revolving drum truck mixers in use today--rear discharge and
front discharge. Both are inclined axis mixers. Perhaps a hundred horizontal
axis truck mixers are still in use. Very few have been produced since 1966
and they have not been a real factor since perhaps the mid to late 1950s.
Pictures of these and other early designs are given in Ref 5. The rear discharge inclined axis mixer predominates. The front discharge unit has been
produced for perhaps 15 years. Most of the early units were built under the
Rite-Way patents. Most of the other major truck mixer manufacturers have
or are producing units of this type. At this time they account for 1 to 5 percent of the total truck mixers in use.
A principal disadvantage of the front discharge has been the difficulty of
adapting it to a standard truck design and the lack of flexibility in locating
the mixer on the truck frame to comply with the weight laws of the various
states. The advantage of the front discharge is that with controls in the cab
the driver can position the truck and chute during placement. Under favorable circumstances this can eliminate the need for a man to handle the chutes.
Chute handling devices have also been developed for rear discharge units.
Table 3 shows that the average size of mixers purchased has increased from
6 to about 7m 3 (8 to about 9 gal 3) since 1966. The predominant sizes sold are
between 6 and 7.6m 3(8 and 10 yd3).
The other big change in truck mixers has been in the method of powering
the mixer. In 1966 the move had just started away from mechanical power
take off. That system of drive shafts and universal joints was a major maintenance item. The use of a hydraulic pump, hydraulic motor, and chain
drive was just beginning.
486
The direct hydraulic drive which eliminates the chain and associated maintenance problems was widely available only 3 or 4 years ago and in that short
time it has become the accepted system. Although initial costs are higher
maintenance and down time are reduced.
The other innovation that has been established is the paving or low slump
mixer. The traditional or standard truck mixer has 92 cm (36 in.) diameter
rear opening. Low slump mixers have discharge openings up to 107 or 122 cm
(42 or 48 in.). Discharge rate of the low slump models is satisfactory at
slumps down to perhaps 3.8 cm (llA in.) but may be marginal at 2.5 cm
(1 in.), depending on the application. The problem with low slump models is
that the volume of high slump concrete they will hold without spillage is
reduced.
Under ASTM Specification C 94 the capacity of a truck mixer is limited to
63 percent of gross drum volume. When used as an agitator the maximum is
80 percent of gross drum volume. The Truck Mixer Manufacturer's Bureau
Standards [18] list standard mixer sizes from 4.6 to 11.5 m 3 (6 to 15 yd3) with
corresponding maximum permissible ratings as agitators from 5.9 to 15.5 m 3
(790 to 201A yd3). Most manufacturers rate capacity as an agitator somewhat
less than the 80 percent maximum because often the unit will not hold that
amount of concrete without spillage. As mixer size increased, the drum became larger and the angle of inclination of the axis of the drum decreased.
This and the increased size of the discharge opening reduced the amount of
high slump concrete a mixer would hold. The degree to which spillage is a
problem depends upon the terrain and concrete slump.
Charging Truck Mixers--Procedures used to charge the ingredients into
a truck mixer are more important to efficient mixing than any of the other
variables such as batch size, mixing revolutions or mixing speed. The effects
of these variables are discussed in detail in Refs 17 and 19.
All loadings must be designed to avoid packing of material, particularly
sand and cement, in the head of the drum during charging [20]. In most
loadings the probability of head packs is decreased by placing some coarse
aggregate and water in the drum before sand and cement.
The handling of the water deserves special attention. Generally about one
fourth to one third of the water must be added to the discharge end of the
drum after all other ingredients have been charged.
Certain variations of ribbon loadings are generally best. However, the
theoretical optimum that blends water, cement, and aggregate from start to
finish may produce headpacks, cement balls, and poor mixing. Coarse
aggregate and some of the water should start ahead of sand and cement.
Handling of the water is reasonably flexible except that it cannot be batched
entirely first or last. In some situations it will be advantageous to add about
one fourth of the water after all solids have been charged.
Cement-last loadings were typical in two-stop plants where the aggregates
were charged first and the truck driven to a second station for cement. If not
487
handled correctly cement-last loadings can cause problems, depending upon

the mixer used and its condition. To be successful, coarse aggregate and
some water should be charged first. It is essential that about one fourth of the
water be batched last after all the cement is in. This water can come from the
water nozzle in the discharge end of the drum.
Cement-first loadings do not work. However, slurry mixing does work well,
although it is time consuming. For slurry mixing all added water is batched
first followed by the cement. The slurry is mixed for perhaps a minute and
then the aggregates added. This loading will give excellent uniformity and
will be free of cement balls [20].
A variety of other loadings involving sandwiching cement between aggregates and double batching work well if the aforementioned principles are
followed.
Mixing revolutions can be either highly important or irrelevant. With
optimum charging sequences most mixers will be able to produce uniformly
mixed concrete in 30 or 40 revolutions. Generally a minimum of 70 and a
maximum of 100 revolutions at mixing speed are specified. ASTM Specification C 94 contains these numbers, but the wording of the section is such
that some members of the responsible subcommittee feel that these limits
apply only when uniformity tests are to be made. In practice a uniform number of revolutions should be used. Many producers mix for 125 or 150 revolutions as their standard procedure, in spite of the evidence that under optimum conditions a half or a third that number should be adequate. If the
charging procedure is poor additional revolutions at mixing speed are generally of little benefit.
The TMMB standards [21] require mixing speeds more than 6, but less
than 18 rpm and agitating speeds more than 2, but less than 6 rpm. Until
recently maximum mixing speeds were at 10 or 12 rpm (or drum speeds of
1.15 m/s (3.75 ft/s) at the maximum diameter). The NRMCA researches
[17] showed that between 4 and 12 or 14 rpm drum speed had little effect,
but at higher speeds up to about 22 or 23 rpm mixing could be improved
dramatically. In no instance was mixing poorer at 18 or 20 rpm than at 12
rpm. This will continue to be true only as long as current maximum highway truck width limits apply. At present the maximum diameter of drums
in use are between 325.5 and 355.8 mm (86 and 94 in.).
High mixing speeds are particularly important in the incorporation of
tempering water. Many specifications require a minimum of 30 revolutions.
Tests show that this is necessary at drum speeds less than about 12 or 14
rpm. At 18 to 22 rpm only 4 or 5 revolutions may be sufficient.
ASTM Specification C 94 limits the total number of revolutions to a maximum of 300. This is to limit grinding of soft aggregates and other undesirable effects on concrete in hot weather. Slump loss tends to be increased
with increased mixing revolutions, particularly when superplasticizing admixtures are used. Another advantage receiving more attention is the atCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:31:15 EDT 2014
488
tempt to limit revolutions to minimize wear on the blades and drum. If this
continues, efficient loading with a short remix on arrival at the job instead of
agitating the concrete during delivery should become more popular.
The final observation on factors that affect mixing uniformity is that batch
size is not a critical variable. If mixing is poor at capacity then reducing
batch size 0.75 or 1.5 m 3 (1 or 2 yd 3) is not likely to improve mixing. Also if
efficient charging sequences are used and mixing is good at capacity then a
0.75 or 1.5m ~ (1 or 2 yd 3) increase over mixer capacity will not adversely
affect uniformity in most available equipment.
Central Plant or Stationary Mixers

The Plant Mixer Manufacturers Division Standards [22] defines four types
of concrete plant mixers.
1. Nontilting Mixer--This is a revolving drum mixer which charges,
mixes, and discharges with the axis of the drum horizontal. Its capacity is
about 30 percent of gross drum volume and is available, generally in sizes
from 0.75 to 3.4 m 3 (1 to 41.,~ yd3).
2. Tilting Mixers--These are revolving drum mixers that discharge by
tilting about a horizontal axis at right angles to the axis of the drum. The
drum axis can be either horizontal or at an angle to the horizontal. Sizes
range from 1.5 to l l . S m 3 (2 to 15 yd3). These are most popular today. Rated
capacity ranges from 30 to 45 percent of gross drum volume.
3. Vertical Shaft Mixers--This mixer has an annular mixing compartment with rotating blades or paddles. The mixing compartment may be
stationary or rotate. This is the "pan' mixer popular in concrete plants 5 to
10 years ago. Rapid wear and difficulties with high slump concrete restrained
its use. Rapid mixing and low overall height are significant advantages. Sizes
up to about 3.4 m 3 (41/z yd 3) are available. It does a better job of mixing
very dry concretes than drum type mixers [23].
4. Horizontal Shaft Mixer~--Although horizontal shaft mixers are not
new, only recently has equipment designed for use in ready-mixed concrete
plants been available. This mixer has a horizontal cylindrical mixing compartment with blades rotating about the horizontal axis. Although it may be
too early to determine, sizes from 0.75 to 7 m 3 (1 to 9 yd 3) and larger may
be feasible.
Charging Central Mixers--Comprehensive studies of effect of charging
sequence have been made by FHWA on the large tilting mixers used for highway concretes [18,24]. They indicate that to minimize mixing time the sequence and rates of material flow during charging must be balanced carefully to approach uniform blending of all ingredients. Particular care must
be taken to blend the water uniformly throughout the charging period.
Procedures for charging central mixers are quite different from those
necessary for truck mixers. The reason is that a revolving drum central mixer
489
is much less full than a truck mixer and therefore blading and mixing action
is quite different. In a truck mixer there is very little folding action. The
spiral blades must force concrete down the outside of the truck mixer drum
and establish a return flow to the discharge end of the drum along the central
axis.
Mixing Time--ASTM Specification C 94 and the Corps of Engineers [25]
require mixing time in a central mixer of I min for the first cubic yard and 15
s for each additional cubic yard. Reduced mixing time is permitted when
acceptable uniformity is demonstrated. Under this formula a 7 m 3 (9 yd 3)
mix would require 3 min. Experience with mixer uniformity tests suggests
that 90 s would be conservative for all large central mixers. Rarely are mixing
times longer than 60 s required and in many instances 45 s is sufficient.
Many highway departments use either 60 or 45 s as the minimum permitted.
Commercial ready-mixed concrete plants tend to use somewhat longer times,
probably to obtain a more stable indication of slump.
Difficulties with excessive air content have been noted where discharge
was delayed more than about 60 s longer than the regular 45 to 75 s mixing
time. This is likely more prevalent in low slump concrete and in concretes
hauled in nonagitating equipment.
Volumetric Batching and Continuous Mixing

The only available equipment in the United States which produces concrete under ASTM Specification C 685 is the "Concrete Mobile." The equipment is truck or trailer mounted and consists of bins for sand, coarse aggregate, cement, and water. The aggregate bins have a belt at the bottom with
calibrated gates. The aggregate belt is geared to the rotation of a vane feeder
for cement, water and admixtures are proportioned with flow meters. Production rate is inversely proportional to cement content. For mixing, materials are fed into a rubber lined auger mixer. The units are available in sizes
from 3 to 9 m 3 (4 to 12 yd3). The mixer holds about 0.06 m 3 (2 ft 3) of
concrete and mixing time is about 15 or 20 s. Slump is readily adjustable and
the unit can be started and stopped as concrete is needed. Under ASTM
Specification C 685 calibration and mixer uniformity tests are required at
6-month intervals. Yield is checked by weighing output of the unit over a
convenient interval and determining the unit weight of the concrete. The
calculated volume is compared to the output meter on the unit. Washout
tests can be used to check quality of fine and coarse aggregates.
The unit is versatile and has been used in a variety of work, generally
where the amount of concrete required is relatively limited. In somewhat
larger work where requirements are less than 23 or 30 m3/h (30 or 40 yd3/h)
it has been used as a job site mixing plant. Although conventional readymixed concrete producers have incorporated these units in their fleets, most
often they have been purchased by companies catering to the small volume
490
market, by city maintenance departments, and by contractors for special

work. Although no data are available, the number of units in the field is
thought to be about 3 or 4 percent of the total truck mixer fleet.
The outstanding advantage of the unit is that it produces freshly mixed
concrete. Therefore, concrete strengths are improved. The units do encounter problems with false setting cements because of the very short mixing
time.
Delivery and Delivery Time

This discussion will cover delivery as it affects quality, but not the other
aspects such as dispatching, truck utilization, and customer service which
are of great importance to the contractor and producer alike.
The quality issues relate to measurement and control of mixing water,
slump, tempering, retempering and their effects on strength and other properties or characteristics.
The fundamental fact is that strength decreases when water is added to
concrete or when the mixing water content is increased. Also, slump decreases continuously with time. There is a gradual increase in strength the
longer concrete is mixed but only if no water is added and workability is
sufficient to mold acceptable specimens. The available research data are
extensive [26-35].
Control of Slump and Water Content--ASTM Specification C 94 contains
tolerances on slumps of less than 51 mm (2 in.) +_ 13 mm (1/2 in.), 51 to
102 mm (2 to 4 in.) -4- 25 mm (1 in.), and over 102 mm (4 in.) -4- 38 mm
(1 1/2 in.). ACI Specification for Structural Concrete for Buildings (ACI 301)
has a more realistic approach. It specifies maximum slumps of 102 or 127
mm (4 or 5 in.) and permits use of batches of up to 25 mm (1 in.) higher
slump if the average slump or the average of the most recent 10 batches is
less than specified. This has created problems in mass concrete where the
minimum slump is 50 mm (2 in.) and excessively low slump batches that
were difficult to consolidate were used to provide a margin for the higher
slump batches that might be encountered later.
The key to slump control is to provide aggregate handling and storage
facilities which provide uniform moisture in the sand at levels less than about
8 percent. Most sands are produced at moisture contents of 10 to 15 percent.
Once sand is stockpiled it will drain to a reasonably stable value between 6
and 8 percent in about 4 to 6 h depending upon sand grading. Most concrete
plants contain electric moisture meters which provide a reasonably accurate
indication of sand moisture. The placement of electrodes in weigh hoppers
or overhead bins deserve close attention. Calibration checks should be made
daily. These meters will not work if the sand contains small amounts of
dissolved salts and when the sand approaches saturation they give spurious
readings.
491
Although not widely used, neutron moisture meters have been available
for 6 or 8 years. Although they are expensive they are much more accurate
and calibration is stable [36] over long periods of time.
Central mix plants with tilting mixers generally are equipped with ammeters or watt meters which give good indications of the slump. Control of
slump is simplified greatly. Readings are affected by both batch size and the
characteristics of the particular mix design.
Measurement and control of reused water and wash water from the previous batch continues to be a problem. Most producers find it necessary to
rinse off the rear fins between loads with the mixer washed out only at the
end of the day. Hot weather and unusual mix designs may require washing
and discharge of wash water after every load. A few producers find that on
short hauls it is not necessary to wash out between loads. Pollution control
regulations make it increasingly difficult to wash out after every load.
The argument of slump control versus water/cement (w/c) ratio control
continues. The solution lies in use of the statistical approach to w/c ratio
specifications which recognizes inherent variability and which encourages
use of an average w/c ratio well below the intended maximum. The situation
is analogous to that for strength were the statistical approach is widely accepted.
The amount of water required to produce a given slump varies from dayto-day and on the same day. Some of the factors are: (1) errors in methods
of measuring moisture (testing errors), (2) accuracy of hatching water, (3)
interdependence of slump and air content, (4) temperature, delivery time,
and amount of mixing, (5) slight changes in cement and admixture properties, (6) variations in time of addition of admixtures, and (7) changes in
aggregate grading within specification limits.
In controlling slump in the field it is generally assumed that the addition
of about 5 litre/m 3 (1 gal/yd 3) will increase slump about 25 mm (1 in.). This
is recognized to vary with mix design and slump range. The middle range
of slumps around 75 to 100 mm (3 to 4 in.) are generally more sensitive. In
air-entrained mixes the problem is complicated greatly by the fact that an
increase in slump is likely to be accompanied by a corresponding increase in
air content. Again this is critical in concretes with slumps up to about 100
mm (4 in.) or even 130 mm (5 in.). The effects are generally manageable
at target air contents of about 4 percent, but when 8 percent is specified the
variation of air content with slump is appreciable [37].
The significance of variations in coarse aggregate grading, within specifications, on slump, water ratio, and strength have been studied in some
detail [38, 39]. Both studies show that under these conditions production
of uniform slump concrete will produce more uniform strength than production of uniform w/c ratio concrete. This results from the fact that coarse
gradings are similar to large maximum size materials. That is, at a given
water ratio, strength decreases with increased maximum size (or a coarser
492
grading), but increases with a decrease in water required to provide a given

slump.
In very tightly controlled jobs the standard deviation of w/c ratio is
likely to be about 0.03. This is roughly equivalent to a standard deviation
of 10 litre/m 3 (2 gal/yda). This means that if the specifications call for a
maximum w/c ratio of 0.50, the average w/c ratio would have to be about
0.45 to avoid rejection of more than about 5 percent of the batches.
Control of Tempering and Retempering

Since 1968, ASTM Specification C 94 has controlled tempering and retempering by the following:
1. Tempering water can be added only (a) on arrival at the job site, and
(b) and when the slump is less than specified.
2. No water may be added at any later time.
3. The producer is responsible for slump for 20 min after arrival at the
job site.
4. Discharge must be complete within 11/2 h or 300 revolutions but (a)
these limits can be waived if the concrete can be placed without addition of
water, and (b) in hot weather a time less than 11/2 h may be specified by
the purchaser.
The subcommittee which has jurisdiction over the standard has discussed,
balloted, and reballoted a variety of changes in these sections. No other
proposal has been acceptable nor is it likely that the present text would be
acceptable to a sufficient majority if it was not already in the standard.
One proposal was to delete Item 4 containing the 11A-h limit since if no
water is added concrete strength improves with time. Another proposal was
to provide reduced times in hot weather and increased time in cold weather
or if special batching procedures were used to keep cement dry. Both readymixed producers and inspection agencies favor reliance on fixed time
limits.
Two changes should be made in the current provisions. The time limit
should be to the start of discharge rather than completion of discharge. The
absolute prohibition of later addition of tempering water should be relaxed
to permit up to about 5 or 71/2 litre/m 3(1 or 11/2 gal/yda).
Meininger [27] provides data and reviews the effect of batching procedures
on available delivery time and retempering. When concrete is central mixed
or truck mixed in the yard, slump loss starts early and there are only two
possible methods of extending available delivery time. One is to use cooled
concrete and the other is to produce an initial slump of 50 or 75 mm (2 or
3 in.) and temper it on arrival at the job to the 130 mm (5 in.) required. Data
[35] suggest that it is better to retemper than to send out high slump in anticipation of the slump loss.
In truck mix operations there are several alternatives, all of which delay
493
mixing until arrival on the job and are designed to keep a portion of the
cement dry. Meininger [28] shows that when cement is ribboned with damp
aggregate it is wetted and results are similar to those that would be obtained
from mixing concrete at the plant. When cement was loaded as the last
ingredient it was expedient to slow the drum to 4 or 5 rpm when the cement
was loaded. In this instance only 3 or 4 drum revolutions were required to
load the cement and delaying mixing was of significant benefit. The implication is that perhaps half of the cement stayed dry. The big change came
when aggregates and part of the water were batched first, the drum stopped
and the cement loaded on top of the aggregate without turning the drum.
This was the standard procedure fnr horizontal axis mixers and is practical
for the small percentage of inclined axis mixers which have quick release
batches. Delays exceeding 12 h reduced strength only 5 to 8 percent.
Later tests [34, 40] show that cement can be loaded into a still drum of
the normal inclined axis mixer and that the cement is not wetted significantly. Precautions to be observed to obtain adequate mixing with cement-last
loading were discussed earlier.
The other obvious alternative that should be considered is the location of
a cement bin or a complete plant on the job site where a long haul is anticipated.
There has been considerable recent interest in the properties of cement,
and cement-admixture combinations that can affect rate of slump loss [23,
32, 33]. Rapid slump loss can be rather complex but is more prevalent in
hot weather and generally due to an unfortunate combination of cement and
admixture properties. The possible solutions include (a) changing brands of
cement, (b) removing the admixture, (c) trying a 1 to 5 min delayed addition of a half dose of retarder, (d) increasing gypsum content, (e) changing
admixture.
Grinding of aggregates during mixing has been discussed, but relatively
little quantitative data are available [26, 41, 42]. Tests show that because of
the surface area, grinding is mostly in the sand sizes. In one instance grinding released active clay minerals which increased mixing water requirement
and greatly accelerated slump loss. An attrition test procedure [42] has been
employed by the NRMCA laboratory to discover other similar instances. On
several occasions researchers have speculated that continued mixing would
grind the cement, but unfortunately no tests have been reported.
The stability of the air content under prolonged mixing has been studied
on several occasions and generally the loss of air has generally been less
than about 20 percent of the initial amount after 11/2 to 2 h of mixing and
agitation. Limited tests show adequate durability and infer maintenance of
adequate air void characteristics. Recently instances of serious rapid loss
of air has been reported in the field on concretes made with fly ash. The
problems have been experienced with fly ashes with ignition losses of about
5 percent. Other characteristics of the ash are likely of importance. ConCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:31:15 EDT 2014
494
cretes made with other similar fly ashes have retained adequate air contents
for periods as long as 6 or 7 h.
Hot and Cold Weather Concrete
The report of ACI Committee 305 on Hot Weather Concrete published

in 1977 [40] and the report of Committee 306 on Cold Weather Concreting to
be published in early 1978 as ACI Recommended Practice for Cold Weather
Concreting (ACI 306) provide timely information which will not be covered
here.
Strength problems are more prevalent in hot weather than cold. Producers
who keep good records know that as the weather warms up strength results
decrease. Typically, summer concretes might be 2.75 or 4.15 MPa (400 or
600 psi) lower.
The decreased strength levels in hot weather are likely the combined results of (1) higher concrete temperatures reducing strength levels, (2) higher
temperatures increasing mixing water requirement, (3) increase rate of
slump loss, requiring more tempering water, and (4) nonstandard testing
procedures in part related to increased construction activity: (a) initial curing
exceeding the 15 to 27~ (60 to 80~ specified, (b) increased rate of moisture loss, and (c) failure to start moist curing promptly on receipt at the
laboratory at one day.
Although all are important the first three can be solved only by using cool
concrete, reducing delivery time or increasing the potential strength level
by using admixtures, fly ash, or additional cement. It should be noted that
for high strength mixes where long delivery times cannot be avoided, an
increase in cement content may actually reduce strength levels.
Increasingly, specifiers are imposing 32.2~ (90~ temperature limits
for usual construction and 15 to 24~ (60 to 7S~ temperatures for massive
concrete in hot weather. In very large jobs, ice or cooled water is practical
since the refrigeration facilities can be included in the plant. In small jobs
where the demand is more sporadic it is much more difficult.
Block ice is in short supply during hot weather in all parts of the country.
To provide ice the producer must install a slinger to crush or chip the block
ice, refrigerated storage at the site, and one or more refrigerated trailers
to transport the ice. It is not unusual to have to haul ice 160 km (i00 miles)
or more.
In certain massive structures it may be more economical to use moderate
heat cement and use reinforcing to control crack widths. The report of ACI
Committee 207 [43] contains much useful information.
Installation of water chillers for mixing water may be a more economical
solution where a reduction in concrete temperature of 4 to 6~ (about 7 to
10~ will be sufficient.
Several years ago it appeared that use of liquid nitrogen to cool mixing
495
water would be economic. However, with increased cost and demand few
installations have been made in recent years.
In cold weather it is generally necessary to heat aggregates and water by
one of several available methods [44]. One method of heating aggregates in
silos that does not appear to have been previously described is the use of a
burner or pipe or pipes located down through the center of the silo. A blower
at the lower end of the pipe provides the draft with the burner installed near
the top of the silo.
Reuse of Wash Water as Mixing Water in Concrete
Disposal of returned concrete has always been a problem for the industry.
In recent years producers have not been able to discharge wash water into
streams or from their property. This has renewed interest in devices for
reclaiming and reusing returned concrete and wash water. The paper by
Parker and Slimak [45] summarizes the situation.
Some of the solutions include:
1. Working closely with job superintendents to minimize overordering.
2. In large mass pours preparation of smaller alternative work that can be
placed if there is a breakdown or form failure after eight or ten trucks
have been batched and are on the way to the job.
3. Incorporating returned concrete into subsequent batches of lower
strength concrete. This requires control and centralized authority in
the plant to avoid using too much returned concrete or concrete that is
too old or using it in important construction. Reshipping returned
concrete seems to be a fact of life since between 2 and 4 percent of the
total volume shipped is going to be returned. It should be done only
with the full knowledge of the purchaser.
4. Construction of waste separation facilities which separate aggregates
and reuse wash water for washing out equipment and for reuse as mixing water in concrete.
One of the most advanced systems originated in California [46]. In it,
returned concrete is screened and washed to remove aggregate. Fresh water
is used in these operations and the aggregates are returned for use in concrete. The cement, water, and some sand fines are collected in a circulatory
system to be reused as mixing water in subsequent batches. In this portion
of the system the attempt is to produce as dilute a slurry as possible by
adding the maximum amount of clean water. Control is assured by maintaining a specific gravity of about 1.03.
Other less elaborate systems include pits and sloping slabs, drag tanks,
and various rotary scrubbers. Rarely do these systems solve the troublesome
problems of disposal of cement slurry or aggregates, unless additional modifications are made. Often it will not be possible to reduce water use sufficiently to avoid the need for disposal of at least some water.
496
Recently revisions to ASTM Specification C 94 were adopted to permit

reuse of wash water as mixing water for concrete. Tests for mortar strength
and time of set are required weekly for the first month and monthly thereafter as long as test results are satisfactory. Included also is a list of optional
chemical requirements that may be specified. The limits are: (1) chlorides
(prestressed--500 ppm, reinforced in moist environments--1000 ppm), (2)
sulfates--5000 ppm, (3) total solids--50 000 ppm, and (4) alkalies--600
ppm.
The available data indicate no adverse effects due to the use of wash water
[47, 48]. However, care must be exercised in batching admixtures. If they
are batched into wash water required dosages may increase dramatically.
The admixture can be batched on the sand or into a limited quantity of clean
regular water.
StrengthTesting
When strength is used as a basis for acceptance for the concrete, ASTM
Specification C 94 and ACI standards require tests of standard moist cured
152 by 305 mm (6 by 12 in.) cylinders. Job cured specimens are often used to
assess the adequacy of curing, protection, and strength development in the
structure, but not for acceptance of the potential strength of the concrete
delivered. The application of the normal probability distribution and statistical procedures are accepted widely. Results of strength tests vary because
of variations in materials, manufacturing, and in testing procedures. For
this reason the average strength must exceed the specified strength, f j ,
to avoid an excessive number of tests less than f ' . The amount higher
ranges from 2.75 MPa (about 400 psi) to perhaps 8.27 MPa (1200 psi) or
more.
Under ASTM Specification C 94 and ACI 318 trucks are sampled a minimum of once a day for each class of concrete furnished and once for each
115 m 3 (150 yd 3) of concrete delivered each day. There are two acceptance
criteria in ASTM Specification C 94 which are similar to those which appeared in ACI 318 and ACI 301 prior to 1971 or 1972. For structures designed by the ultimate strength method or for prestressed concrete not more
than 10 percent of individual tests are to have values less than f " and the
average of any three consecutive tests must exceed f ' . For working stress
and other structures the averages of six consecutive tests must exceed fc'
and not more than 20 percent of tests can be less t h a n f L
In ACI 301 and ACI 318 these earlier requirements were changed to require averages of three consecutive to exceed fc' with no individual tests
more than 3.45 MPa (500 psi) belowfc'.
The probability of tests less than fc' under current ACI 318 procedures
ranges from about 3.5 to 9 percent or in other words the required overdesign
497
is somewhat larger than in ASTM Specification C 94 for ultimate strength

or prestressed structures. The current ACI procedures which produce from
3.5 to 9 percent of tests less than fc' are in reasonable agreement with the
recommendations of the CEB-CIB-FIP-Rilem Committee on Statistical
Quality Control of the Quality of Concrete [49]. The recommendation is that
the average strength of the concrete be at that level which will limit the
probability tests less than the design strength to 5 percent.
ASTM Specification C 94 provides information on the average strength
necessary to meet the acceptance criteria depending upon the coefficient
of variation of the operation. No formalized mix design submission or approval procedure is included in ASTM Specification C 94. In 1971 and 1972
the ACI standards adopted the acceptance criteria described above and
required mix submission procedures that emphasized the use of prior experience with similar mixtures. The generalized procedure for mix approval
is as follows.
1. Determine the standard deviation from one or more groups of a total
of 30 tests on "similar" concretes. The ACI standard then specifies over
design ranging from 2.75 to 8.27 MPa (400 to 1200 psi) for standard deviation from 2 to 4 MPa (300 to 600 psi). If no suitable record of 30 tests is
available the required overdesign is 8.27 MPa (1200 psi).
2. Identify a mix design which will provide the average strength determined above. The average can come from the recording of 30 tests or can
be determined by preparatory laboratory batches.
These mix qualification procedures only recently are being used widely.
They are subject to great differences in interpretation. The recent article
by Peters [50] discusses some of these difficulties.
Failure to Meet Strength Requirements

The consequences of a failure to meet strength requirements varies
greatly. ASTM Specification C 94 suggests that the manufacturer and purchaser confer and if they cannot agree, appoint an arbitration panel.
Since ACI 318 is a code document it addresses the question of structural
adequacy and public safety. A 1977 revision requires steps to be taken to
improve the average strength level when either of the two criteria are not
met--average of three consecutive or individual tests more than 3.45 MPa
(500 psi) belowfL Further when individual tests are more than 3.45 MPa
(500 psi) below f~' "steps shall be taken to assure that the load-carrying
capacity of the structure is not jeopardized." Tests of cores may be required
"if the likelihood of low strength concrete is confirmed and computations
indicate that the load-carrying capacity may have been significantly reduced." Cores can be tested wet or dry depending on moisture condition in
the structure and results are considered satisfactory if the average of three
498
cores equals 0.85f" with no cores less than 0.75fc'. Load tests of the structure are suggested as an additional step if structural adequacy remains in
doubt.
A large body of significant information has been published in the past 15
years on the testing and interpretation of core tests. References 51 and 52 are
useful and an extensive annotated bibliography is being prepared by ACI
Committee 214.
Future Developments in Acceptance and Quality Control Testing

Significant developments are being made in the application of statistical
principles to the problems of judging compliance with strength specifications, determining the practical consequences of a failure to meet a strength
specification and controlling the risk associated with any specified level of
quality. The Federal Highway Administration and a number of state transportation departments have been moving vigorously toward something which
comes under the general label of "Statistically Oriented End Result Specifications" [53]. Although each application seems to be somewhat different
at least two factors are common. First, instead of relying on fixed levels of
strength that are either acceptable or unacceptable they have instituted
graduated reduced payment clauses for strengths less than some specific
level. The theory is that a minor deviation below specified levels will reduce
the service life or increase required maintenance and, therefore, the contractor should be penalized in proportion to that loss in value. Naturally, at
some strength levels the concrete will be unacceptable and must be removed.
This objective is rarely achieved because of the difficulty of quantifying the
unknown. A number of graduated penalty functions have been used, but
ultimately they are more or less arbitrary. Willenbroc and Kopac [54] discuss
four approaches and recommend procedures for developing three of them.
Under traditional specifications, engineering judgment was relied upon to
determine the nature and amount of corrective action required in any given
instance. For the highway department it presumably reduces the decisions
to a mathematical formula and removes them from the political or judgmental arena. For the contractor and ready-mixed concrete producer it
creates a great many problems. Too often the concrete producer is asked
to accept reduced payments based on the cost of concrete in place which is
often three to eight times the price he receives at the end of the chute. The
potential advantage for the concrete producer that is rarely realized is that
it permits a calculation of the level of risk associated with the quality level
supplied. Unfortunately this holds little appeal for the producer since in most
circumstances the uncertainties in the testing and in his ability to control
constituent materials and production do not permit predictable control of
the risk involved. The only reliable response is to furnish an excessive strength
margin.
499
The second issue in the move to oriented statistically end result specification is to require specific quality control by the contractor. One practical
result is to shift the burden of the routine testing to the contractor. In some
instances, the contractors quality control tests, made under surveillance of
highway department inspectors, are also used for acceptance. In other instances separate acceptance testing is done by the specifying agency.
Another important development is the "Authorization Scheme" of the
British Ready Mixed Concrete Association (BRMCA) [55]. The "Authorization Scheme" covers personnel, materials, plant and equipment, operating
procedures, and quality control procedures.
Control charts, based on the Cusum method developed by Ready Mixed
Concrete Ltd., England, are maintained for: (1) the difference between the
target 28-day strength and the 28-day strength predicted from the accelerated tests, (2) the difference between the assumed standard deviation and
that estimated from successive pairs of results, and (3) the difference between
the predicted and actual 28-day strengths.
Cumulative sum or Cusum charts are developed by plotting the accumulated differences between the actual values and the target or expected average values. A rise or fall in the accumulated sum indicates actual values
consistently above or below the target values. When significant changes
occur, cement content is modified to maintain target or acceptable values.
Predicted 28-day strengths are based either on the 7-day results or 1-day
accelerated tests [56]. Recent developments greatly reduce the volume of
testing that has to be done to establish the correlation between results of
accelerated and standard 28-day procedures. The BRMCA "Scheme" does
quantify the risk of failure to meet specification requirements, but as in commercial work in this country the consequences of a failure are not quantified.
The emerging reduced payment clauses in U.S. highway department attempt
to address this problem.
It remains to be seen if it will ever be possible to develop penalty functions
which are equitable to both the owner and contractor.
Closure
The quality of ready-mixed concrete continues to improve; however, the

problems of obtaining accurate testing and of deciding what to do when test
results fail to conform to specifications continue to create problems for specifiers, contractors, and producers. Increasingly sophisticated automation
and recording devices will be more reliable and efficient and reduce batching
errors to the practical minimum. Developments in specification, quality
control, and accelerated strength testing have great potential for reducing
and quantifying the risks assumed by producers, contractors, and owners
alike. With well conceived educational programs, improved testing, and
500
improved quality control of constituent materials the incidence of defective

concrete can be reduced.
References
[1] "Cement and Concrete Terminology," Publication SP 19, American Concrete Institute,
1967, p. 103.
[2] "Standard Industrial Classification Manual-1972," SIC No. 3273, p. 141.
[3] "Pictorial History of the Ready Mixed Concrete Industry," National Ready Mixed
Concrete Association, 1964, p. 44 (out-of-print).
[4] "NRMCA Production and Value Survey," National Ready Mixed Concrete Association,
(annually through 1970).
[5] "NRMCA Industry Data Survey," National Ready Mixed Concrete Association, 1972
through 1975.
[6] Levine, Sid, "Economic Indicators for the Concrete Industry," Modern Concrete, June
1976, p. 41.
[7] NRMCA Check List for "Certification of Ready Mixed Concrete Production Facilities,"
3rd. Revision, 1 Jan. 1976, National Ready Mixed Concrete Association, p. 21.
[8] Concrete Manual, Bureau of Reclamation, 8th ed., 1975.
[9] Warden, W. B., "Stockpiling of Aggregate for Gradation Uniformity, NSGA Circular
No. 99, May 1966, pp. 33.
[10] Miller-Warden Associates, "Effects of Different Methods of Stockpiling and Handling
Aggregates," NCHRP Report 46, Highway Research Board, 1967, pp. 102.
[11] NBS Handbook 44, Sec. T.3.5 and T.2.8., National Bureau of Standards.
[12] "Concrete Plant Standards of the Concrete Plant Manufacturers Bureau," Sixth Revision,
Dec. 1977, p. 5.
[13] "CPMB Does Cost Survey on Central Versus Transit Mixing," Concrete Products,
Nov. 1974, pp. 41.
[14] "A Study of Economic Factors of Central Mixing in the Production of Ready Mixed
Concrete," CPMB Publication No. 103, 1976.
[15] Strehlow, R. W., "Concrete Plant Production," CPMB, 1973, pp. 112.
[16] Bloem. D. L. and Gaynor, R. D., "Factors Affecting the Homogeneity of Ready Mixed
Concrete," National Ready Mixed Concrete Association, Phase 1, June 1969, p. 25.
[17] Gaynor, R. D. and Mullarky, J. I., "Mixing Concrete in a Truck Mixer," NRMCA
Publication No. 148, National Ready Mixed Concrete Association, Jan. 1975, pp. 24.
[18] Bozarth, F. M., Granley, E. C., and Grieb, W. E., "A Study of Mixing Performance of
Large Central Plant Concrete Mixers," Research and Development Report, U.S. Department of Commerce, Bureau of Public Roads, Office of Research and Development,
Washington, D.C., July 1966.
[19] Mullarky, J. I. and Gaynor, R. D., "Abstract of Mixing Concrete in a Truck Mixer,"
NRMCA Technical Information Letter No. 296, National Ready Mixed Concrete
Association, 16 March 1973.
[20] "Cement Balls or Lumps in Freshly Mixed Concrete--Causes and Cures," NRMCA
Technical Information Letter No. 320, National Ready Mixed Concrete Association,
11 April 1975.
[21] TMMB Standards of the Truck Mixer Manufacturers Bureau, 9th rev., 1 Nov. 1971.
[22] Concrete Plant Standards of the Plant Mixer Manufacturers Division, Concrete Plant
Manufacturers Bureau, 4th rev., 1 July 1973.
[23] Johansson, Arne, "The Relationship Between Mixing Time and Type of Concrete Mixer,"
Proceedings, Swedish Cement and Concrete Research Institute, Vol. 42, 1971, p. 92.
[24] Bozarth, F. M., "Case Study of Influence of Imbalances in Charging of Cement and
Water on Mixing Performance of an 8-Cubic Yard Central Plant Mixer," Research and
Development Report, U.S. Department of Transportation, Federal Highway Administration, Bureau of Public Roads, Washington, D.C., July 1967.
[25] "Standard Practice for Concrete," EM 1t10-2-2000, Change 3, June 1974, Department
of the Army, Office of the Chief of Engineers.
501
[26] Gaynor, R. D., "Effects of Prolonged Mixing on Properties of Concrete," NRMCA

Publication 111, National Ready Mixed Concrete Association, 1963, p. 18.
[27] Meininger, R. C., "Study of ASTM Limits on Delivery Time," NRMCA Publication 131,
National Ready Mixed Concrete Association, 1969, pp. 17.
[28] Meininger, Richard C., "Effect of a Delay Between Batching and Mixing on Concrete
Strength," National Ready Mixed Concrete Association, 900 Spring Street, Silver Spring,
Md. 20910.
[29] Ravina, Dan, "Retempering a Prolonged-Mixed Concrete with Admixtures in Hot
Weather," Journal. American Concrete Institute, Vol., 72, No. 6, June 1975, p. 291-295.
[30] Beaufait, F. W. and Hoadley, P. G., "Mix Time and Retempering Studies on Ready
Mixed Concrete," Journal, American Concrete Institute, Vol. 70, No. 12, December
1973, p. 810 and Discussion, Journal, ACI, Vol. 73, No. 4, April 1976, p. 233.
[31] Tuthill, L. H., Adams, R. F., Bailey, S, N., and Smith, R. W., "Case of Abnormally Slow
Hardening Concrete for Tunnel Lining," Journal, American Concrete Institute, Vol. 57,
No. 9, March 1961, pp. 1091-1109.
[32] "Concrete Admixtures," TRB Record 564, Transportation Research Board, 1976 (papers
by Green, Young and Daugherty and Kowaleski).
[331 Hershey, A. T., "Slump Loss Caused by Admixtures, Problems-Practices," Journal,
American Concrete Institute, Vol. 72, No. 10, October 1975, p. 526.
[34[ Gaynor, R. D., "Shrinkage Compensated Cement and the Ready Mixed Concrete
Producer," NRMCA Technical Information Letter No. 294, National Ready Mixed
Concrete Association, 6 Feb. 1973.
[35] Previte, R. W., "Concrete Slump Loss," Journal, American Concrete Institute, Vol. 74,
No. 8, Aug. 1977, pp. 361-368.
[36] Ullman, G. R. and Gaynor, R. D., "Nuclear Moisture Meters for Sand," NRMCA
Technical Information Letter No. 295, National Ready Mixed Concrete Association,
8 Feb. 1973, p, 7.
[37] Gaynor, R. D., "Control of Air Content in Concrete--An Outline," NRMCA Technical
Information Letter No. 333. National Ready Mixed Concrete Association, July 18, 1977.
p. ll.
[38] Hudson, S. B. and Waller, H. F., "Evaluation of Construction Control Procedures,"
NCHRP Report No. 69, Highway Research Board, 1969. (Chapter 2 "Study of Effect of
Variations in Gradation of Coarse Aggregate on Characteristics of Portland Cement
Concrete," pp. 4-12.)
[39] Baker, S. D. and Scholer, C. F., "Effect of Variations in Coarse Aggregate Gradation
on Properties of Portland Cement Concrete," Highway Research Record 441, Highway
Research Board 1973, pp. 97-110. (Also Joint Highway Research Project, Purdue University and Indiana Highway Commission, No. 8, May 1973, p. 107.)
[40] ACI Committee 305, "Hot Weather Concreting," Journal, American Concrete Institute,
Vol. 74, No. 8, Aug. 1977, pp. 317-332.
[41] Slater, W. A., "Tests of Concrete Conveyed from a Central Mixing Plant," Proceedings,
American Society for Testing and Materials, Vol. 31, Part 2, 1931, pp. 510-533.
[42] Davis, R. E., Mielenz, R. C., and Polivka, M., "Importance of Petrographic Analysis
and Special Tests Not Usually Required in Judging Quality of Concrete Sand," Journal of
Materials, American Society for Testing and Materials, September 1967, p. 461.
[43] "Effect of Restraint, Volume Change, and Reinforcement on Cracking of Massive Concrete," Journal, American Concrete Institute, Vol. 70, No. 7, July 1973, pp. 445-470.
[44] "Cold Weather Ready Mixed Concrete," NRMCA Publication No. 130, National Ready
Mixed Concrete Association, Sept. 1968, p. 19.
[45] Parker, L. C. and Slimak, M. W., "Waste Treatment and Disposal Costs for the Ready
Mixed Concrete Industry," Journal, American Concrete Institute, Vol. 74, No. 7, July
1977, pp. 281-287.
[46] Harger, H., "A System for 100 Percent Recycling of Returned Concrete," NRMCA
Publication No. 150, National Ready Mixed Concrete Association, March 1975.
[47] Unpublished memorandum from R. D. Gaynor to ASTM Subcommittee C09.03.09
Methods of Testing and Specifications for Ready Mixed Concrete, "C 94 and C 685
Requirements for Quality of Mixing Water," 16 June 1975, pp. 9.
[48] Pistilli, M. F., Peterson, C. F., and Shah, S. P., "Properties and Possible Recycling of
502
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
Solid Waste from Ready Mixed Concrete," Cement and Concrete Research, Vol. 5, No.
3, May 1975, p. 249.
Recommended Principles for the Control of Quality and the Judgement of Acceptability
of Concrete," Materials and Structures, RILEM, No. 47, Sept.-Oct. 1975, pp. 387-403.
Peters, D. J., "Concrete Quality--A Producer Looks at the Building Code Requirements," Journal American Concrete Institute, Oct. 1977, p. 501.
Petersons, N., "Recommendations for Estimation of Quality of Concrete in Finished
Structures," Materials and Structures, RILEM, Vol. 4, No. 24, Nov.-Dec. 1971.
Gaynor, R. D., "One Look at Concrete Compressive Strength," NRMCA Publication
No. 147, National Ready Mixed Concrete Association, Nov. 1974.
"Statistically Oriented End-Result Specifications," NCHRP Synthesis No. 38, Transportation Research Board, p. 40, 1976.
Willenbrock, Jack H. and Kopac, Peter A., "A Methodology for the Development of
Price Adjustment Systems for Statistically Based Restricted Performance Specifications,"
Research Report No. 1, Pennsylvania Transportation Institute, Project No. 74-27 (1),
Oct. 1976, p. 169.
BRMCA Authorization Scheme, Part IV of BRMCA Code for Ready Mixed Concrete,
British Ready Mixed Concrete Association, May 1975.
Grant, N. T. and Warren, P. A., "A Cusum-Controlled Accelerated Curing System for
Concrete Strength Forecasting," ERMC077 Congress, Stockholm, June 1977.
D. W . L e w i s 1
Chapter 30--Lightweight Concrete

and Aggregates
Introduction
Both "lightweight concrete" and "lightweight aggregate" are general
terms which include a wide variety of products and are frequently subject
to varying definitions. Concrete lighter in weight than that usually obtained
with "normal weight" aggregates (gravel, air-cooled blast-furnace slag,
and crushed stone) may be produced by several methods:
1. Lightweight aggregate concrete, which is comparable to conventional
concrete except for the use of lightweight materials as all or part of the
aggregate.
2. Cellular or foam concrete, where the light weight is caused primarily
by inclusion of a large amount of air or gas, usually 25 percent or more. A
suitable foam or foaming agent is combined with cement and water, either
with or without other ingredients such as sand or pozzolanic fines.
3. No fines or "popcorn" concrete, with the light weight produced by
high air contents or specially graded aggregates having little or no fine
material or both. Either lightweight or normal weight aggregates may be
used.
Lightweight concretes produced by these various methods may range in
unit weight from 240 to 320 k g / m 3 (15 to 20 lb/ft 3) to weights only slightly
less than those of conventional concretes. Applications include: (a) structural
concrete for use in buildings and bridge decks, where weight reduction
produces design economies and other characteristics desired are the same
as for conventional concretes made with normal weight aggregates, (b)
masonry units requiring moderate strength with light weight and good
insulation properties, and (c) insulation or fill uses where light weight and
a high degree of insulation are required, with strength being of lesser importance.
Lightweight aggregates of various types have been used for many years,
1Chiefengineer, National Slag Association, Alexandria, Va. 22314.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
504
and in the past 25 years they have become an important factor in the concrete construction industry. Both mineral (inorganic) and vegetable (organic)
materials of many varieties have been considered and tested for use as
lightweight aggregates. The inorganic or mineral materials are the only
ones that have been used to any significant extent and are the only materials covered by current ASTM specifications for lightweight aggregates.
The most common method of making lightweight concrete in the United
States is the use of lightweight mineral aggregates, with some use of cellular
concretes for insulating and fill purposes.
Lightweight concrete, as discussed in this paper, is that having air-dry
unit weights of 1840 kg/m 3 (I15 lb/ft a) or less and made with lightweight
aggregates as all or a major part of the aggregate. Discussion of lightweight
aggregates is confined to the major types of inorganic aggregates covered
by current ASTM specifications.
Types of Lightweight Aggregates

Various methods have been used to classify lightweight concrete aggregates into general types. These are frequently based on original source, as
either natural or artificial materials [1]. 2 The artificial aggregates are either
by-products of an industrial operation or specially processed by calcining,
expanding, or sintering. In this classification system, such materials as
pumice, scoria, and tuff are natural aggregates; by-product materials include cinders; processed lightweight aggregates are produced from clay,
shale, slate, perlite, vermiculite, blast-furnace slag, and fly ash. Some of
the materials may be placed in more than one of the type groups. For
example, pumice may be used either in its natural state or after heat processing (calcining); blast-furnace slag and fly ash are by-products that are
processed for use as lightweight aggregates by expanding and sintering,
respectively; clays, shales, and slates are natural materials that are always
heat processed during production of lightweight aggregates.
It seems most reasonable to classify lightweight aggregate materials for
discussion purposes on the basis of their end use under ASTM specifications. Therefore, the various kinds of lightweight aggregates are discussed
here on the basis of the ASTM specifications in which they are included.
Lightweight Aggregates for Structural Concrete

ASTM Specification for Lightweight Aggregates for Structural Concrete
(C 330) covers two general types of lightweight, cellular, and granular inorganic material as follows: "aggregates prepared by expanding, calcining,
or sintering products such as blast furnace slag, clay, diatomite, fly ash,
LEWIS ON LIGHTWEIGHT CONCRETE AND AGGREGATES
505
shale or slate, and aggregates prepared by processing natural materials,

such as pumice, scoria, or tuff."
Expanded shales, clays, slates and blast-furnace slags, and the natural
scorias and pumices are the materials usually considered for structural
lightweight concrete usage. Of these aggregates, the expanded shales, clays,
and slates are the most readily available in many geographic areas and,
therefore, are used in the majority of current structural applications. Slags
and the natural aggregates are more limited in their areas of availability.
Some use has been made of diatomite. Commercial production of fly ash
as a lightweight aggregate has been developed in the past decade.
Expanded slag is produced by controlled processing and cooling of molten
blast-furnace slag to obtain a lightweight cellular "clinker" or pellet, which
is subsequently crushed and screened to the required size. ASTM Definition
of Terms Relating to Concrete and Concrete Aggregates (C 125) defines
expanded blast-furnace slag as "the lightweight cellular material obtained
by controlled processing of molten blast-furnace slag with water, or with
water and other agents such as steam or compressed air or both." A number of different methods for accomplishing and controlling the expansion
have been used [2-4]. No significant difference in product characteristics
results from the various processes except the pellet form which has lower
internal absorption and rounded, smooth particle shapes. Expanded slags
have been produced in North America since the late 1920s, with production
of the pelletized materials beginning in 1969.
Expanded shales, clays, and slates are generally treated as a single type,
since no significant differences appear to exist due to the particular classification of the raw material used. Lightweight aggregates of this type were
introduced in the United States about the time of World War I and later
came into widespread use due to rising demands for lightweight concrete
and widespread distribution of suitable raw materials. The aggregates are
produced either by the rotary kiln or sintering processes.
In the rotary kiln process, the raw material is heated to a plastic state in
the kiln, and gases released within the material cause it to expand to form
a lightweight, cellular structure. Following expansion, the material is cooled,
crushed, and screened. In a variation of this process, material is sized before expansion, producing aggregate with relatively smooth particle surfaces,
referred to as "coated" aggregates.
Sintered materials are produced on a traveling grate over which suitable
raw material mixed with small amounts of fuel has been spread evenly.
The fuel is ignited under an "ignition hood" and continues to burn as the
grate moves over blowers. The clinker is allowed to cool, then crushed
and screened.
Fly ash is processed for use as a lightweight aggregate by mixing with
sufficient moisture to permit it to be either pelletized or extruded to form
spherical or cylindrical shapes that can then be sintered. Carbon present
506
in the fly ash forms all or a large p a r t of the fuel required after ignition.
Fine aggregate sizes can be produced by crushing after sintering and cooling.
Pumice and scoria are cellular, lightweight materials of volcanic origin.
True pumice is a spongy lava from which steam or gas escaped while it
was still molten. Commercially, volcanic ash naturally cemented together
(more properly designated as tuff) is sometimes referred to as pumice.
There are many deposits of pumice in the Western states, widely varying in
strength and porosity characteristics. The weaker materials (including tufts)
are sometimes benefited by calcination at a temperature near incipient
fusion. Scoria is a volcanic "cinder" with its pores chiefly in the form of
vesicules instead of the more tubelike, interconnected pores of the pumices.
It is available in a number of localities in the Western states, crushed and
screened to proper size, and is imported from Mediterranean sources for
occasional use in East Coast areas.
Lightweight Aggregates for Concrete Masonry Units

ASTM Specification for Lightweight Aggregates for Concrete Masonry
Units (C 331) includes all lightweight aggregates described above and, in
addition, cinders and the expanded perlites and vermiculites. Since the primary use for perlite and vermiculite is in insulating concrete, they are
covered in the following section on insulating concretes.
Cinders (referred to as "clinker" in Great Britain) are the residue of
high temperature combustion of coal or coke. Known as "industrial cinders," "boiler cinders," or "steam cinders," these materials from industrial,
power plant, and steam locomotive boilers were probably the first lightweight aggregates used in the United States. Christensen [5] summarized
their characteristics, uses, and performance in 1931, including possible
effects from deleterious materials present in some sources. Dwindling
supplies have resulted in greatly decreased usage, and limited applications
in concrete masonry units constitute the only commercial use for cinders in
concrete at the present time.
Lightweight Aggregates for Insulating Concrete

ASTM Specification for Lightweight Aggregates for Insulating Concrete
(C 332) covers those aggregates considered satisfactory for the manufacture
of lightweight concrete not exposed to the weather and for which the primary consideration is the thermal insulating properties of the concrete.
They are divided into two groups: Group I includes such materials as expanded perlite and vermiculite, which generally produce concrete weighing
from 240 to 800 kg/m 3 (15 to 50 lb/ft a) and having a thermal conductivity,
k, of 0.06 to 0.22 W / m . K (0.45 to 1.50 Btu 9in./h 9ft 2. ~
while Group
507
II includes the same materials as those covered under ASTM Specification

C 330 for structural concrete (described above). When usec~ in concrete primarily for insulation purposes. Group II aggregates are expected to result
in concrete weighing from 720 to 1440 k g / m 3 (45 to 90 lb/ft 3) and having
thermal conductivities of 0.15 to 0.43 W / m - K (1.05 to 3.00 B t u - i n . /
h" ft2~
Perlite is a volcanic glass which contains sufficient combined water that,
when it is heated quickly, generates steam causing disruptive expansion
and breakage into small expanded particles. Under controlled kiln conditions an aggregate of high internal void content and very low unit weight
is produced.
Vermiculite is a micaceous mineral that forms another extremely lightweight aggregate when subjected to sudden heating. Conversion to steam
of layers of combined water in the laminated structure of the mineral produces an "exfoliated" or expanded lightweight aggregate.
Lightweight Aggregate Specifications and Tests

The characteristics of lightweight aggregates covered by the current
ASTM specifications are discussed in the following sections. ASTM Specifications C 330 and C 331 are quite similar in requirements. Both include
limitations on deleterious substances and various concrete-making properties. Insulating concrete (ASTM Specification C 332) is presumed to be
protected from weathering, and durability requirements are omitted; however, thermal conductivity is specified. Specific requirements are discussed
below under the individual characteristics.
Composition
All specifications stipulate that "the aggregates shall be composed predominantly of lightweight cellular and granular inorganic material," thus
excluding from consideration such materials as seaweed, straw, cork, and
sawdust, which have not been extensively used in this country [6]. General
types of materials considered to be acceptable are enumerated and the
aforementioned aggregates are listed as examples. The specifications are
not restricted to these particular aggregates, however.
Grading
Each specification tabulates requirements for gradation of the aggregates.
Coarse aggregate gradings up to 25.0 mm (1 in.) nominal maximum size
are given in ASTM Specification C 330, while ASTM Specifications C 331
and C 332 restrict top size to 12.5 mm (1A in.) for the same aggregates.
Most lightweight aggregates are available only in sizes smaller than 19.0 mm
508
( 90 in.) while perlites and vermiculites are normally supplied only in 4.75 mm
(No. 4) or smaller top size gradations. Special gradings are stipulated in
ASTM Specification C 332 for these two aggregates. In general, the gradings in ASTM Specifications C 330 and C 331 are similar to those for normal
weight aggregates (ASTM Specification for Concrete Aggregates (C 33))
except for somewhat larger amounts of material passing the smaller sieve
sizes. ASTM Specification C 331 provides that the gradation requirements
may be waived to produce special characteristics in masonry units, such as
texture, strength, weight, acoustical, or thermal insulating properties.
Grading is determined in accordance with ASTM Test for Sieve or
Screen Analysis of Fine and Coarse Aggregates (C 136), except that the
test sample of fine aggregate is reduced in size, and the sieving time is
limited to 5 min. A table provides weights for the test samples as determined
by the fine aggregate unit weight. Size of the test sample for coarse aggregate is set at 2830 cm 3 (0.1 ft a) or more, of the material used for the determination of unit weight. These modifications are intended to prevent blinding of the smaller sieves with an excessive volume of material and breakage
of the more friable aggregates during the sieving operation.
Uniformity of Grading
Variations in grading of lightweight aggregates may not only affect water
demand and strength characteristics of concrete but will also change the
unit weight because of differing specific gravities for various sizes of particles. The ASTM specifications for lightweight aggregates provide that the
maximum variation in fineness modulus shall not be greater than 7 percent
from that of the sample submitted for acceptance tests, unless it can be
demonstrated that the resultant concrete will have the required characteristics. This provision applies to both the coarse and fine aggregates.
Unit Weight
All lightweight aggregates are subject to unit weight limitations. Expanded
slags, clays, shales, slates, fly ashes, cinders, and natural lightweights must
have dry loose unit weights not greater than 880 kg/m 3 (55 lb/ft 3) for the
coarse aggregates, 1120 k g / m 3 (70 lb/ft 3) for the fine aggregate, and 1040
kg/m 3 (65 lb/ft 3) for combinations of the coarse and fine when marketed
as a single grading. Maximum dry loose unit weights for perlite and vermiculite are 196 and 160 k g / m 3 (12 and 10 lb/ft3), respectively. The tests are
conducted in accordance with the ASTM Test for Unit Weight of Aggregate
(C 29) using the shoveling procedure, except that oven-dry aggregate is
tested. Minimum unit weights of 120 and 96 kg/m 3 (7189 and 6 lb/ft 3) are
stipulated for perlite and vermiculite, respectively, to avoid excessively expanded, weak, and friable materials. Maximum values only are applied to
509
the materials used in structural concrete and masonry units, since excessively
light and weak aggregates of these types would be excluded from use either
by failure to comply with other specification requirements or by the need
for excessively high cement contents and the high concrete unit weights
resulting from degradation.
The unit weight test furnishes a control over the maximum unit weight
or density of lightweight aggregates and, therefore, provides some idea of
the weight characteristics of concrete in which they might be used. However, correlations of aggregate and concrete unit weights are very general in
nature, since particle shape and gradation differences may cause large
variations in the voids between aggregate particles in the test.
Uniformity of Unit Weight

Unit weight of a lightweight aggregate will be affected by changes in
gradation, specific gravity or density of the mineral material, and in the
particle shape or surface texture. Of course, the variations will be reflected
in the water requirements, yield, unit weight, and strength of the concrete
in which the aggregate is used. To provide control of uniformity in use, the
specifications require that the unit weight of successive aggregate shipments shall not differ by more than 10 percent from that of the sample
submitted for acceptance tests.
Deleterious Substances
A number of materials are considered to be potentially deleterious in
lightweight aggregates used in structural concrete or masonry units. Accordingly, they are subject to limitations in ASTM Specifications C 330
and C 331, but are not included in the insulating concrete aggregate specifications (ASTM Specification C 332).
Organic impurities are tested for by ASTM Test for Organic Impurities
in Sands for Concrete (C 40). Aggregates producing a color darker than
the standard are subjected to rejection, "unless it can be demonstrated
that the discoloration is due to small quantities of materials not harmful to
concrete."
Staining of concrete surfaces may result from use of lightweight aggregates that contain excessive amounts of iron or iron compounds. Although
it is not a serious problem with the commonly used materials, the specifications provide limits and specify a test procedure to evaluate the potential
degree of staining from the aggregate.
The test is conducted in accordance with ASTM Test for Staining Materials in Lightweight Concrete Aggregates (C 641), and consists essentially
of exposing aggregate samples, enclosed in filter paper cups, to steam for
16 h. The degree of staining is evaluated either by visual comparison of the
510
stains on the filter paper with photographic reference standards provided

in the test method or by chemical determination of the amount of ferric
oxide (Fe203) deposited on the filter papers from 200 g of the aggregate
tested. Aggregates having a visual classification of "heavy stain" or darker
are required to be tested by the chemical procedure and rejected if the Fe203
is 1.5 mg or more. The test method and specification limits are based entirely on extensive studies of cinder aggregates by Seaton [7], but appear to
be serving a useful purpose applied to all lightweight aggregates.
Clay lumps are limited to 2 percent by weight of the aggregate in both
ASTM Specifications C 330 and C 331. The test method used is ASTM
Test for Clay Lumps and Friable Particles in Aggregates (C 142). Originally
intended as a limit on the amount of "unburned or underburned lumps in
expanded clay, shale or slate aggregates," it is doubtful that such lumps
are sufficiently friable or soft after 24-h immersion to be broken by finger
pressure as stipulated in the test method. The specification wording was
changed from "unburned or underburned" to "clay" lumps, and the restriction is probably useful only for detecting contamination of the aggregate.
Adequate evaluation of the effects of unburned or underburned lumps and
development of a suitable test procedure for identifying them remains an
unsolved problem. ASTM Recommended Practice for Petrographic Examination of Aggregates for Concrete (C 295) is useful in this regard.
Loss on ignition of lightweight aggregates is limited to S percent for all
materials covered by ASTM Specifications C 330 and C 331, except cinders
which are limited to 35 percent (ASTM Specification C 331). No specific
test procedure is required, nor is any reference made to such tests. Attention
is called to the fact, in notes, that some lightweight aggregates may be
partially hydrated during production or naturally contain innocuous carbonates or water of crystalization that would contribute to the loss on ignition.
Such materials are not considered to be deleterious, and the notes state
that "consideration should be given to the type of material when evaluating
the product in terms of ignition loss." The limitation is based apparently
on unproven fears that residual unburned fuels in sintered or rotary kiln
products might be deleterious.
Concrete-Making Properties
Properties of the lightweight aggregates, other than those described previously, are specified indirectly in terms of their performance in concrete.
These specification requirements are varied, dependent upon the primary
use and desirable characteristics of the concrete. Requirements for drying
shrinkage, popouts, and durability are listed in ASTM Specifications C 330
and C 331, since both structural concrete and masonry unit performance
could be adversely affected by lack of desirable properties in these respects.
In ASTM Specification C 330, the importance of both strength and lightness
511
of weight are recognized by suitable limitations. For insulating concrete,

lightweight and low thermal conductivity are of primary importance; limits
are imposed on these characteristics in ASTM Specification C 332.
Drying shrinkage is limited to 0.10 percent in ASTM Specifications C 330
and C 331 when determined in accordance with procedures outlined in the
specifications. ASTM Test for Length Change of Hardened Cement Mortar
and Concrete (C 157) is followed using fixed proportions of one part cement to six parts aggregate by dry, loose volumes, with sufficient water to
produce a slump of 50 to 76 mm (2 to 3 in.). Initial length measurements
are made after 7 days moist storage with final shrinkage measurements at
the age of 100 days after storage at 50 percent relative humidity.
No data correlating results of this test with actual performance are known
to exist. The concrete proportions used bear no resemblance to actual mixes
of structural lightweight concrete; while the cement/aggregate ratio could
be similar to that of some concrete masonry units, the slump (and the
water/cement (w/c) ratio) are far from any used in this application. A
much more realistic test procedure and limits can be found in the American Association of State Highway and Transportation Officials (AASHTO)
Specification for Lightweight Aggregates for Structural Concrete, M 195 [8].
The AASHTO test uses proportions similar to those proposed for field use
and limits shrinkage to 0.07 percent after 28 days storage at 37.8~ (100~
and 32 percent relative humidity.
Under some conditions, lightweight aggregate concretes have been found
to have drying shrinkage values no greater than those of some natural
aggregate concretes [9]. The need for specifications of this nature should
be examined carefully, and more realistic procedures adopted if they are to
be used for acceptance testing. The present requirement for a 100-day
test (after fabrication of the specimens) is excessively time consuming as
well as unrealistic with respect to the concrete used.
Popouts in lightweight concrete are comparatively rare, but may be
caused occasionally by such materials as unhydrated pieces of hard-burned
lime or magnesia, calcium sulfate, or unstable iron compounds. The popouts, when encountered, occur as a result of hydration or oxidation or
both, rather than as a result of freezing and thawing action which is responsible for most popouts with natural aggregates. To test for such materials, concrete specimens are prepared in the same manner as outlined
above for the shrinkage test. The specimens are cured and tested in accordance with ASTM Test for Autoclave Expansion of Portland Cement
(C 151). No popouts, as determined by visual inspection of the autoclaved
specimens, are permitted under either ASTM Specifications C 330 or C 331.
Durability of both structural concrete and masonry units is important
for those structures and structural elements exposed to weathering. To
date, no specific procedures or specification limits have been developed to
measure this attribute. ASTM Specifications C 330 and C 331 stipulate
512
that "in the absence of a proven record of satisfactory durability" the

lightweight aggregates "may be required to pass a concrete freezing and
thawing test satisfactory to the purchaser." Details of the test procedures
and the evaulation of results are left to the discretion of the purchaser.
The specification, as originally adopted, included limitations on the
absorption of the lightweight concretes, with the permissible values varying
with the unit weight and strength values. Later investigations [10] demonstrated that absorption did not correlate with resistance to freezing and
thawing and, therefore, the requirements were deleted.
Strength and unit weight requirements for structural concretes made
with the lightweight aggregates are included in ASTM Specification C 330.
Both compressive and splitting tensile strengths are specified, with the
minimum values required increasing as the 28-day air dry unit weight increases. The requirements were originally on the basis of all lightweight
aggregate only, but have more recently been modified to include use of
natural sand fine aggregate with lightweight coarse aggregates.
Unit weight is determined by ASTM Test for Unit Weight of Structural
Lightweight Concrete (C 567), using 150 by 300-mm (6 by 12-in.) cylinders,
moist cured for 7 days and stored for 21 days at 23~ (73~ and 50 percent relative humidity. The unit weight thus obtained is an air-dried value
that more closely corresponds to the weight in actual service than would
the plastic unit weights. Air dry weight ranges in ASTM Specification C 330
are 1600, 1680, and 1760 kg/m a (100, 105, and 110 lb/ft 3) for concrete
made with all lightweight aggregate; 1680, 1760, and 1840 kg/m a (105, 110,
and 115 lb/ft 3) for sand/lightweight aggregate concretes.
Compressive strengths (28 days) required range from 17 to 28 MPa (2500
to 4000 psi) for the lightest to heaviest weight range in both all lightweight
and sand/lightweight mixtures. Specimens are made in accordance with
ASTM Making and Curing Concrete Test Specimens in the Laboratory
(C 192). Curing may be either in accordance with ASTM Method C 192 or
the drying procedure used in the unit weight test may be used. Investigations
have shown that these differences in curing have no significant effect on
the 28-day compressive strength of structural lightweight concretes [11].
Required 28-day splitting tensile strength values range from 2.0 to 2.2
MPa (290 to 319 psi) for all lightweight aggregate mixes and from 2.1 to
2.3 MPa (305 to 334 psi) for the sand/lightweight mixes. The test is run in
accordance with ASTM Test for Splitting Tensile Strength of Cylindrical
Concrete Specimens (C 496). The method requires the testing of air-dry
specimens which has been shown by extensive testing to produce lower
strengths than continuous moist curing [12,13], differing from compressive
strength in this respect. This research has also shown higher splitting tensile
strengths for the mixes containing natural sand fine aggregates, leading to
the higher requirements in ASTM Specification C 330 than for the all
lightweight mixes. However, the specifications do note that materials that
513
do not meet the splitting tensile strength requirement may still be used if
the structural design is suitably modified.
The lightweight aggregates are not required to meet each of the strength
and unit weight levels when used in concrete. The specification stipulates
that it "shall be possible to produce structural concrete" that will satisfy
the requirements of "one or more" of these categories.
Thermal conductivity values specified in ASTM Specification C 332 for
insulating concretes are limited to a maximum of 0.22 W / m . K (1.50
Btu 9in./h 9ft2~ for concrete having an oven-dry unit weight of 800 kg/m a
(S0 lb/ft a) or less, and to 0.43 W / m . K (3.0 Btu in./h-ft2OF) for those
weighing up to 1440 kg/m 3 (90 lb/ft3). The lighter concretes are those
made with perlites and vermiculites; the heavier weights are those usually
resulting from use of expanded shales, slags, and natural lightweight aggregates.
Thermal conductivity values are determined in accordance with ASTM
Test for Thermal Conductivity of Materials by Means of the Guarded Hot
Plate (C 177). Oven-dried specimens are used for both thermal conductivity
and unit weight tests on the insulating concretes. Curing is for 7 days in
the moist room followed by 21 days at 50 percent relative humidity with
subsequent oven drying. Unit weight is determined by weighing and measuring the dimensions of specimens having a volume of not less than 2360 cm 3
(144 in.3).
Moisture content of insulating materials affects both the thermal conductivity and unit weight, and the use of oven-dried specimens does not
duplicate conditions of actual use. The test conditions serve principally to
permit general classification and comparison of materials and to provide a
standardized reference condition. Under conditions prevailing in actual use,
some increases in both weight and thermal conductivity should be expected
from residual moisture in the job concretes.
Miscellaneous Test and Properties

In addition to the tests and properties included in the lightweight aggregate specifications and described previously several others are of interest
either because they are utilized in mix design and control or in specifications
for normal weight aggregates. The report of ACI Committee 213, "Guide
for Structural Lightweight Aggregate Concrete," includes discussion of
many such properties and comparisons with those of normal weight concretes.
Strength of the aggregate particles undoubtedly influences the properties
of concrete made with them but is not usually included in specifications.
The lightweight materials used for structural concrete and masonry units
have demonstrated through experience that the degradation or breakdown
during mixing and placing is not excessive and need not be a matter of
514
concern with aggregates meeting the specification requirements for other

properties. Attempts to correlate abrasion or crushing strength tests [6] of
the aggregates with properties of the resulting concretes have been successful only in a very general manner. For a given type of lightweight aggregate (produced from similar raw materials) the strength of the aggregate
varies roughly with the unit weight. In general, the weaker lightweight
aggregates are not capable of making the strength requirements for concrete
included in ASTM Specification C 330. Perlites and vermiculites are much
weaker structurally than the expanded or sintered aggregates but are employed only in the very light insulating concretes where high strength is not
a requirement.
Although the strength of lightweight aggregates covers a wide range and
undoubtedly overlaps the strength range for natural aggregates, the average
is probably somewhat less than that for natural aggregates. Therefore,
breakdown and degradation in handling is likely to be somewhat greater
for the lightweights.
Chemical reactions of lightweight aggregates have not been a problem,
and the specifications do not require that tests for chemical reactivity be
applied to them. Mielenz [14] reports potentialities for alkali-aggregate
reaction with some natural lightweight aggregates and expanded perlites
but notes that the "volume change will not necessarily cause structural distress if appropriately accommodated in the design." The inherent characteristics of the aggregates and of the concretes made with them may also
serve to reduce effects of such reaction--high porosity, low modulus of
elasticity, and high creep in low-strength insulating concretes might preyent structural damage in situations where normal weight concretes would
be in distress. The fines in expanded shales, clays, and slates may react as
natural pozzolans and could act to inhibit alkali aggregate reactions. Glassy
phases of blast furnace slag are not deleteriously reactive [14] and in various
investigations [15,16] have been shown to be effective inhibitors of such
reactions.
Absorption and specific gravity test results for lightweight aggregates are
quite different from those for normal weight aggregates. Bulk specific
gravities of lightweight aggregates are much lower than those of normal
weight aggregates, while absorptions are higher. Ranges of both bulk
specific gravity and absorption for various types of lightweight aggregates
have been summarized by Washa [6]. However, the highly irregular and
vesicular surfaces of most lightweight aggregates makes surface drying
difficult and uncertain, with resulting questionable validity of the absorption
and specific gravity values. In addition to the methods provided in ASTM
Test for Specific Gravity and Absorption of Coarse Aggregate (C 127) and
Test for Specific Gravity and Absorption of Fine Aggregate (C 128), special
techniques have been used by the U.S. Bureau of Reclamation [6], Saxer [17]
and the Portland Cement Association [11] for determining the absorption
515
and specific gravity of lightweight aggregates. Other test methods have

been proposed by Helms and Bowman [18] and Landgren [19]. However,
no method is widely used and such tests are not usually conducted on
lightweight aggregates.
While the natural aggregates do not usually exhibit any significant
changes in specific gravity with varying size of particle, it has been shown
[20] that specific gravity of lightweight aggregates increases with decreasing
particle size. Therefore, changes in gradation would be expected to change
the specific gravity and absorption values. Higher specific gravities for the
small sizes may also increase tendencies toward segregation during transportation and handling. Keeping the aggregate damp and handling in
separated sizes are often recommended to minimize these tendencies.
Surface moisture tests, such as ASTM Test for Surface Moisture in Fine
Aggregate (C 70), are of little value with lightweight aggregates, since
accurate information on bulk saturated-surface dry specific gravity is seldom
available. Instead, the total moisture content is used in mix control and
can be determined by procedures outlined in ASTM Test for Total Moisture
Content of Aggregates by Drying (C 566).
Plastic Lightweight Concrete
Mix Proportioning
The principles of mix proportioning usually applied to normal weight
concretes are generally difficult to use with lightweight aggregates. The
lack of accurate values for absorption, specific gravity, and free moisture
means that net w/c ratios cannot be established with sufficient reliability
for use as a basis of mix proportioning. Variations in specific gravity with
size of particle lead to the need for higher percentages of fine material
than are usually considered desirable for normal weight concretes. Optimum
air contents are usually somewhat higher than for normal weight concretes
to avoid workability problems from the angular, vesicular-surfaced particles,
to provide improved resistance to weathering, or to decrease unit weights
in insulating and fill concrete uses.
Lightweight concrete mix designs are established usually by trial mixes,
proportioned on the basis of cement content at the required consistency.
ACI Committee 211 report, "Recommended Practice for Selecting Proportions for Structural Lightweight Concrete," describes such procedures in
detail. The proportions of fine and coarse aggregates and the cement and
water requirements are estimated or based on previous experience with the
particular aggregate. The dry loose unit weights and moisture contents of
the aggregates are determined and the first mix made on the basis of the
estimated proportions and sufficient water to obtain the desired slump.
Unit weight and air content of the mix are determined. Since weights of all
516
materials in the mix and the specific gravities of all except the lightweight
aggregates are known, a "specific gravity factor" for the aggregate can be
calculated. This is not a true specific gravity value, since the proportions of
free and absorbed water are unknown, but it provides a useful value for
adjusting mix proportions for additional trial batches. A pycnometer method
for determining a specific gravity factor developed and reported by Land~
gren et al [21] may be used. Final mix design is on the basis of dry, loose
volume of aggregate per cubic metre (yard), total water necessary to provide the desired slump, and the cement content necessary to obtain the
required strength.
Mixing Procedures and Control

Mixing procedures for lightweight concretes may vary with the different
types of aggregates, and the producers' recommendations should be followed. A common practice for structural and masonry unit concretes is to
mix the aggregates and about two thirds of the required mixing water for
periods up to 1 min prior to the addition of the cement, air-entraining
agent, and the balance of the mixing water. Mixing is then continued as
required for homogeneity, usually two or more minutes for structural mixes
and four or more minutes for masonry unit mixes. For some insulating
concretes, the aggregate is added last to minimize degradation.
Control of lightweight mixes is obtained by keeping the cement content,
slump, and volume of dry aggregate per cubic metre (yard) constant. Unit
weight, slump, and air content tests of the fresh concrete will determine if
changes in aggregate weight have occurred and if the proper cement content
is being used. If a change in aggregate weight has occurred, checks of
moisture content, gradation, and unit weight of the aggregate can be
made to determine causes of the change.
The slump should not only be maintained as consistent as possible but
also kept as low as is practical for the placing conditions. Overwet lightweight mixes tend to segregate, and the lightness of the larger particles
causes them to float to the surface where they may interfere with finishing
operations. Overmanipulation, excessive vibration, and so on should be
avoided to minimize such tendencies.
Plastic Concrete Tests

Tests of plastic or freshly mixed concretes conducted by standard procedures are confined principally to the structural lightweight concrete mixes.
Workability is determined by means of ASTM Test for Slump of Portland Cement Concrete (C 143). The test results are not directly comparable
to those obtained with normal weight aggregates; that is, a 75-mm (3-in.)
slump does not indicate the same workability or placeability for a lightCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:43:38 EDT 2014
517
weight concrete as it would for a normal weight mix. For the same workability, a lightweight mix would have a lower slump, due to the lesser weight
of concrete causing distortion of the mass upon removal of the slump cone.
Unit weight of freshly mixed structural lightweight concrete is important
in determining yield and controlling mix proportions. ASTM Test for Unit
Weight, Yield, and Air Content (Gravimetric) of Concrete (C 138) is used.
With lightweight concretes, tapping the sides of the measure to remove air
bubbles is particularly important, since holes left by rodding do not fill as
readily as in the case of normal weight mixes. The calculations shown in
the method for gravimetric air content are not applicable to lightweight
mixes, because the total absolute volume of the mix ingredients cannot be
determined accurately.
Air content of lightweight concretes is determined by ASTM Test for Air
Content of Freshly Mixed Concrete by the Volumetric Method (C 173). This
method is applicable to any type of aggregate, whereas pressure methods
are accurate only for denser aggregates where proper correction for air
within the aggregate particles can be made readily. Porosity characteristics
and difficulties in specific gravity determinations of lightweight aggregate
make both the pressure and gravimetric methods very inaccurate when used
on lightweight concretes. ASTM Test C 173 volumetric method gives the
most reliable results for air content of lightweight concretes and, therefore,
is the recommended procedure.
Hardened Lightweight Concrete

Earlier published information on the characteristics and properties of
lightweight concretes (and lightweight aggregates) was summarized by
Washa [6], Kluge [22], Carlson [23], and Valore [24], the last three of
these covering lightweight structural concrete, masonry units, and insulating
concretes, respectively. In many cases, tests for properties of lightweight
concretes are the same as those for concretes of normal weight, while in
others special tests particularly suited to lightweight concrete characteristics
are used. Structural lightweight concretes are covered in detail in the ACI
Committee 213 report, "Guide for Structural Lightweight Concrete."
Unit Weight
Except in special design circumstances, the unit weight of hardened
normal weight concrete is seldom of interest. However, lightweight concretes are often economical in use principally because of the weight savings
involved. Specifications for lightweight structural concrete usually stipulate
a maximum unit weight in an air-dry condition, presumed to be comparable to weight in service. ASTM Test for Unit Weight of Structural
Lightweight Concrete (C 567), provides a procedure for determining this
518
value. The usual range of air-dry unit weight values for structural lightweight concrete is from 1360 to 1840 kg/m 3 (85 to 115 lb/ft3). In some
cases, densities above these values may be economical for use.
Unit weights of insulating concretes are usually determined in an ovendry condition, using the oven-dry weight and dimensions of companion
specimens to those subjected to either thermal conductivity or compressive
strength tests. Insulating concretes made with perlite or vermiculite aggregates or cellular concretes may range from 240 to 800 k g / m 3 (15 to 50 lb/ft 3),
while those made with other types of lightweight aggregates are usually in
the range of 800 to 1440 kg/m 3 (50 to 90 lb/ft3).
Unit weight of concrete in masonry units is an oven-dry basis for all
aggregates. ASTM Sampling and Testing Concrete Masonry Units is used,
with lightweight units having concrete weights usually in the range of 1200
to 1760 kg/m a (75 to 110 lb/ft3).
Strength Properties
The strength of lightweight concretes varies with unit weight, aggregate
characteristics, cement content, and all other factors that affect strength of
any concrete. It is quite low for insulating and fill concretes while structural
concretes compare quite favorably with excellent quality conventional concretes.
Compressive strength of both lightweight structural and masonry unit
concretes are determined by the same procedures used for testing normal
weight concretes. Although tests in an air-dry condition are stipulated in
ASTM Specification C 330 for evaluation of unit weight and splitting tensile
strength and may be used for compressive strength tests to evaluate lightweight aggregates, these procedures are not intended for use in job control
or acceptance testing of lightweight concretes. Standard procedures for
making, curing, and testing normal weight concretes are used for these
purposes. Specifications usually require the lightweight concretes to have
the same compressive strengths as would be required for normal weight
concretes in structural and masonry unit uses. Although higher cement
contents are required with some lightweight aggregates to obtain these
equivalent strengths, many of them have essentially the same cement factorstrength relationships as do high-quality normal weight aggregates.
In contrast, insulating concretes are not required to produce high compressive strengths and frequently are in the range of 1.4 to 5.6 MPa (200 to
800 psi). ASTM Test for Compressive Strength of Lightweight Insulating
Concrete (C 495) covers the procedures for testing fresh concretes weighing
less than 800 kg/m 3 (50 lb/ft3). Specimens are 75 by 150-mm (3 by 6-in)
cylinders, with curing consisting of 7 days moist, 18 days at 50 percent relative humidity, and 3 days in an oven at 60~ (140~
Moisture content
519
has a large effect on the strength of insulating concrete, and complete saturation would not represent its "in service" condition. The oven-dry test
requirement was selected as a reproducible condition that would provide
reasonably realistic values. Tendencies of some insulating concretes to
crack if dried rapidly led to the use of relatively low-temperature oven drying preceded by a period of air drying. A similar test condition is stipulated
for determining compressive strength of hardened insulating concrete,
using cubes cut from structures in place (ASTM Securing, Preparing, and
Testing Specimens from Hardened Lightweight Insulating Concrete for
Compressive Strength (C 513)).
Flexural strength tests are sometimes conducted on structural lightweight
concretes, using the standard test procedures employed for normal weight
materials. The ratio of flexural strength to compressive strength is usually
somewhat higher for the lightweight concretes than for regular concrete
mixes [25].
Tensile strength of lightweight structural concrete is determined by ASTM
Test C 496. For lightweight concretes, the method provides that the test be
run on air-dried cylinders for evaluation in accordance with ACI Standard
318, Building Code Requirements for Reinforced Concrete. Better correlation of diagonal tensile strengths with air-dry tensile strengths than with
tests of moist cured specimens has been shown [26].
Plastic Properties and Creep

These tests on lightweight concretes are confined usually to the structural
concretes. The standard test procedures, ASTM Test for Static Modulus
of Elasticity and Poisson's Ratio of Concrete in Compression (C 469) and
ASTM Test for Creep of Concrete in Compression (C 512), are used. The
modulus of elasticity is usually lower than that of normal weight concrete
of the same strength. The value varies with both unit weight and strength.
Pauw [27] developed a relationship between modulus of elasticity, unit
weight, and strength that appears applicable to both lightweight and
normal weight concretes. The current ACI Standard 318 uses this relationship to obtain design values for lightweight structural concretes in the
absence of actual test values. The influence of unit weight is apparently the
reason for the lower E values of lightweight concretes.
Some investigations [11] have shown lightweight structural concretes to
have higher creep values than normal weight mixes. However, such comparsions have usually been based on comparative tests using only one normal weight aggregate, frequently of higher than average quality. Although
the average lightweight concrete may have greater creep than the average
normal weight concrete, the variations found in both types result in considerable overlap in the ranges of values.
520
Volume Change
The volume change of lightweight concretes is usually somewhat greater
than that of normal weight concrete having similar cement contents.
However, some of the structural lightweight concretes have drying shrinkage
characteristics that fall in the range found in concretes made with regular
aggregates.
High drying shrinkage alone is not indicative of increased cracking
tendencies, since resistance to cracking is also dependent upon modulus of
elasticity, tensile strength, and creep characteristics. Lightweight concretes
are generally capable of undergoing greater deformation before cracking
than regular concrete. Shrinkage tests may be conducted in accordance
with ASTM Test C 157 although, as noted previously, this test has been
modified in some details for use in determining the concrete making properties of lightweight aggregates (ASTM Specification C 330).
Volume changes due to temperature variations are quite similar to those
of regular concretes. The coefficient of thermal expansion is about the
same for lightweight and normal weight concretes, except in the case of
lightweight masonry units where the value is usually smaller.
Thermal Conductivity
The thermal conductivity of lightweight concrete varies primarily with
unit weight; however, as shown by Foxhall [28], type of aggregate is also a
significant factor. Tests are conducted usually by ASTM Test C 177, as
discussed previously under the specifications for lightweight aggregates for
insulating concrete. Values of thermal conductivity of lightweight concretes
indicate that structural concretes have about one third, masonry units
about one fifth, and insulating concretes about one tenth of the thermal
conductivities of regular concretes.
Fire Resistance
Lightweight concretes offer greater fire resistance than do normal weight
concretes in structural uses, fireproofing, and masonry units [29,30]. Comparison of the results of standard fire tests on lightweight and normal weight
concretes indicates that 50 mm (2 in.) of lightweight concrete is the equivalent of about 75 mm (3 in.) thickness of normal weight concretes. In both
cases, of course, the type of aggregate has a significant effect upon the fire
resistance of the concrete.
Bond with Reinforcing Steel

Bond characteristics of lightweight structural concretes may be evaluated
by means of ASTM Test for Comparing Concretes on the Basis of the Bond
521
Developed with Reinforcing Steel (C 234). Tests reported by Shideler [11]

show bond strengths of lightweight concrete to be essentially the same as
those of normal weight concrete.
Corrosion of Reinforcing Steel

Both natural and manufactured lightweight aggregates are free from any
corrosive effects on reinforcing steel. Cinder concrete used in the past was
questioned in this respect, but excellent records in many places led Christensen [5] to the conclusion that corrosion of the reinforcing steel was associated principally with the use of porous and permeable, low-quality concrete
exposed to moist conditions. Lightweight aggregates of the types now commonly employed in structural concrete have been in use many years without
any evidence of ill effects on the reinforcing steel. Some lightweight concrete
exposed to seawater for long periods and with less than 25 mm (1 in.) of
concrete over the steel [31] has shown no evidence of corrosion. As a result
of corrosion tests and experience, Short [32] concluded that aggregate type
(dense or lightweight) had little, if any, influence on rusting of the reinforcement, and that thickness of cover, quality of the concrete, and exposure
conditions constituted the important factors.
Absorption
The water absorption of lightweight concretes is greater than that of concretes made with normal weight aggregates. For valid comparison of
absorption values they must be expressed as percentages by volume rather
than by weight, due to the wide differences in unit weights of the concretes.
No correlation of absorption with any significant or desirable property of
the concretes has been shown to exist, and no limits are placed on lightweight structural or insulating concretes. However, specifications for concrete masonry units do include such limitations where exposure to weathering is involved. The maximum permissible absorptions, on a volume basis,
are varied with both severity of exposure and with the unit weight of concrete in current specifications. Masonry units made with the lighter weight,
generally more absorptive aggregates, are permitted higher absorption
values than heavier units containing denser aggregates.
Acoustic Properties
Sound absorption properties of building materials are of interest in control of air-borne noise in many structures. Porosity is an important factor
in the sound absorbing properties, and concrete masonry units are employed frequently in such uses. Lightweight aggregates apparently contribute to sound absorbing efficiency in masonry units [23], generally
522
showing higher efficiency than normal weight concrete units. However,

continuous pores through such units tend to provide lower resistance to
sound transmission through walls. There would appear to be little difference between sound transmission through walls of lightweight structural
concrete, or of lightweight masonry units sealed by painting or plastering,
and walls of normal weight concrete.
Nailability and Sawability

Lightweight concretes are more readily nailable and can be sawed more
easily than normal weight concretes, particularly compared to those made
with very hard, natural aggregates. In general, the lower the unit weight,
the better the nailability and sawability. Thus insulating concretes are very
easily sawed with the lightweight structural concretes being somewhat
more difficult to saw or nail.
Durability
Although laboratory studies have shown that some lightweight structural
concretes with inadequate entrained air [10] and lightweight masonry units
made with lean mixes [33] can be destroyed by artificial cycles of freezing
and thawing, properly proportioned mixes show excellent durability. The
field performance of all types of lightweight concrete appears to have been
generally excellent from the standpoint of resistance to the effects of
weathering. The literature seems devoid of any authenticated field cases
where lack of durability has been caused by use of lightweight aggregate
concrete.
It seems that the possible use of lightweight aggregates containing excessive amounts of unburned or underburned particles of clay or shale is the
greatest potential danger to durability in lightweight structural concretes.
Such particles could retain sufficient clay-like characteristics to cause high
drying shrinkage and detrimental volume changes from alternate wetting
and drying in structural concrete. The drying shrinkage test included
in the lightweight aggregate specifications, ASTM Specification C 330,
should preclude the use of such aggregates. However, this phase of lightweight concrete durability is deserving of additional research and investigation.
References
[1] Davis, R. E. and Kelly, J. W., "Lightweight Concrete and Aggregates," Significance of
Tests and Properties of Concrete and Concrete Aggregates, A S T M STP 169, American
Society for Testing and Materials, 1955, pp. 238-249.
[2] Miller, R. W., "Expanded Blast Furnace Slag for Use as a Lightweight Concrete
Aggregate," General Meeting, American Iron and Steel Institute, New York, N. Yo,
27-28 May 1953.
523
[3] Lewis, D. W., "Lightweight Concrete Made with Expanded Blast Furnace Slag,"
Journal American Concrete Institute, Nov. 1958; Proceedings, Vol. 55, pp. 619-633.
[4] Emery, J. J., Cotsworth, R. P,, and Hooton, R, D., "Pelletized Blast Furnace Slag,"
Proceedings, Seminar on Energy and Resource Conservation in the Cement and Concrete
Industry, Department of Energy, Mines and Resources of Canada, 8-9 Nov. 1976,
Paper 4.1.
[5] Christensen, Einar, "Cinders as Concrete Aggregate," Journal American Concrete
Institute Feb. 1931; Proceedings Vol. 27, pp. 583-646.
[6J Washa, G. W., "Properties of Lightweight Aggregates and Lightweight Concretes,"
Journal, American Concrete Institute, Oct. 1956; Proceedings, Vol. 53, pp. 375-382.
[7] Seaton, S. G., "Study of Causes and Prevention of Staining and Popouts in Cinder
Concrete," Journal American Concrete Institute, Jan. 1948; Proceedings, Vol. 44,
pp. 361-380.
[8] American Association of State Highway and Transportation Officials, "Standard
Specifications for Transportation Materials," Sept. 1974, pp. 434-437.
[9] Reichard, T W., "Creep and Drying Shrinkage of Lightweight and Normal-Weight
Concretes," National Bureau of Standards Monograph 74, U. S. Department of Commerce, 4 March 1964.
[10] Klieger, Paul and Hanson, J. A., "Freezing and Thawing Tests of Lightweight Aggregate
Concrete," Journal, American Concrete Institute, Jan. 1961; Proceedings, Vol. 57,
pp. 779-796.
[ll] Shideler, J. J., "Lightweight-Aggregate Concrete for Structural Use," Journal, American
Concrete Institute, Oct. 1957; Proceedings, Vol. 54, pp. 299-328.
[12] Pfeifer, D. W., "Sand Replacement in Structural Lightweight Concrete-Splitting
Tensile Strength," Journal, American Concrete Institute, July 1967; Proceedings, Vol.
64, No. 7, pp, 384-392.
[13] Pfeifer, D. W., "Fly Ash Aggregate Lightweight Concrete," Research and Development
Bulletin, RD003.01T, Portland Cement Association, 1969.
[14] Mielenz, R. C., "Petrographic Examination," Significance of Tests and Properties of
Concrete and Concrete Aggregates, ASTM STP 169. American Society for Testing and
Materials, 1955, pp. 253-273.
[15] Barona de la O, Federico, "Alkali-Aggregate Expansion Corrected with Portland-Slag
Cement," Journal American Concrete Institute, March 1951; Proceedings, Vol. 47,
pp. 545-552.
[16] Pepper, Leonard and Mather, Bryant, "Effectiveness of Mineral Admixtures in preventing Excessive Expansion of Concrete Due to Alkali-Aggregate Reaction," Proceedings, American Society for Testing and Materials, Vol. 59, 1959, pp. 1178-1202.
[17] Saxer, F. L., "A Direct Method of Determining Absorption and Specific Gravity of
Aggregates," Rock Products, Vol. 59, No. 5, May 1956, pp. 78-79, 87.
[18] Helms, S. B., and Bowman, A. L., "Extension of Testing Techniques for Determining
Absorption of Fine Lightweight Aggregate," Proceedings, American Society for Testing
[19] Landgren, Robert, "Determining the Water Absorption of Coarse Lightweight Aggregates for Concrete," Proceedings, American Society for Testing and Materials, Vol. 64,
1964, pp. 846-865.
[20] Sweet, H. S., "Variations in Density of Lightweight Concrete Aggregates," Bulletin 184,
Sept. 1952, pp. 44-47.
[21] Landgren, R., Hanson, J. A., and Pfeifer, D. W., "An Improved Procedure for Proportioning Mixes of Structural Lightweight Concrete," Journal, Portland Cement
Association, Research and Development Laboratories, Vol. 7, No. 1, May 1965, pp.
47-65.
[22] Kluge, R. W., "Structural Lightweight-Aggregate Concrete," Journal, American
Concrete Institute, Oct. 1956; Proceedings, Vol. 53, pp. 383-402.
[23] Carlson, C. C,, "Lightweight Aggregates for Concrete Masonry Units," Journal, American Concrete Institute, Nov. 1956; Proceedings, Vol. 53, pp. 491-508.
[24] Valore, R. C., Jr., "Insulating Concretes," Journal. American Concrete Institute, Nov.
1956; Proceedings, Vol. 53, pp. 509-532.
[25] Price, W. H., and Cordon, W. A., "Tests of Lightweight-Aggregate Concrete Designed
524
for Monolithic Construction," Journal American Concrete Institute, April 1949;

[26] Hanson, J. A., "Tensile Strength and Diagonal Tension Resistance of Structural Lightweight Concrete," Journal. American Concrete Institute, July 1961; Proceedings. Vol.
58, pp. 1-37.
[27] Pauw, Adrian, "Static Modulus of Elasticity of Concrete as Affected by Density,"
Journal American Concrete Institute, Dec. 1960; Proceedings, Vol. 57, pp. 679-687.
[28] Foxhall, W. B., "Heat Transmission of Lightweight Aggregate Concretes," Heating
and Ventilating, Vol. 47, No. 5, May 1950, pp. 84-85.
[29] Benjamin, I. A., "Fire Resistance of Reinforced Concrete," Symposium on Fire Resistance of Concrete, Publication SP-5, American Concrete Institute, 1962, pp. 25-39.
[30] Copeland, R. E., "C/M and Fire Resistance," Concrete Products. Vol. 67, No. 11,
Nov. 1964, pp. 38-42.
[31] Willson, C., "Concrete Ship Resists Sea Water for 34 Years," Concrete. Vol. 62, No. 1,
Jan. 1954, p. 5.
[32] Short, A., "The Use of Lightweight Concrete tbr Reinforced Concrete Construction,"
The Reinforced Concrete Review, Reinforced Concrete Association, London, Vol. 5,
No. 3, Sept. 1959, pp. 141-175.
[33] Wendt, K. F. and Woodworth, P. M., "Tests on Concrete Masonry Units Using
Tamping and Vibration Molding Methods," Journal American Concrete Institute, Nov.
1939; Proceeding, Vol. 36, pp. 121-163.
A. C. Carter ~
Chapter 31--Packaged, Dry,

Combined Materials for
Mortar and Concrete
The Products
Packaged, dry, combined materials for mortar and concrete are combinations of cementing materials and aggregates which require only mixing with
proper quantities of water to be ready for use. The packages, varying in mass
from 5 to 41 kg (10 to 90 lb) generally are available at the producing plant,
building supply, and hardware stores, as well as some supermarkets. The
general trend is to mix the various ingredients thoroughly before packaging
in moisture-resistant multiwall paper bags. Some products introduce the ingredients into the package separately without prior mixture. Still others
package the cement separate from the aggregates. Since the original publication of ASTM Specification for Packaged, Dry, Combined Materials for
Mortar and Concrete (C 387) in 1956 and during subsequent revisions of
that specification covering these materials which will be discussed later in this
paper, the buying public has become increasingly specification conscious.
Use of the Products
The original developer of the packaged materials for concrete and mortar
visualized the principal user as the "do-it-yourself' homeowner. While this
represents an important market, the scope of use has expanded far beyond it.
Industry makes use of the products for repairs and maintenance requiring
relatively small quantities of materials. Highway departments and railroads
use them for patching, setting signs, signals, and so on. Public utilities are
regular customers, using the concrete or mortar for a variety of purposes including patching of sidewalks and pavements after installation of pipe or
conduit, anchoring poles carrying telephone cables and transmission lines,
1Manager of quality assurance, Texas Industries, Inc., Arlington, Tex. 76010.
Universidade
do Gois pursuant
Agreement.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
526
and for other work too numerous to list. In other cases, concrete in quantities
of several cubic metres up to several hundred cubic metres has been used in
locations inaccessible to ordinary transport.
Quality Control
A most important step in the direction of standardization and quality control was taken when Subcommittee C09.03.17 on Method of Testing and
Specifications for Packaged, Dry, Combined Materials for Mortars and Concrete of ASTM Committee C-9 on Concrete and Concrete Aggregates
developed ASTM Specification C 387. This was published as tentative in
1956, adopted as standard in 1960, and further revised in subsequent years.
This specification now requires that cements, admixtures, and aggregates,
and also test methods conform with applicable ASTM standards. They
cover: (1) two classes of lightweight concrete, (2) two classes of normalweight concrete, (3) a high-strength concrete-sand mortar, and (4) four
classes of mortar for unit masonry. Minimum compressive strength is
specified for each product. For masonry mortars, minimum water retention
is also specified (see Table 1).
TABLE 1--Physical requirements.
Compressive Strength, min, psi (MPa)
Water RetenKind of Material
Concrete
high-early strength
normal strength
normal weight
lightweight using
normal weight sand a
lightweight
High-strength mortar
Mortar for unit masonry
Type M
Type S
Type N
tion, min, %
3 days
7 days
28 days
2500(17.2)
3500(24.1)
...
2500(17.2)
3500(24.1)
2500(17.2)
2500(17.2)
3000(20.7)
3500(24.1)
3500(24.1)
5000(34.5)
...
...
...
2500(17.2)
1800(12.4)
750(5.2)
70
70
70
aLightweight concrete using normal weight sand may contain some portion of lightweight
fines.
The specification also covers container construction, and requires that all
packages be identified as conforming to ASTM Specification C 387, and that
there be printed on the bag:
(a) kind and type of material used,
(b) net weight in each bag,
(c) yield in cubic metres or square metres per 25.4 m (1 in.) of thickness,
and
(d) amount of water recommended for mixing.
CARTER ON MATERIALS FOR MORTAR AND CONCRETE
527
In 1975 the scope of ASTM Specification C 387 was revised to include

descriptions of intended uses of dry-packaged mortars and concretes as an
aid to the consumer. In addition, minimum strength requirements for
lightweight concretes were increased to equal those for normal-weight concretes in Table 1.
Significance of ASTM Specification C 387

In ASTM Specification C 387, the producer and consumer of packaged
materials for mortar and concrete have a useful tool for bringing about standardization and improvement of products. Frequent revisions of the original
1956 specification indicate the widespread interest in the subject. The principal purpose of such revisions has been to clarify statements in the specification, rather than to make any great changes in its original requirements. Further changes undoubtedly will be made in the future.
Both producer and consumer should scrutinize ASTM Specification C
387 to determine what changes or additions seem to be desirable. ASTM
is so organized that recommendations from any source will be considered
by the sponsoring committee. The responsible producer may be presumed
to comply with the basic requirements of ASTM Specification C 387 and
to use the procedures outlined to study his products and to develop higher
and more uniform quality. The consumer should be encouraged to require
that the product be certified to comply with ASTM Specification C 387.
A product made with the letter and spirit of this specification will be a
good product. There is now little need to accept anything less.
Acknowledgments
The original author of this chapter was Prof. A. W. Brust, Department of
Civil Engineering, Washington University, St. Louis, Mo., now retired, and
longtime member of ASTM Committee C-9. The writer has endeavored to
update the chapter in this revision.
B. A . L a m b e r t o n 1
Chapter 32--Preplaced Aggregate

Concrete
Introduction
The term "preplaced aggregate" (PA) concrete describes a method of placing gap graded concrete. As an alternative to blending the fine and coarse aggregate, cement, and water and placing the mixture in a form, coarse aggregate is preplaced in the form and a structural grout is injected into the
coarse aggregate mass in such a way as to fill void spaces where it hardens to
form dense homogeneous concrete. The process was developed by Wertz to
whom a U.S. patent [1] 2 was issued in 1943.
The drying shrinkage of PA concrete is less than that of conventionally
placed concrete with similar material proportions [2]. This is due apparently
to the point-to-point contact of coarse aggregate particles. Because of this
low shrinkage characteristic, it was first used as a method of repairing concrete and masonry structures, particularly railroad bridge piers and tunnel
linings [3, 4]. The method is now widely accepted for other specialized applications such as placement of mass concrete under water [5], in heavily
reinforced structures [6], and for high density biological shielding [7].
While the two-step placement procedure would increase costs over that of
conventional methods in a great majority of routine concrete work, it has the
advantage of permitting placement of coarse aggregate, representing roughly
60 percent of the total concrete volume, in a timed sequence independent of
mixing and placing cementitious constituents. The coarse aggregate fraction
can be preplaced either very rapidly with bulk handling equipment or slowly
and carefully by hand labor, depending on the application, followed by grout
injection at a time convenient to the overall construction schedule. The
method is particularly applicable to placement of concrete in structures containing a profusion of inserted fixtures [8].
IVice President, Engineering, Intrusion-Prepakt, Inc., Cleveland, Ohio 44114.

Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
LAMBERTON ON PREPLACED AGGREGATE CONCRETE
529
Materials
The slurry containing a mixture of portland cement, sand, admixtures,

and water which is injected into the coarse aggregate mass is referred to as
"grout." Actually a highly fluent structural mortar, it bears no resemblance
to the low strength, generally high water/cement (w/c) ratio grouts which are
pressure injected into soil or rock formations to increase strength or reduce
permeability. Grout used in PA concrete work is a structural material. Injected into coarse aggregate, it produces concrete comparable in strength
and other physical characteristics to that mixed and placed by conventional
methods. The same working stresses used in conventional concrete design are
applicable to PA concrete.
Cementing materials suitable for use in conventionally placed concrete
may be used with equal confidence for placement by the PA concrete
method. It is customary practice to substitute pozzolanic quality fly ash for
part of the portland cement. Substitution in the range of 15 to 35 percent by
weight of portland cement is common. In addition to the beneficial properties normally imparted by any pozzolan to portland cement concrete such as
lower heat generation, reduced permeability, and improved resistance to
chemical attack, fly ash is particularly beneficial in the PA concrete process
in that it improves pumpability and retards initial set. Fly ash should meet
the Class F requirements of the ASTM Specification for Fly Ash and Raw or
Calcined Natural Pozzolan for Use as a Mineral Admixture in Portland Cement Concrete (C 618).
Fly ash is often omitted when concrete is to be used for high density
biological shielding. Since the specific gravity of this material is appreciably
less than that of cement, partial replacement of cement with fly ash causes a
slight reduction in unit weight of the concrete. Fly ash is also less effective
than cement in chemically fixing water and so is undesirable in shielding subject to high neutron flux which is more effectively attenuated by hydrogen
atoms.
Fine aggregate, properly graded, is essential to successful execution of the
work. Oversize material will cause obstruction of the void channels in the
coarse aggregate mass. Excess fines will increase water requirement and so
reduce compressive strength and increase drying shrinkage. Gap graded
fine aggregate may cause excessive bleeding. A typical gradation appears
in Table 2, Grading 1 of ASTM Specification for Aggregates for Radiation-Shielding Concrete (C 637). Although described in this specification in connection with high-density aggregate for biological shielding, the
grading is equally suitable for use with normal-density aggregates. Either
natural or manufactured sand may be used, although the former is usually
preferable from the standpoint of pumpability. As indicated in Table 2,
Grading 2 of ASTM Specification C 637, somewhat finer grading is required
where high-density sand is used or where coarse aggregate grading is finer
than normal.
530
Coarse aggregate grading is far less critical than is fine aggregate grading.
The only absolute requirements are that it be (a) free of surface dust which
would prevent bond of grout to the aggregate particles, (b) sufficiently
saturated that it will not absorb water from the grout and so cause premature
thickening, and (c) of such a grading that grout will flow readily by gravity
alone through the void system. The particles should be of such toughness and
hardness that they do not fracture or degrade during transport and placement into the forms. Extra caution must be exercised during placement to
avoid degrading friable aggregate.
Normal aggregate grading limits for most structural applications are
shown in Table 2, Grading 1 of ASTM Specification C 637. Aggregate
graded within these limits will exhibit a void content in the range of 43 to 48
percent. In the absence of very closely spaced reinforcing or restricted form
configuration, it is common practice to scalp coarse aggregate on a 20-mm
(3/4-in.) wash screen. A trommel type screen is generally more effective than
a deck screen in washing the aggregate particles to the required surface
cleanliness.
There is no practical limit for maximum aggregate size other than that imposed by handling equipment, section thickness, and spacing of reinforcing
bars and embedments. Aggregate for mass concrete work such as bridge
piers may be graded from 25 to 150 mm (1 to 6 in.), for example, while mansized pieces of waste rock may be used for deep mine concreting. Where
minimum grout consumption is desired, consideration should be given to extreme gap grading of the coarse aggregate such as a blend of 20 to 40-mm ( 90
to 11A-in.) aggregate with 200 to 300-m (8 to 12-in.) aggregate. Although
the economies of aggregate manufacture often preclude such a blend, void
contents as low as 32 percent have been reported for such a grading.
A wide variety of high density aggregate, both fine and coarse as described
in ASTM Specification C 637, have been placed by the PA concrete method
which is suited particularly for construction of high-density biological
shielding. The advantage of the method for this type of work is two-fold. The
heavy particles of mineral ore or steel punchings or a blend of these coarse
aggregates is placed in the form with no possibility of segregation such as
may occur when coarse aggregate and grout is combined in a plastic mass
and placed by conventional methods. Second, the method is well suited to
concrete placement around multiple embedments, closely spaced reinforcing, and within the tight confines of complex formwork characteristic of most
shielding installations.
Groutfluidifier is required on all PA concrete work to promote fluidity of
the grout mixture and, most importantly, to cause expansion of the grout
prior to initial set in an amount sufficient to more than offset setting
shrinkage of the grout which would otherwise take place beneath coarse aggregate particles. Grout fluidifier is customarily used as a commercially
preblended material conforming to U.S. Army Corps of Engineers SpecificaCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:31 EDT 2014
531
tion for Grout Fluidifier (CRD-C 566). It contains a water-reducing agent, a

suspending agent, aluminum powder, and a chemical buffer to assure properly timed reaction of the aluminum powder with the alkalies of the cement.
This reaction results in the formation of hydrogen gas to expand the fluid
grout, usually in the range of 5 to 8 percent. This gas-forming reaction
should be so timed that, at 21~ (70~
it is approximately S0 percent complete in 1 h and 95 percent complete in 3 h. At a slow expansion rate,
bleeding tends to increase. If the expansion time is more rapid, grout mixing
and pumping time is shortened unnecessarily.
Since the expansive characteristic of the grout is dependent on the alkali
content of the cement, it may be depressed when the standard grout fluidifier
formulation is used with low alkali cements. Performance of grout fluidifier
in combination with job cement should be determined, therefore, and the
product custom formulated by the manufacturer if necessary when use of
cement is anticipated, the alkali content of which is less than 0.40. The
release of hydrogen bubbles and the expansive reaction have an effect similar
to that of an air-entraining agent with respect to freeze-thaw durability. An
air-entraining agent or air-entrained cement should not be used routinely in
the PA concrete process. However, where freeze-thaw conditions are severe
and a specific minimum air content is required, an air content determination
should be made using grout fluidifier and job cement and an air-entraining
agent incorporated in the grout mix if required.
Grout mix proportions may be selected in accordance with U.S. Army
Corps of Engineers Method of Selecting Proportions for Intrusion Grout
Mixtures (CRD-C 85). For typical structural applications, proportions are
normally in the range of 2:1:3 to 4:1:5 by weight in which the ratio
represents, respectively, the weights of cement, fly ash, and sand. The ratio
of water to cementing materials (cement plus fly ash) is in the range of 0.43 to
0.47. Grout fluidifier is added in an amount equal to 1 percent by weight of
the cementing materials. For mass concrete applications the ratio of
cementing materials to sand may be reduced to as low as 1:2. Further
decrease in this ratio is usually precluded by pumpability limitations.
Specifications
Specifications for PA concrete placement vary widely in detail depending
on job complexity. For routine structural repair of concrete or masonry, performance type specifications normally are employed, requiring only that a
minimum stipulated compressive strength be exhibited by test cylinders and
that good construction practice be followed in accordance with such accepted
publications as ACI Standard 304 [9]. For difficult projects such as large
heavily reinforced structures or biological shielding, specifications may be
entirely of the prescription type, setting forth not only compressive strength
required but aggregate composition and grading, grout insert spacing, grout
532
surface monitoring procedures, placement temperature, and in-place density.

High-density aggregates typically used in PA concrete construction are
described in the ASTM Descriptive Nomenclature of Constituents of Aggregates for Radiation-Shielding Concrete (C 638) and typical specifications
for these aggregates are presented in ASTM Specification 637. All aggregates
described in ASTM Method 638 are suitable for placement by the PA concrete method, although particular care should be exercised with the softer
natural minerals such as geothite and limonite. Ferrophosphorous varies
widely in hardness and some of it may be so soft and friable as to be unsuitable for placement by the PA concrete method. All of these materials are
obviously far more expensive than normal density aggregate and purchase
specifications for these materials should be written with particular care. They
should not only call for specific gravity limitations but also carefully describe
sampling procedure with respect to quantity represented by each sample, frequency of sampling, and the location where tests are to be performed.
Aggregates should always be tested prior to shipment, with conformance
sampling and testing performed at the job site. If the aggregate is relatively
friable and breakage during transit is high, consideration should be given to
reclaiming undersize material and incorporating it in concrete to be placed
by conventional methods.
Where blended aggregates are to be used, such as a combination of iron
ore and steel punchings, the two materials must be inspected separately and
the final blend again inspected for particle size gradation and homogeneity of
the blend. Aggregates may be blended continuously by strear/aing through
metering gates onto a belt ahead of a washing screen which completes the
blending process, or they may be batch blended by a few rotations in a concrete mixer followed by washing and screening. Although the latter method
may appear to be somewhat more accurate, in practice either type of blending operation causes some size degradation of the more friable of the two aggregate fractions, and allowance must be made for a loss of some of this
friable material in the blending operation.
All high-density aggregate should be weighed before placement and
sampled frequently to determine unit density and void content. Based on
these determinations, calculations can be made to determine required grout
density. Where watertight steel forms are used, void content of the entire
coarse aggregate mass may be determined in-place by metering water into
and out of the aggregate-filled forms. Using careful testing and sampling
procedures it is possible to specify in-place density of concrete placed by the
PA concrete method as close as ___1 percent, although a tolerance of _+2 percent is a more common limitation.
Temperatures at the time of grout injection are often of great importance
where minimum thermal shrinkage is an important consideration. The aggregate may be precooled in the form by flooding it with chilled water or
533
covering the coarse aggregate with flaked ice and allowing the ice to melt and
drip down through the aggregate mass. Using this technique, in combination
with precooling the grout ingredients and using ice in the mixing water, permits placement temperatures to be achieved within the range of 5 _ 3 ~ (40
_____5~
Grout surface monitoring is of particular importance in ensuring complete
penetration of all voids within the coarse aggregate mass and complete contact with embedments and penetrations. For routine repair work and for
many structural concrete applications, location of grout surface can be determined with sufficient accuracy by simply observing grout or moisture seepage
through the forms or through insert or inspection holes drilled in the side of
the forms. Slotted sounding wells may also be used, particularly for deep
mass concrete placements, in which a float on the end of a' line is dropped
into the sounding well and the depth to the grout surface physically
measured. Using such methods, location to the grout surface can be determined with an accuracy of about ___300 mm (12 in.). For critical applications, location and frequency of sounding wells may be specified, as well as
location and spacing of grout insert pipes and vent pipes. The Time Domain
Reflectometer provides a more accurate means of determining grout surface
location in a nonflooded aggregate mass, Short voltage pulses are transmitted on electrically calibrated detector wires located within the aggregate mass
and the pulse reflections assembled and displayed on an oscilloscope. Grout
or water around the sensor wires alters the impedance permitting an experienced operator to locate the grout surface with an accuracy of about ___80
mm (_-+-3 in.) [9].
Fluid grout characteristics are always routinely tested during any placement by the PA concrete method. Measurements of grout viscosity or flow,
measuring the time in seconds for discharge of 1725 cm 3 of grout through a
12.7-mm (1A-in.) orifice in accordance with U.S. Army Corps of Engineers
Method of Test for Flow of Grout Mixtures (CRD-C 79) should be made
routinely at about 2-h intervals. This test is analagous to the slump test for
conventionally placed concrete. Typical consistencies range from 21 to 30 s
with the upper end of the range suitable for use with plus 20 mm ( 90 in.)
coarse aggregate and the lower end of the range required for plus 12 mm (189
in.) coarse aggregate. Specifications usually require grout consistency to be
controlled within a range of 2 of a stipulated value.
The test for expansion in accordance with U.S. Army Corps of Engineers
Method of Test for Expansion of Grout Mixtures (CRD-C 81) is performed
less frequently, often no more than once a day. This test requires only observation of total grout expansion 3 h after mixing. The utility and significance
of the test procedure is greatly enhanced if observations are made of both
total grout expansion and accumulated bleed water at four 15-rain intervals
following mixing and at 30-min intervals thereafter until all expansion has
ceased. Although with normal cement and at normal temperature, expansion
534
is virtually complete at 3 h, expansion may continue as long as 4 h at lower

temperatures and with some low alkali cements.
Following the procedures of CRD-C 81, approximately 800 cm 3 of fresh
grout are placed in a 1000-cm 3 graduate. The observer should then note and
report not only the total volume of the specimen, including grout and accumulated bleed water, at the aforementioned intervals, but the total volume
of grout and the total volume of bleed water. Expansion, expressed in percent, is calculated as total volume minus initial volume divided by initial
volume. Bleeding, expressed in percent, is reported as the volume of bleed
water (total volume minus grout volume) divided by initial volume.
Grout bleeding, grout expansion, and grout consistency are all strongly influenced by the effectiveness of mixing equipment and procedures. Before
any adjustment is made to the mix design, adequacy of the mixing procedure
should be evaluated carefully.
Density of hardened concrete is normally determined by casting test
cylinders in accordance with Corps of Engineers Specifications CRD-C 84.
From these test cylinders, 100 by 100 by 150 mm (4 by 4 by 6-in.) prisms are
cut on which final acceptance tests are performed. Density determinations on
the standard 150 by 300-mm (6 by 12-in.) test cylinder will result in an erroneously low figure. This procedure is particularly important in counterweight construction where excessive density may be as objectionable as
deficient density.
Hardened concrete placed by the PA method is evaluated by the same
methods employed with conventionally placed concrete. However, two notes
of caution are in order. In accordance with good practice, cores extracted
from conventionally placed concrete should be of a diameter equal to at least
four times that of the maximum coarse aggregate particle size. This limitation on minimum core diameter is even more important with PA concrete by
reason of the fact that there is a higher percentage of large aggregate particles with a correspondingly greater possibility that these particles will be
torn loose from the grout matrix. Secondly, nondestructive testing procedures such as the Schmidt hammer and the Windsor probe should be used
and interpreted with particular caution. Because of the fact that densely
packed large coarse aggregate particles are in close proximity to the formed
surface, covered only with a thin layer of structural mortar, such tests may
show a substantial variation in strength between what is essentially a test on
an aggregate particle and another test, fractions of an inch away, performed
on the structural mortar. Discrepancies in test results might be expected, for
example, on attempting to determine 28-day strength on concrete placed by
the PA method using very dense aggregate and grout with a relatively high
percentage of fly ash replacement of the cement.
535
Closure
The PA concrete process is an internationally accepted method of concrete
placement appropriate to certain specialized construction problems. Test
methods for evaluating the properties of the fresh grout and coarse aggregate
fractions employed in this method are well established. Several additional
test methods and specifications were prepared by a subcommittee of ASTM
Committee C-9 on Concrete and Concrete Aggregates and will be published
in the Annual Book of A S T M Standards, Part 14.
References
[1] Wertz, L. S., "Process for Filling Cavities," Patent No. 2,313,110, 9 March 1943.
[2] "Investigation of the Suitability of Prepakt Concrete for Mass and Reinforced Concrete
Structures," Technical Memorandum No. 6-330, U.S. Army Corps of Engineers, Oct. 1951,
p. 34.
[3] Davis, R. E., "Prepakt Method of Concrete Repair," Journal, American Concrete Institute; Proceedings Vol. 57, No. 2, Aug. 1960, pp. 155-172.
[4] Keats, B. D., "The Maintenance and Reconstruction of Concrete Tunnel Linings with
Treated Mortar and Special Concrete," Journal, American Concrete Institute; Proceedings
Vol. 43, No. 7, March 1947, pp. 813-826.
[5] Davis, R. E., Jr. and Haltenhoff, C. E., "Mackinac Bridge Pier Construction," Journal,
American Concrete Institute; Proceedings Vol. 28, No. 6, Dec. 1956, pp. 581-595.
[6] Klein, A. M. and Crockett, J. A. J., "Design and Construction of a Fully VibrationControlled Forging Hammer Foundation," Journal American Concrete Institute, Proceedings Vol. 24, No. 5, Jan. 1953, pp. 421-444.
[7] Davis, S., "High-Density Concrete for Shielding Atomic Energy Plants," Journal, American
Concrete Institute; Proceedings Vol. 29, No. 11, May 1958, pp. 965-977.
[8] Tirpak, E. G., "Report on Design and Placement Techniques of Barytes Concrete for Reactor Biological Shielding," Report No. 1739, Oak Ridge National Laboratory, May 1954.
[9] ACI Committee 304, "Preplaced Aggregate Concrete for Structural and Mass Concrete,"
Journal, American Concrete Institute; Proceedings Vol. 66, No. 10, Oct. 1969, pp. 785-797.
Bibliography
Akatsuka, Yuzo and Moriguchi, H., "Strengths of Prepacked Concrete and Reinforced Concrete Beams," Journal, American Concrete Institute; Proceedings, Vol. 64, No. 4, April 1967,
pp. 204-212.
Concrete Manual, U.S. Department of the Interior, 8th Ed., 1975, pp. 446-449.
Davis, H. S., "How to Choose and Place Mixes for High Density Concrete," Nucleonics,
June 1955, pp. 60-65.
Davis, R. E., Jamesen, C. E., and Nulands, W. T., "Restoration of Barker Dam," Journal,
American Concrete Institute; Proceedings, Vol. 44, No. 8, April 1948, pp. 633-667.
Davis, R. E. Jr., Johnson, G. D., and Wendell, G. F., "Kemano Penstock Tunnel Liner
Backfilled with Prepacked Concrete," Journal, American Concrete Institute; Proceedings,
Vol. 52, No. 3, Nov. 1955, pp. 287-308.
Downs, L. V., "Floating Caisson Facilitates Repair of Grand Coulee Spillway Bucket," Civil
Engineering, April 1950, pp. 35-39.
Lamberton, B. A., "Placing Concrete in Deep Mines," Civil Engineering, 1956, pp. 37-39.
Olds, F. C., "QA Put to Work on the PCRV," Power Engineering, March 1970, pp. 45-48.
Steinman, D. B., Miracle Bridge at Mackinac, Wm. B. Eerdmans Publishing Co., 1957,
pp. 93-103.
R. C. Mielenz
Chapter 33mPetrographic
Examination
Introduction
Petrographic examination of concrete aggregate is visual examination

and analysis in terms of both lithology and properties of the individual
particles. The procedure requires use of a hand lens and petrographic and
stereoscopic microscopes; less commonly, X-ray diffraction, differential
thermal analysis, or electron microscopy is used to supplement the examination with optical microscopes. By petrographic examination, the relative
abundance of specific types of rocks and minerals is established; the
physical and chemical attributes of each, such as particle shape, surface
texture, pore characteristics, hardness, and potential alkali reactivity, are
described; coatings are identified and described; and the presence of contaminating substances is determined.
As will be discussed subsequently, petrographic examination contributes in
several ways to the investigation, selection, testing, and quality control of
aggregates. Consequently, the method is progressively being applied more
widely. Since 1936, all aggregates used in concrete construction by the
Bureau of Reclamation, U.S. Department of the Interior, have been
examined petrographically as a part of the basis for their selection [1].2
The method has been applied similarly by the Corps of Engineers since
before 1940 [2]. Coarse aggregates proposed for use in either portlandcement concrete or bituminous concrete are examined by petrographic
methods in laboratories of the Ministry of Transportation and Communications-Ontario, and specifications governing acceptance are based
upon the results [3]. Petrographic examination of aggregates is performed
also by several other agencies of the U.S. government and state highway
departments, and may be obtained through some commercial laboratories.
In 1952, the ASTM accepted a Tentative Recommended Practice for
1Vice president, research and development, Master Builders Division, Martin Marietta,
Cleveland, Ohio 44118.
Copyright9 1978by ASTMInternational
www.astm.org
540
Petrographic Examination of Aggregates for Concrete (C 295), which was

adopted as a standard in 1954. Minor modifications were made in 1964
and 1965. Petrographic examination is cited in the ASTM Specification for
Concrete Aggregates (C 33).
The abundant data obtained and the rapidity with which petrographic
examination can be completed justify more general use of the method in
investigation, selection, manufacture, and use of concrete aggregate.
This paper summarizes the objectives and applications of petrographic
examination of aggregates with reference to gravel, sand, crushed stone,
blast-furnace slag, and the most common types of lightweight aggregate.
Techniques of the examination are treated briefly because instructions have
been published [1-10] and are included in ASTM Recommended Practice
C 295.
Purpose of the Petrographic Examination of Concrete Aggregates
Preliminary Determination of Quality

Preliminary petrographic examination of concrete aggregate is performed either in the field or in the laboratory as an adjunct to geologic
examination, exploration, and sampling. The examination assists the
geologist or materials engineer in determining the extent to which
consideration of an undeveloped deposit is justified. Also, the preliminary
petrographic examination indicates the relative quality of aggregates from
alternate sites. By revealing variations in the material, examination of exposures or core from pilot drill holes establishes the minimum program of
exploration and sampling necessary for acceptance or rejection of the
deposit.
Establishing Properties and Probable Performance

Petrographic examination is primarily a supplement to the acceptance
tests. Probable performance of concrete aggregate is estimated in two
general ways by petrographic examination. First, the examination reveals
the composition and physical and chemical characteristics of the
constituents. From this information, the probable response of the aggregate
to such phenomena as attack by cement alkalies, freezing-thawing,
wetting-drying, and heating-cooling, usually can be estimated. The rapidity
with which the petrographic examination predicts potential alkali reactivity of aggregate is especially valuable because of the long time
commonly required by tests of concrete or mortar.
Second, petrographic examination establishes the fundamental nature of
aggregates so that aggregates from unfamiliar sources can be compared
MIELENZ ON PETROGRAPHIC EXAMINATION
541
with aggregates upon which information is available. This application is

discussed below.
Correlation o f S a m p l e s with Aggregates Previously Tested or Used
Detailed petrographic examination is the only test that permits comparison and correlation of samples with aggregates previously used or tested.
Thus, data and experience previously obtained by use and long-time tests
of similar aggregates can be applied in the selection of materials proposed
for current work, even though the materials come from new sources.
By relating the sample to aggregates previously used in construction,
aggregate indicated to be unsound by standard tests may be found
adequate, or conversely, aggregate indicated to be sound in standard tests
might be found unsatisfactory for the intended use.
For example, gravel in certain deposits near Jackson, Mich. meets usual
specification requirements for soundness, abrasion resistance, and content
of soft particles, yet produces objectionable popouts in pavements after
exposure for two winters (Fig. 1). Petrographic examination of the gravel
Courtesyof MichiganState HighwayDept.

FIG. 1--Popouts produced by claystone, shale, and chert in concrete pavement near
Jackson, Mich.
and of particles producing popouts during service has identified the
unsound rock types. Examination of proposed materials will demonstrate
the presence or absence of these rock types and thus will indicate, in the
light of other data, whether the aggregate should be accepted, rejected, or
subjected to special tests. Petrographic examination of aggregate may be a
part of an investigation of concrete constructions as described in the
542
ASTM Recommended Practice for Examination and Sampling of

Hardened Concrete in Constructions (C 823). For example, the examination
may be applied to determine the source of aggregate used in constructions,
such as whether it was obtained from an approved or disapproved source.
Selecting and Interpreting Other Tests

All properties of aggregates influencing performance of concrete ordinarily
are not evaluated by test before selection of the aggregate to be used in the
work, primarily because of cost and time required. Such factors as thermal
properties and volume change with wetting-drying rarely are determined.
Other properties, such as chemical reactivity or effect of the aggregate on
the freezing-thawing resistance of concrete, usually are not determined.
Consequently, it is worthwhile to apply a test by which the relative
significance of such properties can be determined and the need for supplementary quantitative tests indicated.
Petrographic examination aids interpretation of other tests. For example,
are the particles identified as clay lumps in accordance with ASTM Test
for Clay Lumps and Friable Particles in Aggregates (C 142) indeed clay
lumps or are they merely friable or pulverulent particles? What is the cause
of unexpected failure of concrete specimens in freezing and thawing? Is it
the presence of unsound particles which do not disintegrate in the surface
soundness test, yet expand in freezing and thawing? Is it the result of
alkali-aggregate reaction? Is failure in the soundness or abrasion test the
result of complete breakdown of a small proportion of unsound or soft
particles or partial disintegration of the greater proportion of the aggregate?
Detection of Contamination
Petrographic examination is the best method by which potentiallly
deleterious and extraneous substances can be detected and determined
quantitatively. Inadvertent contamination with natural substances, industrial products, or wastes, such as from overburden or from trucks or
railroad cars not properly cleaned of previous cargo, may decrease the
quality' of aggregate markedly. Contamination introduced by containers
may invalidate samples. Such substances as clay, soil, coal, vegetable
matter, petroleum products, or refractories containing calcium or magnesium
oxides are especially important. Incomplete processing of synthetic
aggregates may contaminate the finished product with raw or partially
fired raw materials or coal (Fig. 2).
Undesirable substances inherent in the material, such as coatings, clay,
plant remains, coal, and soluble salts, are detected easily and can be
determined quantitatively by petrographic methods.
543
Courtesy of Bureau of Reclamation, U.S. Department of the Interior

FIG. 2--Coal in sintered clay before crushing ( X2. 7). Reduced one half for reproduction.
Determining Effects of Processing

Petrographic examination aids in production and processing of
aggregate. The relative merit of alternative processing methods and
equipment can be determined quickly by comparison of the original
material with the processed aggregate. Comparison can be based on
particle shape, content of unsound or chemically reactive constituents,
presence of coatings, or production of rock dust.
The feasibility of beneficiation by removal of unsound or deleterious constituents depends on the properties of the particles and their abundance.
By petrographic examination, the undesirable particles can be identified
and separated. Their properties then can be evaluated and compared with
properties of the remainder of the aggregate. If the undesirable particles
are unusually soft, friable, dense, lightweight, or high in magnetic
susceptibility, separation may be feasible on a commercial scale.
Petrographic examination can be used to control the manufacture of
synthetic aggregate, such as expanded shale or clay, perlite, slag, and
other types. Microscopical examination will reveal the presence of raw or
underburned materials, alkali-reactive phases, and contaminants, such
as coal. X-ray diffraction analysis is the most dependable means to identify
and determine quantitatively the proportion of crystalline phases, such as
periclase (MgO), free lime (CaO), and clays whose crystal structure was
not destroyed by the calcination. Differential thermal analysis can be used
to estimate the effective temperature to which clays, shales, slate, and
similar materials have been fired; the trace of the differential thermal
544
curves for the raw and fired product will coincide above the effective temperature achieved in the firing operation.
Performance of the Petrographic Examination

Samples for Petrographic Examination
Samples of aggregate for petrographic examination should be representative of the source. Recommended procedures for sampling and for
preparation of the sample for analysis are covered by ASTM Methods of
Sampling Aggregates (D 75) and ASTM Recommended Practice (C 295).
Examination in the Field

Petrographic analysis of samples in the field usually is qualitative or only
semiquantitative because lack of facilities makes detailed work difficult.
However, detailed examination in the field may be warranted if aggregate
retained on the 75-ram (3-in.) sieve is to be used in the work, inasmuch as
a representative sample of this size fraction necessarily weighs a hundred
or more kilograms and transportation of the sample to the laboratory is
costly. Chips from cobbles not adequately identified in the field should be
taken to the laboratory for further study.
If the deposit or rock exposure is variable, samples should be selected
from each zone, and detailed notes made at the site should relate each
sample to a particular zone and portion of the deposit or rock ledge. The
relative proportion of unsound, fractured, or chemically deleterious
materials should be estimated from measurements made at exposures.
These notes and the results of the tests on the samples will be the basis for
operation of the deposit inasmuch as it may be desirable to waste or avoid
zones or portions containing inferior, unsuitable, or alkali reactive
materials.
Examination in the Laboratory

Petrographic examination of aggregate in the laboratory may be brief
or detailed. Brief examination indicates the relative merit of materials from
alternate sources and supplies justification for abandonment or continued
investigation of undeveloped deposits. The preliminary examination should
not replace the quantitative analysis included in the program of acceptance
tests.
Samples supplied to the laboratory comprise: (a) granular materials;
such as gravel, sand, crushed stone, slag, or synthetic aggregate; (b) stone
in quarried blocks and irregular pieces; or (c) drilled core. For granular
materials, the examination should be performed on all size fractions inCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:34 EDT 2014
545
cluded in the aggregate. The sample of each size fraction should comprise
a minimum of 300 particles. For natural sand and gravel and crushed
stone, the minimum representative samples are shown in Table 1.
TABLE 1--Minimum representative sample size.
Size Fraction, mm (in.)
90 to 38
38 to 19
19 to 9.5
9.5 to 4.75
4.75 to 2.36
2.36 to 1.18
1.18 to 0.60
0.60 to 0.30
0.30 to 0.15
(3 to 11/2)
(11/2 to 3/4)
(3/4 to 3/8)
(3/8 to 3/16)
(No. 4 to 8)
(No. 8 to 16)
(No. 16 to 30)
(No. 30 to 50)
(No. 50 to 100)
Weight of 300 Particles

26 kg
8.6 kg
1.2 kg
0.34 kg
(57 lb)
(19 lb)
(2.6 Ib)
(0.75 lb)
15 g
2.1 g
0.28 g
0.033 g
0.0066 g
Samples for petrographic examination are obtained by sieving in accordance with the ASTM Test for Sieve or Screen Analysis of Fine and
Coarse Aggregates (C 136) and ASTM Test for Materials Finer Than No.
200 (75-/~m) Sieve in Mineral Aggregates by Washing (C 117). The
fractions are quartered or, for fine aggregate, split repeatedly on an appropriate riffle. Specially calibrated and fabricated cups holding 100 grains in
the various size fractions have proved convenient in the petrographic
laboratory of the Bureau of Reclamation for obtaining the small samples of
fine aggregate for the analysis.
Details of procedure are outlined by K. Mather and B. Mather [2] and
in ASTM Recommended Practice C 295.
During the analysis, helpful clues to identity and physical condition can
be obtained by noting such features as: (1) friability or pulverence in the
fingers; (2) resonance when struck; (3) ease of fracturing; (4) nature of the
fracture surface and fracture fillings; (5) odor on fresh fracture; (6) color
and its variation; (7) internal structure, such as porosity, granularity, or
lamination; (8) reaction to water, such as absorption of droplets on fresh
fracture, evolution of air on immersion, capillary suction against the tongue,
slaking, softening, or swelling; and (9) differential attack by acids or
other media.
Fractions retained on the 600-~m (No. 30) sieve are best identified,
examined, and counted under the stereoscopic microscope. The analysis
can be made conveniently by traversing a representative portion of each
fraction by means of a mechanical stage and microscope assembly like that
employed in the point-count procedure according to ASTM Recommended
Practice for Microscopical Determination of Air-Void Content and
Parameters of the Air-Void System in Hardened Concrete (C 457) (Fig. 3).
546
FIG, 3--Performing the petrographic examination of fine aggregate with stereoscopic

microscope, mechanical stage, and tally counter.
The mineralogic composition of finer fractions usually is determined most

easily and accurately in immersion oils under the petrographic microscope.
However, better continuity in description of physical characteristics of the
particles is obtained if analysis of fractions passing the 600-/zm (No. 30)
sieve and retained on the 150-/~m (No. 100) sieve is performed under the
stereoscopic microscope.
Thin sections occasionally are necessary in examination of natural
aggregate. They usually are employed in the study of quarried stone.
However, preparation of plane surfaces by sawing and lapping, with or
without etching or staining, commonly is preferable because of the larger
area made available for examination or analysis. Such examination should
be supplemented by microscopical examination of grain mounts in
immersion oils. Thin sections or polished surfaces, supplemented by use of
oil immersion mounts of granular material, usually are used in analysis of
blast-furnace slag. X-ray diffraction and differential thermal analysis may
be required to identify or determine quantitatively constituents that are
very finely divided or dispersed through particles of aggregates (Fig. 4).
Ordinarily, analysis is performed on at least three size fractions of coarse
aggregate and five size fractions of the fine aggregate. The results may be

L
547
'..-Sandstone w~th kaolin,te

I Sisga Oom,Colombio,SA
b canto,nine
L
9Altered Diabose
nontion~
--I
X/x~
:i
.j,.."~~e
w l t h Organ,~ matter
k,./]
near Kansas City, Kansas
I
..Limestone with Drganic
..' ChickemaugoDam,Tenn
er
l..Phon]l,te with pyri!e

/" Cripple Creek,Cola
i
i
J
Heating Rate II~

IO0
200
300
400
500
600
700
TEMPERATURE-DE6 C
800
900
IOOO
ilao
Courtesy of the Bureau of Reclamation, U.S. Department of the Interior

FIG. 4--Typical differential thermal analysis records obtained on concrete aggregates.
Kaolinite is indicated by the endotherm (downward shift) at 990 to 1025 ~ Nontronite is revealed by endotherms at 100 to 350~ and 450 to 550~
lllite produces the small endotherms at 100 to 200~ and 500 to 615~
Organic matter produces large exotherms at
430 to 500~ or 440 to 600~ Pyrite develops a marked exotherm at 400 to 485~
used to compute the petrographic composition of the aggregate in any

gradation comprising the analyzed fractions. Occasionally, one analysis
may be performed on a graded aggregate. For such an analysis, the
composition by count usually differs greatly from the composition by
weight.
Numerical results of the analysis may be expressed by weight or count
for particles retained on the 600-#m (No. 30) sieve, but for the finer fractions, the results are based only upon count of grains unless a correction
factor is applied. Consequently, consistent analyses of all fractions of fine
and coarse aggregate can be reported only by count. Analysis by count is
the more appropriate determination technically because the influence of
548
particles of given type upon performance of concrete depends primarily

upon their frequency and distribution in the mass. When applied to
coarse aggregate only, the petrographic analysis is more rapid if the
relative proportion of the several rock types or facies is determined by
weight. However, the accuracy and precision of analysis by weight are
affected adversely by unavoidable loss of rock substance when the particles
must be broken to allow identification of rock and mineral types and description of physical condition. Specifications on lithologic composition
preferably are expressed as percentage by count of particles.
Details of calculating and reporting are summarized in ASTM Recommended Practice C 295.
Observations Included in the Petrographic Examination

In reporting the results of the petrographic examination, the petrographer should supply information on the following subjects as necessary to
evaluation of the aggregate:
Mineralogic and lithologic composition
Particle shape
Surface texture
Internal fracturing
Coatings
Porosity, permeability, and absorption
Volume change, softening and distintegration with wetting or wetting
and drying
Thermal properties
Strength and elasticity
Density
Hardness
Chemical activity
Solubility
Oxidation, hydration, and carbonation
Alkali-silica reactivity
Alkali-carbonate reactivity
Sulfates
Sulfides
Cation exchange reactions
Organic Substances
The significance of these properties is discussed by Rhoades and Mielenz
[7,8], Swenson and Chaly [9], Hansen [11], and Mielenz [10,12]. A simple
test to detect forms of pyrite that are likely to oxidize while enclosed in
concrete was devised by Midgley [13] and is discussed more generally by
Mielenz [12]. Petrographic criteria useful in detecting dolomitic rocks that
are potentially susceptible to a deleterious degree of alkali-carbonate reacCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:34 EDT 2014
MIELENZ ON PETROGRAPHIC EXAMINATION 549
tivity are described by Hadley [14], Gillot [15] and others [16-20]. DolarMantuani [21] describes alkali-silica reactivity of argillites and graywackes.
It is intended that the aforementioned properties be determined qualitatively by observation of the mineralogic composition and texture of the
particles or by simple tests, if appropriate. If other test data are available or if the particles can be compared petrographically with previously
tested materials, the properties can be evaluated semiquantitatively or
quantitatively.
Condition of the Particles

The following classification of properties is useful in cataloging the
physical and chemical condition and particles constituting and aggregate.
Physical condition is defined by three terms: (1) satisfactory, (2) fair, and
(3) poor; chemical stability in concrete is designated by two terms: (1)
innocuous and (2) deleterious.
Satisfactory--Particles are hard to firm and relatively free from
fractures, capillary absorption is very small or absent, and the surface
texture is relatively rough.
FairwParticles exhibit one or two of the following qualities: firm to
friable, moderately fractured, capillary absorption small to moderate,
surface relatively smooth and impermeable, very low compressibility, coefficient of thermal expansion approaching zero or being negative in one or
more directions.
Poor--Particles exhibit one or more of the following qualities: friable to
pulverulent, slake when wetted and dried, highly fractured, capillary
absorption high, marked volume change with wetting and drying, combine
three or more qualities indicated under "fair."
Innocuous--Particles contain no constitutents which will dissolve or
react chemically to a significant extent with constitutents of the
atmosphere, water, or hydrating portland cement while enclosed in
concrete or mortar under ordinary conditions.
Deleterious--Particles contain one or more constituents in significant
proportion which are known to either react chemically under conditions
ordinarily prevailing in portland-cement concrete or mortar in such a
manner as to produce significant volume change, interfere with the normal
course of hydration of portland cement, or supply substances that might
produce harmful effects upon mortar or concrete.
Coatings should be reported and evaluated separately because coatings
usually are confined to portions of a deposit and, for crushed stone, the
nature and abundance of coatings vary with processing methods and
equipment. Similarly, particle shape should be considered apart from
other aspects of physical quality because particle shape commonly is
subject to control or modification by processing of the aggregate. Unless
550
otherwise defined by applicable specifications, particles whose length is

more than five times their width should be designated as "elongated
pieces," and those having a ratio of width to thickness greater than five
should be designated as "flat pieces." The measurement should be made
in accordance with definitions of these terms as they appear in ASTM
Definition of Terms Relating to Concrete and Concrete Aggregates (C 125).
Summarizing the Petrographic Examination

Tables 2 through 5 exemplify a variety of forms in which the petrographic analysis may be reported. All are based upon samples received as a
part of engineering investigations. The tabulations always are accompanied
by appropriate discussion and supplementary description. Table 2 is in the
form generally employed by the Bureau of Reclamation [22]. It departs
somewhat from the format recommended in ASTM Recommended Practice
C 295, but the inclusion of pertinent description facilitates interpretation of
the analysis. The tabulation could be simplified by cataloging the rock
types into major and secondary classifications, such as "granite: fresh,
moderately weathered or fraetured, deeply weathered." The summary of
physical and chemical quality is included in a separate tabulation.
Table 3 is the analysis of the sand produced with the gravel whose
composition is summarized in Table 2. The format accords with ASTM
Recommended Practice C 295. Table 4 is similar except that the designations of quality are not used because they are inappropriate for description of lightweight aggregate.
Table 5 is an analysis of a sample representing a commercial crushed
stone coarse aggregate passing the 38-mm (11A-in.) sieve. The analysis was
obtained because the aggregate apparently delayed or prevented
development of specified strength by the concrete under certain conditions.
In the tabulation, the denotation of "innocuous" and "deleterious" is
restricted to potential deleterious alkali reactivity because the significance
of the sulfides and organic matter in the stone could not be evaluated.
Petrographic Examination of Natural Aggregates

Examination of Natural Aggregates in the Field
Sand and gravel result from weathering and natural abrasion and impacting of rock and the deposition of the resulting particles along streams,
in lakes or marine basins, or by wind or glaciers on the earth's surface.
Consequently, sand and gravel are more or less complex mixtures of different kinds of rocks and minerals. Moreover, deposits of sand and gravel
usually vary vertically by stratification and laterally because of the lenticular nature of zones or strata. The concrete-making qualities of the aggreCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:34 EDT 2014
551
gate are influenced by these changes. Examination in the field also should
reveal the variability of the sand and gravel with reference to unsound or
deleterious particles, clay, and organic matter.
Deposits of sand and gravel commonly are changed by deposition of
mineral matter from ground water or by weathering of particles. Examination in the field should indicate the lateral and vertical extent and the
physical nature of the coatings. Areas of the deposit free from coatings
and zones so heavily coated as to preclude processing should be delineated.
Weathering of gravel and sand after formation of the deposit is common
on terraces and at lower levels of deposits along existing stream channels.
The examination should indicate the extent and distribution of such
weathering.
Close observation of gravel and sand exposed on the surface of the
deposit commonly will reveal unsound particles which slake or fracture
with freezing-thawing or wetting-drying (Fig. 5). Identification of such
Courtesy of the Michigan State HighwayDept.

FIG. S--Cobbles of argillaceous limestone disrupted by freezing-thawing in the deposit,
near Charlesvoix, Mich. Observation of natural disintegration forewarns of possible difficulty in service.
particles will aid evaluation of the petrographic examination performed in

the laboratory. In interpreting the effects of natural freezing and thawing,
the consequences of particle size should be evaluated [23]. Representative
specimens of particles affected by freezing and thawing should be
packaged separately and transmitted to the laboratory with samples of the
gravel and sand. Water-soluble salts in coatings or ground water also may
be revealed by efflorescence at or near the surface of the deposit. Their
presence forewarns of the need for quantitative determination of their concentration in the aggregate.
0.2
.o
Weathered andesite
porphyry
Basalt
2.2
Andesite porphyry
...
0.3
1.2
0.2
0.8
Rhyolite porphyry
6.1
0.9
6.4
Coarse-grained granite
0.5
17.7
Fractured coarse-grained
granite
o.
Deeply weathered granite
12.0
Weathered granite
40.0
3/4 in.
29.5
3/4 to
3/8 in.
11/2 to
Granite
Rock Types
0.6
...
0.1
1.1
...
8.4
...
17.2
48.6
3/8 to
3/16 in.
Size Fraction, percent

by weight a
as above, fractured and

weathered
weathered, fractured, black,
microcrystalline
pink, rounded, includes some

free quartz
mierocrystalline, porphyritic,
white to brown
microcrystalline, porphyritic,
tan, to green
medium- to fine-grained,
rounded to fragmental
fractured, weathered,
fractured, slightly friable,
pink, rounded, includes some
free quartz
Description of Rock Types
TABLE 2--Example of tabulation of a petrographic analysis of gravel.

Plant,
Company, near Denver, Colo.
Sample No.
fair
fair
satisfactory
satisfactory
fair
satisfactory
poor
fair
satisfactory
Physical
Quality
innocuous
innocuous
innocuous
innocuous
innocuous
innocuous
innocuous
innocuous
innocuous
Chemical
Quality
rn
),
fi-)
&-)
n'i
(3
0
Z
(.-)
"O
m
"I~
(D
-q
0.6
Milky quartz
0.2
.~
0.2
Shale
Rhyolite
0.5
...
...
0.5
fine-grained, hard, massive

to schistose
massive, hard, brittle, dense,
smooth
fine-grained, massive, firm
to hard
porous, brown, platy,
quartzose
soft, absorptive, rounded,
gray
cryptocrystalline, porphyritic,
pink to gray
medium- to fine-grained,
hard, massive
hard, banded, fine- to
medium-grained
as above, fractured to
slightly friable
as above, intensely fractured
to friable
hornblende schists, hard,
rounded
as above, fractured
aBased upon analysis of 19.0 lb of 11/2 to 3/4-in., 2.7 Ib of 3/4 to 3/8-in., and 0.80 lb of 3/8 to 3/16-in. aggregate.
1 in. = 25.4 m m , and

1 lb = 0.45 kg.
Con version f a c t o r s - -
0.4
Ferruginous sandstone
. . . . . .
0.1
Quartzose sandstone
1.0
6.2
2.8
Quartzite
2.0
0.2
2.2
Schist
0.4
. . . . . .
0.2
Deeply weathered gneiss
Fractured schist
2.7
7.8
10.3
Weathered gneiss
2.2
15.8
14.3
32.2
Granite gneiss
2.3
0.1
0.4
0.4
Diorite
innocuous
Innocuous
innocuous
Innocuous
fair
poor
satisfactory
fair
satisfactory
poor
fair
satisfactory
fair
deleterious
Innocuous
Innocuous
Innocuous
Innocuous
Innocuous
innocuous
satisfactory
satisfactory
innocuous
satisfactory
03
01
Ga
"l-
"O
N
O
Z
"O
m
:IJ
O
O
:D
m
rm
100.0
100.0
"016
0.7
41.2
3.3
1.0
33.6
28.7
0.9
No.
8-16
100.0
No.
50-100
No.
100200
100.0
100.0
16.9
I00.0
16.8
1.1
48.8
23.8
1.2
. . . . . . .
10.2
813
li12
43.0
18.7
100.0
9.6
51.1
...
28.8
10.5
...
Passing
No.
200
4.8
. . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
No.
30-50
...
"915
59.2
65.0
16.7
11.1
. . . . . . .
...
313
0.8
6.3
21.0
2.0
0.3
No.
16-30
82.0
2.3
4.7
51.6
..
"
"1.2
...
0.2
13.5
8.2
0.3
Sc
pc
17.0
......
""
'3.8
11.4
......
. . . . .
1.0
iii
012
...
1.0
0.6
0.8
0.2
. . . . .
Fc
Physical Q u a l i t y
98.5
2.3
"3.8
4.7
51.6
11.4
0.2
15.1
9.2
ic
1.5
...
i12
...
...
...
...
...
013
...
Dc
100.0
2.3
4.7
51.6
11.4
0.2
1.2
3.8
0.2
15.1
9.2
0.3
Tc
C h e m i c a l Quality
In W h o l e S a m p l e b
a Based on count of 500 particles in each sieve fraction.

b Based on g r a d a t i o n of the s a m p l e received, a n d on the d i s t r i b u t i o n of constituents by size fractions shown at the left above.
c s = satisfactory; F = fair; P = poor; I = chemically innocuous; D = potentially chemically deleterious; T = total of c o n s t i t u e n t in whore
sample.
Total
2.6
0.5
...
......
24.6
2.4
...
Basalt
Sericite schist
Quartz and quartzite
Feldspar
Claystone
Chalcedonic chert
Mica
H o r n b l e n d e , garnet, zircon,
etc .
.
.
.
.
.
.
34.9
34.2
0.8
G r a n i t e a n d g r a n i t e gneiss
Pegmatite
Rhyolite tuff
Constituents
No.
4-8
In Size F r a c t i o n s I n d i c a t e d
A m o u n t , as N u m b e r of Particles, percent
T A B L E 3--Example of tabulation of a petrographic analysis of natural sand.

Plant,
C o m p a n y n e a r Denver, Colo.
S a m p l e No,
---I
m
(D
:D
m
0
0
z
0
:D
m
--4
m
o~
O
--i1
0
-o
m
z
o
-o
--4
m
O~
--4
(D
01
t.n
9.8
2.8
0.6
...
Gray to pink, bricklike particles
Gray to black, friable

particles
100.0
100.0
o.1
0.4
0.2
1.7
100.0
"b:8
0.4
2.0
0.4
7.7
32.3
20.2
36.2
No.
4-8
'oi1
...
...
8.8
22.0
25.0
100.0
44.1
No.
8-16
100.0
"i11
1.4
0.8
9.2
25.6
15.4
46.5
No.
16-30
'21s
2.8
0.3
3.0
12.4
100.0
24.0
SS.0
No.
30-50
I00.0
"119
4.7
100.0
'210
7.0
1.1
4.9
20.6
64.4
Passing
No. 100
. . . . . .
1.6
7.8
21.8
62.2
No.
50-100
100.0
1.4
0.1
0.8
1.0
0.4
7.1
21.3
0.8
23.5
1.2
42.4
In
whole
sample b
Remarks
not vesicular, contain

coal
hard, dense
hard, fine-grained
hard to friable
not vesicular, friable,

many slake in water
not vesicular, fragile, many

slake in water
not vesicular, firm to fragile
vesicular, friable
vesicular, hard
vesicular, friable
vesicular, hard
a Based on count of 500 particles in each size fraction.

b Based on gradation of the sample received, and on the distribution of constituents by size fractions shown at left above.
Total
...
0.2
23.6
4.9
1.8
7.7
29.5
26.1
Granite
Sandstone
Coal
1.7
13.1
1.0
30.2
44.6
3/8 to
3/16 in.
Black to gray, vesicular particles

Black to gray, vesicular particles
Red to tan, vesicular
particles
Red to tan, vesicular
particles
Red to brown, bricklike particles
Gray to pink, bricklike particles
Constituent
3/4 to
3/8 in.
In Size Fractions Indicated a
Amount, as Number of Particles, percent
TABLE 4--Example of tabulation of a petrographic analysis of expanded clay aggregate.

Plant,
Company
Sample No.
ol
ida
O!
x
/:
ffl
-o
a-
m
:13
O
fi3
20
z
N
O
z
m
/-rtl
same as above except c o n t a i n i n g one or

more seams of iron sulfides a n d organic
matter
white to gray, coarse- to m e d i u m - g r a i n e d
l a m i n a t e d , fine-grained, iron sulfides
a n d organic m a t t e r a b u n d a n t
white to gray, particles of zones of
chalcedony evident
conchoidal fracture, dense
includes also g r a i n s of q u a r t z a n d feldspar
gray to buff, contains sparse organic

m a t t e r with pyrite a n d m a r c a s i t e
g r a y to buff, soft to friable, slightly
porous, sparse organic m a t t e r with
pyrite a n d m a r c a s i t e
Description of the Rock Type or Facies
59.7
"0.1
1.0
2.4
...
...
56.2
Sc
40.3
0.1
...
...
...
2.1
4.2
33.9
...
Fe
...
pc
98.9
0.1
2.4
2.1
100.0
0.1
0.1
0.1
...
1.1
1.0
1.0
4.2
33.9
2.4
2.1
...
33.9
56.2
Tc
..
..
...
56.2
4.2
Dc
Ic
C h e m i c a l Quality
a S a m p l e g r a d e d in accordance with specifications of the

State Highway Dept. for 11/2 in. to No. 4 aggregate for concrete
highway pavement.
b Based u p o n analysis of 25.9 lb of a g g r e g a t e split from the sample.
c s = satisfactory; F = fair; P = poor; I :-- not deleteriously a l k a l i reactive; D = potentially deleteriously alkali reactive; T :- total of constituent in the sample.
Total
Chalcedonic chert
Sandstone
Chalcedonic limestone
Limestone
L a m i n a t e d limestone
Soft, organic d o l o m i t i c limestone
Soft dolomitic limestone
Dense dolomitic limestone
Rock Type or Facies
Physical Quality
A m o u n t , percent by weight b
T A B L E 5--Example o f tabulation o f a petrographic analysis o f crushed stone coarse aggregate, a

Plant,
Company
S a m p l e No.
m
t..D
"11
ITI
m
-4
ITI
0
z
0
O
"11
:IJ
O
"u
m
:lJ
-t
-u
-I
-I
m
O1
t~
O~
557
Examination of Natural Aggregates in the Laboratories

Samples and data from the field should be examined to determine: (1)
the abundance of individual lithologic or mineralogic types; (2) the abundance of particles in various physical conditions and degrees of chemical
reactivity; (3) the composition, frequency, abundance, and physical nature
of coatings; (4) particle shape; and (S) the possible qualitative contribution
of particles of the several types of properties of concrete (Tables 1 and 2).
Natural aggregate may contain more than 20 rock and mineral types. Consequently, petrographic examination commonly is time consuming. Based
upon the similarity in composition and probable performance in concrete,
two or more rock types commonly can be combined into a single category
with considerable saving in time and without loss in validity of the analysis.
For example, granites, quartz monzonites, and granite gneisses--or
rhyolites, dacites, and latites--of similar composition and physical
condition might be combined, thus eliminating the need for tedious
examination sufficient to effect a separation.
Soft and altered particles may be original constituents of sand and
gravel, or may be developed by weathering in the deposit. Weathering in
the deposit is especially significant because the alteration affects most or
all particles, causing softening and absorptivity in the superficial portion of
the particles. This action decreases both the bond with cement and the
strength and durability of the concrete.
Coatings on gravel and sand vary from minute spots and films to a
cement which produces zones of sandstone or conglomerate in the deposit.
Coatings usually are composed of silt, clay, and calcium carbonate; but
organic matter, iron oxides, opal, manganiferous substances, alkali and
alkali-earth sulfates, and soluble phosphates have been identified [8,24].
The petrographic examination should reveal the composition, abundance,
physical properties, probable potential chemical reactivity, and the ease
with which the coatings are removed by impact and abrasion.
Gravel and sand potentially susceptible to the alkali-silica reaction occur
along many important rivers in the United States and have been responsible
for serious distress in many concrete structures (Fig. 6). The location of
many known deposits of alkali reactive natural aggregate in western United
States has been reported by Holland and Cook [25]. The known alkali
reactive substances are the silica minerals, opal, chalcedony, tridymite,
and cristobalite; glassy to cryptocrystalline rhyolites, dacites, latites, and
andesites, and their tufts; at least certain artificial siliceous glasses; and
certain phyllites, slates, argillites, and graywackes. Occasional basalts
contain glass whose index of refraction is less than 1.535 and palagonite
may be partially altered to opal. Fraser et al [26] have provided a detailed
study of composition of ehalcedonic cherts in relation to their reactivity
with alkaline solutions. Alkali-reactive dolomitic rocks may occur as
558
! #7,. _o.S ,,,.:._.,___

9 ,o.
AmlZ I
r ....
I
~,
"~c')~q 7 " ~
L ....
)-__
A,,~l.. . . .
'S<A_
'
- -
_ Section of river carrying reactive sand or gravel
Source of reactive crushed stone

9 Structure affected by alkali-silica reaction
O Structure affected by alkali-carbonate reaction
FIG. 6--Location of structures in the United States known to be affected by alkali-aggregate reaction in concrete, sources of sand or gravel subject to alkali-silica reaction, and
sources of crushed stone subject to alkali-carbonate reaction.
constituents of natural aggregates. Any rock containing a significant proportion of these substances is potentially deleteriously reactive.
Petrographic E x a m i n a t i o n of Crushed Stone
Examination
o f S t o n e in t h e F i e l d
Rock formations are massive or stratified; the strata can occur in any
attitude with respect to the horizontal; and rock may vary widely in porosity, hardness, toughness, degree of fracturing, or chemical reactivity. For
example, the alkali reactivity of dolomitic limestones typically varies widely
even within the same quarry when deleterious facies are present [14,15].
Rock formations commonly contain zones of faulting, jointing, or local
shearing, within which the materials are fractured or chemically
decomposed. Certain zones may contain deleterious or unsound substances,
such as chalcedonic or opaline chert and clay or shale in limestone or
dolomite. Especially in warm humid areas, rock formations commonly are
fractured, leached, and partially decomposed near the surface.
559
These features should be discovered by geologic and petrographic

examination of natural exposures, quarried faces or other excavations, and
drilled core. Petrographic examination in the field should include exposures
of geologic formations and stone used in the area for fill, ballast, riprap,
and masonry. This survey may reveal lithologic varieties which have failed
in the natural exposure to freezing-thawing or wetting-drying or to
oxidation and hydration (Fig. 7). The significance of size of fragments
Courtesy of the Bureau of Reclamation U.S. Department of the Interior

FIG. 7--Disintegration of argillaceous facies of limestone used as riprap at Chickamauga
Dam, Tenn. Observation of such disintegration indicates unsoundness of stone in at least
portions of a quarry.
should be recognized in any evaluation of such observations [23]. Disintegration of rock on natural exposure commonly does not coincide with
results of the sulfate soundness test [27]. Shale, claystone, and argillaceous
rocks, including deeply altered basalts and diabases containing nontronite,
may slake and fracture during brief exposure even though they appear
sound when excavated. Similar effects occur if the zeolites leonharditelaumontite constitute a significant proportion of the rock; they are especially common in diorites and quartz diorites. Soluble salts usually
will be revealed by efflorescence at exposed surfaces.
E x a m i n a t i o n o f C r u s h e d S t o n e in the L a b o r a t o r y
As is indicated above, crushed stone aggregates commonly are complex

petrographically. Consequently, they should be examined in the detail
required for natural aggregate (Table 5).
560
Petrographic examination of stone in the form of quarried blocks or

irregular pieces should include inspection of the entire sample. The examination should establish the relative abundance of individual rock types or
facies (Fig. 8). Specimens representative of each type should be obtained
by sawing or coring of typical pieces; these specimens serve for special
tests, detailed petrographic examination, and reference.

FIG. 8--Crushed limestone aggregate containing seams and pieces of illite shale, Webster
Dam, Kan. ( X 89 ). The physical quality was indicated as follows: satisfactory, 61 percent;
fair, 34 percent; poor, 5 percent by weight.
Quantitative petrographic analysis can be made on calcareous rocks by

preparing plane surfaces by sawing and lapping, followed by etching.
Thirty seconds of immersion in 10-percent hydrochloric acid (HC1) will
provide adequate differential attack on calcite and dolomite and will
expose noncalcareous constituents, such as clay, chert, and siliceous sand
or silt. The etched surface can be analyzed by point-count or linear
traverse in general accord with procedures given in ASTM Recommended
Practice C 457. This technique can be applied to calcareous aggregates in
hardened concrete (Fig. 9).
Quantitative analysis can be made similarly on prepared surfaces of
quartzose and feldspathic rocks following etching by 1189 to 2 min
immersion in hydrofluoric acid (HF) (Fig. 10).
If the sample of stone was submitted for crushing tests in the laboratory,
the manufactured aggregate produced should be examined to determine
561
FIG. 9--Concrete containing dolomitic limestone coarse aggregate (a) beJbre and (b) after
etching for 30 s in lO-percent HC1. Scattered crystals in the etched areas on the lapped
surfaces of the aggregate particles are dolomite. ( X 10). Reduced one half for reproduction,
the effects of natural fracturing, jointing, and internal texture on particle

size and shape; frequency of fractures and seams within the particles;
distribution of unsound or deleterious substances in the size fractions;
and the abundance and composition of crusher dust. These features should
be correlated with the processing equipment and methods employed.
Inspection of the stone prior to processing is important because only
thus can the examination of the finished aggregate be interpreted fully. For
example, if unsound or deleterious particles constitute 10 percent of the
finished aggregate, was this proportion derived from approximately one
piece in ten of the original sample, or does a typical piece of the stone contain approximately 10 percent of unsound material? The former possibility
suggests that the quarry should be examined to determine whether the
unsound zones can be avoided or wasted; the latter suggests that the
material should not be used as aggregate in permanent construction.
If the samples are in the form of drilled core, the entire length of core
should be examined and compared with logs available from the driller and
geologist. Especial attention must be given to sections in which core loss
was high or complete, inasmuch as such zones commonly represent
fractured, altered, or otherwise unsound rock. The core should be
562
FIG. lO--Augite gabbro (a) before and (b) after 189 rain etching by concentrated HF.
White areas are plagioclase feldspar, dark areas are augite in the etched surface ( tO). Reduced one half for reproduction.
examined by means of the hand lens, stereoscopic microscope, and petrographic microscope, as necessary, to establish variations in lithology;
frequency and intensity of fracturing; content of clay and shale, regardless
of rock type; and presence of deleterious substances, such as sulfides,
soluble salts, and alkali reactive substances. The examination is facilitated
by sawing and lapping of the core along the length, with or without etching
or staining. These observations should be correlated from hole to hole so
that the variation in lithology or quality of the rock, both in depth and
laterally, is established.
The quality of the aggregate to be expected from the formation represented by the cores also will be indicated by petrographic examination of
aggregate produced from cores in the laboratory.
Petrographic Examination of Blast.Furnace Slag
Blast-Furnace Slag
Blast-furnace slag is the nonmetallic product, consisting essentially of
silicates and aluminosilicates of calcium and of other cations, which is
563
developed simultaneously with iron in a blast furnace (see ASTM

Definition C 125). Properties of blast-furnace slag as concrete aggregate
have been reported by Gutt et al [28,29]. Three general types of blastfurnace slag are used for concrete aggregates, namely, air-cooled slag,
granulated slag, and lightweight slag [30]. Petrographic examination of
lightweight or expanded slag will be discussed in a later section.
Performance o f the Petrographic E x a m i n a t i o n o f Blast-Furnace Slag
The procedure for petrographic examination of blast-furnace slag is not

included specifically in ASTM Recommended Practice C 295. However,
the instructions provided for examination of ledge rock, crushed stone, and
manufactured sand are applicable. In addition to the indicated
microscopical methods, examination of polished and etched surfaces in reflected light is a valuable technique, being preferred by some petrographers
because the surfaces are easy to prepare in sizes larger than thin sections,
the two-dimensional aspect simplifies quantitative estimation of composition, and microchemical tests can be used to identify various phases
in the section under study.
Air-cooled slag is more or less well crystallized, depending primarily
upon the method of disposal employed at the steel plant. Such slag crushes
to angular and approximately equidimensional pieces, the surface texture
of which is pitted, rough, or conchoidal. Crystals range from submicroscopic to several millimetres in size. Abrasion resistance relates to glass
content and the condition of internal stress [30]. Well-granulated slag is
substantially all glass; crystals occur individually or in clusters scattered
through the glass matrix. Incipient crystallization produces brown or
opaque areas in thin sections (Fig. 11).
Courtesyof the Bureau of Reclamation,U.S. Department of the Interior

FIG. 11--Photomicrograph of granulated blast-furnace slag. Note the vesicles (bubbles) in
glass phase (white) (XS0). The dark areas are concentrations of microcrystalline melilite
and merwinite. Reduced one halffor reproduction.
564
The petrographic examination of blast-furnace slag aggregate should

include description of the various types of slag as well as of contaminating
substances. The slag constituent usually can be segregated into two or
more varieties, depending upon particle shape, surface texture, color,
vesicularity, crystallinity, or presence of products of weathering. Each type
should be studied in some detail to assure identification of potentially
deleterious compounds.
More than 20 compounds have been identified in blast-furnace slag
(Table 6), but even well-crystallized slag rarely contains more than five
compounds (Table 7). The typical constituent of blast-furnace slag is
melilite, a compound of variable composition between akermanite (2CaO.
MgO.2SiO2) and gehlenite (2CaO.AI203-SiO2). Pseudowollastonite and
anorthite are of common occurrence. High-lime blast-furnace slags commonly contain one or more forms of calcium disilicate (~, fl, or 3, forms of
2CaO.SiOD. Magnesium-containing blast-furnace slags usually include
monticellite, forsterite, or merwinite. Calcium sulfide almost always is
present in small proportion. Sulfides of manganous manganese and ferrous
iron are common. Properties and techniques for identification of these
compounds are summarized by Rigby [31], Nurse and Midgley [32], Insley
and Frechette [33], and Snow [34,35], and in standard works on mineralogy.
X-ray diffraction methods are necessary if crystalline phases are submicroscopic and are a great aid to a petrographer who is developing experience independently in this field.
Several original constituents of blast-furnace slag may be deleterious to
the performance of concrete. Sulfides released into the cement paste matrix
produce innocuous greenish staining of the interior of concrete; occasionally, a mottled aspect is produced on the surface of damp concrete
[12]. Presence of colloidal sulfides is suggested by yellow or brown coloration of the glass phase [30]. Gypsum commonly forms a blast-furnace
slag by weathering and may result in sulfate attack on the cement paste
matrix in concrete in service. Inversion of fl-dicalcium silicate to the
~,-dicalcium silicate, with the accompanying 10-percent increase in volume
of the crystals, causes "dusting" or "blowing" of slag [30,36]. The inversion ordinarily takes place before the slag has cooled, and the disintegrated material is removed by screening in the production of coarse aggregate. In less severe occurrences, the disintegration takes place slowly,
producing pieces that are partly or wholly weak and friable, and thus unsuitable for concrete aggregate. This action of dicalcium silicate can be
avoided by maintaining a ratio of CaO to SiO2 in the slag sufficiently low
to prevent formation of the compound, or by chilling the molten slag so
that the compound does not crystallize [30]. If air-cooled slag is poured in
thin layers, rapid cooling ordinarily arrests the compound in the fl modification. Dicalcium silicate can be identified microscopically in slag by
SiO 2
CaO- A120 3
2 MgO. 2A120 3"5SiO 2
cristobalite
calcium aluminate
cordierite
siUimanite
mullite
madisonite
CaO- (Mg, Fe)O. 2SiO 2

MgO- SiO 2
MgO. SiO 2
A120 3" SiO 2

3A120 3.2SiO 2
2CaO. 2MgO. A120 3" 3 SiO 2
CaS
FeS
MnS
(Mg, Fe)O. A120 3
CaO. A120 3"2SIO 2
MgO
CaO
Chemical Formula
oldhamite
ferrous sulfide
m a n g a n o u s sulfide
spinel
anorthite
periclase
lime
Compound
2CaO. A120 3" SiO 2

2CaO. MgO- 2SiO 2
ctCaO. SiO 2
flCaO. SiO 2
~ ' 2 C a O . SiO 2
fl2CaO. SiO 2
72CAO 9SiO 2
2(Mg, Fe)O-SiO2
3CaO. MgO. 2SiO 2
3CaO- 2SiO 2
CaO. MgO- SiO 2
Chemical Formula
a Compiled from several sources, primarily Nurse and Midgley [32], McCaffery et al [38], and American Concrete Institute Committee 201 [40].
Gehlenite
Akermanite
Pseudowollastonite
Wollastonite
Bredigite
Larrlite
7-dicalcium silicate
Olivine
Merwinite
Rankinite
Monticellite
Pyroxene
Diopside
Enstatite
Clinoenstatite
Compound
TABLE 6--Compounds occurring in blast-furnace slag. a
Grl
O)
O~
.-I
-o
I
z
N
O
Z
-u
Ill
..q
:]0
O
Q
m
f...
frl
X
X
X
x
X
X
X
X
x
X
f melilite
C2MS2 = a k e r m a n i t e J
C2S : d i c a l c i u m silicate
CS = wollastonite or
pseudowollastonite
X
X
X
X
X
C2MS2
C~AS
a After Nurse a n d Midgley [32].

b Key: C2AS = gehlenite -~
Dolomite
Limestone
F l u x Stone
X
X
C2S
X
C3S2
X
X
X
X
C3MS2
CMS2 = diopside
CMS = monticellite
CAS2 = a n o r t h i t e
C352 = r a n k i n i t e
CAMS2 = merwinite
M A = spinel
CS
X
X
X
X
X
x
MA
X
X
X
CMS
MxS = forsterite
MS : e n s t a t i t e
M g O = periclase
X
X
CMS2
C o m b i n a t i o n of C o m p o u n d s b
X
X
X
x
X
CAS2
X
X
X
M2S
MS
MgO
T A B L E 7--Most frequently occurring combinations o f compounds of CaO. M g O . A l 2 0 3 9Si02 produced by crystallization of blast furnace slag. a
--4
I"11
00
fi3
Q
20
m
60
0
-11
0
0
z
0
20
m
-4
X
C:7
"13
20
0
"o
m
20
--4
m
o9
Go
.-4
567
special techniques [36]. Scattered crystals of/3-dicalcium silicate commonly

are stable in blast-furnace slag and so are innocuous.
Free lime (CaO) and magnesia (MgO) are extremely rare as constituents
of air-cooled blast-furnace slag and are not likely to form either as a primary phase or devitrification product in granulated, blast-furnace slag
[32,37]. 3 Nevertheless, their absence or presence and abundance should
be established by the petrographic examination. These compounds are
deleterious because of the increase in solid volume resulting from hydration or carbonation in place. Cristobalite has been reported as a constituent of blast-furnace slag [38]. This compound is potentially alkali reactive; if identified, its abundance should be determined. The glass phase
of normal blast-furnace slag is not deleteriously reactive with cement
alkalies.
Contaminating substances whose presence or absence should be established by petrographic examination are metallic iron, iron carbide,
coke, and incompletely fused fluxstone. The last is important because
delayed hydration and carbonation of free lime and magnesia may produce
expansion of the concrete and popouts. Metallic iron and iron carbides
rust by oxidation and hydration if exposed at the surface of concrete.
Petrographic Examination of Lightweight Concrete Aggregates
Expanded Clay, Shale, and Slate

Petrographic examination of lightweight aggregate should include
segregation of the particles into as many categories as required to describe
the sample adequately (Table 4). Particles of expanded material may be
segregated on the basis of particle shape, surface texture, development of a
coating or "skin," vesicularity, and friability. Incompletely expanded
particles and particles not expanded should be distinguished from vesicular
ones. These should be separated on the basis of porosity, absorptivity,
density, friability, softness, occurrence of efflorescence, and reaction to
water (softening, slaking, or swelling). The petrographic examination supplements standard tests in distinguishing clayey particles from "clay
lumps" determined in accordance with ASTM Test C 142. Other materials
to be identified and determined quantitatively are underburned or raw
material, coal, and rock particles. The composition and content of raw
material commonly can be established most easily and accurately by
X-ray diffraction or differential thermal analysis.
The individual types of particles also should be analyzed petrographically
3Unlike blast-furnace slags, basic open-hearth slags commonly contain free oxides that
are subject to hydration in portland-cementconcrete, namely, free lime (CaO) and magnesiowl]stite, a solid solution of magnesium oxide (MgO), manganese oxide (MnO), and ferrous
oxide (FeO).
568
to establish the presence of free magnesia or lime. These result mainly

from decomposition of calcium and magnesium carbonates in the feed
during firing. They may produce distress or popouts in concrete or concrete p r o d u c t s unless the aggregate is water- or steam-cured prior to
use [39].
Petrographic examination also assists selection of raw materials, development of manufacturing methods and equipment and process control.
Cinders
Cinders used as concrete aggregate are the residue from high-temperature
combustion of coal and coke in industrial furnaces. Petrographic examination should determine the physical nature of the cinder particles on the
basis of composition, friability, softness, particle shape, and surface texture.
Especial attention should be given to identification of sulfides, sulfates,
coal, and coke.
Expanded Blast-Furnace Slag

Expanded blast-furnace slag is produced by carefully controlled intermingling or molten slag and water or steam in one of several ways [30].
The petrographic examination should describe the aggregate in terms of
the nature of the expanded particles, including their particle shape, surface
texture, friability or softness, effects of weathering in stockpiles, and content of contaminating substances, such as dense slag.
Pumice, Scoria, Tuff, and Volcanic Cinder

Pumice, scoria, tuff, and volcanic cinder used for lightweight aggregate are naturally occurring porous or vesicular volcanic materials. Pumice
is a very highly porous and vesicular volcanic rock composed largely of
natural glass drawn into approximately parallel or loosely entwined fibers
and tubes. Scoria is a highly porous and vesicular volcanic rock in which
the vesicles typically are rounded or elliptical in cross section, the interstitial
glass occurring as thin films (Fig. 12). Tuff is a general term designating
consolidated volcanic ash of any lithologic type or physical character. Volcanic cinder is a loose accumulation of highly vesicular (scoriaceous)
fragments of lava, predominantly ranging from 4 to 32 mm in diameter.
Petrographic examination of these types of aggregate includes segregation of the particles on the basis of particle shape, surface texture, porosity
or vesicularity, fracturing, friability or softness, weathering, specific gravity,
secondary deposits in voids, coatings, and potential alkali reactivity. Extraneous or contaminating substances are primarily dense particles of
volcanic rock and organic matter. In production of lightweight aggregate,
569

FIG. 12--Typical basaltic scoria aggregate (natural size). The pieces are black, gray, and
reddish brown. Note the rounded vesicles. Reduced one half for reproduction.
two types of volcanic materials are intermixed occasionally for economy or

to control gradation or unit weight. The type and relative proportion of the
materials can be established by petrographic examination.
Volcanic glass with an index of refraction less than 1.535 is potentially
deleteriously reactive with cement alkalies; glass whose index is in the
range 1.535 to 1.570 probably is alkali reactive. Opal, chalcedony, tridymite, and cristobalite are also common alkali-reactive constituents of
volcanic aggregates. However, in spite of alkali-silica reaction and formation of alkalic silica gel in highly porous pumice and tuff, expansion
usually is prevented by the abundant voids into which the hydrating gel
can escape without development of excessive stress in the mortar. For
example, a very highly porous rhyolite tuff from Hideaway Park, Colo.,
containing abundant tridymite caused only 0.041 percent expansion of a
high-alkali cement mortar during one year of moist storage in accordance
with ASTM Test for Potential Alkali Reactivity of Cement-Aggregate
Combinations (Mortar-Bar Method) (C 227), yet the specimen contained
abundant alkalic silica gel [41]. A similar but dense tuff from near Castle
Rock, Colo., produced an expansion of 0.400 percent under the same
conditions.
Perlite
When heated rapidly to fusion, certain obsidians and pitchstones release
gases which, being trapped within the molten glass, vesiculate the rock,
and cause disruption into small pieces. The product is known commercially
as perlite.
570
Petrographic examination should indicate the composition of the aggregate in terms of particle shape, surface texture, composition, density,
friability or fragility, and potential alkali reactivity. Perlite may contain
particles of dense volcanic rock or individual crystals.
Being composed of volcanic glass, typical perlite is potentially reactive
with cement alkalies, although significant expansion may not occur because of the porosity of the particles. However, laboratory tests demonstrate that certain perlites produce significant expansion of mortar stored
in accordance with ASTM Test C 227 in combination with either highalkali or low-alkali cement [42,43]. Such volume change will not necessarily cause structural distress if appropriately accommodated in the design.
Exfoliated Vermiculite
Exfoliated vermiculite is produced by rapid heating of the micaceous
mineral, vermiculite. Release of combined water expands with crystals-like an accordian--increasing the volume to as much as 30 times its original
size. The degree of expansion varies widely, depending upon mineralogic
properties and purity and the conditions of firing.
During petrographic examination, the particles of vermiculite are segregated by degree of expansion, elasticity of brittleness of the flakes, and
fragility of the expanded crystals. These differ significantly within individual
samples from some sources, especially from marginal deposits where the
vermiculite grades into hydrobiotite or biotite. Also to be reported is intermixture of the vermiculite with particles of rocks and minerals occurring
with the vermiculite in the deposit.
Diatomite
Crushed and sized natural diatomite typically is soft, porous and absorptive, and ranges from firm to pulverulent. Finely divided opal and
opaline skeletons of diatoms are the predominant constituents. Fine sand,
silt, clay, and volcanic ash are present in widely differing proportions. At
least certain diatomites produce significant expansion of mortars stored in
accordance with ASTM Test C 227, with both high- and low-alkali cement
[42,43].
Conclusion
Petrographic examination should be included in the investigation and
testing of concrete aggregate for use in permanent construction. Applied
in the field, the method aids exploration and sampling and permits preliminary evaluation of materials from alternative sources. Detailed exCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:34 EDT 2014
571
a m i n a t i o n of aggregate in the laboratory s u p p l e m e n t s the s t a n d a r d acceptance tests, especially by (1) detecting adverse properties, (2) c o m p a r i n g
the aggregate with aggregates for which service records or previous tests
are available, (3) e x p l a i n i n g results of tests a n d justifying special tests as
required, (4) detecting c o n t a m i n a t i o n , a n d (5) d e t e r m i n i n g the efficiency
a n d relative merit of processing a n d m a n u f a c t u r i n g methods.
Validity of the results depends u p o n the t r a i n i n g a n d experience of the
petrographer. However, with proper t r a i n i n g a n d the adoption of u n i f o r m
techniques a n d n o m e n c l a t u r e , subjective elements in the e x a m i n a t i o n are
not significant.
The methods can be applied effectively to sand, gravel, crushed stone,
slag, a n d n a t u r a l or synthetic lightweight aggregate.
References
[1] Mielenz, R. C. and Witte, L. P., Proceedings, American Society for Testing and
Materials, Vol. 48, 1948, p. 1071.
[2] Mather, Katharine and Mather, Bryant, Proceedings, American Society for Testing
and Materials, Vol. 50, 1950, p. 1288.
[3] Bayne, R. L. and Brownridge, F. C., "An Engineering Method of Petrographic Analysis
for Coarse Aggregate Quality," presented at the 36th Convention Canadian Good
Roads Association, Banff, Alberta, 13-16 Sept. 1955.
[4] Mielenz, R. C., Bulletin, Geological Societyof America, Vol. 5"7, 1946, p. 309.
[5] Rexford, E. P., Transactions, American Institute of Mining, Metallurgical, and
Petroleum Engineers, Vol. 187, No. 3, 1950, p. 395.
[6] Mielenz, R. C., Proceedings, American Society of Testing and Materials, Vol. 54,
1954, p. 1188.
[7] Rhoades, Roger and Mielenz, R. C., Proceedings, American Concrete Institute, Vol. 42,
1946, p. 581.
[8] Rhoades, Roger and Mielenz, R. C., Mineral Aggregates, A S T M STP 83, American
Society for Testing and Materials, 1948, p. 20.
[9] Swenson, E. G. and Chaly, V., Proceedings, American Concrete Institute, 1956, p. 987.
[10] Mielenz, R. C., "Petrographic Examination of Concrete Aggregate to Determine Potential Alkali Reactivity," Research Report 18-C, Highway Research Board, 1958, p. 29.
[11] Hansen, W. C., "Anhydrous Minerals and Organic Materials As Sources of Distress in
Concrete," Research Record No. 43, Highway Research Board, 1964, p. 1.
[12] Mielenz, R. C., "Reactions of Aggregates Involving Solubility, Oxidation, Sulfates or
Sulfides," Research Report No. 43, Highway Research Board, 1964, p. 8.
[13] Midgley, H. G., Concrete Research, Vol. 10, 1958, p. 75.
[14] Hadley, D. W., Proceedings, Highway Research Board, Vol. 40, 1961, p. 462.
[15] Gillot, J. E., Bulletin, GeologicalSociety of America, 1963, 13. 759.
[16] Swenson, E. G. and Gillot, J. E., Concrete Research, Vol. 19, 1967, pp. 95-104.
[17] Smith, P., "15 Years of Living at Kingston With a Reactive Carbonate Rock," Record
[18] Ryell, J., Chojnacki, B., Woda, G., and Koniuszy, D., "The Uhthoff Quarry Alkali
Carbonate Rock Reaction: A Laboratory and Field Performance Study," Record
[19] Ozol, M. A., and Newlon, H. H., Jr., "Bridge Deck Deterioration Promoted by AlkaliCarbonate Reaction: A Documented Example," Record No. 525, Transportation
[20] Buck, A. D., "Control of Reactive Carbonate Rocks in Concrete," Technical Report
C-75-3, U.S. Army Engineer Waterways Experiment Station, 1975, p. 36.
[21] Dolar-Mantuani, L., "Alkali-Silica-Reactive Rocks in the Canadian Shield," Record
572
[22] Concrete Manual, Bureau of Reclamation, U.S. Department of the Interior, 8th ed.,
1975.
[23] Verbeck, G. and Landgren, R., Proceedings, American Society for Testing and Materials, Vol. 60, 1960, p. 1063.
[24] Goldbeck, A. T., Proceedings, Highway Research Board, Vol. 12, Part 1, 1933, p. 305.
[25] Holland, W. Y., and Cook, R. H., "Alkali Reactivity of Natural Aggregates in Western
United States," Mining Engineer, 1953, p. 991.
[26] Fraser, G. S., Harvey, R. D., and Heigold, P. C., "Properties of Chert Related to
Its Reactivity in an Alkaline Environment," Circular 468, Illinois State Geological
Survey, 1972.
[27] Loughtin, G. F., Rock Products, Vol. 31, 17 March 1928, p. 50.
[28] Gutt, W., Kinniburgh, W., and Newman, A. J., Concrete Research, Vol. 19, 1967,
pp. 71-82.
[29] Everett, L. H. and Gutt, W., Concrete Research, Vol. 19, 1967, pp. 83-94.
[30] Josephson, G. W., Sillers, F., Jr., and Runner, D. G., "Iron Blast Furnace Slag:
Production, Processing, Properties, and Uses," Bulletin No. 479, Bureau of Mines,
U.S. Department of the Interior, 1949.
[31] Rigby, G. R., The Thin Section Mineralogy of Ceramic Materials, British Ceramic
Research Association, 2rid ed., 1953.
[32] Nurse, R. W. and Midgley, H. G., "The Mineralogy of Blast Furnace Slag," Silicates
Industriels, Vol. 16, No. 7, 1951, p. 211.
[33] Insley, H. and Frechette, V. D., Microscopy of Ceramics and Cements, Academic Press,
Inc., New York, N.Y., 1955.
[34] Snow, R. B., Technical Publications No. 2167, American Institute of Mining, Metallurgical, and Petroleum Engineers, 1947.
[35] Snow, R. B., Journal, American Ceramic Society, Vol. 36, No. 9, 1953, p. 299.
[36] Parker, T. W. and Ryder, J. F., Journal, Iron and Steel Institute, Vol. 146, No. 2, 1942,
p. 21.
[37] Stutterheim, N., "The Risk of Unsoundness Due to Periclase in High Magnesia Blast
Furnace Slags," Monograph 43, Proceedings of the 4th International Symposium on
Chemistry of Cement, National Bureau of Standards, Vol. 2, 1960, p. 1035.
[38] McCaffery, R. S., Jr., Oesterle, R., and Schapiro, Leo, Technical Publication No. 19,
American Institute of Mining, Metallurgical and Petroleum Engineers, 1927.
[39] Nordberg, Bror, Rock Products, Vol. 57, 1954, p. 104.
[40] "Committee 201 Report on Blast Furnace Slag as Concrete Aggregate," Proceedings,
American Concrete Institute, Vol. 27, 1930-1931, pp. 183,661.
[41] Mielenz, R. C., Greene, K. T., and Benton, E. J., Proceedings, American Concrete
Institute, Vol. 44, 1948, p. 193.
[42] Hickey, M. E., Cordon, W. A., and Price, W. H., "Properties of Concrete Made With
Typical Lightweight Aggregates--Housing and Home Finance Agency Research Program," Laboratory Report No. C-385, Bureau of Reclamation, U.S. Department of
the Interior, 1948.
[43] Price, W. H. and Cordon, W. A., Proceedings, American Concrete Institute, Vol. 45,
1949, p. 581.
W. H. Price I
Chapter 34--Grading
Introduction
The influence of aggregate gradings on the properties of concrete has been
extensively discussed in the technical literature during the past 125 years and
many methods for arriving at " o p t i m u m " or "ideal" gradings have been
presented. However, none of these has been accepted as being universally applicable, because of economic consideration, differences in particle shape
and texture of the aggregates, and the effects of entrained air and amount of
cementitious material contained in the concrete [1-25]. 2
Experience has demonstrated that either very fine or very coarse sand, or
coarse aggregate having a large deficiency or excess of any size fraction, is
usually undesirable, although aggregates with discontinuous or gap grading
have sometimes been used to advantage [26-28].
This paper is concerned mainly with normal weight aggregates, but the
grading limits discussed are applicable to lightweight and heavyweight aggregates as well, because aggregate, cement, water, and entrained air, of
which concrete is composed, fit together as absolute volumes and in no other
way.
Nomenclature
The particle size distribution of aggregate as determined by separation
with standard sieves is known as its gradation. Sieve analysis, screen
analysis, and mechanical analysis are terms used synonymously in referring
to the gradation of aggregate.
A sieve is defined as a metallic plate or sheet, a woven wire cloth, or other
similar device, with regularly spaced apertures of uniform size, mounted in a
suitable frame or holder for use in separating material according to size; in
mechanical analysis, an apparatus with square openings is a sieve, one with
circular apertures is a screen.
1Consulting engineer, Laguna Hills, Calif. 92653.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
574
ASTM Specification for Wire-Cloth Sieves for Testing Purposes (E 11) lists
dimensions for the standard sieves in metric units. Metric units for sieve
openings will be used in this paper, with the U.S. customary units shown in
parentheses.
A convenient system of expressing the gradation of aggregate is one in
which the consecutive sieve openings are constantly doubled, such as 9.5, 19,
and 38.1 mm (3/8, 3A, and 1189 in.) Under such a system employing a log
scale, lines on a graph can be spaced at constant intervals to represent the
successive sizes and the fineness modulus can be computed.
The fineness modulus is an empirical factor used to indicate the coarseness
or fineness of the material. It is obtained by adding the total percentages of a
sample of the aggregate retained on each of a specified series of sieves, and
dividing the sum by 100. The U.S. standard sieve sizes of this series in
millimetres are: 0.15 (No.100), 0.3 (No.50), 0.6 (No.30), 1.18 (No.16), 2.36
(No.8), 4.75 (No.4), 9.5 (3/8 in.), 19 ( 90 in.), 38.1 (1 t/2 in.), 76 (3 in.), and 150
(6 in.). The fineness modulus of fine aggregate is determined by using the
cumulative percentages of material retained on sieves below the 4.75 (No.4)
size. The fineness modulus of fine aggregate usually falls between 2.3 and
3.1, with its coarseness increasing as the fineness modulus becomes larger.
Other sieves, also used for separating coarse aggregate are: 12.5 (1/2 in.),
25 (1 in.), 50 (2 in.), 63 (21A in.), and 90 (31/2 in.).
Size Separation
The gradings of sand and gravel deposits and the crusher product of quarried material seldom occur in desirable percentages to produce uniform
workability and economical strong concrete. The fine aggregate must be
separated from the coarse aggregate and the coarse aggregate separated into
sizes as listed in Table 2 of ASTM Specifications for Concrete Aggregates (C
33). This separation for recombining in the desired proportions in the concrete mixture pays dividends through the ease with which the resulting
uniform workable mixture can be placed and the savings in cement that is
realized.
Although many different gradings have been found to produce suitable
concrete, the gradings cannot vary suddenly or widely during concreting
operations, because of their pronounced effect on the slump and workability
of the concrete. Uniformity of concrete is of such importance that some
specifications for large jobs require that the grading of fine aggregate shall be
controlled, so that the fineness moduli of at least 9 out of 10 consecutive test
samples of the fine aggregate as delivered to the mixer shall not vary more
than 0.20 from the average of the 10 samples [28].
Perfect separation of the aggregate at the processing plant cannot be accomplished at reasonable cost. This is recognized in grading specifications.
Table 2 of ASTM Specification C 33 lists the amount of undersize and overCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:34 EDT 2014
PRICE ON GRADING
575
size permitted in each size fraction of coarse aggregate. It has been found
that where there is an excessive amount of either undersize or oversize,
segregation will occur in the stockpiling and batching operations, and slugs
of fine or coarse material will reach the mixer. Because of the segregation
which usually occurs in handling, it is desirable that the ratio of the size
openings of the upper and lower screens which determine any size fraction be
not greater than 2 to 1. This practice is generally followed except for the
finest size of coarse aggregate where the ratio is usually increased above the 2
to 1 figure, and in this fraction it is common practice to extend the size from
4.75 to 19 mm (3/16to 90 in.). Segregation in such finer size fractions may be
very pronounced, with portions of the material that reach the mixer containing a preponderance of fines below the 9.5 mm (3/s in.) size. Under such conditions it is difficult to obtain uniformly workable concrete. The condition is
further aggravated by breakage and segregation which occurs in the larger
sizes of coarse aggregate during stockpiling and handling. Lenses of fines
which form in large stockpiles are another constant source of trouble.
Breakage and segregation can be minimized through careful and proper
handling of the aggregate. Some have found it economical to finish-screen
the aggregate just before it goes into the batching bins, and where this is
done, acceptable separation can be assured.
Maximum Size Aggregate

The amount of water per unit volume of concrete required to produce a
given slump decreases as the maximum size aggregate increases, as shown in
Table 1.
For compressive strengths below 21 MPa (3046 psi), the cementitious content of the concrete can be progressively reduced as the maximum size aggregate is increased while maintaining a constant water/cement ratio and
strength. At the 28 Mpa (4061 psi) level, no cementitious material can be
saved by using aggregate above 76 mm (3 in.) in size. As the strength level is
increased, smaller maximum size aggregates must be used for most efficient
use of cement with 37.5 mm (11/2 in.) for the 35 MPa (5076 psi) level, 19 mm
( 90 in.) for the 42 MPa (6092 psi) level, and 9.5 to 12.5 mm (3/s to 189 in.) for
strength above 42 MPa (6092 psi) [28,29].
In the production of high strength concrete above 42 MPa (6092 psi), there
seems to be no advantage in using maximum size aggregate below the 12.5
mm (1/2 in.) size [30, 31]. A large particle of aggregate has less area for bonding with the mortar than an equal weight of smaller aggregate particles and it
is for this reason it is not possible to obtain high strengths with large maximum size aggregate [32]. Also, it is not possible to obtain strengths above a
practical level of 35 MPa (5076 psi) with the entrained air contents listed in
Table 1 [33].
In Section 3.3.2 of Building Code Requirements for Reinforced Concrete
276
520
100
6
245
462
100
13
M a x i m u m Size Aggregate
Water, k g / m 3"
Cement, k g / m 3b
Fine aggregate, %c
Entrapped air, %d
Water, k g / m 3a
Cement, k g / m 3b
Fine aggregate, %c
Total air, %d
201
380
58
8
228
430
62
3
9.5 m m
19 m m
202
382
49
2
192
362
S0
7
177
334
45
6
Air-Entrained Concrete
216
406
54
2.5
Nonair-Entrained Concrete
12.5 m m
169
318
40
5
192
360
44
1.5
25 m m
157
296
37
4.5
177
334
40
1
37.5 m m
148
279
34
4
168
316
37
0.5
50 m m
139
262
31
3.5
160
302
34
0.3
76 m m
118
222
25
3
139
262
28
0.2
150 m m
~Approximate amount of mixing water in kilograms per cubic metre required for 7S-mm slump with well shaped angular coarse aggregate. Quantities listed can be reduced significantly through the addition of a water-reducing admixture.
bCement required in kilograms per cubic metre for 0.53 water/cement ratio by weight.
CApproximate percentage of fine aggregate of total aggregate by absolute volume.
dRecommended average total percentage of entrained air required for frost resistance from Table 5.2.3 of ACI Recommended Practice for Selecting
Proportions for Normal and Heavy Weight Concrete (ACI 211.1).
4.75 m m
(sand)
TABLE 1--Mortar requirements for workable concrete with various maximum aggregate sizes.
--I
rrl
o)
rl-I
"-i1
m
-4
m
C')
0
Z
C)
0
"0
m
-n
.-I
Z
0
"0
-.4
(,r)
--4
I11
t.,rl
O~
PRICE ON GRADING
577
(ACI 318), it is stated: "The nominal maximum size of the aggregate shall
not be larger than one fifth of the narrowest dimension between sides of
forms, one third of the depth of slabs, nor three fourths of minimum clear
spacing between individual reinforcing bars or bundles of bars or pretensioning tendons or post-tensioning ducts. These limitations may be waived if, in
the judgment of the engineer, workability and methods of consolidation are
such that the concrete can be placed without honeycomb or void." From the
limitation on maximum size aggregate set by ACI 318, it can be seen that, except for massive sections of concrete, the maximum size aggregate will usually be below 38.1 mm (11/2 in.). For the lean, low-strength concrete used in
the interior of thick concrete dams, some economies can be realized, with
other advantages of lower heat generation and shrinkage, by using 150 mm
(6 in.) maximum size aggregate. The use of cobbles larger than 150 mm (6
in.) accomplishes little or no saving in cost or improvement in the
characteristics of the concrete.
It can be seen from the foregoing discussion that grading of aggregate has
a marked influence on the strength of concrete. Also, the physical
characteristics of the aggregate have a pronounced effect on the strength of
concrete. But, for any given set of materials and air content, and for a constant grading, the strength will vary in accordance with the water/cement
ratio law.
Grading of Fine Aggregate

The grading of fine aggregate has a much greater effect on workability of
concrete than does the grading of the coarse aggregate [34]. If the mortar is
workable and satisfactory in other respects, it is usually possible to secure
workable concrete by using enough mortar to fill the voids among the coarse
aggregate particles and separate them so that there is ample room for them
to move in the mortar without interference. Experience has shown that usually very coarse sand or very fine sand is unsatisfactory for concrete mixtures.
The coarse sand results in harshness, bleeding, and segregation; and the fine
sand requires a comparatively large amount of water to produce the
necessary fluidity and also tends to cause segregation. Fine aggregate
gradings falling within the specification limits of ASTM Specification C 33,
shown in Table 2 of this paper, should be satisfactory for most concretes.
Under these specifications the minimum percentage of material passing
the 0.3 and 0.15 mm (Nos. 50 and 100) sieves may be reduced to 5 and 0,
respectively, if the aggregate is to be used in air-entrained concrete containing more than 251 kg/m 3 of cement, or in nonair-entrained concrete containing more than 307 kg/m 3 of cement, or if an approved mineral admixture is
used to supply the deficiency in fines passing these sieves. ASTM Specification C 33 also requires that fine aggregate shall have not more than 45
percent retained between any two consecutive sieves of those shown in the
100
Percent passing
by weight a
95 to 100
4.75 (No.4)
80 to 100
2.36 (No.8)
aThese percentages can be modified as explained in the text.
9.5(3/8 in.)
Sieve Size, m m
50 to 85
1.18 (No.16)
TABLE 2 - - G r a d i n g limits f o r f i n e aggregate.
25 to 60
0.6 (No.30)
10 to 30
0.3 (No.50)
2 to 10
0.15 (No.100)
fi)
-n
m
fi)
m
.-d
m
~)
0
Z
0
CO
-11
0"u
m
-n
z
o
t..q
~
Co
PRICE ON GRADING
579
table, and that the fineness modulus shall be not less than 2.3 nor more than
3.1.
Fine aggregates having gradings outside ASTM Specification C 33 have
produced satisfactory concrete in some instances, but where the behavior of
an available sand having a grading falling outside accepted limits is
unknown, and where the size of the job does not warrant the expense of trial
mixes before specifications are written, it is recommended that an accepted
specification such as ASTM Specification C 33 be used. On large jobs involving many thousands of cubic metres (yards) of concrete, it is recommended
that the average grading as produced from the deposit or quarry without
blending or wasting to bring it within some predetermined limits be tried in
trial mixes to determine whether the aggregate can be used without expensive
processing. However, as has been discussed under Size Separation, it is
stressed that regardless of the grading selected it must be maintained within
certain limits without sudden variations if satisfactory economical concrete is
to be produced.
It has already been mentioned that ASTM Specification C 33 permits a
reduction in the minimum amount of material passing the 0.3 and 0.15 mm
(Nos. 50 and 100) sieves from that shown in Table 2 in the ease of rich or airentrained mixes. On the other hand, increased percentages of these sizes may
be desirable in the ease of the lean concrete used in mass construction.
For the high cement contents used in the production of high-strength concrete, it has been found that a coarse sand with a fineness modulus of
around 3.0 produced the best workability and highest compressive strength
[30]. In general, manufactured sands require more fines than natural sands
for equal workability [21].
Concrete for pumping must be very workable, with high sand contents and
preferably with aggregate graded below the 25 mm (1 in.) size. The sand
must be well graded and somewhat on the fine side, with 15 to 30 percent
passing the 0.3 mm (No.S0) sieve.
It is sometimes difficult to finish a floor because of the grading of the sand
which causes "chattering" or waves to form under the trowel. In one case this
"chattering" was eliminated by increasing the amount of material passing
the 0.3 mm (No.50) sieve from 12 to 18 percent and the amount passing the
0.075 mm (No.200) sieve from 1 to 4 percent.
Coarse Aggregate
As a rule it is less difficult to provide satisfactory grading in coarse aggregate than in sand. It is practicable to make size separation of the coarser
material by screening at comparatively small expense and to recombine the
separated sizes in desired proportions. Table 3 shows the approximate ranges
over which it will usually be practicable to vary the relative proportions of size
Cobbles
75 to 150
0
0
0
20 to 35
Maximum Size Aggregate in

Concrete, mm
19 (% in.)
38.1 (1% in.)
75 (3 in.)
150 (6 in.)
0
0
20 to 40
20 to 32
Coarse
38.1 to 75
0
40 to 55
20 to 40
20 to 30
Medium
19 to 38.1
55
30
15
12
to
to
to
to
73
35
25
20
9.5 to 19
Fine
Percentages Reduced to Cleanly Separated Basis
27
15
10
8
to
to
to
to
45
25
15
15
4.75 to 95
TABLE 3--Approxirnate ranges in grading o f gravel coarse aggregates f o r various m a x i m u m sizes (in millimetres).
"-4
m
Go
"n
m
-i1
rn
--4
rn
I"11
O~
O
-o
::]o
0
"o
rn
:]0
.-I
or)
-I
m
Go
O1
tX~
0
PRICE ON GRADING
581
fractions of coarse aggregate for concretes having various maximum sizes of

aggregates and for optimum sand percentages.
Workable and placeable concretes have been made with aggregates graded
up to every maximum size between 9.5 and 225 mm (3/8 and 9 in.); however,
as already mentioned, the use of cobbles larger than 150 mm (6 in.) accomplishes little or no saving in cement and involves difficulties in handling
and placing procedures.
Air Entrainment and Grading

Purposefully entrained air in concrete materially improves the resistance of
concrete to frost action, improves its workability, and reduces segregation in
the mix. The effect of the finely dispersed air bubbles obtained through the
use of air-entrained agents is entirely different from that of the larger air
voids which result from poorly graded aggregates; the larger voids do not improve the quality of the concrete and contribute to poor workability and
segregation. Table 1 shows the reduction in fine-aggregate percentages which
may be realized through the use of air entrainment in concrete. In addition to
a reduction in the amount of fine aggregate which may be accomplished for a
given set of materials and grading, it is sometimes possible to use a coarser or
finer grading or a jump grading and obtain concrete of satisfactory workability with the proper amounts of entrained air. ASTM Specification C
33 recognizes the effect of entrained air on workability and permits a smaller
percentage of fines for air-entrained concrete. In general, more poorly
graded aggregates can be used for air-entrained concrete, but it is recommended where such aggregates are being considered that they be tried in trial
mixes before the usual specifications are relaxed or changed.
Conclusions
1. Many methods have been proposed for arriving at an "ideal" grading,
but none has been universally accepted because of economic considerations,
the effect of particle shape and surface texture of the aggregate, air content,
and cement content on the resulting characteristics of the concrete.
2. Although grading of aggregate has a pronounced effect on the strength
of concrete and small maximum size aggregate must be used for efficient
production of high-strength concrete, the strength of concrete for any given
set of materials and grading will follow the water/cement ratio law.
3. Coarse, fine, and gap gradings have produced satisfactorily workable
concrete under certain conditions, but unless trial mixes are made
beforehand to determine whether they can be used, it is recommended that
gradings be required to meet a specification such as ASTM Specification C
33.
4. Economical, uniform concrete cannot be produced with pit-run or
582
crusher-run aggregate, and it is necessary that the aggregate be separated

into its component sizes so that it can be combined in the concrete mixture
within the limits of variation permitted by the specification.
References
[1] Price, W. H., "Grading of Mineral Aggregates for Portland Cement Concrete and
Mortars," Mineral Aggregates, A S T M STP 83, American Society for Testing and
[2] Wright, W. H., Treatise on Mortars, 1845.
[3] Feret, R., Bulletin de la Socidtd d'Encouragement pour l'Industrie Nationale, 1897.
[4] Mercer, L. B., The Law of Grading for Concrete Aggregates, Melbourne Technical
College Press, 1951.
[5] Fuller, W. B. and Thompson, S. E., "The Laws of Proportioning Concrete," Transactions, American Society of Civil Engineers, Vol. 59, 1907, p. 67.
[6] Wig, R. J., Williams, G. M., and Gates, E. R., "Strength and Other Properties of
Concrete as Affected by Materials and Methods of Preparation," Technical Paper No. 58,
National Bureau of Standards, Washington, D. C.
[7] Talbot, A. N. and Richart, F. E., "The Strength of Concrete and Its Relation to the
Cement, Aggregate, and Water," Bulletin No. 137, University of Illinois, Oct. 1923.
[8] Edwards, L. N., "Proportioning of Mortars and Concretes by Surface Area of Aggregates,"
Proceedings, American Society for Testing and Materials, Vol. 18, Part 2, 1918, p. 235.
[9] Young, R. B., "Some Theoretical Studies on Proportioning Concrete by the Method
of Surface Areas of Aggregates," Proceedings, American Society for Testing and
Materials, Vol. 19, Part 2, 1919, p. 444.
[10] Abrams, D. A., "Design of Concrete Mixtures," Bulletin No. 1. Structural Materials
Research Laboratory, Lewis Institute, Chicago, Ill.
[11] Swayze, M. A. and Gruenwald, E., "Concrete Mix Design--A Modification of the Fineness Modulus Method," Journal, American Concrete Institute, March, 1947; Proceedings,
Vol. 43, p. 829.
[12] Walker, S. and Bartel, F., discussion of paper by Swayze and Gruenwald, (Ref 11),
Journal, American Concrete Institute, Dec. 1947; Proceedings, Vol. 43, pp. 844-51.
[13] Weymouth, C. A. G., "Effects of Particle Interference in Mortars and Concretes,"
Rock Products, 25 Feb., 1933, p. 26.
[14] Weymouth, C. A. G., "A Study of Fine Aggregate in Freshly Mixed Mortars and Concretes," Proceedings, American Society for Testing and Materials, Vol. 38, Part 2, 1938,
p. 354.
[15] Weymouth, C. A. G., "Designing Workable Concrete," Engineering News--Record,
29 Dec., 1938, p. 818.
[16] Walker, S., "Relation of Aggregates to Concrete," Bulletin No. 2, National Sand and
Gravel Association, 1928.
[17] "Standard Guide Specifications for Concrete," Department of the Army, Corps of Engineers, 1963.
[18] Walker, Stanton and Bloem, D. J., "Effect of Aggregate Size on Properties of Concrete,"
Journal American Concrete Institute, Sept. 1960.
[19] Kellerman, W. F., "Effect of Type and Gradation of Coarse Aggregate Upon Strength
of Concrete," Public Roads, June 1929.
[20] Blanks, R. F., Vidal, E. N., Price, W. H., and Russell, F. M., "The Properties of Concrete
Mixes," Journal, American Concrete Institute, April 1940; Proceedings, Vol. 36, p. 433.
[21] Tyler, I. L., "Concreting of Norris Dam," Bulletin, Engineering and Construction Department, Tennessee Valley Authority, Sept. 1937.
[22] Gilkey, H. J., "Size, Shape, Surface Texture, and Grading of Aggregates," A S T M STP
22.4, American Society for Testing and Materials, 1943, p. 92.
[23] Higginson, E. C., Wallace, G. B., and Ore, E. L., "Effect of Aggregate Size on Properties
of Concrete," Publication SP-6, Symposium on Mass Concrete, American Concrete
Institute.
PRICE ON GRADING
583
[24] Price, W. H., "Grading and Surface Area (of Aggregates)," ASTM STP 169. American
[25] Price, W. H., "Grading and Surface Area (of Aggregates)," ASTM'STP 169,4, American
[26] Mercer, L. B., "Gap Grading for Concrete Aggregates," Journal, American Concrete
Institute, Feb. 1943; Proceedings, Vol. 39, p. 309.
[27] Li, Shu-tien, "Proposed Synthesis of Gap-Graded Shrinkage-Compensating Concrete,"
Journal American Concrete Institute, Oct 1967; Proceedings, Vol. 64, p. 654.
[28] Concrete Manual 8th ed., Bureau of Reclamation, U.S. Department of the Interior,
Washington, D.C.
[29] Cordon, W. A. and Gillespie, H. A., "Variables in Concrete Aggregates and Portland
Cement Paste Which Influence the Strength of Concrete," Journal, American Concrete
Institute, Aug. 1963; Proceedings, Vol. 60, p. 1029.
[30] Schmidt, William and Hoffman, E. S., "9000 psi Concrete--Why? Why not?" Civil
Engineering, Vol. 45, No. 5, May 1975, p. 48.
[31] Mather, Katharine, "High-Strength, High-Density Concrete," Journal American Concrete Institute, Aug. 1965; Proceedings, Vol. 62, p. 951.
[32] "Tentative Interim Report on High Strength Concretes," Journal. American Concrete
Institute, Sept. 1967; Proceedings, Vol. 64, p. 556.
[33] Gaynor, R. D., "High Strength Air Entrained Concrete," Publication No. 17, National
Sand and Gravel Association and National Ready Mixed Concrete Association, March
1968.
[34] Guide and Recommendations on Aggregates for Concrete for Large Dams, Bulletin 18,
Commission Internationale Des Grand Barrages, 11, Rue de Teheran, Paris VIII, France.
M. A. Ozol 1
Chapter 35 Shape, Surface Texture,

Surface Area, and Coatings
Introduction
The equidimensionality and angularity of coarse and fine aggregate

particles, the shape of the relief and the degree of relief on their surfaces,
the amount of surface area caused by the kind and degree of relief, and the
chemical and physical nature of coatings, if present, are all characteristics
that significantly affect the properties of both fresh and hardened concrete.
Differences in particle shape and surface texture, as between different
rock types, will affect the bulk void contents and frictional properties of
the aggregates and, in turn, the ingredient proportions, mechanical properties, and economy of the unhardened concrete mixtures made from them.
In the hardened concrete, the compressive and flexural strength, the
elasticity, and the distribution of stresses within the concrete are influenced
by: particle orientation, the degree of adhesion between the particles and
cement paste, and the mechanical behavior of the cement-aggregate interface. Those properties are influenced individually, and in combination, by
the equidimensionality, angularity, surface texture, surface area of, and
coatings on, the coarse and fine aggregate particles.
Coatings can be responsible for various detrimental physical or chemical
effects in both plastic and hardened concrete, depending on their physical
nature and mineral, or chemical, composition. Coatings can also be innocuous or, conceivably, beneficial. Their effects may relate to adhesion
between particles and matrix, chemical stability of the aggregates, and
contribution of excess material to the plastic concrete mixture.
ASTM Specification for Concrete Aggregates (C 33) does not explicitly
contain any limitations on shape, surface texture, surface area, and coatings. However, the specification does have limitations on certain substances
which, prior to the application of the test procedure for their determination, might have been present originally as coatings. 2
1Principal research scientist, Martin Marietta Laboratories, Baltimore, Md. 21227.
2Organic impurities, materials finer than 75 /~m (No. 200) sieve, clay lumps and friable
particles, and deleteriouslyalkali-reactivesubstances.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
OZOL ON SHAPE
585
The ASTM Test for Index of Aggregate Particle Shape and Texture
(D 3398) is used for determining a numerical index of aggregate particle
shape and (surface) texture based on the weighted average void content of
specified sizes which have been subjected to two different compactive
efforts. There are no ASTM standard test methods for determining the
surface area of particles, individually or in bulk. ASTM Recommended
Practice for Petrographic Examination of Aggregates for Concrete (C 295),
calls for the determination of whether exterior coatings are present on
aggregate particles and, if they are, their identification by standard mineralogical techniques.
Although present ASTM specifications are silent regarding any quantitative limits to be placed on, for example, particle shape, as measured by any
present direct (for example, caliper [1,2] 3) or indirect (for example, void
content, ASTM Test D 3398) method, it does not necessarily follow that
this aspect of the specification will, or ought to be, changed when sufficient
accumulated research and testing information are available from which to
extract and organize numerical limits on those properties for specification
purposes. Rather, investigations in past years, in which, for example, the
shape of the aggregate particles has been intentionally or incidentally
a factor, demonstrate that concrete suitable to a purpose can be made with
coarse and fine aggregates of very different sphericity, roundness, surface
texture, surface morphology, and surface area [3-9].
Concrete made with coarse or fine aggregates of rough surface textures
with less equidimensionality and greater angularity than, for example,
most siliceous natural sands and gravels, may need mixture proportions
(that is, combining admixtures as needed) tailored to the aggregate characteristics. The exercise of that tailoring for the determination of the most
suitable grading and proportions of coarse aggregate, fine aggregate,
cement, water, and entrained air for the most economical mixture for a
particular purpose will work toward the fullest utilization of the natural
geological materials resource. Moreover, where conditions favor, it will
lead to the best use of substitute, alternative, solid waste, or other unconventional materials, which often are of irregular texture and shape. Thus,
although present ASTM standards mention almost no tests useful for the
comparative performance evaluation of aggregates differing in particle
shape, surface texture, and surface area, the development of such tests
would further the widest and best utilization of aggregates that vary in
those properties. The background research and comparisons, which will
have been prerequisite to test development, will in turn help to elucidate
further the relationships between concrete properties and the aggregate
properties of surface area, surface texture, and particle sphericity and
angularity. Taken altogether, the foregoing will be germane and comple3 The italic numbers in brackets refer to the list of references appended to this paper.
586
mentary activites, which will address the concerns of interested agencies

and individuals that "good" aggregates are becoming less available and
that therefore aggregate materials, both natural and artificial, heretofore
considered marginal in various physical and chemical properties, need to
be reexamined with respect to those properties in light of their anticipated
future value [I0].
The fundamental attributes of particle geometry are shape and surface
texture. The surface area of an individual particle, or the total surface
area of an aggregate, is a quantity that is fixed by shape and surface texture when the sizes of the particles are fixed.
Coatings are a separate issue, in practice unrelated to questions of particle shape, surface texture, and area (especially with regard to crushed
stone), but in principle having a subtle correlation: the kind of intrinsic
coating present on a particle and its shape, surface texture, and area can
each be related back to the geology of the deposit (especially with regard to
sands and gravels). The surface texture and the amount of surface area
present may then have controlled both the volume of the coating and the
degree of adherence of the coating to the particle at the same time that
other geological factors specific' to the deposit determined the mineralogy
or chemistry of the coating.
Shape
Definitions, Theoretical Discussions
The concept of particle shape incorporates three geometrical ideas,
namely, sphericity, roundness, and form, which are distinct and separately
definable properties in the abstract or mathematical sense. They may be
linked properties in the geological sense, in that a process which affects the
expression of one property may, concurrently, promote or inhibit development of the others.
One investigator [11] includes a fourth property, surface features ( = surface texture) under the broad heading of particle morphology. The discussion of shape herein is concerned with shape from the standpoint of
particle proportions and overall object geometry. And, insofar as the surface texture of a rock particle is more particularly a function of the way
the intrinsic mineralogical properties cause the material to fracture or to be
modified by chemical or physical weathering, the subject of surface texture
is treated separately. That is, in principle, the stone or aggregate material
could be carved or "sculpted" into any shape, all of which would have the
same surface texture--if shaped by the same tools and methods.
Sphericity is a measure of how nearly equal are the three axes or dimensions of a particle, based on the degree to which the volume of a par-
OZOL ON SHAPE
587
ticle fills the volume of a circumscribed sphere whose diameter is the maximum dimension of the particle.
Roundness is a measure of the sharpness of the edges and corners of a
fragment, or the degree to which the contour of a particle fits the curvature
of the largest sphere that can be contained within the particle.
Sphericity and roundness can be visualized conveniently, but not defined
rigorously, by the analogy of an irregular solid within which is a sphere of
the largest possible size and around which is a sphere of the smallest possible size. The spheres may or may not be concentric. The congruence of the
particle boundary to the inner sphere is an indication of roundness and its
coincidence to the outer sphere, the sphericity. When the spheres coincide,
the particle itself is a sphere with both a roundness and sphericity of 1--the
maximum value.
The form of Folk [11] or the alternative shape factor of Aschenbrenner
[12] is a measure of the relation between the three dimensions of a particle
based on ratios between the proportions of the long, medium, and short
axes of the particle, or of the smallest circumscribing ellipsoid. Form or
shape factor distinguishes between particles of the same numerical sphericity
but of different axial proportions.
Various alternative mathematical expressions, almost all based on the
geometric properties of a sphere as the standard reference form, have been
proposed and used for sphericity. If s is the surface area of a hypothetical
sphere of the same volume as a particle, and S is the actual surface area of
the particle itself, then, since a sphere has the least surface for a given
volume, the ratio s / S is the sphericity [13] ( = 1 for a spherical particle
and ---- 0.806 for a cube).
Proportions or dimensions are more directly determined than surface
areas, and an expression more practically suited to measurement is: Sphericity ff = D , / D c s , where D, is the diameter of a sphere of the same volume
as the particle (nominal diameter), and De, is the diameter of the circumscribing sphere (equals the largest dimension of the particle) [14]. That
relation derives from the expression--developed by Wadell [13] and taken
as the fundamental equation--relating the volumes in question, 4 that is,
3
sphericity = "~/-Vp/Vcs, where lip = actual volume of the particle, and Ve,
is the volume of the circumscribing sphere of diameter De,.
Vp may be measured by displacement in water, and Vc, calculated from
the direct measurement of Des. D, may be calculated from Vp ( D , =
4The diameter of the sphere of volume Vp is D,; the diameter of the sphere of volume
Ves is Des.
S(r = ~Vp/V~s --
r/6D, 3
~. 7r/6D ~s3
= D n/Dc~
588
~rVJ0.806). Essentially, the same measurement may be accomplished

without direct physical determination of V;. Krumbein [15] used the
product of the long (l), intermediate (i), and short (s) dimensions (as if
the particle were enclosed in a circumscribing triaxial ellipsoid rather than
a sphere) to approximate D , 3 and to define sphericity as
S,~b = 3~ (Tr/6)l'i's(Tr/6)13 = 3 ~
i.sl2
Literal interpretation of the numerical value of sphericity as the simple

proportion to which the volume of an irregular object fills its circumscribing sphere is, by the above definitions, faithful only as Vp approaches V,.
For low relative values of Vp, the sphericity is "high" for particles that
occupy only a small fraction of Vc,. Use of the cube root of the volumetric
ratio gives a sphericity value of 0.46 for a particle that is only one tenth of
the volume of the circumscribing sphere.
Table 1 illustrates that for some values of Vp relative to Vcs at the unit
volume.
T/~BLE l--Wadell sphericity values corresponding to the ratio Vp/Vcs.
Vp
Vcs
D,,"
Dcsb
0.1
0.5
0.9
1.0
1.0
1.0
0.575
0.984
1.197
1.24
1.24
1.24
Wadell Sphericity
S = Dn/D~s
Vp/vcs
0.464
0.794
0.96,5
0.1
0.5
0.9
"Diameter of sphere of volume Vp.

bDiameter of sphere of volume Vcs.
Note that from S = D , / D , , or ~Vp/V,.,, a cube has a sphericity of
0.716 since De, is equal to v~, the long diagonal of the cube, and D, =
~f6Y~r. By the Krumbein formula, the sphericity of a cube is around 0.9 if
l, i, and s are considered to be determined by direct measurement (that is,
I = x/3, s : x/2, i is ~ 1.6) and is 1.0 if l, i, and s are measured from the
circumscribing ellipsoid which in this case is a sphere.
Sneed and Folk [16] define another measure of sphericity that relates
particularly to settling behavior in a fluid, but may possibly be of some
relevance regarding particle behavior in plastic concrete. Settling velocity
is supposed to increase with particle sphericity, but Folk [11] points out
that a rod settles faster than a disk of the same volume although the respective Wadell sphericity values might indicate the opposite. Since particles tend to orient, in hydrodynamic behavior, with the maximum projection area, the plane with the l and i axes perpendicular to the direction
of motion of the fluid, they define maximum projection sphericity as:
OZOL ON SHAPE
589
~ss27~.i. This formula compares the maximum projection area of a given

particle with the maximum projection area of a sphere of the same volume.
Maximum projection sphericity was used in a convenient method developed
by Burke and Freeth [17] for determination of shape, sphericity, and size
of fragments, and by Ehrlich and Weinberg [18], and others [19,20], for a
mathematically detailed exact method for characterization of grain shape.
A projection, or cross section, of a particle, as in a petrographic thinsection, may be used to calculate the Riley Sphericity = D~Di/Dc where
Di is the diameter of the largest inscribed circle, and Dc is the diameter of
the smallest circumscribing circle [21].
The evaluation of sphericity ( = equidimensionality) by the volumetric
approach does not yield, especially at low numerical sphericities, unique
values for particles of decidedly different geometric proportions. A flat
oblate form with two dimensions very much longer than the third may
have the same numerical sphericity value as a long prolate form with two
dimensions very much shorter than the third (Fig. 1). Distinction between
the prolate, oblate, equant, and bladed forms is made possible by means
of manipulations of the axial lengths or proportions of the particle, which,
when used in conjunction with the Wadell Sphericity, can uniquely define
1.0
__
2~m
III
-- TABouR~
L~I
IV
~!ii~!i::e d
Roller
]0
Rod
,=.,
I
Bladed;
-Fl~:
='andElong~
12~
~.0
jSHORT I _I~ERMEDlATE
FIG. 1--Shape categories of Zingg (~-IV) defined by ratio of two thirds for i/l and s/i.
Synonomous descriptive terms are given for the four shape categories. The hyperboloid arc is
a line of equal (Wadell) sphericity.
590
the particle geometry. Thus, the long (l), intermediate (i), and short (s)
axes of the particle, or of the circumscribing triaxial ellipsoid of the Krumbein formula, may be used in various ways to define the related measures
of f o r m of Sneed and Folk [16], the shape factor of Aschenbrenner [12],
and the four shape classes of Zingg [22].
Zingg showed that the ratio of the intermediate to the long intercept
(i/l), when plotted versus the short to the intermediate intercept ratio (s/i),
can define four shape categories when fixed values are assigned to those
ratios. The ratio value used by Zingg, 2/3, and the resulting shape classes
are illustrated in Fig. 1.
Note in Fig. 1 that rounded and unrounded ( = zero roundness) versions
of the illustrated forms may have the same sphericity and shape values,
which would be defined by the intersection of a sphericity curve with a
point of given i/l and s/i values. Although the forms within the upper right
cell of the shape category diagram are designated by some authors as
equant, equidimensional, or equiaxial, and a cube is used as the illustration
of the zero roundness but high sphericity analog of the sphere, a cube is
not, strictly speaking, equidimensional; that is reflected in its Wadell
Sphericity value as calculated by the volumetric method.
Asehenbrenner [12] designed i/l the elongation ration, q, and s/i the flatness ratio, p. The ratio p / q is defined as the shape factor, F, which then is
equal to ls/i 2. Values of F > 1 represent prolate forms with i approaching
s; values o f f < 1 represent oblate forms with i approaching l.
This additional idea permits the gross particle geometry, excluding
roundness and surface texture, to be completely specified by (Wadell)
Sphericity and (Aschenbrenner) shape factor (see Fig. 2).
Sneed and Folk [16], using a triangular diagram, plot s/l versus (l -- i)/
(l -- s) (Fig. 3), producing a different scheme for the combined specification
of form and sphericity.
Roundness, or angularity, ideally is the ratio of the average radius of
curvature of the corners and edges of a particle to the radius of the maximum inscribed sphere, or to the nominal radius of the particle. For measurements, it is more convenient to use the two-dimensional projection or
section of the particle (as in Riley Sphericity), and that is the basis of the
definition. That is, roundness, as defined by Wadell [13], is the average
radius of curvature of all the corners divided by the radius of the largest
inscribed circle: p = E (ri/R)/N. A sphere has a roundness of 1. Pettijohn [23] points out that a cylinder capped by two hemispheres also has a
roundness of 1.
All orthogonal forms have a roundness of zero because of their infinitely
small radii of curvature. And non-orthogonal forms with planar surfaces
intersecting in sharp obtuse or acute angles also have roundnesses close to
zero with radii of curvature that are small and, more importantly, not easily
measured accurately. For that reason, quantitative roundness measurements
OZOL ON SHAPE
Disks IFiat)
Spherical
(Neither Flat nor Elongated).
"~
\ 1
,; o., l//U('
i'L'}(\\\~ \ ~ 't/\
~ '
-~ o,
~ /
I/'X\\\k"x
/%
'U
tJ6
(Fiat and ,-------i(~
EJongaled) 0'
591
~F
: 1.00
.~.aF : 1.50
----- F = 3.00
.,.,,Rods (Elongated)
~______...~(~l~
~ ~
,2:
0.4
~,------,--- ~
0.6
:0.8
:].01
'FLATNESS RATIO, p = sii
FIG. 2--Particle shape as deft'ned by Wadell sphericity (~b), and Asehenbrenner shape factor (F).
COMPACT
/ <>-/ ....
PLATY
.73
BLADES
.67
ELONGATE
L-'I/L- S- - ~
FIG. 3--Form triangle. Shapes o f particles falling at various points on the triangle are
illustrated by a series of blocks with axes of the correct ratio; all blocks have the same volume.
Independence o f the concepts o f sphericity and f o r m may be demonstrated by following an
isosphericity contour f r o m the disklike extreme at the left to the rod-like extreme at the
right. From Sneed and Folk [16]. ~ -= sphericity.
592
TESTS AND
PROPERTIES OF CONCRETE
AGGREGATES
at a level of accuracy and reproducibility that can discriminate realistically

small differences in roundness would best be done on sections or projections
that do not have sharp corners but do have a degree of roundness to begin
with. Then the definition might be better worded as: " . . . the ratio of the
average radius of curvature of the protrusions and convexities of the particle to the radius of its maximum inscribed circle."
Roundness and sphericity evaluations of individual particles are obtained
more commonly by use of visual guides which may assign numerical values
or descriptive terms, or both, to silhouette charts or photographic charts
for both pebbles and sand grains (Fig. 4).
0.9
>0
t~
hi
"1-
IP
(~
0.3
~
o.i
,0'
03
lID
-'l-
~
o.5
or
0.9
ROUNDNESS
FIG. 4--Visual chart for estimating roundness and sphericity of particles [24].
In principle, sphericity and roundness are unrelated geometrically, but

the same natural forces that produce roundness also produce sphericity,
and, in natural sediments--of the same sort as natural sand and gravel
deposits used for aggregate--investigations have found sphericity and
roundness to be correlated closely and also each to be a function of size,
though not necessarily to the same degree. Generally, coarser grains round
easier than finer ones, and roundness increases with size [11,23]. The
better rounded grains are the more spherical, but the relation statistically
is such that, in plotting roundness versus sphericity, a small change in
sphericity is correlated with a large change in roundness, for example, a
sphericity increase of 15 percent might coincide with a roundness increase
of 125 percent [23].
OZOL ON SHAPE
593
Sphericity increases with size but the relationship in naturally occurring

materials is less pronounced than with crushed materials. Experiments by
Thiel [25] on the abrasion of coarse and fine sand-sized materials, all at
the same starting sphericity, showed at the conclusion a clear correlation
between size and sphericity. In a study of the shape of various, as supplied,
coarse aggregates, including two gravels and three crushed stones, Conway [26] found that sphericity for all increased with increasing sieve size.
The form (including shape factor of Aschenbrenner) and sphericity of
both sands and gravels and crushed stone are primarily the result of: (1)
the degree of anisotropy of the material (the result of bedding planes,
schistosity, mineral cleavage, etc.) and (2) the original shape of the particle,
that is, the shape prior to the action of agents inducing roundness and
sphericity. Secondarily, they are, for sands and gravels, the result of factors
associated with transportation and abrasion and, for crushed stone, the
result of the action of crushing and sizing equipment on the shot rock of
the muck pile.
In the case of gravels, the original particle shape is controlled by largerscale structural features of the bed rock, such as joints, fractures, faults,
and bedding planes, which were then exploited by glacial action, stream
action, and mechanical (frost) and chemical weathering. In the case of
natural sands, the original particle shape, that is, its shape prior to transportation, was probably most strongly controlled by its mineral grain shape
in the parent material.
With crushed stone, the analogous "original" particle shape is that produced by blasting of the bedrock. Those shapes are a complex function of
the influence of the large-scale structural features of the bedrock but,
additionally, are a function of the size and frequency of the cracks initiated
by the action of shock waves on the smaller-scale flaws and discontinuities
of the rock--approximately of the order of magnitude of the grain sizes.
Most of the shot rock requires further size reduction by crushing, and the
final shapes of the particles are influenced strongly by preferred orientations
of any sort contributing to anisotropy. Rocks with bedding, cleavage,
schistosity, shaly partings, mineral cleavage, etc., may yield particles more
elongated or flattened than equidimensional. Equigranular rocks with no
preferred orientations tend to yield more cubical particles on crushing. The
characteristics of crushing equipment also influence the particle shape of
the rock being crushed. Generally, the greater the reduction ratio, the
more flat or elongated the product--slightly more so if the machine is of
the compression (jaw, gyratory, or cone crusher) type. That effect is less
pronounced with impact, or impeller, type machines, which would produce
more nearly cubical particles at equivalent throughputs.
Rounding of sand and gravel particles is chiefly the result of those geologic factors that were of secondary importance in contributing to their
form and sphericity, that is, the abrasion, attrition, chipping, rubbing, and
594
(possibly) solution incident to their transportation and to the site of deposition. The roundability of particular mineral or rock fragments depends
directly on their hardness and toughness and inversely on the presence of
cleavage or cracks, which would tend to induce fracturing, negating what
rounding had been accomplished.
Measurement
Methods that have been used at large for obtaining a particle shape
evaluation, or index, include: (1) visual comparison of the particle silhouette
or photograph with standard shapes; (2) measurement of the percentage of
voids at one or more compactive efforts; (3) sieving, by rectangular openings
(actually or in effect), of a fraction that has been sieved by square openings;
(4) measuring the time required for material to pass through an orifice;
(5) determination of the surface area to volume ratio; (6) measurement of
the radii, radii of curvature, and diameters of particles, or photographs of
particles.
When it has been desired to evaluate the shape of individual aggregate
particles in concrete research and technology, the methods used mostly
have been based on measurement of particle dimensions rather than their
volumes. When it has been desired to obtain an evaluation of the overall,
or statistical, shape characteristic of an aggregate, most of the methods
used are based on the principle that the volume of voids in the aggregation
changes according to the axial ratios of the individual particles. The volume
also changes according to particle angularities and ~urface textures, and
that aspect has been examined by Tons and Goetz [27] and Ishai [28] and
is discussed later.
The sphericity may be determined by the application of Wadell's defining equation, with individually measured long dimensions and weights
of the particles and volume calculated from the bulk specific gravity of the
aggregate. Conway [26] used this procedure on coarse aggregates, following
an earlier reported procedure of Huang [29]. Chamberlin [30] determined
sphericities according to Krumbein's formula for 25 sands by using a petrographic microscope to measure the lengths of the principle axes of 100
particles from the No. 14 28 size fraction of each. He determined the
average roundness value for the same fraction of the sands by assigning
roundness values of 0, 33, 66, and 100 according to whether the particles
had, respectively, all edges sharp, <50 percent of edges rounded, >50
percent of edges rounded, or all edges rounded. The range of sphericity
was from 0.68 to 0.78 and of roundness from 33.1 to 55.7. The sand with
the highest sphericity had the lowest roundness; the sand with the lowest
sphericity had a roundness of 49.7; the highest roundness sand had a
sphericity of 0.74.
The proportional calipers, as developed by the Corps of Engineers [1]
OZOL ON SHAPE
595
and Schulze [2], measure particles individually. In those devices, of slightly

different construction, when one dimension of the particle is gaged by one
set of jaws, the other automatically closes to a multiple of the measuring
jaw. For fine aggregate, there is a Corps of Engineers method [31] that uses
a stereoscopic microscope and requires the determination of length, width,
and thickness of 300 particles in each mesh size. Particles with length to
width (l/w) or width to thickness (w/t) greater than three are classed as
elongated or fiat, respectively; the same criteria are applied to the results
of the proportional caliper tests on coarse particles. Laughlin [32] evaluated
sand "shape" by a straightforward application of Wadell's roundness
formula, using enlarged photographs of grains retained on 2.36 mm, 1.18
mm, 600/~m, 300 #m, and 150 /zm (Nos. 8, 16, 30, 50, and 100) sieves.
His values ranged from 0.34 to 0.55 for a crushed chert and an artificially
rounded limestone, respectively.
The method for particle shape determination in British Standard 812,
Methods for Sampling and Testing of Mineral Aggregates, Sands and
Fillers, based on Markwick [33], is essentially an individual particle measurement and defines both a flakiness and elongation index. Particles in
each size fraction of the sieved aggregate are gaged for thickness (flakiness)
by determining whether whey will pass through an elongate slot 0.6 times
the mean of the upper and lower sieves of the particular fraction. They
are tested likewise for length (elongation) against a distance of 1.8 times
the mean of the sieves using pegs which are set the prescribed distance
apart on a board. The weight percent of the material passing the flakiness
gage is the flakiness index, and the weight percent not passing the elongation gage is the elongation index. The specification requires that the sum
of the flakiness and elongation indexes not exceed 35 percent. A method
used by the Ohio Department of Highways [34] is similar in concept. Flat,
thin, and elongated particles in, for example, the No. 8 by No. 16 (square
openings 2.36 by 1.18 mm (0.0949 and 0.0474 in.), respectively) fraction
are determined by passing the material over a sieve with 0.7112 by 3.556 mm
(0.028 by 0.14 in.) rectangular openings.
Burke and Freeth [17] used projected silhouettes of gravel particles
placed on a grid to measure dimensions for calculation of maximum projection sphericity. The length of the minor axis was obtained by rotating
the fragment about its major axis until the minimum silhouette width was
seen. Heigold and Lamar [35] experimented with a novel method, which
depends on availability of computer data processing. They measured the
intercepts of photographic silhouettes of 10 by 14 mesh grains of calcite
and 17 different limestones on a radial grid of 16 equally spaced rays (diameters) and used a computer to compare data obtained with reference
data on a variety of geometrical standards. The results were expressed as
the statistical correlation of each grain to the most similar standard shape,
as well as to the shape that the total sample most nearly resembled. CorreCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:36 EDT 2014
596
lation of the results with other existing measures of sphericity, roundness or

form was not attempted.
A related and mathematically more elaborate approach was developed
by Ehrlich and Weinberg [18] to describe grain shape. They carried out a
Fourier series expansion of the radius about the center of the mass, utilizing
coordinates of peripheral points. The measurements were made on the twodimensional maximum-projection-area-grain shape; the result is expressed
as a shape equation. Czarnecka and Gillot [19,20] developed a further
modification of the Fourier m e t h o d of Ehrlich and Weinberg to express
more precisely the total roughness of the particle profile as the sum of
two separately measurable shape and texture factors. The method was
applied to determine the effect of particle orientation on the measurement
of the shape and texture of particles exposed in precisely oriented sections
of two concrete prisms made with elongated glass rods and natural gravel,
respectively, as the coarse aggregates. Where the aggregate has an "extreme," that is, elongated and smooth shape, the orientation of the particle
cross section affects the values of the shape and roughness coefficients.
When the shape of the particles does not induce their preferred orientation,
the values do not differ significantly for differently oriented sections of the
concrete prism.
A different graphical approach was employed by Mackey [36] who compared the outline of the particle to that of a perfect ellipsoid and assessed
the angularity of its periphery by assigning values to interior angles of
points and angles subtended by arcs. The shape factor of the reference
ellipsoid is 1.0 by definition, and values accumulated for the various curves
and angles of the particle outline are subtracted from that value to give
the shape factor for the grain in question. Since a gross particle geometry,
that is, sphericity and shape factor, is assumed, the method assesses mainly
angularity and roughness.
Computer processing for automated analysis of grain images is described
by Mueller and Hunn [37] and others [38-40]. The necessary equipment is
marketed by several microscope or instrument manufacturers. Automatic
image analysis is applicable to sand- (or smaller-) sized materials, coarse
aggregate sizes, or to photographs and silhouettes of coarse or fine aggregates. Depending on the object sizes, the image for processing is obtained
by a television camera attached to a microscope or to a low-magnification
macro lens. The image is displayed on a cathode ray tube, and measurements of both shape and volumetric parameters are obtained by a variety
of electronic manipulations; scanning lines and the gray scale of the image
are the basic measuring tools. The estimation of roundness by this technique presents no theoretical problems; the estimation of sphericity and
form would require the same assumptions as the estimation of sphericity
from, for example, a two-dimensional chart, such as Fig. 4, or the standard shapes of Rittenhouse [41].
OZOI.. ON SHAPE
597
Direct physical measurements on individual particles, for practical reasons, are done most conveniently on coarse aggregates.. Bulk methods for
particle shape measurement are applied with equal convenience to both
coarse and fine sizes. Illustrating that category of test method are those
used by Gray and the National Crushed Stone Association [42], Wills [6],
Huang [29], Rex and Peck [43], Shergold [44], and Ishai and Tons [28].
On the basis that the interparticle friction of aggregates of irregular
particle shape (low roundness and high roughness) is great compared to
the internal friction of aggregations of smooth, spherical particles, Rex
and Peck [43] developed a test for comparing the flow time of a fixed
quantity of unknown sand through a 9.525-mm (3/8-in.) diameter orifice
with that of a standard smooth, rounded sand of the same size. The specific
gravity of the sands must be known or determined. The flow rate of the
sands (R), in s/100 cm 3 of absolute sand volume, is determined and expressed as a time index: Runk. . . . /Rstandard. A representative suite of time
index values was obtained with this method by Malhotra [45] for nine
natural and seven crushed sands. The lowest time index, excluding the
standard (Ottowa I11. sand) which was 1.0 by definition, was 1.05 (flow
rate = 13.3), for a rounded natural sand; the highest was 1.57 (flow rate =
19.9), for a laboratory-crushed coarse grained granite. Bennett and Katakkar [46] used the method in a study of 12 sands. Their results, expressed as s/lO0 cm 3 of solid sand volume rather than as time index,
ranged from 15.4 to 23.7, for a natural silica sand and a crushed stone
sand, respectively.
In the National Crushed Stone Association method, the percentage of
voids in specified size fractions (each tested separately) in a loosely compacted condition--produced when the sand falls into a cylindrical cont a i n e r - i s used as an index of particle shape. The greater the departure
from equidimensionality and roundness, the higher the void content. The
percent voids = 100 (1 -- (w/v.g)) where w = weight of sand in cylinder, v
= volume of cylinder (cm3), and g : bulk specific gravity of the aggregate,
determined independently on coarse particles of the same rock. The size
fractions used are 8 by 16, 16 by 30, and 30 by 50; the value reported is the
average of the determinations on the three fractions. Values for various
kinds of sands--natural sand, stone sand, and experimentally prepared
sands--ranged from 48 to 59 percent [42].
A similar method, applied to both coarse and fine aggregates, was used
by Wills [6], who collected the samples for void content determinations by
means of the orifice flow method of Rex and Peck [43]. For the fine aggregate tests, Wills used equal amounts of three sizes: 8 by 16, 16 by 30, and
30 by 50, combined. Void contents were determined by weighing each size
first at loose compaction, produced by the fall of the material from the
orifice into the container, and then at dense compaction, produced by
vibrating the loosely compacted sample, replenishing with additional aggreCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:36 EDT 2014
598
gate, and again striking off and weighing. Void values at loose compaction
ranged from 50 percent for a partially crushed mica schist sand to 40 percent for a well-rounded smooth glacial sand. At dense compaction, the
same materials were 41 percent and 32 percent, respectively. For assessing
particle shape and surface textures of coarse aggregates, a large cone
(571.5 mm (22 1/2 in.) high), sufficient to hold approximately 59 kg (130 lb)
of sample, was constructed. The coarse aggregate samples for testing contained equal amounts of four sizes: 25.4 by 19.1 mm, 19.1 by 12.7 mm,
12.7 by 9.525 ram, and 9.525 to 4.75 mm (1 by 3/4 in., 3/4 by 1/2 in., 1/2
by 3/8 in., and 3/8 to No. 4). Void content at dense compaction was determined independently by ASTM Test for Voids in Aggregate for Concrete
(C 30). Void contents at loose compaction ranged from 47 percent for
crushed limerock to 39 percent for a smooth-surfaced, well-rounded gravel;
voids at dense compaction were 42 and 33 percent, respectively.
Wills expressed orifice flow rate results in terms of time/absolute volume
of material rather than as the time index of Rex and Peck [43]. The correlation between the two methods was excellent (coefficient >0.9); roughtextured, angular aggregates had the highest orifice flow times and void
contents; whereas, the opposite was true for smoother rounded materials
for both coarse and fine aggregates.
A test method based on void content and orifice flow similar to the NCSA
and Wills methods is used in New Zealand [47] and employs the complete
gradation of the sand.
In ASTM Test D 3398, generally after Huang [29,48], the index (Ia)
obtained is a function of the original percentage of voids and the change in
the percentage of voids produced in a single-sized sample at two compactive
efforts. After the sample is separated into nine sizes between 19.05 mm
(3/4 in.) and 75 #m (No. 200) and the bulk dry sp gr of each size is determined, the percentage of voids for each fraction is obtained by filling the
mold in three courses, each rodded by 10 strokes, and determining the net
weight of the aggregate. The same procedure is repeated with 50 strokes to
obtain the higher degree of compaction. Percentage of voids is calculated
from: V, percent = [1 -- ( W / S . V)] 100, where W is net weight of aggregate at either 10 or S0 tampings per layer, S ---- bulk sp gr, and V = volume of mold (ml). The particle index,/~, relates the two void percentages
by the expression Ia = 1.25 Vlo -- 0.25 Vs0 - 32.0. The constant 32 was
derived empirically and represents the porosity of smooth uniformly sized
spheres at zero compactive effort.
A related technique by Shergold [44] computes angularity number:
A N = 33-- 100 [1 -- (w/c.g)], where w = weight of dry aggregate in the test
cylinder, c ---- weight of water to fill cylinder, and g is apparent sp gr of
the aggregate. The A N thus indicates the percentage voids ratio in excess
of 33 percent, the value found by Shergold for very well-rounded, singlesized beach gravel. Bennett and Katakkar [49] measured the (Shergold)
OZOL ON SHAPE
599
angularity number for twelve sands of different geological and shape characteristics, but all evidently more angular than the most angular reported
by Shergold. The A N values ranged from 15 to 30, for natural quartz sand
and crushed stone sand (presumably igneous rock), respectively.
In British Standard 812, the method for A N is drived from Shergold's
work but calculates the angularity number by the formula:
A N = 67--100 [A/(V S)]

where
A = mean mass of compacted aggregate in cylinder,
V = volume of cylinder, and
S = sp gr of aggregate.
Although it nominally expresses aggregate angularity, the Shergold procedure basically assesses the same sort of property as the previously described
particle shape tests which used void content or flow time. No matter what
the form of the results of all these methods, the product always includes
contributions traceable to angularity and surface texture of the particles, as
well as to their sphericity and form (or shape factor). However, there is
evidence that the dominant characteristic affecting void volume in the
Shergold procedure, and by analogy in the Huang method, is particle
angularity [50,44,27]. Bennett and Katakkar [49] stated that the three
measures--time of flow, angularity number, and specific surface--were
closely interrelated, and any one may be considered to be a satisfactory
index of particle shape.
An advantage of measurements of individual particle-shape characteristics
is that the kind of deviation from equidimensionality may be isolated and
identified as to size fraction and lithology, should that be desired. The
bulk methods, basically void content methods, do not conveniently allow
that distinction. Numerous combinations of sphericity, form (or shape
factor), roundness, and surface texture may yield the same numerical void
content. An aggregate of rough-textured but equidimensional particles
could conceivably have the same void content as an aggregate of flat and
elongated but smooth-surfaced particles. However, bulk methods of particle
shape evaluation are germane to the practical task of proportioning the
ingredients for concrete because the voids ratio of the particular coarse
aggregate, or combined coarse and fine aggregate, bears directly on the
mortar or paste requirement.
Explicit relationships quantifying the extent to which surface texture and
the three attributes of particle geometry separately influence the voids ratio
of an aggregate do not yet exist. However, Tons and Goetz [27], recalling
the definition for sphericity of Krumbein [15], state, "Geometry of irregular
particles can be characterized satisfactorily by an ellipsoid." They illustrate
600
mathematically that one-size ellipsoids pack like spheres, thus producing

voids percents i d e n t i c a l to those of spheres for each of the three ideal states
of packing. Data excerpted from Kaplan [ 5 0 , 5 l ] , who measured the elongation, flakiness, angularity numbers, and surface texture for a suite of 13
coarse aggregates, may be examined to deduce the stronger influence of a
factor other than particle proportions on the aggregate void contents (see
Table 2). It is evident from comparing aggregates K and E that both equiTABLE 2--Angularity number versus properties of aggregates.
Aggregate
K, crushed granite
A, natural quartzite gravel
E, crushed flint gravel
H, crushed basalt
F~
Eb
E + F
AN c
RoughnessFactord
9
8
34
31
!8
24
42
42
27
32
76
73
9
1
9
9
16.6
4.4
2.2
13.1
aAverage flakiness index.

bAverage elongation index.
CAngularity number (percent voids minus 33) ranged from 1 to 10 in [50,51].
dMethod of Wright [52] ranged from 2.2 to 16.6 in Refs 50,51.
dimensional and flat and elongated particles may produce equivalent void
contents. Aggregate A is more flat and elongated than K, which has next
to the highest void content in the suite, but A produced the lowest void
content in the suite because it is more rounded. Aggregates H and E have
almost the same characteristics except for roughness factor, indicating that
surface texture did not contribute to A N of H. It follows that flatness,
elongation, and rough surface coincide with high void contents only when the
flat, elongated, and rough particles are also angular.
Shergold's measurements [44] showed that decreasing the average sphericity of an aggregate by increasing the amount of flat particles from 20 to
40 percent produced little change in the percentage of voids, but changing
roundness had a direct effect.
On the basis of the above information, it may be reasoned that particle
angularity--the presence of (sharp) edges and corners--is a more important
determinant of bulk void content than axial proportions or surface texture.
Surface Texture and Surface Area
Definitions, Theoretical Considerations
Surface texture of an aggregate particle is the degree to which the surface

may be defined relative to arbitrary numbers as being rough or smooth
(loosely referring to the height of the asperities) or coarse grained or fine
OZOL ON SHAPE
601
grained (loosely referring to their spacing). Additional elements of surface

texture, not easily incorporated into a concise comprehensive definition,
include the lateral and vertical periodicity or irregularity of the roughness,
that is, the statistics of the height distribution of the population of asperities
and their frequency of occurrence over an area, as well as their morphology.
Two independent geometric properties are the basic components of surface texture: (1) the degree of the surface relief, also called roughness or
rugosity, and (2) the amount of surface area per unit of dimensional or
projected area. The latter property, although it is the ratio of areas, has
been defined by Wenzel [53, 54] as the roughness factor, R = A / a , where
A is the true (real) surface area, and a is the apparent or projected area. A
rough surface (in the sense of degree of relief) does not, ipso facto, have
more surface area than a smooth surface of equivalent dimensions--it may
or may not. The two properties are not related functionally but depend on
the relative amplitude and frequency of the asperities on the surfaces being
compared. For example, in Fig. 5, the roughness of segment L~ is greater
I,
12
LI,
~I"
~T
LT
hi
FIG. 5--Independence of roughness and surface area: illustration of roughness factor.

Ll = L2, P = profile line.
than L2, that is, the asperities are twice as high, but the lengths of their
profile lines (P1 and P2) are exactly the same. By projecting the sides of the
smaller triangles to intersect the legs of the larger enclosing triangle (left)
it is seen by inspection that the total profile length of the smaller triangles
equals the profile of the enclosing triangle. If the smaller triangles were
microscopic, and the base of the large triangle appeared as a smooth line,
the result would be exactly the same. Profile P2 is both rougher and has
more profile line (surface area) than P3. If the total length of the profile of
the larger triangles represents A, the true surface area, and L represents
the projected, geometric, or cross-sectional area, a, then roughness factor
R =- P / L : A / a . In this illustration R : 1.414 ( : x/2, since all triangles
are 45 deg). A more elaborate demonstration of the independence of roughness and surface area, using three dimensional models, is in Ref 55.
The criterion by which one surface is designated rougher than another,
in terms of their reliefs, is variously defined depending on the intended use
of the information, but almost invariably involves some measure of the
deviation of a profile of the surface from a hypothetical reference surface.
602
For example, the mean surface line is defined to be parallel to the surface
and at a depth such that it divides the peaks and valleys of the profile into
equal areas. Lines connecting the peaks of the asperities, or both the peaks
and valleys separately, have been used for other numerical definitions
(Fig. 6). A roughness value is determined from the mean surface line by
P,
Swedish gondord oppoims KI =0,0~5
dx~
(C) ["
P,
K z = 0.10
/~'T
=~Lyd-x
Avg Depth
FIG. 6--Mathematical definitions for numerical evaluation of surface texture [55].
averaging the height and depth of numerous equally spaced ordinates. The
simple arithmetic or center line average (CLA) may be used. Alternatively,
the use of the root mean square (RMS) average will emphasize the maximum peak to valley relief and will identify a surface with a wide statistical
range of ordinate measurements (peak to valley heights) as being rougher
than one with a smaller range but with the same CLA height.
Another analytical treatment of a roughness profile is used by Wright
[52] and Orchard [56] generally after the method of Kramrisch [57], and is
essentially the two-dimensional analog of the roughness factor of Wenzel
[53,54]. They distinguish between waviness ( : unevenness) and roughness:
OZOL ON SHAPE
603
(cx)
FIG. T--Analytical treatment of roughness profile.
waviness is the undulatory or long wave irregularity (Fig. 7(a)) on which is

superimposed the short wavelength irregularity or "roughness" (Fig. 7(b)) to
produce the resulting texture (Fig. 7(c)). The lengths of the unevenness line
and of the profile, that is, the roughness line, are obtained by intersecting the
magnified profile with chords of varying lengths. Alternatively, the
roughness line may be obtained with a map measuring wheel. The difference
between the two lengths, the ratio of the difference to the unevenness line, or
the ratio of the roughness line to the unevenness line, is used as a measure of
texture. The latter ratio corresponds to the roughness factor of Wenzel
[53,54] as the unevenness line flattens, or if its length is considered to
correspond to the projected or geometric surface. Orchard [56] favors the
profile length per unit of center line ( = mean surface line), and the profile
length per asperity, as easily obtained criteria for routine determinations.
Profiles of widely differing surfaces, markedly different in asperity shape
and coarseness, can have the same average depth, CLA or RMS height,
and therefore those, or related common measures, do not provide a unique
quantitative assessment of the surface roughness [58,59,56]. In view of
that, further refinement and sophistication of the mathematical treatment
of roughness profiles has been done which, it is stated by the authors,
permit unique numerical descriptions of surface topography [59, 58].
Tons and Goetz [27] and Ishai and Tons [28] present theoretical and
experimental approaches for expressing the geometric irregularity attributable to shape, angularity, and surface texture in a unified characteristic
related to individual particles and to their packing behavior in bulk. A
theoretical particle volume, the packing volume, Vp, is defined as the
volume enclosed by a dimensionless membrane that connects the peaks of
the surface asperities (Fig. 8). Loosely packed particles are in contact with
each other primarily at the high spots of their individual surface roughnesses. Therefore, the more mechanically significant space occupied by an
irregular particle within a mass of adjacent particles is not just its solid
volume, but is that plus the volume of the depressions between the high
604
TESTS A N D PROPERTIES OF CONCRETE AGGREGATES
Macro-Surface
Voids
Volume "
Membrane
' ~
~"
FIG. 8--Particle packing volume and macro-surface voids and micro-surface voids enclosed
by packing volume membrane [28].
spots or asperities, which are the points of contact with the neighboring
particles. There may now be defined a packing specific gravity, Gp =
W/Vp, where W equals the dry weight of the particle. The Gp of a solid,
smooth-surfaced sphere is equal to its apparent specific gravity, Gap. An
unknown packing specific gravity (Gpx) for an aggregation of one-size
particles is found from the relation G,x = (r~Wx/r,W,) 9 G,,, where Gp, is
the known packing specific gravity of a standard smooth-surfaced spherical
aggregate (marbles), and r~Wx and ~W, are the respective weights of the
unknown and the standard materials which occupy the same volume in
bulk.
The test which was developed, "the pouring test," is similar in operation
to the void content determination for a one-size material. The Gp, as
defined and obtained by this method, is numerically the lowest of all the
commonly used specific gravities. If G, is known, Vp may be calculated
from Gp = W/Vp, where W is the dry weight of the particle. The Gp is
further utilized, with the G,p, in the calculation of the specific rugosity,
S~, which expresses the volume of the surface voids of the particle as a
percent of the Vp of the particle or, volumetrically, the degree to which test
particles are irregular compared to smooth uniform spheres. The authors
[27,28] state that in this way all the geometric irregularity factors are accounted for and unified as the total surface voids under the packing volume
membrane. Sty can be partitioned into macrosurface voids and microsurface voids, Sty = S.,. + S.,i, to provide a better and more refined distinction between different types of geometric characteristics.
The total surface area of an aggregate of particles is controlled by their
size, shape (d/), and the roughness factor, R, (Wenzel [53,54]) component
of surface texture. The surface of the gross geometric particle shape, that
OZOL ON SHAPE
605
is, that due to sphericity (~b), corresponds to the dimensional or projected

area in the definition of R. For R of 1, a hypothetically smooth surface,
holding size constant, the surface area, that is, the geometric surface area
by definition, diminishes as the particle shape approaches equidimensionality. That area, expressed per unit weight (cm2/gm) or volume (cm2/cm3),
is termed the specific surface area (Powers [60]) or the "membrane" surface area (Tons and Goetz [27]). Specific surface is also applied to results
of direct physical determinations of true, or available, surface area by
sorption (BET) or permeability (Carman-Kozeny) methods.
Measurement
Profiles for analysis can be generated by electromechanical stylus devices
used for the investigation of metal surfaces, for example, the Talysurf
machine. An electromechanieal stylus apparatus especially for rock and
aggregate surfaces, which directly determines profile length per unit of
center line length, has been described by Orchard et al [61].
Wright [52] makes a thin section of an aggregate particle embedded in
resin and traces the profile of the magnified ( 125) image of the interface. This method was used by Kaplan [51] on 13 coarse aggregates used in
concrete. Scrivener and Hudson [62] describe a method for obtaining the
profile length of coarse surface textures using a device with spring-loaded
probes.
It would appear that direct assessment and comparison of roughness
could be accomplished by electron microscopical techniques which can
both construct the surface roughness and concurrently analyze it by computer [63]. The technique used by Boyde [64] on canine dental enamel
and sandpaper would be applicable to rock surfaces.
Surface texture may be evaluated by indirect methods. Tons and Goetz
[27] determined the volume between asperities, that is, between the packing volume membrane and the particle surface, by coating particles with
asphalt and then removing the excess down to the roughness peaks. The
result, termed rugosity, equals the volume of asphalt remaining on the
particle (cubic centimetres), divided by its "membrane", or geometric,
area (square centimetres) calculated from particle dimensions. Rugosity, by
this method, also may be influenced by angularity. The authors found
that, for crushed limestone and gravel, the rugosity increased with the
particle Vp. For a rounded gravel, the rugosity increased very little with
increase in Vp. Bikerman used a similar, that is asphalt residue, approach
on flat sawn rock slabs [65]. The weight of a fine powder used to "level" a
unit area of surface roughness also may be used. Measurement of the
resistance to air flow between a rock surface and an elastic membrane
held against it at a given pressure has been used as a measure of roughness
[66].
606
Based on the theory of Carman [67], specific surface area may be physically determined directly from measurements of permeability--the loss
of head and rate of flow--of a liquid through a column of single-sized
particles. Shacklock and Walker [68], using this method, determined
specific areas (square centimetres/cubic centimetres) for nine sizes each
of crushed granite, rounded gravel, and irregular gravel, which were also
used in concrete. The granite had a specific surface - 1 1/2 times that of
the rounded gravel for any particular size. The coarse sizes of the irregular
gravel had values similar to the rounded gravel. The same method was
used by Bennett and Katakkar [49] on one size each (B.S. --7, q- 14) from
twelve fine aggregates which were then used in concrete. Illustrative values
of specific surfaces obtained from [49] and [68] are shown in Table 3.
TABLE 3--Specific surfaces obtainedfrom Refs 49 and 68.
Specific Surface, cm-1
B.S. Size mm (in.)

19 by 12.7 mm
(aA by 1/2)
2.8 by 1.4 mm
(7 by 14)
Rounded
Gravel a
Irregular
Gravel a
Crushed
Granite ~
4.12
5.47
6.63
56.5
50
98.5
Crushed
Sandstone b
Natural Silica Sand b
. . . . . .
60
42
a Ref 68.
b Ref 49, the sands with the highest and lowest values.
Direct determination of true surface area may be done by adsorption

methods. Patat's method, using Gibbs Adsorption Isotherm, is based on
the principle of weighing the amount of adsorbed substance directly on the
substrate [69]. It is applicable to larger samples and can be used for finding Wenzel's roughness factor. Standard BET methods [70] can be, and
usually are, applied to fine particles, with the result expressed as area/unit
weight.
The above applications concern external surface; the adsorption and
permeability methods are also used for determination of the internal surface area of the total pore volume. Determination of true surface area of
rock specimens by the differential capacity method, used in electrochemical research for determination of true surface area of electrodes [71] might
be possible. The rock surface (for example, cross section of a small cylinder) would need to be coated with platinum, as for scanning electron
microscopy, and surface not to be included in the measurement (sides and
bottom) would need to be masked or insulated.
Without benefit of direct physical determination, the (geometric) specific
area may be estimated from the particle size, or sieve analysis, and bulk
OZOL ON SHAPE
607
specific gravity (neglecting angularity and assuming a spherical shape).

Stamenkovic [72] gives a table for the surface area, in' square foot per
pound, of various size fractions. If !, i, and s, axes for a particle, are measured, the simplified equation for area of a prolate spheroid may be used
[27], or the formula used by Chamberlin [30]: cm2/cm 3 = 2/i (i/l + i/s
+ 1). If sphericity (~b) is known, specific surface (cmZ/cm 3) can be estimated from sieve analysis by: cm2/cm a : (558/~b) (Pl q- 1/2P2 d- 1/4P3
+ 1/8P4..) where P~, P2, P3, are the solid volume fractions or the weight
fractions if the specific gravity is the same for all sizes. The specific surface
of spheres for Group 1 (100 200 B.S. sieves) is 558 [60].
Surface Textures of Rocks
As used in the petrographic sense to describe an interior cross section
of a rock, texture involves the following factors: (1) degree of crystallinity
( = mineral composition in the context of surface texture); (2) grain size;
and (3) the shape and relative sizes and arrangement of the mineral grains.
The surface texture should be considered to be very much controlled by
those same factors.
Also contributing to surface texture and therefore surface properties,
are the size, shape, and arrangement of places in the rock where there is
no mineral matter, that is, the voids or pores intrinsic to the rock or artificial aggregate that become voids or pores in the surface when the rock is
broken. Most rocks that are used for concrete aggregate are not intrinsically very porous, for example, porosities < 5 percent, but the surface
porosity will have the same numerical value and directly affect the surface
texture of the aggregate. Again, as with other textural elements, the particular size, shape, and distribution of voids in a porous surface may be
more important for the bond and other mechanical properties of the interface than the absolute numerical value of the porosity. Certain rocks,
slags, or expanded slags, shales, or slates, which are used in various locales
for aggregate, may have very rough or rugose surfaces, the character of
which is mainly attributable to their intergranular voids, for example, as in
pumice, scoria, tuff, and bioclastic limestones.
The (micro) surface texture of aggregates may be random, or various-for example, a granite--reflecting the irregular grain boundaries and size
and shapes of several minerals, each with its own characteristic cleavage or
fracture. In contrast, a fair amount of regularity or periodicity may be
present on broken surfaces of sedimentary or metamorphic rocks (Figs. 9
and 10), with strong cleavages, equigranular textures, or other sorts of
preferred orientations of the minerals.
The petrographic factors shaping surface texture are perhaps more
germane to the production of texture on crushed stone surfaces. The surface features of sands and gravels are rather different and bear the results
608
FIG. 9--Fracture surface of crystalline dolomite rock used as aggregate. Ledger fbrmation.
Pennsylvania. Field is 17. 5 #m across, Note sharp asperities and ridges Jormed by intersections
and edges of cleavage planes.
of preferential solution, etching, leaching, precipitation, and so forth.
Although having less vertical relief, it is obvious from Figs. 11 and 12 that,
on a microscale, quartz sand grains can have surfaces with abundant
projections, indentations, pores, pits, grooves, etc., and thus with surfaces
which, in concrete, are also significant for adhesion and, to some extent,
for mechanical interlock.
Effects and Significance of Properties

Plastic C o n c r e t e
S h a p e a n d T e x t u r e - - I n fresh concrete the particle shape, angularity,
and surface texture of the aggregate affect the volume of cement paste
that must be used first to fill the voids in the aggregate and then to provide
sufficient extra volume of cement paste to "float" the aggregate and provide the mass with enough plasticity to achieve thorough mixing and dispersion of the ingredients. Powers [74] has noted that the volume of concrete exceeds the volume of the compacted aggregate by between 3 and 10
percent, and usually by not more than 3 percent when no air-entraining
OZOL ON SHAPE
609
FIG. lO--Fraeture surface of pure. fine-grained limestone used as aggregate. St. Paul
Formation. Pennsylvania. Field is 17.5 ~m across. Note equigranular texture.
agent is used. Therefore, the "sufficient extra volume of paste" needed

for workability would be that volume equal to 3 or somewhat greater
percent of the outside volume of the aggregate mass. Concretes with angular, rougher, and irregularly shaped aggregates would be those requiring
the relatively higher excess paste volumes, because fewer coarse aggregate
particles can be packed, without interference, in concrete when the particle
shape is irregular than when it is spherical.
From data in Wills [6], Shergold [44], and Kaplan [50,51], it may be
estimated that, at a given grading of combined aggregate, the maximum
range in void content at dense compaction between aggregates at the extremes of the shape and texture spectrum might be around 8 to 9 percentage points. In practice, the average difference encountered should be
less because of the unlikelihood that two aggregates being compared would
be at the opposite extremes of the shape and surface texture spectrum.
The effect of the shape characteristics of the aggregate on the workability of the concrete and the concomittant effects on the hardened concrete, depending on the amounts and proportions of the cement and water,
may be investigated experimentally or empirically for either the workability
as the variable or the properties of the concrete at a given workability and
cement paste volume as the variable. In any case, whatever the paste
610
FIG. 11--Deep surjace etching on silt grain from consolidated deposit, USSR. Field
68 gm across. Note extensive surface pitting and etching with occasional deeper pores. Atlas
of Quartz Sand Surface Textures [73] used with permission.
volume needed for satisfying the two requirements of void filling and workability, it is evident that it may be achieved by adjusting the quantities of
cement and water to either maintain the same water/cement (w/c) ratio or
to increase or decrease it.
For example, working to the same w/c ratio (0.50), it may be calculated
from quantities reported by Blanks [ 75] that there was about an 8 percentage point difference in paste volume required--the undifferentiated
amount for both void filling and workability--between two concretes with
the same maximum size coarse aggregate 38.1 mm (11/2 in.), but one made
with a "well rounded" and the other with a "harsh and angular" sand.
The first concrete needed 99.8 kg (220 lb) of water and 261.6 kg/m 3 (4.69
bags/yd 3) of cement, the second 136.1 kg (300 lb) and 356.4 kg/m 3 (6.39
bags/yd3).
OZOL ON SHAPE
611
FIG, 12--Quartz crystal growth on sand grain f r o m beneath till, Roslyn, l"4,Y, USA. Field
is ~ 150 tzm across. Crystal faces have Jormed as upturned cleavage plates, by precipitation.
Atlas of Quartz Sand Surface Textures [73] used with permission.
Alternatively, Laughlin [32] determined the varying water requirement

to provide a constant mortar flow (100 percent) and concrete slump (3 in.)
according to the measured (Wadell) roundness values of his sands. Five
sands, one natural, three crushed limestones (one artificially rounded),
and one hammermilled chert were graded to the same fineness modulus
and used in concrete with fixed coarse aggregate and cement 334.6 kg/m 3
(6 bags/yd 3) content. The roundness values ranged from 0.55 (rounded
limestone) to 0.34 (chert); the respective w/c ratios of the concretes at
7.62-cm (3-in.) slump were 5.57 and 6.00. The water requirement to produce the same mortar flow (100 percent) for the most rounded sand (0.55)
was 84 percent of that for the least rounded (0.34).
Not varying anything, but allowing the effect of particle shape and surface texture on the plasticity of a mortar mix at constant sand gradation
612
and mix proportions to be expressed, Malhotra [45], using the time index
of Rex and Peck [43] as a measure of the particle shape and surface texture of nine natural and seven crushed sands, found that the flow of mortars (of exactly the same mix proportions) was related inversely to the time
indexes of the sands. The flow was 125 percent for the mortar with the
standard (highly spherical and well-rounded) Ottawa sand, time index
1.00; 117 percent for Leighton Buzzard (a rounded natural) sand, time
index 1.05; and 76.5 percent for a crushed coarse-grained granite, time
index 1.57. He concluded that the water requirement for constant flow
would vary as the time index, and that that index is a satisfactory method
for determining shape and water requirement.
Allowing the w/c ratio to change, Wills [6] showed the direct influence
of aggregate particle shape on the void content of the aggregate and, in
turn, on the concrete mixing water demand at constant cement content
and, through the mixing water demand, on the strength of the concrete.
The nine aggregates used by Wills had companion coarse and fine fractions from the same source. One, a natural quartz sand with a companion
gravel, was used for control aggregates, coarse and fine, with which the
other materials were combined as coarse and fine categories. (Incidentally,
void content and orifice flow measurements demonstrated that the fine and
coarse aggregates from the same source had similar particle shapes.) The
experimental design permitted a noteworthy distinction to be made between the separate effects on water demand and strength of the fine and
coarse aggregates. Overall, at constant grading and cement content, the
mixing water required to maintain a given slump increased directly with
the percent voids of the aggregates, and the strengths of the hardened
concretes decreased as the water requirement increased. Data for the concretes containing the coarse aggregates with the highest and lowest void
contents, with their companion fine aggregates, are shown in Table 4.
The fine aggregates had a much greater effect on concrete mixing water
demand and strength than did their coarse companions. A 4-percent increase in fine aggregate voids caused a 14.9 litre/m 3 (3 gal/yd 3) increase in
water demand--which lowered compressive strength 6.9 MPa ( - 1000
psi)--while a 3.8 to 5.7 litre (1 to 11A gal) increase was associated with a
similar voids increase of the coarse aggregate. Fine aggregates influenced
compressive strength almost entirely through their effect on mixing water
demand, whereas, for the coarse aggregates, only a part of the variation in
compressive strength was attributable to variation in mixing water requirement. Other factors, for example, elasticity, mineralogy, or bond strength
with cement paste of the coarse aggregate, were suggested as being as
important as the particle shape which had controlled the percent voids and
water demand.
Previously cited works and others lead naturally to the question of which
of the shape or surface properties of aggregates has the greater effect on
Cement,
kg/m 3 (bags/yd 3)
306.2 (5.49)
304.0 (5.45)
Voids percent
CA (1 in max)
dense compaction
41.6
33.4
190.0 (38.2)
160.0 (32.2)
Water,
kg/m 3 (gal/yd 3)
33.1
41.0
Sand, percent
of Aggregate
88.9 (3.5)
91.4 (3.6)
Slump,
mm (in.)
36.3 (5270)
39.3 (5700)
Compressive
4.4 (640)
4.8 (690)
Flexural
Strength, 28 days, MPa (psi)
TABLE 4 - - D a t a f o r the concretes containing the coarse aggregates with the highest a n d lowest void contents [6].
o~
6o
ol
0
z
o~
"1"
t'-
0
N
0
614
workability. It was earlier noted that angularity appears to have the

greatest effect on void content.
Holding grading, and aggregate to cement, and w/c ratios constant, and
using the same natural sand fine aggregate in all mixes, Kaplan [50] investigated the separate effects of flakiness, elongation, angularity, and
surface texture of 13 coarse aggregates, on the workability of concrete
as measured by the compacting factor [British Standard on Methods of
Testing Concrete (1881)]. All the concretes had sufficient mortar to fill the
voids. Although there was a wide range in the surface texture of the aggregates (roughness factors from 2.1 to 16.4 by the method of Wright [52])
no statistical relation could be established between it and the compacting
factor, and Kaplan concluded that changes in surface texture have no
appreciable effect on the workability of the concrete. Rather, he found
that changes in the angularity of coarse aggregates, as measured by the
AN, have the greatest effect on workability, with changes in flakiness and
elongation having decidedly less effect. There was some correlation between flakiness s and angularity but when the two were analyzed together
statistically, 29 percent of the variation in workability was due to flakiness
and 59 percent to angularity. If angularity and flakiness are varied separately: at constant flakiness an increase of 10 in the A N would cause a
reduction of about 0.07 in compacting factor, while at constant angularity,
an increase of 10 in the flakiness index would cause a reduction of about
0.01.
Kaplan suggested that, if workability is a concern, it is more important
to use aggregates that have low angularity than low flakiness; he gave
approximate quantitative relationships for estimating the compacting
factor from flakiness index and angularity number, separately or together.
From studies of hardened concrete, Popovics [76] reached a similar
conclusion regarding the greater importance of angularity than shape.
Based on the average mortar intercept--the measured distance along a
random line of the mortar layer between coarse aggregate particles--as a
criterion of workability, his results indicated that: there exists a minimum
value for the average mortar layer intercept 3.8 mm (0.15 in.) below which
the workability of the concrete is inadequate; the minimum intercept
depends on several factors, such as the method of compaction and grading
of the sand, and seems to be independent of the particle shape of the
coarse aggregate. It was found that the average mortar intercept was related inversely to the angularity number in concretes made with aggregates
of (visibly) different particle shapes--crushed reef shell, crushed stone, and
gravel--but with the same grading and mixture proportions. Popovics
SElongation was found to be well correlated with flakiness (0.922) and therefore only
flakiness was used in the statistical analyses; but the conclusions regarding flakiness also
apply to elongation.
OZOL ON SHAPE
615
considered that the results supported the Shergold test method for the
measurement of angularity number.
Surface Area--The relation between the physically-measured (permeability method) specific surface of aggregate and concrete workability was
examined by Shacklock and Walker [68] for mixes made with rounded
gravel, irregular gravel, and crushed granite. They found that the "effective" w/c 6 ratio for a particular workability was directly proportional to
the mean specific surface for mixes that were oversanded. But for mixes
of normal gradings, the relation was less direct. And for mixes of coarser
gradings, the relationship appeared to be inverted: a higher w/c ratio
being required, at a constant workability, as the specific surface decreased.
Bennett and Katakkar [49] similarly measured the specific surface of
twelve sands which were used in concrete at a fixed gradation with a standard coarse aggregate and fixed w/c and aggregate/cement ratios. They also
determined void contents (angularity numbers) and orifice flow times for a
single size of the twelve sands. They concluded that "the workability of the
various mixes was closely related to the specific surface of the fine aggregate," that is, the compacting factor increased as specific surface decreased. However, since AN and flow time were very closely correlated with
specific surface and varied directly with it, a similar conclusion could be
drawn with respect to those properties.
For 25 sands for which he had measured sphericity, and roundness, and
calculated (geometric) specific surface, Chamberlin [30] determined the
water required to bring mortars, mixed to constant volumetric proportion
between cement and saturated surface dry sand, to a predetermined consistency as measured by penetration of a standard cone. He found that
variations in water requirement were correlated only partially with specific
surface and sphericity (coefficients of 0.49 and 0.63, respectively). Introduction of roundness improved correlation with sphericity only slightly. He
suggested that, even allowing for substantial error in the determination of
surface area and sphericity, variations in surface texture of the sands
exerted a considerable influence on the water requirement variation. Similarly, Abrams and others earlier had observed that surface area of aggregate varied widely without appreciable difference in concrete strength, and
had noted that water required to produce a given consistency depended
largely on the character of the aggregate apart from its surface area. Excerpted data from Chamberlin [30] show, for example, that the sand with
the highest water requirement did not have the highest specific surface but
had the lowest roundness. The sand with the highest roundness had next to
the lowest water requirement (in source data as well) but had a mid-value
for specific surface (see Table 5).
~The "effective" w/c ratio takes into account the absorption and specific gravity of the
aggregate.
616
TABLE S - - W a t e r requirement versus sand surface and geometry.
Sand No. ~
Sphericity
Roundness
Specific Surface
Water
Requirement, ml
25
24
18
7
2
0.78
0.68
0.72
0.74
0.77
33.1
49.7
51.7
55.7
35.7
8.16
8.59
7.14
7.67
7.36
233
215
201
191
189
~Ref 30. The table is constructed from the highest and lowest values for each property
independently, with the corresponding values for the other properties.
Regarding differences in surface area between aggregates of the same

size, it appears that the higher paste requirement associated with higher
specific surface for particles of the same size fractions, where the particles
are of different kinds and also differ in angularity and sphericity, could be
attributed just as easily to increase in void volume.
Regarding increase of surface area as a consequence of decreasing particle size, it is experienced in the mixing of concrete that, as a practical
necessity, the paste requirement increases as the particle sizes of aggregates
of the same kind decrease. This is said to occur because of the increased
surface area of the aggregate to be coated. It has never been shown that
the extra paste requirement, if for coating of surface area, increases (after
void volume is deducted) in the same proportion as surface area increases,
that is, according to decrease in particle diameter. The void volume of an
aggregate of one-size spheres is the same for the same mode of packing
(for example, rhombohedral ~ 26 percent) whether the spheres are small
or large. Although the surface area of the smaller spheres will be greater,
all of it must be coated if the 26 percent void volume of the aggregation is
filled completely. It is not clear then how surface area alone leads to the
requirement for greater paste volume. It is suggested rather that it is the
greater number of points of contact between the greater number of particles, comprising the same total solid volume of material that require separation and therefore demand the higher paste requirement. The point
contact area between two (ideal) spheres is the same whether the spheres
are small or large. If it may be presumed that the minimum paste thickness requirement for separation is proportional to that contact area, then,
other things being equal, the minimum paste thickness requirement per
contact is the same whether the spheres are small or large. But, there are
many more such contacts in an aggregate of smaller spheres, each requiring separation and lubrication and thereby constituting a greater total
paste volume demand, than an aggregate of larger spheres of the same
total solid volume of material.
OZOL ON SHAPE
617
Hardened Concrete
Apart from the effects they exert on the strength of hardened concrete,
by means of their effects on the required paste volume and the proportions
of cement and water for a given workability, differences in particle shape
and surface texture directly influence the strength of hardened concrete of
the same mixture proportions. And, according to Kaplan's results [51],
flexural strength is more affected by those properties than compressive
strength; the effect becomes more significant as the strength level of the
concrete increases. Kaplan [51] used 13 coarse aggregates with both differences and similarities in angularity numbers, surface textures, and
flakiness and elongation indexes. These included three natural gravels,
which were used both uncrushed and crushed, and four types of crushed
stone, which were incorporated into three basic mixtures of different aggregate/cement and w/c ratios. For the intermediate mixture (cement/aggregate by weight : 1:7.53; w/c = 0.60) at 28 days, the highest and lowest
compressive (with corresponding flexural) and flexural (with corresponding
compressive) strengths with the aggregate shape and texture values are
shown in Table 6.
From regression analyses on all of the data, it was concluded that surface
texture was the most important aggregate property influencing compressive
strength, that is, the rougher the texture, the greater the strength, with the
effect increasing as strength increased. The elasticity of the aggregate was
the single most important factor affecting flexural strength, followed by
shape and surface, except that surface texture had the dominant effect for
the concrete at the highest strength level. The angularity number was used
in the regression analyses to represent the shape properties because it had
been shown that elongation was well correlated with flakiness, and flakiness
with angularity. Flakiness and angularity together accounted for only
slightly more variance than angularity alone. The variance due to each of
the properties as a percentage of the total variance accounted for by their
effect on the concrete strength, was given for the average of all concrete
mixes, as shown in Table 7.
For the concretes with the crushed and uncrushed natural gravels, Kaplan's data showed that, out of a total of 54 pairs of tests (27 each) of compressive and flexural strength covering different mixture proportions and
ages, the crushed condition of the same gravel produced higher strength
in 41 cases, lower strength in eleven, and equal strength in two. These
isolated data illustrate effects on concrete strength due to shape and texture
independent of the intrinsic lithologic properties of mineral chemistry and
modulus of elasticity, while altogether the data demonstrate the same effects
caused by lithology controlling the shape and texture.
It must be presumed that explanations as to why rougher texture and
increased angularity contribute to concrete strength at equivalent mixture
A,
N,
H,
E,
40.7 (5910)
44.8 (6500)
47.6 (6910)
37.8 (5480)
3.65
5.48
5.10
4.55
(530)
(795)
(740)
(660)
Compressivea
Flexural a
Strength, MPa (psi) Strength, MPa (psi)
a28 day.
b British Standard 812.
CAngularity number (percent voids--33) ranged from 1 to 10 for the 13 CA's.
dMethod of Wright [52] ranged from 2.2 to 16.6 for the 13 CA's.
natural quartzite gravel

crushed limestone
crushed basalt
crushed flint gravel
Aggregate
31
34
26
Flakiness b
24
33
42
42
Elongation b
TABLE 6 - - C o n c r e t e strength versus coarse aggregate shape a n d surface texture.
AN C
4.4
10.0
13.1
2.2
Roughness Factor d
(Surface Texture)
O~
-4
m
(O
:IJ
~3
h3
m
-4
I'll
0
0
z
0
O
-I1
l[
-o
~o
O
-o
rn
z
~D
-4
co
-4
m
Oo
OZOL ON SHAPE
619
TABLE 7--Proportional contribution of shape, texture and modulus of elasticity to concrete

strength [51].
Relative Effect, percent, of the FollowingProperties
Flexura] strength
Compressivestrength
Shape
SurfaceTextUre
Modulusof Elasticity
31
22
26
44
43
34
proportions will involve the contributions of: (1) mechanical interlock (due
to the surface relief) and (2) total surface area available for adherence of
the cement paste, due both to particle shape (geometric area) and roughness factor (true surface area per unit of geometric area).
Patten [77] investigated the relative contributions of adhesion and "keying," or mechanical interlock, by compressive and tensile strength measurements of concrete at five different ages in which the same coarse aggregate
had been used in both a "bondless," or surface-treated, and an untreated
condition. The object was to eliminate adhesive bonds without affecting the
mechanical interlock, or physical keying, between the aggregate and the
mortar; the coarse aggregate was coated with a mold release agent. The
elimination of adhesion between the mortar and the coarse aggregate reduced compressive strength by an average of about 23 percent. Compared
to the control, the strength of the bondless concrete decreased with age
from 19 percent difference, or reduction, at seven days to 27 percent difference at six months. Tensile strength was less affected, ranging from 6
to 28 percent difference, with an average reduction of about 17 percent;
but in this case with the strength of the bondless concrete increasing with
age as a percent of the control. At 6 months, in the high strength series,
the splitting tensile strength of the concrete with the "bondless" aggregate
was 92 percent of the strength of the concrete with the untreated aggregate,
whereas at 7 days it was 79 percent. Investigations along similar lines have
been conducted by Darwin and Slate [78] and others [79,80].
Studies focusing further on the components of the adhesive force making
up the percent contribution of adhesion to bond strength endeavor to test
the interface in as pure a state of tension as possible to avoid mechanical
contributions. Various specimen configurations and techniques for measuring aggregate-cement bond strength have been reviewed by Alexander
et al [81]. Under conditions of tension, the total adhesive force is the product
of: (1) the specific adhesion, that is, the strength of the adhesive force per
unit area of surface from whatever physicochemical mechanism it derives,
and (2) the amount of surface area available for bonding over which the
adhesive force acts. Different rocks have been observed to have different
tensile bond strengths to cement paste [81-84]. A possible explanation is
620
that the bond is chemical and its specific adhesion differs significantly
according to the particular chemistry of the mineral surface [85]. Alternative
possibilities are that (1) the bond is chemical but its specific adhesion on
the smallest unit area basis does not differ greatly between minerals, or (2)
the bond is physical [86], deriving from the same sorts of forces that hold
materials together in general. For either of these possibilities, an alternative
hypothesis to explain differences in bond strength between different rock
types is that surfaces of different lithologies have different roughness factors
and therefore present different true surface areas available for bonding [87].
Undesirable effects of flatness, elongation, and angularity of aggregates
have been noted. Flat particles may cause uneconomic grading and, based
on experimental work, may induce structural weakness within the concrete
if they are so oriented as to be loaded as columns in compression [88]. As a
consequence of their influence on the water content necessary to obtain
suitable consistency, they thereby affect, in addition to compressive strength,
bleeding, drying-shrinkage, and durability of the concrete--especially
freeze-thaw durability if excessive bleeding has caused the formation of
channels. Highly angular and flat and elongated particles are more subject
to fracturing and change in gradation by generation of fines than similar
amounts of rounded spherical particles.
Elongated or flaky particles may affect durability adversely if they are
oriented parallel to the concrete, for example, pavement, surface with
bleed water, and air voids having been trapped underneath. Good bond
may be inhibited [89] and any pores formed may be exploited subsequently
by weathering [90]. The mortar cover over fiat coarse aggregate particles
oriented parallel to the pavement surface may "fleck o f f ' when the particle
surface is micaeous or contains other minerals, for example, calcite, with
good cleavage directions oriented parallel to the surface. The flecking,
similar to the failure of a poorly adhering coating, is due to the stronger
bond between the mortar and the upper mineral surface than within the
mineral itself [91].
In certain instances where, for exploratory or practical reasons, concrete
has been fabricated with aggregates of low sphericity, and some degree of
angularity and surface roughness, authors have reported achieving satisfactory workability at acceptable w/c ratios or strengths. Mercer [92], using
aggregates of which 50 to 90 percent were fiat or elongated or both, made
concretes of adequate workability and a w/c ratio of 0.6. Popovics [3]
made concretes using, for example, (by weight) 60 percent crushed reef
shell coarse aggregate and 40 percent beach sand fine aggregate with 334.6
k g / m a (6 bags/yd a) cement and suitable amounts of flyash and air entrainment. Adequate workability was achieved, and strengths of 17.2 MPa
(2500 psi) compressive, 3.79 MPa (550 psi) fiexural, and 2.07 MPa (300 psi)
splitting-tensile were obtained at 28 days. The particle shape of the shell
coarse aggregate was described as highly unfavorable because of extreme
angularity: angularity numbers determined were in the range of 26 to 30.
OZOL ON SHAPE
621
Concretes were made by MeGhee [5] using a crushed slate coarse aggregate which had flakiness and elongation indexes of 57..2 and 68.3, respectively [26]. Adequate workability was achieved without deviation from
the proportions derived from the standard ACI method. The w/c ratios
were varied between 0.4 and 0.5. Corresponding compressive and flexural
strengths were 34.5 and 5.52, and 29.1 and 5.31 MPa (5000 and 800, and
4220 and 770 psi), respectively. Corresponding slumps were between 3.3
and 4.8 cm (1.3 and 1.9 in.) without a water-reducing admixture and 7.1
cm (2.8 in.) with the admixture (2.1 ml/kg (3 oz/bag)).
Coatings
Definition
Coatings are adhering materials which may be cemented strongly or
weakly to the particle surface and which are present as layers, blebs,
patches, or individual grains over large or sometimes only very small portions of the surface of coarse or fine aggregate particles. Coatings may
occur on sand and gravel, crushed stone, or artificial aggregates alike.
Their chemical and mineralogical nature may differ according to those
generic categories of aggregate.
The same substance may be both a coating and a granular component,
to the extent that it is present as both adhering and discrete material.
Further, the same, or closely related species of minerals may be at the
same time major components of the rock comprising the aggregate and
also coatings on its particles. Minerals in this category are those which
are both major rock-forming minerals and occasional "coating formers"
on particles of the same or of different mineralogy. Probably only two
materials, calcium carbonate (CaCO3) and silica (SiOz), are volumetrically
important in this dual behavior. For example, secondarily deposited opal,
chemically SiO2 may occur as a coating on pebbles of quartz or quartzite,
which are also chemically SiO2. Similarly, secondarily deposited CaCO3
may occur as an overgrowth on pebbles or grains of limestone or quartz, or
on other rock types.
In sands and gravels, what is for all intents and purposes a kind of
coating may result from weathering and leaching of the particle surface,
causing the formation of a "rind" which may be softer, more absorptive,
and more prone to degrade than the particle interior.
Identification and Composition

The most specific method of identifying the nature and extent of coatings
present on concrete aggregate particles is by direct, that is, petrographic
examination [93]. In general, of the two main categories of aggregates
most commonly used, namely, sand and gravel and crushed stone, the
622
former would be the more likely to have any coatings present that are
intrinsic to the materials--as a consequence of factors related to the particular geology of their formation.
Substances in this class, that is, those which might occur as coatings on
sand and gravel due to deposition or precipitation from waters, are therefore
mostly mineral in nature and may include calcium carbonate, iron oxides,
gypsum, and sulfates other than gypsum, for example, alkali sulfates,
misenite (KHSO4), thenardite (Na2SO4), opal, and phosphates. Other
substances such as clay, silt, and organic matter may be intrinsic coatings
on particles of the deposit in situ, but also, through carelessness or oversight, may become coatings on those aggregates at some stage of manufacturing, storage, or transportation subsequent to extraction.
In the case of crushed stone, it is in the nature of the integral bedrock
deposit and the subsequent quarrying and crushing operations that continually expose fresh surfaces, that intrinsic mineral coatings will not be
present as they may be in the unconsolidated or loosely consolidated deposits
from which sand and gravel are processed. Therefore, materials found as
coatings on crushed stone are more likely to be silt, clay, and organic
matter, or other substances not intrinsic to the deposit which have come in
contact with the material after extraction and crushing.
In the case of both coarse and fine crushed stone aggregates that are
processed incompletely or improperly, it may happen that dust of fracture-individual particles of which may be in the clay and silt sizes--adheres to
and coats portions of the aggregate particles. Although undesirable because
it may become an unanticipated and unwelcome contribution to the gradat i o n - a n d increase water demand--dust of fracture is not clay in the
mineralogical sense but rather, by definition, of the same mineralogy as
the rock deposit being produced. Goldbeck [94] reported that a decrease
of 1.5 percent flexural and up to 2 percent compressive strength can be
caused per percent of dust.
Both crushed stone and sand and gravel after manufacturing, but especially during transportation, are vulnerable to contamination by materials
which may adhere to portions of individual particles and form coatings.
In this case, it is less likely that the coatings will be of the natural mineral
type as for sand and gravel in situ. Later contamination which may form
coatings on crushed coarse and fine aggregate and sand and gravel can be
portland cement, asphalt, oils, solvents, or other organic materials.
Sigrlig~cance
Adherent cemented coatings can be detrimental physically if, in the
hardened concrete, they are bonded more strongly to the cement paste
than they are to the host particle. In that case, the coating/particle interface may be predisposed to separate as in the case of bond failure between
OZOL ON SHAPE
623
the particle/paste (or mortar) interface. In a related manner, poorly adhering dust, dirt, clay, and so forth, will impair the bond between cement
paste and aggregate. Wright [95] says that this effect is more likely to
occur with natural sands than with crushed stone sands.
Soluble, or chemically reactive, materials present as coatings will cause
the same variety of effects as if they were present other than as coatings.
Iron sulfides and other iron compounds can cause rust stains on concrete
surfaces, as well as interfering with bonding as previously described. Chlorides and sulfates can cause efflorescence and staining on concrete surfaces
and corrosion of reinforcing steel. Siliceous coatings of, for example, opal,
may be susceptible to alkali-silica reaction, depending on the alkali content
of the cement and other factors [96]. Organic materials may retard the
setting of cement. Soft coatings may cause surface pitting and reduced
abrasion resistance of the hardened concrete.
To summarize the behavior of coatings on coarse or fine aggregate particles, they may be: (1) composed of a deleterious or innocuous soluble
substance which dissolves during mixing of the concrete, (2) composed of
an innocuous or tolerable substance that is insoluble, but removable, thereby
altering the gradation somewhat and, if clayey, increasing water demand,
(3) innocuous chemically but bonded less strongly to their host particle in
the hardened concrete than to the enclosing paste or mortar, thereby constituting a bond or joint more prone to failure under physical stress than
other paste aggregate interfaces without interposed coatings, and (4) innocuous chemically and as strongly bonded to the host particle as is the
coating to the paste or mortar--in which case they are innocuous physically
as well.
Conclusion
Insofar as the naturally rounded mineral aggregates, sands and gravels,
are products derived from parent bedrocks, it is the case geologically that
the primary material is the more abundant and the derived material the
more exhaustible. It is doubtful that nature is manufacturing sand and
gravel at the quite remarkable rate at which convenient sand and gravel
deposits are being used up in rapidly developing parts of the world. They
are in the long r u n - - a n d in the short run for certain localities at present-a fixed resource. Bedrock, suitable for the manufacture of portland cement
and for the production of mineral aggregate is, geologically speaking,
virtually inexhaustible.
Despite the investment of energy in the manufacture of portland cement,
portland cement concrete requires the minimum of energy consumption for
its production relative to all other major construction materials, except
lumber. Therefore, it is reasonable to forecast increasing use of concrete
as a building material with increasing use of aggregates other than sand
624
a n d gravel. (Historically, in t h e U n i t e d States, t h e yearly t o n n a g e o f sand

a n d gravel has exceeded the p r o d u c t i o n of c r u s h e d stone. A b o u t 1972, a
crossover p o i n t was reached: in 1972 a n d 1973 the totals for s a n d a n d
gravel p r o d u c t i o n were, respectively, 822.6 a n d 885.6 metric tons (914 a n d
984 million tons); for crushed stone, 827.1 a n d 952.2 metric tons (919 a n d
1058 million tons).
It is the case t h a t the n a t u r a l l y r o u n d e d aggregates are easier to work
with a n d have p r e s e n t e d little technical challenge with respect to shape,
texture, a n d area. By their n a t u r e , they exhibit less variety in the expression
of those p r o p e r t i e s a n d raise fewer questions. T h e increasing use of crushed
stone for b o t h coarse a n d fine aggregate, along with p e r h a p s recycled concrete a n d other recycled materials, forecasts t h a t the effects a n d significance
of shape, texture, a n d surface a r e a will be m o r e p r o m i n e n t considerations
in the future.
Acknowledgments
T h e a u t h o r wishes to express his g r a t i t u d e a n d a p p r e c i a t i o n to Bryant

M a t h e r for his review of the m a n u s c r i p t a n d v a l u a b l e suggestions.
References
[1] "Method of Test for Flat and Elongated Particles in Coarse Aggregate," Designation
CRD-C 119-48-52-53, Handbook for Concrete and Cement, U.S. Army Engineers Waterways Experiment Station, Vicksburg, Miss., 1949, 1952, 1953.
[2] Schulze, K., "Rapid Method for Particle Shape Determination and Suggestions for
Definition of Ordinary and High-Quality Chippings," Road Abstracts, Vol. 21, No. 4,
April 1954.
[3] Popovics, Sandor, "Reef Shell Beach Sand Concrete," Living with Marginal Aggregates,
A S T M STP 597, American Society for Testing and Materials, 1976, pp. 97-113.
[4] Johnston, C. D., "Waste Glass as Coarse Aggregates for Concrete," Journal of Testing
and Evaluation. Vol. 2, No. S, pp. 344-350.
[5] McGhee, K. H. and Ozol, M. A., "An Evaulation of A Slate Aggregate for Use in Concrete," Virginia Highway Research Council, Charlottesville, Va., Nov. 1966.
[6] Wills, M. H., Jr., "How Aggregate Particle Shape Influences Concrete Mixing Water
Requirement and Strength," Journal of Materials. American Society for Testing and
Materials, Vol. 2, No. 4, Dec. 1967, pp. 843-865.
[7] Wills, M. H., Jr., "Lightweight Aggregate Particle Shape Effect on Structural Concrete," Journal. American Concrete Institute, Vol. 71, No. 3, March 1974.
[8] Imbert, I. D. C., "Influence of the Grading of Aggregates on Concrete Mix Proportions," HRB Record No. 441., Highway Research Board, 1973.
[9] Buck, A. D., "Recycled Concrete as a Source of Aggregate," Journal, American Concrete Institute, May 1977.
[10} Living with Marginal Aggregates. A S T M STP 597, American Society for Testing and
Materials, 1976.
[11] Folk, R. L., Petrology of Sedimentary Rocks. Hemphill's, Drawer M, University Station,
Austin, Texas, 1968.
[12] Aschenbrenner, B. C., "A New Method of Expressing Particle Sphericity," Journal of
Sedimentary Petrology. Vol. 26, 1956, pp. 15-31.
[t31 Wadell, H., "Volume. Shape and Roundness of Rock Particles," Journal of Geology.
Vol. 40, 1932, pp. 443-451.
OZOL ON SHAPE
625
[14] Wadell, H., "Volume, Shape and Roundness of Quartz Particles," Journal of Geology,
Vol. 43, 1935, pp. 250-280.
[15] Krumbein, W. C., "Measurement of Geological Significance of Shape and Roundness of

Sedimentary Particles," Journal of Sedhnentary Petrology, Vol. 11, No. 2, 1941, pp. 6472.
[16] Sneed, E. D. and Folk, R. L., "Pebbles in the Lower Colorado River, Texas--A Study
in Particle Morphogenesis," Journal of Geology, VoI. 66, No. 2, 1958, pp. 114 - 150.
[17] Burke, K. and Freeth, S. J., "'A Rapid Method for the Determination of Shape, Sphericity
and Size of Gravel Fragments," Journal of Sedimentary Petrology, Vol. 39, 1969, pp.
797-798.
[18] Ehrlich, R. and Weinberg, B., "An Exact Method for Characterization of Grain Shape,"
Journal of Sedimentary Petrology, Vol. 40, No. 1, March 1970, pp. 205-212.
[19] Czarneeka, E. T. and Gillott, J. E., "A Modified Fourier Method of Shape and Surface
Texture Analysis of Planar Sections of Particles," Journal of Testing and Evaluation,
American Society for Testing and Materials, Vol. 5, No. 4, July 1977, pp. 292-298.
[20] Czarnecka, E. T. and Gillott, J. E., "The Effect of Orientation on the Analysis of Shape
and Texture of Concrete Aggregates by the Modified Fourier Method," Journal of Testing
and Evaluation, American Society for Testing and Materials, Vol. 5, No. 4, July 1977,
pp. 299-302.
[21] Riley, N. A., "Projection Sphericity," Journal of Sedimentary Petrology, Vol. 11, No. 2,
pp. 94-97, 1941.
[22] Zingg, T., "Bertrag zur Schlotteranalyse," Sehweiz. Mineralog. Petrog. Mitt,, Bd. 15,
1935, pp. 39-140.
[23] Pettijohn, F. J., Sedimentary Rocks, 2nd ed., Harker and Bros., New York, 1957.
[24] Krumhein, W. C. and Sloss, L. I., Stratigraphy and Sedimentation, W. H. Freeman,
1956.
[25] Thiel, G. A., "The Relative Resistance to Abrasion of Mineral Gains of Sand Size,"
Journal of Sedimentary Petrology, Vol. 10, pp. 103-124.
[26] Conway, C. D., "Measurement of Shape Characteristics of Coarse Aggregates," Virginia
Highway Research Council, Charlottesville, Va., 1968.
[27] Tons, E. and Goetz, W. H., "Packing Volume Concept for Aggregates," HRB Record
No. 236, Highway Research Board, NRC 1968, Wash., D.C.
[28] Ishai, Ilan and Tons, E., "Concept and Test Method for a Unified Characterization of
the Geometric Irregularity of Aggregate Particles," Journal of Testing and Evaluation,
American Society for Testing and Materials, Vol. 5, No. 1, Jan. 1977, pp. 3-15.
[29] Huang, E. Y., "A Test for Evaluating the Geometric Characteristics of Coarse Aggregate
Particles," Proceedings, American Society for Testing and Materials, Vol. 62, 1962, pp.
1223-1242.
[30] Chamberlin, W. P., "Influence of Natural Sand Fine Aggregate on Some Properties of
Hardened Concrete Mortar," HRB Record No. 124, Aggregates for Concrete, Highway
Research Board, 1966.
[31] "Method of Test for Flat and Elongated Particles in Fine Aggregate," Designation
CRD-C 120-55, Handbook for Concrete and Cement, U.S. Army Engineers, Waterways
Experiment Station, Vicksburg, Miss., 1955.
132] Laughlin, G. R., "Limestone Fine Aggregate in Portland Concrete," Highway Materials
Research Laboratory, Lexington, Ky., Feb. 1960,
[331 Markwiek, A. H. D., "'The Shape of Road Aggregates and its Measurement," Road
Research Bulletin No. 2, HMSD, London, 1937.
[34] State of Ohio Department of Highways, Construction and Materials Specifications,
Section M-2.11.4, Particle Shape for Sand Used in Portland Cement Concrete, 1961,
p. 414.
[35] Heigold, P. C. and Lamar, J. E., "Two-Dimensional Shape of Sand Made by Crushing
Illinois Limestones of Different Textures," Industrial Minerals Notes 41, Illinois State
Geological Survey, Aug. 1970.
[36] Mackey, R. D., "The Measurement of Particle Shape," Civil Engineering and Public
Works Review. Vol. 60, No. 703, Feb. 1965, pp. 211-214.
[37] Mueller, W. and Hunn, H., "Texture Analyzer-System," lndustrialResearch, Nov. 1974.
[38] Ryland, V. N. and Stevens, R. E., "Image Analysis Automates Microscopy," Industrial
Research, June 1977, pp. 91-93.
626
[39] Haun, W., "A New Electro-Optical System for Quantitative Image Analysis," The Microscope, Vol. 19, April, 1971.
[40] Kanin, G., Kluge, N., Miller, W., and Rzeznik, J., "An Instrument for Automatic Image
Analysis," Scientific and Technical Information, G. Wangorsch ed., Ernst Leitz, Wetzlar.
[41] Rittenhouse, G., "A Visual Method of Estimating Two Dimensional Sphericity," Journal
of Sedimentary Petrology, Vol. 3, 1943, pp. 79.
[42] Gray, J. E. and Bell, J. E., "Stone Sand," Engineering Bulletin No. 13. National
Crushed Stone Association, 1964.
[43] Rex, H. M. and Peck, R. A., "A Laboratory Test to Evaluate the Shape and Surface
Texture of Fine Aggregate Particles," Public Roads, Vol. 29, No. 5, Dec. 1956, pp.
118-120.
[44] Shergold, F. A., "The Percentage Voids in Compacted Gravel as a Measure of Its
Angularity," Magazine of Concrete Research, Vol. 5, No. 13, Aug. 1953.
[45] Malhotra, V. M., "Correlation Between Particle Shape and Surface Texture of Fine
Aggregates and Their Water Requirement," Materials Research and Standards, American Society for Testing and Materials, Vol. 4, Dec. 1964, pp. 656-658.
[46] Bennett, E. W. and Katakkar, S. H., "A Comparative Study of Twelve Types of Sand
used as Fine Aggregate in Concrete," Journal, British Granite and Whinstone Federation, London, 1965.
[47] Clelland, J., "Sand for Concrete--A New Test Method," New Zealand Standards Bulletin
No. 14, Ministry of Works, Central Laboratories, Wellington, New Zealand, 1968.
[48] Huang, E. Y., "An Improved Particle Index Test for the Evaluation of Geometric
Characteristics of Aggregates," Journal of Materials, American Society for Testing and
Materials, Vol. 2, No. 1, 1967, pp. 81-111.
[49] Bennett, E. W. and Katakkar, S. H., "A Comparative Study of Twelve Types of Sand
Used as Fine Aggregate in Concrete," Journal, British Granite and Whinstone Federation, London, Spring, 1965.
[50] Kaplan, M. F., "The Effects of the Properties of Coarse Aggregates on the Workability
of Concrete," Magazine of Concrete Research, Vol. 10, No. 29, Aug. 1958, pp. 63-74.
[51] Kaplan, M. F., "Flexural and Compressive Strength of Concrete as Affected by the
Properties of Coarse Aggregates," Journal, American Concrete Institute, Vol. 30, No. 11,
May 1959, pp. 1193-1208.
[52] Wright, P. J. F., "A Method of Measuring the Surface Texture of Aggregate," Magazine
of Concrete Research, Nov. 1955.
[53] Wenzel, R. N., "Resistance of Solid Surfaces to Wetting by Water," Industrial and
Engineering Chemistry, Vol. 28, 1936, p. 988.
[54] Wenzel, R. N., "Surface Roughness and Contact Angle," Journal of Physical and Colloid
Chemistry, Vol. 53, 1949, p. 1466.
[55] Marian, J. E., "Surface Texture in Relation to Adhesive Bonding," Symposium on
Properties of Surface, A S T M STP 340, American Society for Testing and Materials,
1963.
[56] Orchard, D. F., "Properties and Testing of Aggregates," Concrete Technology, 3rd ed.,
Vol. 3, Wiley, N.Y., 1976, p. 93.
[57] Kramrisch, F., "Zur Rauhigkeitzbestimmung yon Gesteinbruchflacher," Geologie und
Bauwesen, Vol. 7, No. 2, 1935, pp. 33-59.
[58] Wallach, J., "Surface Topography Description and Measurement," Surface Mechanics,
Symposium Volume, American Society of Mechanical Engineers 1-23, 1969.
[59] Myers, N. O., "Characterization of Surface Roughness," Wear, Vol. 5, 1962, pp. 182189.
[60] Powers, T. C., "Topics in Concrete Technology: 1, Geometric Properties of Particles and
Aggregates," Journal, Portland Cement Association Research and Development Laboratories, 6, No. 1, Jan. 1964, pp. 2-15.
[61] Orchard, D. F., Yandell, W. G., and Lye, B. R. X., "A Quick Method of Measuring
the Surface Texture of Aggregate," Proceedings, 5th Conference of the Australian Road
Research Board, Vol. 4, Part 5, pp. 325-341.
[62] Scrivener, F. H. and Hudson, W. F., "A Modification of the AASHO Road Test Serviceability Tridex Formula," HHR No. 46, 1963.
[63] Boyde, A., "A Stereo-Plotting Device for SEM Micrographs and a Real Time 3-D System
for the SEM," Proceedings, 7th Annual SEM Symposium, IITRA, Chicago, I11., 1974.
OZOL ON SHAPE
627
[64] Boyde, A., "Photogrammetry of Stereo Pair SEM Images Using Separate Measurements
from the Two Images," Proceedings, 7th Annual SEM Symposium, IITRA, Chicago, Ill.,
1974.
[65] Bikerman, J. J., "Adhesion of Asphalt to Stone," Research Report R64-3, Massachusetts
Institute of Technology, Civil Engineering Department, 1964.
[66] Gur, Y., Shklarsky, E., and Livneh, M., "Influence of Particle Shape on the Properties
of Crusher Graded Aggregates," Technion, Israel Inst. of Technology, Report No. CV-223
(1967) in Hebrew.
{67} Carman, P. C., "The Determination of the Specific Surface of Powders," Journal. Society
of the Chemical Industry, Vol. 57, No. 7, July 1938, pp. 225-234.
[68} Shacklock, B. W. and Walker, W. R., "The Specific Surface of Concrete Aggregates
and its Relation to the Workability of Concrete," Research Report No. 4, Cement and
Concrete Association, 1958.
[69] Patat, F. and Schliebener, C., "Die Adsorption yon Makromolektilen. I. Uber Eine
Neue Meszmethode," MakroMolekulare Chemie, Vol. 64-66, 1961, pp. 643-68.
[70] Brunauer, S., Emmett, P. H., and Teller, E., "Adsorption of Gases in Multi-molecular
Layers, Journal, American Chemical Society, Vol. 60, 1938, p. 309.
[71] Brodd, R. J. and Hackerman, N., "Polarization Capacity at Solid Electrodes and True
Surface Area Values," Journal, Electrochemical Society, Vol. 104, No. 12, 1957.
[72] Stamenkovic, H., "High Strength and Water Impermeability of Concrete as a Function
of Surface Area Aggregate," Materiaux et Constructions, Vol. 3, No. 14, 1970, pp. 91-98.
[73] Krinsley, D. A. and Doornkamp, J. C., Atlas of Quartz Sand Surface Textures, Cambridge University Press, 1973.
[74] Powers, T. C., "The Nature of Concrete," Significance of Tests and Properties of Concrete and Concrete Making Materials, A S T M STP 169A. American Society for Testing
[75] Blanks, R. F., "Good Concrete Depends on Good Aggregate," Civil Engineering, Vol.
122, No. 9, Sept. 1952, pp, 651-655.
[76] Popovics, Sandor, "Aggregate Grading and the Internal Structure of Concrete," HRB
Record No. 441, Highway Research Board, Washington, D.C., 1973.
[77] Patten, B. J. F., "The Effects of Adhesive Bond Between Coarse Aggregate and Mortar
on the Physical Properties of Concrete," Civil Engineering Transactions, The Institution
of Engineers, Australia, 1973.
[78] Darwin, D. and Slate, F. O., "Effect of Paste-Aggregate Bond Strength on Behavior of
Concrete," Journal of Materials, Vol. 5, No. 1, March 1970, pp. 86-98.
[79] Shah, S. P. and Chandra, S., "Critical Stress Volume Change and Microcracking of
Concrete," Journal, American Concrete Institute, Vol. 65, No. 9, Sept. 1968, pp. 770781.
[80] Nepper-Christensen, P., and Nielsen, T. P. H., "Modal Determination of the Effect of
Bond Between Coarse Aggregate and Mortar on the Compressive Strength of Concrete,"
Journal. American Concrete Institute, Vol. 66, No. 1, Jan. 1969, pp. 69-72.
[81] Alexander, K. M., Wardlaw, J., and Gilbert, D. J., "Aggregate-Cement Bond, Cement
Paste Strength and the Strength of Concrete," Proceedings, International Conference on
the Structure of Concrete, London, Sept. 1965, pp. 59-81.
[82] Hsu, T. T. C. and Slate, F. O., "Tensile Bond Strength between Aggregate and Cement
Paste or Mortar," Journal, American Concrete Institute; Proceedings, Vol. 60, No, 4,
April 1963.
[831 Valenta, O., "The Significance of the Aggregate-Cement Bond for the Durability of
Concrete," Colloque International, Durabilite des Betons, Prague, Nakladatelstoi
Ceskoslovenske Akademie red, 1961, pp. 53-87,
[84] Alexander, K. M., "Strength of the Cement Aggregate Bond," Journal, American Concrete Institute, Vol. 56, No. 5, Nov. 1959, pp. 377-390.
[85] Farran, J., "Mineralogical Contribution to the Study of Adhesion Between the Hydrated
Constituents of Cement and the Embedded Material," Revue des Materiaux de Construction, July-Aug. and Sept. 1956, pp. 155-172, and pp. 191-209.
[86] Chatterji, S. and Jeffery, J. W., "The Nature of the Bond Between Different Types of
Aggregates and Portland Cement," Indian Concrete Journal, Vol. 45, Aug. 1971, pp.
346-349.
[87] Ozol, M. A., "The Portland Cement Aggregate Bond--Influence of Surface Area of the
628
Coarse Aggregate as a Function of Lithology," Report No, 71-R40, Virginia Highway Research Council, June 1972.
[88] Mercer, L. B., "Effect of Particle Orientation Upon Compressive Strength Results,"
Pit and Quarry, Vol. 45, No. 3, Sept. 1952, p. 212.
[89] Blanks, R. F., "Modern Concepts Applied to Concrete Aggregates," Transactions,
ASCE, Vol. 115, 1950.
[90] Scholer, C. H,, "Durability of Concrete," Proceedings, Highway Research Board, Vol.
10, 1931, p. 132.
[9l] Ozol, M. A., "Observations on the Use of Foliated Rocks for Coarse Aggregate in Paving Concrete," Proceedings, 2nd Forum on Geology of Industrial Minerals, 1966.
[92] Mercer, L. B., "Aggregate Particle Shape Determination," Pit and Quarry, Vol. 44,
No. 11, June 1952, pp. 111-112.
[93] Mielenz, R. C., "Petrographic Examination," Significance of Tests and Properties of
Concrete and Concrete Making Materials, A S T M STP 169A, American Society for Testing and Materials, 1966, pp. 381-403.
[94] Goldbeck, A. T., "A Digest of a Report on Effect of Stone Dust on the Properties of
1929, p. 301.
[95] Wright, P. J. F., "The Properties and Specification of Aggregates for Concrete," The
Quarry Managers 'Journal, Sept. 1962.
[96] Hansen, W. C., "The Chemical Reactions," Significance of Tests and Properties of Concrete and Concrete Making Materials, A S T M STP 169A, American Society for Testing
and Materials, 1966, pp. 487-496.
W. G. Mullen i
Chapter 36--Weight, Density,

Absorption, and Surface Moisture
Introduction
Aggregates in stockpiles are collections of various sizes of irregularly

shaped particles often with rough or convoluted surfaces, sometimes with
pores and wrinkles or crevices that hold water. Because the particles do not
fit together nicely, spaces called voids are left between them and the whole
pile contains water in the pores and on the surface of each particle, even
when located on the edge of the African desert. In order to use these aggregates as the largest part of a concrete mixture, five factors must be determined: (1) the volume of voids in order to fill them with paste and to float the
particles apart with an excess of paste, (2) the volume that solids will contribute to the concrete mix, (3) the specific gravity in order to convert volume
to weight, since most concrete is batched by weight, (4) the absorption
because it affects the specific gravity that is used, and (5) the surface
moisture because mixing water must be adjusted to compensate for it.
For construction purposes, including use in portland cement concrete, aggregates often are identified by physical properties and by gradation. To
accomplish this identification, numbers must be found to describe the
aggregates for design and for specification purposes. The tools for this identification are provided to the construction industry through standard
specifications and methods of test, such as those published by The American
Society for Testing and Materials, American Association of State Highway
and Transportation Officials, and others. Generally, tests for aggregate
properties are logical and based upon sound physical principles; they are
fairly well codified to meet the testing and specification needs for most construction work; and the tests that find substantial use are fairly simple.
For most aggregates, the properties that lend themselves readily to determination are specific gravity, density, unit weight, absorption, and grada1professor, Civil Engineering Department, North Carolina State University, Raleigh, N.C.
27650.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
630
tion. For many lightweight or highly porous aggregates, specific gravity and
absorption are not determined as readily for normal or heavyweight
aggregates due to their highly vesicular nature and due to inability to
distinguish between surface roughness or rugosity and surface porosity. Fortunately, it is possible to arrive at workable concrete mix proportions without
knowing aggregate specific gravity through the use of density or yield trial
batching. Gradation and unit weights are determined easily for almost all
aggregates.
The values that are obtained for physical properties vary with aggregate
source, minerology, size, range of gradation, shape and surface texture, but
for a given aggregate the values are relatively stable for design purposes. The
only variable quantity that is needed regularly in concrete mixture production is the free or surface moisture and this value usually is not required until
after mix proportions have been established.
A Philosophy of Aggregate Tests

A "single test" is not worth a thousand "expert opinions." A single test
may be worthless if it determines a value that has no use; it may be harmful if
it furnishes an erroneous value that is believed; or, without companion tests,
it may be insufficient statistically to give a value that is reliable. What is
needed from aggregate tests are numbers describing properties of aggregates
that may be used for design purposes or for acceptance. Such numbers are
unit weight, density, specific gravity, absorption, voids, gradation, and so
on. These numbers should be determined simply, if possible, using simple
equipment and procedures and without unnecessary requirements for
precision.
Testing requirements for research and for production must not be confused. Testing for research is an aid to understanding in the solution of problems. Testing for production is an aid to control of quality in the world of
commerce. An example of confusion is the running controversy over use of
the terms density and specific gravity where the purists would have us use
density in place of specific gravity. It does not really matter which term is
applied as long as the designer or buyer or seller is comfortable with the term
and finds it useful in his work. If the end is accomplished without detriment
to any of the parties participating or affected, then the use of specific gravity
is perfectly legitimate. Unfortunately, such arguments consume valuable
time in standardization committee work and bring to mind the Lilliputian
political parties [1] 2 who were divided into "big enders" and "little enders"
over which end of an egg to open, when, in truth, it did not matter as long as
the egg could be opened.
All aggregates are different when we examine them closely enough, but to
MULLEN ON WEIGHT, DENSITY, ABSORPTION, AND MOISTURE 631
describe these differences may not be necessary for use of the aggregates in
construction. We must stand far enough away in testing to see only those
properties that are important in our work. To do otherwise will result in
unnecessarily complicated tests that probably will not find wide usage. Fortunately most of the ASTM tests that will be discussed here attack the opening of the egg and leave the question of which end of the egg to approach to
be considered only when it truly makes a difference in the value to be obtained
from the test.
Aggregate Configuration
It is generally recognized that aggregate particles come in various sizes
that are described for engineering identification purposes by separation on
square opening sieves. It also is recognized that aggregates having the same
size designation may vary widely in shape, angularity, and surface texture or
roughness.
There are rounded, subangular, and angular aggregates produced as
gravel, crushed gravel, and crushed stones with varying degrees of rounding
of the corners and edges.
Aggregate particle surfaces vary from smooth sometimes glassy to rough
irregular surfaces. Minerology and porosity of the source rock as well as
natural polishing contribute to the final surface texture. As a general rule
natural gravels and sands tend to have smooth surfaces from polishing during wind, water, or glacial transportation. Crushed stones produced directly
from ledge rock by and large have rough surfaces. Sometimes the crystal
structure of ledge rock is such that for a particular size all particles may be
single crystals with cleavage planes as surfaces. Such an aggregate containing
single feldspar crystals was recently encountered in North Carolina [2].
Calcite crystals have been known to break out as single particles [3]. The
textures of slags and lightweight aggregates may vary from glassy to extremely rough depending upon the processing method.
Rugosity has been described by Tons and Goetz [4] as the asphalt holding
capacity or the surface roughness of aggregate particles when an asphalt
coating is applied and then literally scraped off. Tons and Ishai [5] describe a
concept and test method for geometric irregularity of aggregate particles that
includes microvoids, macrovoids, and packing voids as seen in Figs. 1 and 2.
The rugosity is indexed to the volume of asphalt held by the surface which
represents the amount of material that must be applied to "smooth up". the
surface. The rugosity also can be an index of possible surface interlock between adjacent aggregate particles and of the frictional resistance to movement of one surface past another when in contact.
Shape and surface texture of aggregate particles have been discussed by
Mather [6] using the terms sphericity, roundness, smoothness, and
roughness as descripters. Kummer and Meyer [7] in their work on skid resisCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:36 EDT 2014
632
Macro-Surface
Voids
~
-f
Packing
/-/
Volume Membrane
Pocknig
~'~E ~ j " ~ - ~ / - ~
~___.
Porosity
FIG. 1--Packing volume, packing porosity, and geometric irregularity of aggregate particles.
(From Ishai and Tons [5].)
Micro Surface ~
Voids
Macro Surface ~
Voids
~
/ ~
f
"
--
~o\~
..~Ao.,~,~ . _
~
__ ~
MaximumVolume
Membrane (Mix)
PackingVolume
Membrane
Bulk VolumeMembrane
~
- - - EffectiveVolumeMembrane
Asphalt/
~ '~ ~
Apporent Volume Membrane
Absorphon
"~
"~ - - ~ V o i d s ImpermeobletoWoter
Surface Voids r ~ , ~
Impermeable to/
Asphalt
FIG. 2--Volumetric characteristics of an aggregate surface pore. (From Ishai and Tons
IS].)
tance treat surface texture of pavements and, by inference, of aggregates in

certain types of pavements using smooth, fine, and coarse for size or depth of
texture and smooth, rounded, and gritty for sharpness of texture as shown in
Fig. 3.
Shape and texture affect aggregate mobility, the ease with which aggregate
particles move past one another when manipulated. Aggregate mobility affects the energy required for compaction of aggregates and, in a standard
compaction procedure, the degree of compaction obtained. Visual evidence
of differences in aggregate mobility between particles of different shapes and
texture is shown in Fig. 4 where it is possible to form a pile of unconfined
crushed stone several particles deep while unconfined glass marbles can be
piled only one particle deep. The mobility of an aggregate affects the
workability of fresh concrete as well as the quantity of coarse aggregate that
may be used. Goldbeck and Gray [8] in their application of the b/bo concept
for concrete mixture design account for differences in aggregate mobility as
MULLEN ON WEIGHT, DENSITY, ABSORPTION, AND MOISTURE
633
(~ SMOOTH
(~ FINE TEXTURED,
ROUNDED
~) FINE TEXTURED,
GRITTY
(~ COARSE TEXTURED,
ROUNDED
(~) COARSE TEXTURED,
GRITTY
FIG. 3--Surface textures. (After Kummer and Meyer [6].)
FIG. 4--Aggregate mobility affected by shape and surface texture (photo).
reflected in the dry rodded unit weight of coarse aggregate, a measure of

packing density.
If size of aggregate and range of sizes of aggregate are considered as part
of aggregate configuration, then both affect the values obtained for unit
weight, absorption, sometimes for specific gravity, and for free or surface
moisture. For a given weight of aggregate, the surface area will double as the
634
particle size is halved. Put differently, the amount of surface intersecting internal pores allowing absorption is doubled as is the surface area to hold a
given film thickness of surface water. This concept is illustrated in Figs. 5 and
6 by subdividing a one unit cube of aggregate.
Particle shape affects surface area with a sphere having the most efficient
containment of volume and flats and flakes having the least efficient containment. Due to the use of square opening sieves, the relationship between particle size is not quite as straightforward for different shaped aggregates as in
the case of the cubes illustrated in Figs. 5 and 6.
A=24
V=l
A=12
V=I
A=6
FIG. S--Surface area versus particle size at equal volumes.
For geometric shapes of equal sieve size, the following relations for surface
areas for equal absolute volumes hold:
spheres 1
cubes
1
pyramids 1.91
If flat and elongated particles are considered, surface area increases
dramatically.
When packings of these geometric shapes are studied, one size spheres
have voids less than 30 percent while angular particles produce packings with
voids that often exceed 50 percent. It is well known that packings of particles
containing a range of particle sizes result in reduced void contents in keeping
with the presence of particles small enough to fit into the voids between
larger particles. Classic studies in this area are those by Fuller and Thompson [9] and by Weymouth [10].
/,8_
Ld
UJ
24_
h
r
12_
67
1/8
1/4
112
PARTICLE S I Z E
FIG. 6--Size surface area relationships.
That specific gravity may be affected by particle size is evident with some
lightweight aggregates where specific gravity tends to increase as particles
become smaller due to elimination of vesicules by crushing.
Unit Weight and Voids

Historically, unit weights and unit measures have been with us almost
since recorded history. One such reference to a "compacted" unit measure
from Luke 6:38 in the King James translation of the Bible follows: "Give and
it shall be given unto you; good measure pressed down, and shaken together,
and running over, shall men give unto your bosom. For with the same
measure that ye mete withal it shall be measured to you again."
Aggregates, being made up of irregularly shaped particles, do not fit
together nicely with the result that spaces are left between the particles.
These spaces, called voids, are affected by the shape of the particles, the size
distribution of the particles, and the orientation or packing arrangement of
the particles. For aggregate packings, unit weight and voids are complimentary, that is, when one increases the other decreases. Neither unit weight nor
voids are defined in ASTM Definitions of Terms Relating to Concrete and
Concrete Aggregates (C 125), but both are the subject of ASTM Standards.
Voids are usually expressed as percent of volume of a packing arrangement as described above. Brink and Timms [1l] describe unit weight as
follows: "The unit weight of a concrete aggregate is the weight of a unit
636
volume of representative particles." Accordingly unit weight has been expressed in terms of pounds per cubic foot when these units are in use and in
kilograms per cubic metre in the International System of Units (SI) and in
keeping with ASTM Metric Practice (E 380).
"Representative particles" is a synonym for representative sample, a sample of sufficient size to represent the stockpile or barge or bin or other supply
of aggregate to be tested. ASTM Test for Unit Weight of Aggregate (C 29)
provides for determination of unit weights of aggregates of maximum size of
4 in. (100 mm) and smaller. Maximum size is defined in ASTM Definition C
125 as the smallest sieve opening through which the entire amount of aggregate is required to pass.
For the 4 in. (100 mm) maximum size, a volume calibrated container (a
measure) having a nominal capacity of one cubic foot is specified thereby
determining the sample size by volume. The purpose of the test is to determine the weight that will be contained in the measure of known volume. For
smaller maximum sized aggregates volume calibrated measures having
nominal volumes of I/2, %, and 1/10ft 3 are variously specified. In the case of
1 89 in. (40 mm) maximum size aggregate, the 89 ft 3 measure is the smallest
that may be used. The sample size for a 11/~ in. (40 mm) maximum sized aggregate can be compared to a sample size of 16 kg (36 lb) specified for this
same aggregate size in ASTM Test for Sieve or Screen Analysis of Fine and
Coarse Aggregates (C 136) if calculations are made assuming values for
specific gravity and voids. When specific gravity is 2.70 and voids are 45 percent, then the sample weight for the 1/2 ft 3 foot measure is calculated to be
21 kg (or 46.3 lb), somewhat larger than that required for ASTM Test C
136. The sample weight contained in a given measure will vary with the
density of packing and the specific gravity. The number of aggregate particles will vary with the density of packing for a given maximum size and
gradation meaning that the number of particles in the packing will be
affected by the particle shape, as will the unit weight and voids.
ASTM Test C 29 provides for compact and loose unit weight determinations. Compaction is accomplished by rodding for aggregates that are 11/2 in.
(40 mm) maximum size and smaller and by jigging for aggregates that are
larger than 11/2 in. (40 mm) maximum size as these larger aggregates do not
lend themselves to compaction by rodding. Loose unit weights are determined by a shoveling-filling procedure without compaction. In both cases,
the measure is filled in three layers with the first two layers leveled with the
fingers and the third layer placed to overflowing with the surface leveled using a straightedge or the fingers in such a way as to balance projections and
depressions in the final surface. When the test is performed properly, the
results of repeated tests on the same sample by the same operator should be
reproducible within 1 percent.
Standard unit weight measurements are obtained on aggregates dried to
essentially constant weight at 105 to l l 0 ~ (221 to 230~
The unit weight or dry rodded unit weight as it is often called for compacted aggregate is directly useful in concrete mixture design where the b / b o
method as applied by Goldbeck and Gray [8] and its variations are used
[12,13]. For a given aggregate source, changes in unit weight are indicative of
changes in angularity or in gradation or of changes in both. Some measure of
unit weight is particularly useful in grouting of preplaced aggregate where
work may be completely underground or underwater or obscured from visual
control in some other way.
In terms of unit weight, aggregates may be classified into nonstructural or
insulating lightweight, structural lightweight, air cooled slag, normal weight,
and heavyweight. Table 1 contains unit weight ranges for these aggregate
classifications as found in ASTM or other references [12].
TABLE 1--Unit weight classification for aggregates.
Aggregate Classification
Insulating
Lightweight for masonry
Lightweight for concrete
Air cooled slag
Normal weight
Heavy weight
Unit Weight Range Compaction

lb/ft 3 (kg/m 3)
Mode
6 to 12 (96 to 196)
55 to 70 (880 to 1120)
55 to 70 (880 to 1120)
> 70 (> 1120)
75 to 110 (1200 to 1760)
110 to 290 (1760 to 4640)
dry loose
dry loose
dry loose
compacted
compacted
compacted
Unit Weight Source

ASTM C 332a
ASTM C 331a
ASTM C 330a
ASTM C 33 a
PCA b, ASTM C 29c
PcAb
aASTM Specification for Lightweight Aggregates for Insulating Concrete (C 332); ASTM
Specification for Lightweight Aggregatesfor Concrete Masonry Units (C 331); ASTM Specification for Lightweight Aggregatesfor Structural Concrete (C 330); ASTM Specificationfor Concrete Aggregates(C 33), and ASTM Test for Unit Weight of Aggregate (C 29).
bDesign and Control of Concrete Mixtures, Portland Cement Association, July 1968.
r
from balance capacity required by ASTM Test C 29.
For aggregate of a given specific gravity voids will vary inversely as the unit
weight. Due to the presence of aggregate voids, unit weights of concretes using the aggregates will almost always be greater than the aggregate unit
weights. Increases will be greatest when grout or mortar has a higher specific
gravity than the aggregate especially if the b / b o value is less than one. In the
case of aggregates with specific gravities greater than the grout or mortar
specific gravity, greatest increase in unit weight of concrete over aggregate
weight will occur when b / b o = 1 or with preplaced aggregates. As b / b o
decreases below 1, the unit weight of concrete will also decrease for these
higher specific gravity aggregates.
Voids are affected by aggregate shape, gradation and compaction and
some rules of thumb apply as follows: (a) rounded particles pack more closely
and have fewer voids than angular particles, (b) graded aggregates pack
more densely than one sized aggregates, as small particles fill the spaces
between larger particles, and (c) voids decrease with compaction effort that
brings about most favorable particle orientation.
638
Voids expressed in percent are determined using ASTM Test for Voids in
Aggregate for Concrete (C 30). Actually the testing is accomplished using
ASTM Test C 29 for unit weight and either ASTM Test for Specific Gravity
and Absorption of Coarse Aggregate (C 127) or ASTM Test for Specific
Gravity and Absorption of Fine Aggregate (C 128) to obtain the bulk specific
gravity. ASTM Test C 30 is actually a calculation method using the formula
Voids,
% = (A
62.355) -- B
A X 62,355
100
where
A = bulk specific gravity of aggregate,
B = unit weight of the aggregate, and
62.355 = weight in pounds per ft a of water at 62~ (16.7~
Sometimes in construction, unit weights, voids, and specific gravities will
be unknown and field estimates are desired. In cases of this type where normal equipment is not available, all of these quantities can be estimated and
checked frequently if a carpenter and platform scales are available. A tightly
constructed cubic foot box is placed on the platform scales previously leveled.
Weight of box is obtained, followed by weight of box filled with aggregates
compacted in the manner that represents use. The box is then filled level full
with water to determine the voids volume from the added water weight. With
volumes of box, aggregate, and voids known, it is a simple matter to compute
an estimate of specific gravity. During the construction of the cast-in-place
concrete pilings for the main piers of the Tappan Zee Bridge [14] it was
necessary to know the voids of preplaced aggregates in the pilings that extended to rock about --300 ft (--90 m) and were cut off below water at about
--20 ft (--6 m). Aggregate barged to the site from two sources was dropped
into the pilings in increments using a clam shell and a hopper funnel. It was
necessary to know the voids in the aggregates to control the grout level in the
preplaced aggregates to keep the grout pipe below the surface of the grout
pool and the grout pool below the surface of the preplaced aggregates. The
voids were checked barge by barge using a cubic foot plywood box and a platform scale. Voids ranged from 45 to 53 percent for 21/2 in. (63.5 mm)
nominal maximum sized aggregate, a variation that if ignored could have affected by several feet the calculated level of the grout pool in a 280 ft (85 m)
long 29 in. (736 mm) inside diameter pipe piling. The effect of underestimating voids would have been to allow the grout pipe to be pulled
out of the grout pool as pipe sections were removed resulting in mixing grout
and water together in the piling. The effect of overestimating the voids would
have been to allow the grout pool to rise above the aggregate surface, and this
situation would have resulted in dropping aggregate through water and grout
as new increments of aggregate were added. In both cases, churning and
639
mixing of water into the grout would have resulted in weakened concrete at
the affected locations. Within 50 ft (15 m) of the surface, grout levels were
verified using an hydrometer with a specific gravity of 1.5 on a sounding line.
The hydrometer would float on grout having a specific gravity of 2.06 and
sink through river water.
Specific Gravity and Absorption

Specific gravity is defined as the ratio of the weight of a given volume of
material to the weight of an equal volume of water, a dimensionless number.
There are a number of ways to determine specific gravity, but all involve
some measurement of weight and volume. Usually for aggregates, the
method employed involves determination of weight in air and determination
of volume of water displaced. The two basic schemes for making the water
displacement determination are shown in Fig. 7. Overflow devices have been
used also where the water displaced by aggregate may be caught and
weighed, but the two approaches above are the basis for ASTM Tests C 127
and C 128, the two methods for coarse and fine aggregates.
WEIGHT IN
AIR [A]
SPECIFIC G R A V I T Y
WEIGHT IN WATER
[B]
=
A
A-B
/.~
WATER WEIGHT
GRADUATION LINE
AGGREGATE
AGGREGATE
WEIGHT
c
SPECIFIC
AND WATER
D
GRAVITY
E
D
C D-E
FIG. 7--Specific gravity determination schemes.

640
Specific gravity determinations are not quite as simple as indicated in Fig.

7 because of aggregate surface and internal porosity and rugosity and
because water varies in density with temperature. Therefore specific gravity
determinations must be and are temperature related and surface porosity
must be treated to account for volume of permeable pores. As a result of
treatment of permeable porosity and its water content we have not one but
three specific gravity values that are determined for both fine and coarse aggregates.
A schematic aggregate particle is shown in Fig. 8 illustrating the presence
of permeable pores and varying moisture contents. There are two volumes
that may be determined, the volume including permeable pores known as the
bulk volume, and the volume excluding the pores that will be smaller than
the bulk volume by the volume of the permeable pores. However, there are
two moisture conditions that enter into specific gravity calculations though,
bulk dry and saturated surface dry. As shown in Fig. 8, three moisture condi-
~
BULK DRY
ABLE PORE
ABSORBED
WATER
SSD
~_.FREE WATER
(~~FILM
SSD+FREE WATER
9 PORES EMPTY
9 PORES SATURATED
9 PORESSATURATED
9 SURFACE DRY
9 SURFACE DRY
9 SURFACEWET
9 LIGHTER C O L O R
9 LIGHTER COLOR
9 DARKER COLOR
9 FINE A G G R E G A T E
FLOWS
9 FINE A G G R E G A T E
FLOWS
9 FINE AGGREGATE
CLINGS
FIG. 8--Aggregate moisture conditions.
tions may actually exist under test conditions. Porosity as it increases

becomes extremely important in specific gravity determinations to the point
that with some highly vesicular aggregates, specific gravity determinations,
are approximations at best. The reason for this problem is that it is almost
impossible to determine either of the two volumes needed for the
calculations.
Using the volumes and weights shown in Fig. 8 and the schemes illustrated
in Fig. 7, three specific gravities are determined as follows using ASTM Tests
C 127 and C 128.
1. bulk specific gravity--where the oven dry weight is used with the
volume that includes the permeable pores.
2, bulk specific gravity (saturated surface dry basis)--where the weight
of the aggregate including weight of water in permeable pores is used
with the volume including the volume of the permeable pores. If there
are any permeable pores, the bulk specific gravity (SSD Basis) will be
higher than the bulk specific gravity.
3. apparent specific gravity--where the oven dry weight of aggregate is
used with the volume, excluding the volume of permeable pores. If
there are any permeable pores, the apparent specific gravity will be
higher than either of the other two specific gravities.
Both ASTM Tests C 127 and C 128 require that aggregates be oven dried and
resaturated for 24 h before specific gravity determinations are made. Both
methods include a precautionary note indicating that when concrete is proportioned using naturally moist aggregates that oven drying and resaturation
for 24 h may yield specific gravity and absorption values that are too low for
their intended use. The Manual of Concrete Testing [15] contains the same
cautions indicating that the 24-h period is not always long enough for
resaturation to occur.
Absorbed water is the unbound water contained in the permeable pores of
the aggregate particles that may be driven offby oven drying at 100 to 110~
The water carried by an aggregate that is not contained in the pores is surface
water or free water. The difference in the two waters is that absorption is a
property of the aggregate, while the free water is variable and a property of
the aggregate only in that the capacity to carry free water is related to aggregate surface area.
In order to determine absorption, the aggregate must be brought to a
saturated surface dry (SSD) condition. Then the measurement must be made
quickly for the SSD condition is transient in a drying atmosphere. ASTM
Tests C 127 and C 128 prescribe the means for drying aggregates from a free
surface moisture condition back to a SSD condition. For coarse aggregate,
particles are dried in a towel until all visible surface water films are removed.
For fine aggregates, drying is accomplished by gentle manipulation under a
warm air current until particles when molded into a cone shape will not retain the shape when the mold is removed. The key to recognition of the
change from surface wet to SSD is that a color change will occur from dark to
light in all aggregates when the surface water film is gone. Additionally, fine
aggregate particles will no longer cling together when SSD.
In both ASTM Tests C 127 and C 128 an SSD sample is weighed in air,
then in water to determine its SSD volume followed by drying at 100 to 110~
to constant weight to determine the absorbed water weight and by calculation
its volume. Absorption is expressed as a percentage of the oven dry weight.
Because of the surface area per unit volume of material, for similar minerals,
absorption is generally higher for fine aggregates than for coarse aggregates.
642
Crushed angular particles generally have higher absorptions than smooth

rounded particles because the surface area is higher. Obviously, porous aggregates will have higher absorptions than nonporous aggregates. One trend
that is an exception to the general rule is for lightweight fine aggregates to
have lower absorptions than lightweight coarse aggregates. Absorptions may
range from a fraction of one percent for silica gravels and some granitic rocks
to several percent for some limestones. Structural lightweight aggregates may
reach 10 to 15 percent absorption while nonstructural lightweights may exceed 100 percent absorption. The importance of absorption varies with
magnitude of absorption.
In batching concrete, most aggregates are in an SSD plus moisture condition. If this is not the case, aggregates can absorb mixing water and affect the
workability of the mixtures. One has to be particularly careful with
lightweight aggregates and probably the best practice is to presaturate the
aggregates before batching.
On the Tappan Zee Bridge [16], lightweight coarse and fine aggregates
were presaturated by sprinkling the stockpiles. The original specifications
read that the stockpiles would be sprinkled for 24 h immediately prior to use.
Presaturation was successful, but free water was present in such quantity
that it was draining through the stockpile. Water was being batched by
weight and on the first day of concreting, moisture adjustments were made
on practically every batch to compensate for free water changes. Needless to
say, adjustments were made in the presaturated procedure for subsequent
days of concreting. The adjustment required that stockpiles be sprinkled for
24 h but that sprinkling be discontinued at least 24 h before the stockpile was
to be used to allow free water moisture to stabilize before use.
Specific gravities give us the means to convert weights to absolute or solid
volumes for proportioning. Specific gravities also can be used for identification or acceptance purposes. In order to produce concretes of certain unit
weights, aggregates and air contents must be varied as cement and water
specific gravities are constants for all practical purposes. For heavyweight
concretes, high specific gravity minerals such as iron ores and barites in the
3.8 to 5.2 specific gravity range are used. Sometimes steel punchings are
used. Normal weight concretes are made with aggregates in the 2.4 to 3.0
specific gravity range while light aggregate specific gravities, when they can
be determined, range from about 1.7 down to less than one for some expanded slags.
Among normal weight aggregates, lighter weight or lower specific gravity
particles are sometimes of questionable durability in concrete mixes. In
ASTM Test for Lightweight Pieces in Aggregates (C 123), aggregates are
separated at a given specific gravity according to whether they float or sink in
a liquid of given specific gravity. As lighter weight particles are often porous,
the test is conducted with the aggregates in an SSD condition to prevent the
heavy liquid being absorbed by the aggregates. Ropke [17] reported a quick
field laboratory test that he used in prospecting for glacial gravels in an area
of upstate New York where lightweight particles could be a problem. He took
an oven dried sample of gravel, spread it on the floor, and doused it with a
bucket of water. The porous lighter weight particles dried almost immediately as evidenced by color change because they absorbed the surface
water. The percentage of such quick drying particles in the sample was determined and the sample was rejected outright or accepted for further testing on
the basis of this preliminary test.
Density
Scientifically, density is defined as the mass of a unit volume of a material.
A few words were written in the introduction about the use of density instead
of specific gravity. By and large in concrete proportioning, specific gravity is
used and is applied to the individual components of concrete: aggregates, cement, and water. Mixtures of these individual particles into pastes, mortars,
and concretes are described by using weights or weight per unit volume. At
this time, true density or mass per unit volume seldom is used to describe
concrete or aggregates.
Surface Moisture
Surface or free moisture is that water carried on the surface of an aggregate that is available for lubrication or mixing water and for hydration or
chemical reaction with the cement. Surface water is part of the mixing water
in concrete mixes. Free moisture is a variable quantity that must be determined at time of use. All aggregates stored in outside stockpiles will carry
free or surface moisture. The free moisture is almost always expressed as a
percentage of the bulk or oven dry weight. For coarse aggregates, it will
range from near zero to a maximum of 1 or 2 percent on normal weight aggregates. Fine aggregate free moistures range from 3 to 8 or 9 percent with 4
to 6 percent being about normal. If higher free moisture percentages are encountered, an unstable moisture content is indicated with water actually
draining through the stockpile or bin. Fluctuating surface moisture contents
make water adjustments difficult and consistency control exacting. It is best
to allow any stockpile to drain and to reach a stable moisture content before
use in concrete. As noted in earlier paragraphs, lightweight aggregates may
have moisture contents in excess of 100 percent in some cases, mostly absorbed. It is difficult to distinguish between free and absorbed moisture of
lightweight aggregates so that use of total moisture is often a better means of
controlling consistency of lightweight concretes.
ASTM Test for Total Moisture Content of Aggregate by Drying (C 566) is
used for fine and coarse aggregates. The total moisture is determined by drying a moist sample of appropriate size for the aggregate size at 100 to 110~
644
until further drying produces less than 0.1 percent additional loss in weight.
This moisture condition is defined in the method as thoroughly dry. Weights
are determined to the nearest 0.1 percent. Free moisture percent is
calculated from the total moisture percent by subtracting the percent absorption. Using ASTM Test for Surface Moisture in Fine Aggregate (C 70), the
surface or free moisture is determined using a displacement scheme. In effect, the method is used to determine the volume of a weighed wet sample including the surface moisture. From the volume and the weight, the specific
gravity of the mixed aggregate and surface water sample can be calculated. If
the sample specific gravity is less than the bulk specific gravity (SSD Basis) of
the aggregate, then free surface moisture is present and the proportions by
weight of water and aggregate to produce the specific gravity can be
calculated. Once the weights are known, they can be converted to percent
free moisture based upon SSD aggregate. For a given aggregate, this type of
problem calculation can be readily put into nomograph form for rapid graphical solutions. Temperature is not controlled in ASTM Test C 70.
There are other methods of determining free moisture in the field with
varying accuracy. One is the use of the Speedy Moisture Tester that has
powdered carbide to generate gas from the free water on a fine aggregate
sample in a closed chamber. This gas generated causes the pressure in the
chamber to rise and this can be read on a pressure gage as free moisture
expressed as a percentage of the wet sample weight. The carbide does not
generate gas from absorbed water and a test can be completed in a minute
or two. Another method is to dry the sample in a frying pan over a hot
plate, a gas stove, or a gasoline stove. Temperatures in these frying pan
methods often exceed the 100 to l l 0~ drying temperature specified in
ASTM Test C 566 which make it possible to complete a test in 10 to 20
min. Gasoline stoves are not recommended because they are a possible fire
hazard.
More and more batch plants are being equipped with electronic moisture
sensors placed into the aggregate bins with readout on the batching console.
In some cases, aggregate weight and mix water adjustments are made automatically while in less sophisticated plants, manual adjustments are set up
by the batch plant operator. Usually adjustments are made to the nearest
1 percent of free water content. Problems with moisture sensors, and all
reach-in moisture sampling schemes for that matter, are that sampling
may be from dead areas in the bins rather than from the aggregate flow
stream. Location of sensors and sample vents is important to ensure that
representative samples are taken.
Knowledge of and adjustment for surface moisture is the key to successful consistency control in concrete batching and mixing operations. Control
of changes in surface moisture is possible to a certain extent and should
be undertaken where possible. For example, stockpiles will have a moisture
gradient from top to bottom so that systematic removal of material from a
645
stockpile will allow moisture contents to change gradually. Bins left filled
overnight will have a c o n c e n t r a t i o n of m o i s t u r e at the discharge gate m a k i n g
it necessary to treat the first few b a t c h e s for moisture a d j u s t m e n t s at the
start of work on a n individual basis to ensure consistency control. I n freezing weather, unless b i n s are heated, aggregates will freeze together if b i n s
are left filled over night. Avoid the experience of the b a t c h p l a n t operator
in upstate New York who when faced with frozen b i n s poured gasoline into
the tops of the bins a n d ignited it to thaw the aggregates [18]. T h e aggregates were thawed b u t the gasoline r a n from the b o t t o m of the bins, ignited
the b a t c h plant, a n d b u r n e d it to the g r o u n d .
References
[1] Swift, Jonathan, Gullivers Travels.
[2] Mullen, W. G. and Whitfield, J. K., "Surface Wear and Skid Resistance of Portland
Cement Concrete Pavements," Final Report, Highway Research Program, Project ERSD
110-74-3, North Carolina State University, Aug. 1975.
Proceedings, American Concrete Institute, Vol. 38, p. 29.
[4] Tons, Egons and Goetz, W. L., "Packing Volume Concept for Aggregates," BHR Report
No. 235, Highway Research Board, Washington, DC., 1968, pp. 76-96.
[5] Ishai, Ilan and Ton, Egons, "Concept and Test Method for a Unified Characterization of
the Geometric Irregularity of Aggregate Particles," Journal of Testing and Evaluation,
American Society for Testing and Materials, Jan. 1977.
[6] Mather, Bryant, "Shape, Surface Texture, and Coatings," Significance of Tests and Properties of Concrete and Concrete-Making Materials. A S T M STP 169.4, American Society
for Testing and Materials, 1966, p. 415.
[7] Kummer, H. W. and Meyer, W. E., "Tentative Skid Resistance Requirements For Main
Rural Highways," NCHRP Report 37, HighwayResearch Board, Washington, DC., 1967.
[8] Goldbeck, A. T. and Gray, J. E., "A Method of Proportioning Concrete for Strength,
Workability, and Durability," Engineering Bulletin No. 11, National Crushed Stone
Association, 1942.
[9] Fuller, William B. and Thompson, Sanford, E., "The Laws of Proportioning Concrete,"
Transactions. American Society of Civil Engineers, Dec. 1907, p. 67.
[10] Weymouth, C. A. G., "Effect of Particle Interference in Mortars and Concrete," Rock
Products, 25 Feb. 1933, p. 26.
[11] Brink, R. H. and Timms, A. G., "Weight, Density, Absorption, and Surface Moisture,"
Significance of Tests and Properties of Concrete and Concrete-Making Materials, A S T M
STP 169A, American Society for Testing and Materials, 1966, p. 432.
[12] Design and Control of Concrete Mixtures, Portland Cement Association, Skokie, Illinois,
llth Edition, July 1968.
[13] "Recommended Practice for SelectionProportions for Normal and HeavyweightConcrete"
(ACI 211.1-74), Manual of Concrete Practice. American Concrete Institute, Part 1, 1977.
[14] Mullen, W. G., Construction Diaries--Tappan Zee Bridge, Madigan-Hyland Consulting
Engineers, N.Y., 1952-1955.
[15] Annual Book of A S T M Standards, Part 14, 1976, p. 631.
[16] Mullen, W. G., "'LightweightConcrete Deck for Tappan Zee Main Spans," Proceedings,
[17] Ropke, Jack, private communication, 1957.
[18] Mullen, W. G., "Construction Diaries--Berkshire Thruway," Madigan-Hyland Consulting Engineers, N.Y., 1957-1959.
W. L. Dolch
Chapter 37
Porosity
Introduction
Most of the important properties of concrete aggregates are influenced
strongly by the volume and dimensions of the internal pore system of the
material. These properties include density, strength, absorption, and freezethaw resistance.
The purpose of this article is to discuss the parameters related to the pore
system of an aggregate, the methods for their measurement, and the properties of the aggregate and concrete they so strongly influence. Many of these
topics were covered in earlier papers of which this is the successor [i, 2]. 2 It is
not intended to repeat details needlessly, but to emphasize subsequent
developments.
Definitions
Concrete aggregates have a bulk volume that is the sum of the volume of
their solids and that of their voids or pores. The porosity is the ratio of the
volume of voids to the bulk volume. This void volume is that within the piece
of aggregate, not that between pieces, the latter of which is measured by
ASTM Test for Voids in Aggregate for Concrete (C 30). Such confusion in
terminology is unfortunate.
The voids will be either empty or will contain water. The ratio of the
volume of contained water to volume of voids is the saturation of the void
system. The ratio of the weight of water to that of dry solids is the absorption.
The porosity and the degree of saturation control the various densities, which
are defined by ASTM Definitions of Terms Relating to Density and Specific
Gravity of Solids, Liquids, and Gases (E 12).
Other than total volume of voids, the properties that describe a pore
system in a porous solid are the size distribution and shape of the pores.
lProfessor of engineering materials, School of Civil Engineering, Purdue University,
Lafayette, Ind. 47907.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
DOLCH ON POROSITY 647
These are known collectively as the pore structure. The size of a pore cannot
be defined exactly in any way that is useful. In most porous solids the walls of
the pores twist and turn, get larger and smaller, go this way and that. The inability to describe such a system in detail results in the use of a model, a
simpler concept that can be described and analyzed and whose properties
can be determined. The extent to which these properties correspond to their
counterparts measured on the real material is the degree of success of the
model. By far the most commonly used model has been a bundle of round
tubes, sometimes of several sizes. For a discussion of models, see Refs 3
and 4.
If there are various volumes of pores of various sizes, the relationship can
be expressed in two ways. A plot of the volume of voids that are smaller (or
larger) than a given size is the cumulative or integral pore size distribution.
The slope of this curve, plotted against pore size, is the differential pore size
distribution, or the proportion of pore volume of a given size. Generally
speaking, the differential way of plotting the data shows nothing that is not
also evident from the integral curve. The differential curve is also likely to
emphasize unimportant aspects and is, in any event, suitable only for a
relatively small range of pore sizes--say two orders of magnitude. Therefore,
size distributions of aggregate materials usually are presented as a
cumulative curve.
Another way of expressing pore size is by the hydraulic radius, which is
defined as the cross sectional area of a pore space divided by its perimeter.
For example the hydraulic radius of a round pore is one fourth of its
diameter.
The solids in a porous body are more or less finely divided. The specific
surface (area) is the area of the solid surfaces on some unit basis. Most often
that basis is unit weight of solids, but bulk volume, void volume, etc., are
also used.
When a fluid, gas or liquid, flows through a porous solid, the permeability
is defined as the volume flow rate of a fluid of unit viscosity that permeates a
sample of unit cross-sectional area under a unit pressure gradient. So defined, the permeability is a function only of the pore structure and not of the
fluid. If the porous solid is absorbing the fluid by capillarity, the absorptivity
is defined as the square of the ratio of the volume of liquid absorbed to the
cross sectional area that it penetrates in unit time.
Measurement Methods
Porosity
The pore space that is interconnected and connected to the outside surfaces of the piece is sometimes called the effective porosity, to distinguish it
from any portion that is enclosed completely by solids. For practical purposes
648
TEST AND PROPERTIES OF CONCRETE AGGREGATES
with aggregates, this is an unimportant distinction, since most aggregates

have a completely interconnected pore space, and any that is not can be considered part of the solid.
In the past, a common way to measure porosity was to assume that the
volume of absorbed liquid (usually water) equals the volume of voids. This is
completely unwarranted, since the amount of liquid absorbed depends on
how it was taken up and, in any event, the saturation is hardly ever complete.
The absorption of an aggregate is measured by ASTM Test for Specific
Gravity and Absorption of Coarse Aggregate (C 127), or ASTM Test for
Specific Gravity and Absorption of Fine Aggregate (C 128). The two exposures that are most common are a 24-h soaking in water or a period of
evacuation of the dry solid by an aspirator followed by admission of water
and soaking. The former is the "twenty-four hour" absorption and is the
more common. The latter is called the "vacuum absorption"; even this does
not result in complete saturation with most aggregates and cannot be used as
a measure of total porosity.
Pycnometric methods can be used to determine the effective porosity if a
gas, rather than a liquid, is used. The gas is usually air, but helium may be
better, because it is not adsorbed appreciably by the solid surfaces. The
volume of a chamber containing the dry solid is changed by a known amount,
and the pressure on the system is measured before and after the change.
Boyle's law is assumed, and the volume of solids and true density can be
determined. An independent measure of the bulk density permits calculation
of the porosity. Examples of such equipment are given in Refs 5 and 6.
Effective porosity can also be determined with the so-called McLeod gage
porosimeter [7,8] in which a dry sample immersed in mercury has the air in
its voids removed by expansion and collected. The volume of voids is determined directly by measuring the volume of this air.
Porosity also can be determined microscopically by examination of a
polished section or a thin section. The usual stereographic procedures of
point counting, linear traverse, or areal measurement are used [9]. ASTM
Recommended Practice for Microscopical Determination of Air-Void Content and Parameters of the Air-Void System in Hardened Concrete (C 457) is
an adaptation of this method. The method has been used on aggregates [10],
but it is tedious and has the major disadvantage that a lower size limit is imposed by the resolution of the microscope. So pores smaller than about 1/~m
will not be seen at all and cannot be included in the measured porosity. Many
aggregates have major portions of their porosity in sizes smaller than this,
and for most of them the method is unsuitable.
Pore Structure
Pore structure refers to the combined aspects of pore size distribution and
pore shape. An average or "equivalent" pore size can be determined by
measuring the specific surface, the permeability, or the absorptivity.
The equivalent pore size is related to the porosity-to-specific surface ratio.

The form of the relationship depends on the pore model that is used, that is,
cylindrical pores, slits, hydraulic radius, etc. The most important way of
determining the specific surface is by the method of vapor phase adsorption.
The technique involves measuring the amount of vapor adsorbed or taken
up by the solid on its surfaces as a function of the ambient pressure of the
vapor. The experimental techniques can be found, for example, in Refs 11
through 13. The vapors most often used are nitrogen with the sample at the
temperature of liquid nitrogen or water vapor at room temperature. For
solids with a comparatively small specific surface, as are most aggregate
materials, the use of krypton may be advantageous [14].
The sorption data are then interpreted in the light of one of several theories
to allow calculation of the specific surface. The most commonly used theory
is the BET [15]. A general discussion of the sorption method is in the book by
Gregg and Sing [16]. The method obviously measures only the solid surface
that is contacted by the vapor, that is, the surface surrounding the effective
porosity.
The permeability method of determining pore size depends on the retardation of the flow of a fluid (liquid or gas) passing through the material by
drag on the pore walls. The smaller the pore, the greater the retardation
and the smaller the measured permeability. Translations of permeability
measurements into pore sizes depend on the analysis of some model for the
porous medium. One of the most common is that of Kozeny, which has two
properties of most real porous media: an irregular pore shape and a flow
path that is tortuous. The analysis of such a model shows the permeability to
be proportional to the porosity and the square of the hydraulic radius, so this
pore "size" can be determined. The book by Carman [I 7] is a good summary
of the subject.
In the use of the permeability method it is necessary to assume or to
measure the degree of tortuosity of the flow path of the permeating fluid.
There are several ways to measure it; perhaps the electrical analog method is
best [18]. The permeability method has been used on some aggregate
materials [19,20], but the method is not accurate for a material with a wide
range of pore sizes.
The absorptivity was defined by Powers and Brownyard [21]. It is related
to the rate at which a liquid, usually water, is imbibed by capillarity into the
dry porous solid. A round tube or a hydraulic radius model is then used to
calculate a pore size for the sample [19]. The method has the same uncertainties as the permeability method with an additional one that depends on the
influence of the degree of saturation, always less than unity. The method has
the advantage of ease and simplicity. A variant [22], in which the time is
measured for the complete absorption of a small water drop placed on the
surface of the rock, has been shown to correlate well with the conventional
adsorption value.
The vapor adsorption method also is used to determine pore size distribuCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:31:13 EDT 2014
650
tions and to gain some insight into the shape of the pores in a porous
medium. The data are obtained from the comparatively high vapor pressure
portion of the sorption experiment. This is the region in which bulk liquid is
present in the pores by the process of capillary condensation. The relationship between the pore size in which capillary condensate exists and the ambient vapor pressure is given by the Kelvin equation; the lower the vapor
pressure the smaller the capillary size.
These data can be used then to construct a pore size distribution curve.
The sizes of pores that can be investigated range from a lower value of
perhaps several molecular diameters to an upper one of several hundred
micrometres. It is necessary to make a correction for the thickness of film
already adsorbed on the pore walls. The ways to do this have been reviewed
by Linsen and van den Heuvel [23] and by Dullien and Batra [24], who also
covered other methods for the determination of pore structure. Brunauer and
his colleagues have devised a method for the complete analysis of pore size
distributions that does not assume a shape for most of the size range. A summary of this work is given in Ref 27.
The amount of vapor taken up by the solid at a given vapor pressure
depends on whether vapor is being added to or removed from the solid. The
amount is always larger for the latter case, which results in a hysteresis of the
sorption isotherm data. So in calculating a pore size distribution a choice
must be made of which set of data to use. The choice is determined largely by
the pore shape that is appropriate. The subject has been investigated by de
Boer [25], who has analyzed a variety of pore shapes in terms of the kind of
sorption hysteresis they produce. A general review is given by Everett [26].
The second major way of determining the pore size distribution of a porous
solid is by mercury porosimetry. The idea was developed by Ritter and Drake
[28] and has since become almost a standard method, with several forms of
commercial apparatus available.
The method consists of immersing an evacuated porous sample in mercury, and then forcing the mercury into the pores by the application of increasing pressure. The size of the pores is related inversely to the pressure required to cause entry of the mercury, according to the usual capillary
pressure relationship. The volume of pores of any size increment is the
amount of mercury forced in over the corresponding pressure range. And so
a cumulative pore size distribution curve can be generated. The model implicit in the usual analysis is an assembly of round tubes of various sizes. Of
course, it is recognized that this is not really correct for most porous solids
and, specifically, if a void has an entrance that is smaller than its body (a socalled "ink-bottle" pore), the whole void volume will be recorded as being the
size of the entrance. Therefore, a pore size distribution obtained by this
method is weighted towards smaller pore sizes. A general review of the
method is given by Rootare [29].
A problem with the method is that the product of the surface tension of the
DOLCH ON POROSITY
651
liquid and the cosine of the contact angle it makes with the solid must be
known in order to calculate the pore size. Most users of the method have
assumed values of the contact angle that were obtained by early workers,
usually on other materials. In an investigation of the pore structure of
hardened cement paste, Winslow and Diamond [30] measured this product
of the surface tension and cosine of contact angle by drilling small holes of
accurately measured size in the material and measuring the pressure needed
to cause the mercury to enter them.
The mercury porosimetry method has been used on a wide variety of solids
[29]. Applications to concrete aggregates have been reported [31-36].
Significance of Porosity of Aggregates

Broadly speaking, the porosity of a concrete aggregate affects the behavior
of the concrete for one or both of two reasons--either because the pore space
decreases the volume of solids in bulk volume of the aggregate or, more importantly, because the pore space permits the ingress and retention of water
or aggressive solutions.
The effects of the first of these are mostly obvious. The influence on density is of importance in such matters as mix proportioning.
Other things being equal, the strength and, consequently, the abrasion
resistance of the aggregate will be smaller the greater its porosity. This influence is probably not simply one of a reduced cross section to bear the load.
Abrasion resistance is an important property of aggregates that is related to
strength and porosity. ASTM Test for Resistance to Abrasion of Large Size
Coarse Aggregate by Use of the Los Angeles Machine (C 535), and ASTM
Test for Resistance to Abrasion of Small Size Coarse Aggregate by Use of the
Los Angeles Machine (C 131) are part of many aggregate specifications. This
property is recognized to be of importance in determining the degree of
breakdown an aggregate can be expected to undergo during mixing of concrete or compaction of base course, ballast, and so on.
Except for those that are grossly weak, the strength of the aggregate is not
considered to be a primary factor governing the strength of concrete. The
strength of the cement paste, which is related intimately to its own porosity,
and the strength of the paste aggregate bond are more important [37]. Little
has been done on this last question, but it is logical to assume that the
roughness and the porosity of the surface of the aggregate play an influential
part in the development of the interfacial bond with the paste [38].
The porosity of aggregates also affects their other mechanical properties,
such as modulus of elasticity and Poisson's ratio. These properties in turn are
important to the same properties of the concrete, which are additive functions of the properties of its constituents.
Porosity is important to questions of proportioning concrete because of its
control over the absorption of the aggregate. Any mixing water that is
652
abstracted by the aggregate during mixing and placement must be considered in the mix proportions. Usually this is done in an effective, if approximate, fashion by the application of relatively small correction factors. But in
the proportioning of lightweight aggregate concrete the needed correction
factor frequently becomes large and indeterminate because of the large absorption that can occur with lightweight aggregate.
A few aggregates undergo excessive volume change themselves on wetting
and drying and, therefore, can contribute greatly to the shrinkage, basically
because by shrinking the aggregate fails to restrain the shrinking of the paste
[39]. Little work has been done on these aggregates, and, except possibly for
the influence of clay minerals, few generalizations can be stated concerning
the types of materials that may be particularly subject to this difficulty.
Foremost in the list of significant effects that the porosity of aggregates has
on the properties of concrete is freezing-and-thawing durability. Study of
frost durability of aggregates in concrete has been given impetus by the growing scarcity of "good" aggregates and the necessity to use material that
would have been rejected in former days.
It was recognized long ago that the coarse aggregate component had an
important effect on the durability of concrete pavements [40], even though in
those days the effects of freezing and thawing were compounded with those
from other kinds of difficulties.
The so-called "degree of saturation" hypothesis was the earliest attempt to
explain these effects. Since water expands on freezing by about ten percent, a
saturation of greater than about ninety percent should be necessary for trouble to develop. Early work [41] did indeed show a reasonable correlation with
saturation attainable by the aggregate, particularly under vacuum conditions, and field performance of the aggregate in pavement concrete. And,
although this mechanism has been amplified by subsequent work, it still
stands. If the degree of saturation of an aggregate particle is less than
somewhere in the region of 90 percent, there will be no damage to the concrete from this source. It should be remembered that, as saturation is determined ordinarily, an inhomogeneous material, such as most gravels, may
have some individual pieces critically saturated although the value determined on the overall sample may be considerably lower.
The degree of saturation hypothesis is the basis for a widespread
dependence on absorption tests to indicate an aggregate's probable performance. While the absorption value is not, of course, as precise an indicator
as is the saturation, for most aggregates of the usual compositions and
porosities the absorption is good enough, especially considering the simplicity of the test method, to be a fairly useful indication of the likelihood of trouble. Indeed there is some question of how much the subsequent work on test
methods has improved upon it [42].
The modern understanding of the freeze-thaw durability of aggregates in
concrete stems from the analysis of Verbeck and Landgren [43]. They
discussed three classes of behavior. In the first, called elastic accommodation, the porosity is small enough that even if it is saturated and the water
freezes, the expansion will be so small as to cause no difficulty. This critical
porosity is in the region of a fraction of a percent. Some rocks are of such a
low porosity, and they should be immune from freeze-thaw trouble.
The second class of aggregates is those with small pores and, therefore, low
permeabilities. The mode of action is based on the hydraulic pressure
hypothesis of Powers [44] for frost damage in hardened cement paste. The
expansion of the water in forming ice forces unfrozen water ahead of the
freezing front. If this flow must traverse a distance larger than a critical maximum, pressures large enough to fracture the matrix will be developed. The
critical distance in a hardened cement paste is of the order of 0.20 mm (0.008
in.); the critical distances in finely pored rocks, such as many cherts and
shales, are in the small coarse aggregate size range.
Therefore, if a piece of such a material is larger than this size and is
critically saturated, freezing should produce failure within the aggregate.
Such behavior is characteristic of those materials that produce pop-outs; the
failure is within the piece of aggregate, part remaining in the concrete and
the other part coming out with the flake of mortar.
The third class of aggregates is those with larger pores and, therefore, with
permeabilities so large that their critical distances are many centimetres (inches) or even metres (feet). Such materials cannot fail internally by the
hydraulic pressure mechanism, yet they still produce nondurable concrete.
The reason advanced is that unfrozen water forced out of the aggregate piece
has nowhere to go but into the paste, which has a small critical distance, so
failure occurs there, rather than within the aggregate. Analysis of this
mechanism shows that entrained air in the paste provides hardly any help,
and again there is a size effect, with larger-size coarse aggregate being more
deleterious. Recent studies have confirmed this maximum size effect, which
now seems well established [45-47].
Owing to its importance, the predictions of freeze-thaw durability has been
the subject of much research. The upshot of most of the early work was that
absorption and saturation were fair indicators of potential durability performance, and better than sulfate soundness tests. Unconfined freeze-thaw
testing is useful only for those aggregates with small critical distances that exhibit internal failure.
The result of this less-than-satisfactory situation has been that freezing
tests of concrete containing the aggregate in question have been considered
superior to tests on the aggregate alone. The development of such testing is
covered in another chapter in this volume. The current standard is ASTM
Recommended Practice for Evaluation of Frost Resistance of Coarse Aggregate in Air-Entrained Concrete by Critical Dilation Procedures (C 682),
which includes a bibliography of the work that led to its development.
Nevertheless, a test that could be performed on the aggregate alone is still
654
desirable. It has long been considered that the details of the pore structure of
the aggregate should be the most important factor in whether it will be
durable or nondurable to frost in concrete. Logically, a nondurable aggregate would comparatively rapidly acquire a high degree of saturation and
retain it more tenaciously. These properties are related to the absorptivity.
Limited work [8] did show that nondurable aggregates had higher absorptivities and larger rates of saturation increase when imbibing water by
capillarity.
Early studies [10] also showed that nondurable aggregates seem to have a
comparatively large proportion of small pores. It would seem that this matter
could be settled by the determination of pore-size distributions of a variety of
aggregate materials. The tool most used so far has been mercury
porosimetry, but the results of most of the studies that have been done
[31-36] have been neither clear-cut nor unanimous in their conclusions. Recent work has shown a good correlation between laboratory and field freeze
thaw durability of an aggregate in concrete and a factor combining its total
porosity and median pore size, with higher porosities and smaller pore sizes
being worse [51].
Some work has been interpreted as showing that mechanisms other than
ice-expansion and hydraulic pressure are important in freeze-thaw. Dunn
and Hudec [48] measured the frozen water in cold aggregate samples and
found that in nondurable materials only a little water became frozen.
Therefore, they postulated expansion processes that did not depend on conversion of the water to ice. This mechanism has some aspects of theories that
have been advanced for freezing action in hardened cement paste [49], and
may have pertinence for rocks containing large clay components, although
the limited further testing of the idea [50] seems to show the hydraulic
pressure hypothesis to be more important.
More work on the relationship between an aggregate's pore structure and
its durability to freeze-thaw remains to be done. It should be recognized, of
course, that the aggregate's properties are not the only determinant factors,
and that the nature of the concrete plays a role in the durability question,
even assuming the concrete to be properly air-entrained, so failure from this
source does not occur. The primary variable is the permeability of the paste
and the degree to which it permits water to enter and reach critical saturation
in aggregate pieces. The same piece of questionable aggregate near the surface of a slab may not pop out if the overlying paste is dense and impermeable, but will if that paste is more porous and permeable because of a
higher water/cement ratio.
References
[1] Lewis, D. W. and Dolch, W. L., "Porosity and Absorption," A S T M S T P 169, American
DOLCH ON POROSITY
655
[2] Dolch, W. L., "Porosity," A S T M S T P 169A, American Society for Testing and Materials,
1966, p. 443.
[3] Scheidegger, A. E., The Physics of Flow Through Porous Media, Macmillan, New York,
1957.
[4] Proceedings, RILEM/IUPAC Conference on Pore Structure and Properties of Materials,
Vol. 1, Part A, Academia, Prague, 1973.
[5] Washburn, E. W. and Bunting, E. N., "Porosity 6--The Determination of Porosity by the
Method of Gas Expansion," Journal, American Ceramic Society, Vol. 5, 1922, p. 112.
[6] Schumb, W. C. and Rittner, E. S., "A Helium Densitometer for Use with Powdered
Materials," Journal, American Chemical Society, Vol, 65, 1943, p. 1692.
[7] Washburn, E. W. and Bunting, E. N., "Porosity VII--The Determination of the Porosity
of Highly Vitrified Bodies," Journal, American Ceramic Society, Vol. 5, 1922, p. 527.
[8] Dolch, W. L., "Studies of Limestone Aggregates by Fluid-Flow Methods," Proceedings,
[9] DeHoff, R. T. and Rhines, F. N., Quantitative Microscopy, McGraw Hill, New York,
1968.
[10] Sweet, H. S., "Research on Concrete Durability as Affected by Coarse Aggregate,"
[11] Young, D. M. and Crowell, A. D., PhysicalAdsorption of Gases, Butterworth, London,
1962.
[12] Joy, A. S., "Methods and Techniques for the Determination of Specific Surface by Gas
Adsorption," Vacuum, Vol. 3, 1953, p. 254.
[13] Faeth, P. A. and Willingham, C. B., "The Assembly, Calibration and Operation of a Gas
Adsorption Apparatus," Mellon Institute of Industrial Research, Pittsburgh, Pa., 1955.
[14] Beebe, R. A., Beckwith, J. B., and Honig, J. M., "The Determination of Small Surface
Areas by Krypton Adsorption at Low Temperatures," Journal, American Chemical
Society, Vol. 67, 1945, p. 1554.
[15] Brunauer, S., Emmett, P. H., and Teller, E., "Adsorption of Gases in Multimolecular
Layers," Journal, American Chemical Society, Vol. 60, 1938, p. 309.
[16] Gregg, S, J. and Sing, K. S. W., Adsorption, Surface Area and Porosity, Academic Press,
New York, 1967.
[17] Carman, P. C., Flow of Gases Through Porous Media, Academic Press, New York, 1956.
[18] Wyllie, M. R. J. and Spangler, M. B., "Application of Electrical Resistivity Measurements
to Problems of Fluid Flow in Porous Media," Bulletin, American Association of Petroleum
Geologists, Vol. 36, 1952, p. 359.
[19] Dolch, W. L., "Permeability and Absorptivity of Indiana Limestone Coarse Aggregates,"
Thesis, Purdue University, 1956.
[20] Huang, J., Sherwood, W. C., and Wakao, N., "Determination of Mean Pore Radii in
Mineral Aggregates Using Permeability Measurements," Special Report No. 90, Highway
Research Board, 1966, p. 219.
[21] Powers, T. C. and Brownyard, T. L., "Studies of the Physical Properties of Hardened
Portland Cement Paste," Proceedings, American Concrete Institute, Vol. 43, 1947,
p. 865.
[22] Dolar-Mantuani, Ludmila, "Correlation Between Absorption of Single Aggregate Particles by Water-Drop Test and Immersion Tests," RILEM International Symposium
on Durability of Concrete, Prague, 1970, p. A-27.
[23] Linsen, B. G. and van den Heuvel, A., "Pore Structures," The Solid-Gas Interface,
E. A. Flood Ed., Vol. 2, Chapter 35, Marcel Dekker, New York, 1967.
[24] Dullien, F. A. and Batra, V. K., "Determination of the Structure of Porous Media," in
Flow Through Porous Media, Vol. 1, p. 1, American Chemical Society, Washington, 1970.
[25] deBoer, J. H., "The Shapes of Capillaries," in The Structure and Properties of Porous
Materials, D. H. Everett and F. S. Stone, Eds., Butterworth, London, 1958, p. 68.
[26] Everett, D. H., "Adsorption Hysteresis," in The Solid-Gas Interface, E. A. Flood, Ed.,
Chapter 36, Marcel Dekker, New York, 1967.
[27] Brunauer, S., Skalny, J., and Odler, I., "Complete Pore Structure Analysis," Proceedings,
RILEM/IUPAC Conference on Pore Structure and Properties of Materials, Vol. 1,
Academia, Prague, 1973, p. C-3.
[28] Ritter, H. L. and Drake, L. C., "Pore Size Distribution in Porous Materials--Pressure
656
Porosimeter and Determination of Complete Macropore-Size Distributions," Industrial
and Engineering Chemistry, Analytical Edition, Vol. 17, 1945, 782.

[29] Rootare, H. M., "A Short Literature Review of Mercury Porosimetry as a Method of
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
Measuring Pore-Size Distributions in Porous Materials, and a Discussion of Possible

Sources of Errors in This Method," Aminco Laboratory News, Vol. 24, No. 3, 1968.
Winslow, D. N. and Diamond, S., "A Mercury Porosimetry Study of the Evolution of
Porosity in Portland Cement," Journal of Materials, Vol. 5, 1970, p. 564.
Hiltrop, C. L. and Lemish, J., "Relationship of Pore-Size Distribution and Other Rock
Properties of Serviceability of Some Carbonate Aggregates," Bulletin 239, Highway
Meininger, R. C., thesis, University of Maryland, 1964.
Walker, R. D. and Hsieh, T., "Relationship Between Aggregate Pore Characteristics and
Durability of Concrete Exposed to Freezing and Thawing," Record No. 226, Highway
Lange, H. and Modry, S., "Determination ofthe Frost Resistance of Limestone Aggregates
in the Light of Porosity Investigations," RILEM International Symposium on Durability
of Concrete, Prague, 1970, p. B-129.
Sherwood, W. C., Huang, J., Dudash, J. J., and McShee, K. H., Report No. 71-R39,
Virginia Highway Research Council, 1972.
Koh, Y. and Kamada, E., "The Influence of Pore Structure of Concrete Made With
Absorptive Aggregates on the Frost Durability of Concrete," Proceedings, RILEM/IUPAC
Conference on Pore Structure and Properties of Materials, Vol. 2, 1973, p. F-4S.
Hsu, T., Slate, F. O., Sturman, G. M., and Winter, G., "Microcracking of Plain Concrete and the Shape of the Stress-Strain Curve," Proceedings, American Concrete Institute, Vol. 60, 1963, p. 209.
Scholer, C. F., "The Role of Mortar-Aggregate Bond in the Strength of Concrete,"
Record No. 210, Highway Research Board, 1967, p. 108.
Roper, H., Cox, J. E., and Erlin, B., "Petrographic Studies on Concrete Containing
Shrinking Aggregate," Journal of Research, Portland Cement Association Research
and Development Laboratories, Vol. 6, No. 3, 1964, p. 2.
Woods, K. B., Sweet, H. S., and Shelburne, T. E., "Pavement Blowups Correlated with
Source of Coarse Aggregate," Proceedings, Highway Research Board, Vol. 25, 1946,
p. 147.
Lewis, D. W. and Woods, K. B., "Research as Related to the Development of Aggregate
Specifications," Proceedings, 35th Annual Road School, Purdue University, 1949, p. 155.
Gaynor, R. D. and Meininger, R. C., "Investigation of Aggregate Durability in Concrete,"
Verbeck, G. J. and Landgren, R., "Influence of Physical Characteristics of Aggregates
on Frost Resistance of Concrete," Proceedings, American Society for Testing and
Materials, Vol. 60, 1960, p. 1063.
Powers, T. C., "A Working Hypothesis for Further Studies of Frost Resistance of Concrete," Proceedings, American Concrete Institute, Vol. 41, 1945, p. 245.
MacInnis, C. and Lau, E. C., "Maximum Aggregate Size Effect on Frost Resistance of
Concrete," Proceedings, American Concrete Institute, Vol. 68, 1971, p. 144.
Stark, D. and Klieger, P., "Effect of Maximum Size of Coarse Aggregate on D-Cracking in
Concrete Pavements", Record No. 441, Highway Research Board, Vol. 33, 1973.
Klieger, P., Monfore, G., Stark, D., and Teske, W., "D-Cracking of Concrete Pavements
in Ohio," Portland Cement Association, Skokie, Ill., 1974.
Dunn, J. R. and Hudec, P. P., "The Influence of Clays on Water and Ice in Rock Pores,"
Report RR 65-5, New York State Department of Public Works, Part 2, 1965.
Litvan, G. G., "Phase Transitions of Adsorbates IV--Mechanism of Frost Action in
Hardened Cement Paste," Journal, American Ceramic Society, Vol. 55, 1972, p. 38.
Cady, P., "Mechanism of Frost Action in Concrete Aggregates," Journal of Materials,
Vol. 4, 1969, p. 294.
Kaneuji, Makoto, "Correlation Between the Pore Size Distributions and Freeze Thaw
Durability of Coarse Aggregates in Concrete," thesis, Purdue University, 1978.
R. C. M e i n i n g e r ~
Chapter 38 Abrasion Resistance,

Strength, Toughness, and Related
Properties
Introduction
Aggregates for use in concrete, and to a large extent the individual
constituent particles, must possess a reasonably high degree of inherent
strength, tenacity, and stability to resist, without detrimental degradation,
the static and dynamic stresses, impacts, and wearing actions to which it
may be exposed both in concrete production operations and, ultimately, in
concrete in service. In many uses of concrete the roughest treatment to
which an aggregate may be subjected in terms of mechanical forces and
attrition is in the concrete production process. Actions involved in production are fairly predictable; those in service may be less predictable. In
end uses such as beams, columns, covered slabs, walls, footings, and other
mostly structural or architectural elements the only real strength property
needed of the aggregate after the concrete is in place is that necessary to
give the concrete enough strength to resist the distributed service loads.
Other uses may expose aggregate near or at the surface to a variety of
localized impact and abrasive stresses which will be of overriding importance in aggregate evaluation and selection, Examples include pavements and slabs exposed to heavy traffic and hydraulic structures subject
to eroding forces of moving water and sediment material.
Some recognition must be given to the consideration of an aggregate as
a whole versus the properties of the individual particles present. In many
crushed stone aggregates and some sands and gravels, the lithology of the
particles and, consequently, their mechanical properties show little variation. In other cases, great variations may exist among the mineral and
rock particles involved. A small percentage of weak particles in most inI Director of engineering research, National Sand and Gravel Association, National Ready
Mixed ConcreteAssociation, SilverSpring, Md. 20910.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
658
stances is not objectionable and may be thought of as an impurity or deleterious substance and therefore not properly the subject of this discussion.
These weaker or less abrasion resistant particles allowed in small percentages
may cause a slight increase of fines during mixing or may contribute to
occasional surface imperfections not particularly harmful to overall performance. In other instances, eliminating even very small percentages of
friable or weak particles is desired. An example is high velocity hydraulic
structures where high surface strength and very accurate surface alignment
is needed to resist cavitation damage which can spread rapidly once started.
Concrete carrying a distributed static or dynamic load, and not subjected
to local stress concentrations, requires aggregate that will bond with the
surrounding cement paste permitting the transfer of stress through the
aggregate particles. Here the particles are expected to possess enough
strength and rigidity to carry the stresses without mechanical breakdown or
excessive deformation--for usual aggregates and typical stress levels in
concrete this is no problem. The rock, mineral, or synthetic materials
making up aggregates are usually much stronger in compression or tension
than the concrete; and the weakest link in the system is the bond between
the paste and aggregate. In some cases, however, aggregate materials are
used which possess compressive strength of the same order of magnitude as
the concrete in which they are used. Examples are lightweight aggregates,
limerock materials used in Florida, and volcanic cinders sometimes used
in concrete. These weaker materials are used because of economic reasons
or to produce concrete of lighter unit weight. An exception, where high
strength aggregate materials will probably be required, is the production of
very high strength concrete--higher than 35 or 70 MPa (S000 or 10 000 psi).
Impact, abrasion, scuffing, attrition, and other wearing actions, either
dry or in the presence of water, can be an important factor. These loadings
or actions usually are applied to the surface of the concrete so that only
the properties of the aggregate at or near the surface are of any substantial
importance in determining whether the long term performance of the concrete will be acceptable. In some cases special aggregates, concrete mixtures,
or coatings may be used at the surface to improve the surface properties
locally.
Surface actions are of two basic modes--impact and rubbing. Impact is
where hard particles or objects impinge against the concrete surface with
enough momentum to cause shattering, yielding, or debonding of aggregate
particles. Materials with good impact resistance are said to possess good
toughness. In the rubbing or scratching mode, damage can be inflicted by
the movement of particles or objects on the surface under enough load to
cause identation, for the relative hardnesses involved, and therefore cause
scratching or gouging as movement occurs. Scratches and gouges from
relatively large particles or objects are on a macroscopic scale and are
termed wear. In other words, the amount of material removed by such
MEININGER ON PHYSICAL PROPERTIES
659
action, if repeated over and over again, will wear away a significant quantity
of surface mortar exposing more and more coarse aggregate with time.
Rubbing or scratching action from much smaller particles--on a microscopic scale--does not have the capacity to remove much material but it can
cause a polishing action at the surface. In highway uses, and occasionally
other types of service, aggregates that allow rapid polishing can contribute
to low skid resistance of wet surfaces.
The arrangement of the discussion in this chapter is based on the types
of action concrete and aggregate might be expected to resist: (1) distributed
stress of the noninstantaneous type whether it is compression, tension, or
shear stress, (2) instantaneous impact loadings where significant momentum
is involved, and (3) scratching and rubbing actions which on a macro scale
produce wear and on a micro scale involve polishing. The main divisions
which follow are organized with the purpose of discussing the significance
and measurement of aggregate properties, and any aggregate specifications,
relating to the three general classes of mechanical action.
1. Distributed Stress--The sections on "Strength of Aggregate Material"
and "Elastic Properties" apply to distributed stress.
2. Impact--The section on "The Los Angeles Abrasion Test" is included
under impact since the test involves mostly impact with some wearing
attrition. A section on "Impact Resistance and Toughness" is also included.
3. Scratching and Rubbing--A section on "Hardness, Wear, and Skid
Resistance" is included along with a section on "Wet Abrasion and Attrition
Tests" which is a more specialized topic pertaining to testing of aggregates
wet rather than dry. There is an element of impact in these tests as well as
surface attrition and leaching of fines.
Strength of Aggregate Material

Measurements of unconfined and confined compressive (or crushing)
strength of rock and mineral specimens drilled or sawn from rock ledges,
or in some instances cobbles or boulders from gravel deposits, are quite
often made in connection with rock mechanics investigations for foundations, tunnels, and various mining applications. Certain inferences can be
drawn from these data as to the nature and distribution of the strengths of
the materials constituting aggregate particles. However, there are several
problems involved in translating this information to the behavior of aggregate in concrete: (1) the conditions of confinement in concrete are certainly
different than that of unconfined tests and triaxial tests may not truly
represent the situation in concrete, (2) strength tests of rock specimens
usually involve measures to distribute the stress evenly over the entire cross
section, whereas in concrete localized stress concentrations may develop
due to the nonhomogeneous nature of the concrete, and (3) rock properties
are most often reported for specimens that have a minimum of flaws,
660
cracks, and fissures. However, the strength of an individual aggregate

particle can be influenced greatly by defects in the particle.
Udd [1] 2 gives a review of the methods used in testing rock and some
examples of the range and variability of results. Compressive strength increases with confinement and both elastic and plastic deformations occur.
There are many sources for data on compressive strength of rock and in
some cases tensile strength or modulus of elasticity [2-10]. Bond [9] shows
compressive strengths of small cubes and cylinders cut from rock material.
For limestone the values ranged from 90 to 270 MPa (13 000 to 39 000 psi),
for traprock 10S to 235 MPa (15 000 to 34 000 psi), and for granite 85 to
275 MPa (12 000 to 40 000 psi). These values represent ranges from a
number of sources. He also reports compressive strength results from one
gravel source ranging from 165 to 235 MPa (24 000 to 34 000 psi). The
great bulk of the strength data available is for stone materials where specimens of the required size can be obtained and usually the material is less
variable. Gravel deposits on the other hand present difficult problems in
attempting to measure strength. Strength specimens can only be cut from
sizes larger than used in many products and the lithology of many gravel
deposits is quite variable, necessitating the testing of a large number of
specimens to get meaningful results. The assumption can be made that
the strength of the rock materials in gravels is similar to that of the rock
materials from which the gravels were derived. Tests of two types of rock
used as aggregate by Iyer et al [8] yielded, for a limestone source, compressive strengths of 105 to 200 MPa (15 000 to 29 000 psi) and tensile
strengths of 2.1 to 6.2 MPa (300 to 900 psi), and for a quartzite source,
compressive strengths of 340 to 470 MPa (49 000 to 68 000 psi) and tensile
strengths of 7.6 to 15.8 MPa (1100 to 2300 psi). Kaplan [5] tested gravel,
limestone, granite, and basalt aggregates and obtained compressive strengths
ranging from about 160 MPa (23 000 psi) for the gravel and the granite on
the low end to 305 MPa (44 000 psi) for the basalt on the high end. The
tensile strength measured for the gravel, limestone, and granite ranged
from about 6.2 to 11.7 MPa (900 to 1700 psi). The basalt had a tensile
strength of 15.2 MPa (2200 psi). Of the almost 1000 compressive strengths
reported by Woolf, for rock from which crushed stone coarse aggregates
are made, in his state-by-state summary [2], only about 60 results indicated
compressive strength levels of about 70 MPa (10 000 psi) or less.
From the foregoing it is apparent that the great majority of the aggregatemaking rock materials are strong in compression, over 70 MPa (10 000 psi),
and like all brittle materials relatively weak in tension. There is also some,
probably fairly small, proportion of aggregates used for concrete that are
weak enough, less than 70 MPa (10 000 psi), that their strength characteristics may be an important limitation on the properties of concrete made
661
with these materials. Kaplan [5] in using a number of strong aggregate

materials in concrete concluded that "no relationship between the strength
of the coarse aggregate and the strength of concrete was established. This
finding should not be taken to mean that aggregates of low strength will
not affect concrete strength." Johnston [3] using a 48 MPa (6900 psi) sandstone coarse aggregate in concrete and Collins and Hsu [7] using a 28 MPa
(4000 psi) oolite coarse aggregate in concrete were both able to produce
concrete over 40 MPa (6000 psi) at 28 days age. These data suggest, as
does broad experience with lightweight structural aggregates and experience
using volcanic cinders and coral in concrete, [11,12] that aggregate strength
need only be of the same order of magnitude as the concrete strength
needed. Tanigawa [13] in his model analysis with various aggregate and
mortar strengths concluded that the best results in terms of delay in formation of bond cracks at the coarse aggregate-mortar interface were obtained
when the aggregate and mortar were about the same strength.
It is recognized that coarse aggregate to mortar bond is extremely important to concrete strength and that a reasonably high strength aggregate
with excellent bonding characteristics is necessary in the production of very
high strength concrete over 35 or 40 MPa (5000 or 6000 psi). The strength
of individual fine aggregate particles is apparently not too important as
long as some minimum level is achieved, principally so that it does not
break down and create excess fines and surface area during concrete production. Bond is not as important with fine aggregate because of its large
surface area-to-volume ratio. The influence of fine aggregate on concrete
strength is almost exclusively through its shape, texture, and grading
effects on mixing water demand needed for workability.
Fundamental strength tests of specimens cut from aggregate materials
are not used for the specification or qualification of coarse aggregates for
concrete. As a matter of fact, no static load type of strength indicator is
generally used in the United States or Canada. In Britain two strength
tests run on a collection of particles are detailed in the British Standard for
Concrete Aggregates B5812-75. These are the "Crushing Value" and the
"Ten Percent Fines Value" both of which are run on 10 to 14 mm aggregate particles placed in a confining base. Crushing value is simply the percentage of fines generated which pass the 2.36 mm sieve when a load of
400 kN is imposed on the collection of particles. The 10-percent fines value
is conversely the load in kilonewtons needed to produce 10-percent fines as
defined in the crushing value test. The crushing value normally is used for
high-strength aggregate materials where excessive fines are not produced
under the stipulated load. For weak and lower quality aggregates the 10percent fines value is preferred [14]. Figure ! shows that the British crushing value has a reasonably good relationship with rock compressive strength.
Kaplan also found that the relationship also held for rock tensile and flexural strength, at least for the fairly strong aggregate materials that he
662
9 K=plan
| dohnst'on
50,000,
500
40,000
4~
dk~
qJ
30,OOO
. ~kSO
ft.
.~o0
qj
150
~J
u~
v)
0j
t
cL
20,000
o
t,j
I0, O0 0
-~-o
o
0
0
0
I0
British
2.0
Crushlnoj
30
40
EO
Value of A99recjate
F I G . 1--Relationship between rock compressivestrength and the British crushing value of

aggregate derivedfrom the rock.
investigated. Johnston, with the weaker aggregates that he investigated, did

find a trend toward weaker concrete at the same water/cement ratio for
aggregates with lower compressive strength and high British crushing values.
The significance of strength tests of aggregates is not well established.
There are no ASTM test methods or specification requirements in this
area. The usual practice is to evaluate the aggregates' performance in concrete to see if it will produce concrete of the needed strength. Aggregate
strength may be an indicator of other aggregate properties, but relationships
have not been established.
Elastic Properties
Modulus of elasticity and related values such as the Poisson's ratio of
concrete aggregates is not included in aggregate specifications and there
are no ASTM aggregate test methods designed to yield this information.
The influence of aggregate source on concrete modulus of elasticity is determined normally by testing concrete mixtures containing each of the
663
aggregates. As is the case for compressive strength a good deal of information from rock mechanics investigations is available concerning rock deposits [1, 6,15,16], but little is known about the elastic properties of gravel
or sand particles other than what can be inferred from tests of parent rock
material. A number of investigators have also determined rock modulus of
elasticity in connection with researches of aggregate and concrete properties [3,5, 7,8,17]. There has been a good deal of activity in the more theoretical area of developing equations to predict the resulting modulus of
elasticity of a composite material from a knowledge of the elastic properties
of the aggregate and the matrix [18-22].
Figure 2 gives an indication of the range of Young's modulus of elasticity
for various rock types. Both static and dynamic E data are included. Generally, the dynamic values are higher, and there has been shown to be a
good linear relationship between dynamic and static values determined in
the laboratory. A direct linear relationship has also been found between
rock compressive strength and modulus of elasticity [15.16].
Both in compression and tension, stress-strain curves for stone [6] and
for cement paste [23,24] are normally fairly straight lines indicating that
the aggregate and matrix are reasonably elastic. Concrete and mortar, on
the other hand, have a curved stress-strain curve when the stress exceeds
about 30 percent of ultimate strength. This is due to the formation of bond
cracks and slipping at the aggregate paste interface. With a curved stressstrain curve the modulus of elasticity computed for the concrete will vary
depending on which of the several recognized definitions is used. Because
of this and because unknown factors can affect the influence of the pasteaggregate bond on the stress-strain curve there is no simple relationship
between aggregate modulus of elasticity and concrete modulus of elasticity.
As a matter of fact, it does not always hold that an increase in aggregate
modulus of elasticity will increase the E of the concrete even if the properties of the paste are held relatively constant. An example of this is contained
in Houghton's paper [17] where natural rounded aggregate with an E of
about 41 400 MPa (6 x 106 psi) produced concrete with E ranging from
28 000 to 34 000 MPa (4 to 5 106 psi) at 90 days compared to a quarried
aggregate with an E of about 62 000 MPa (9 X 106 psi) which produced
concrete with an E of 21 000 to 28 000 MPa (3 to 4 X 106 psi). Hirsch [22],
however, shows data where aggregates with E values of about 14, 34, 62,
76, and 207 X 103 MPa (2, 5, 9, 11, and 30 X 106 psi) were used in different proportions in concrete indicating "that the modulus of elasticity of
concrete is a function of the elastic moduli of the constituents. An increase
or decrease in the modulus of either the aggregate or matrix constituent
will produce a corresponding effect on the concrete." The equations advanced by Hirsch and others are complex and include, of necessity, various
assumptions.
664
TESTS A N D PROPERTIES OF C O N C R E T E A G G R E G A T E S
AMPHIBOLITE
ARGILLITE
BASALT
BORAX
CHALK
CHERT
CONGLOMERATE
DIABASE
DIORITE
DOLOMITE
EPIDOSITE
GABBRO
GNEISS
..=~
0=o
9149
~ 9
.~.
~.oo~
9
o~
Dee 9
..
* ,,=
9 =,.
~,~;.~9oz~...
GRANITE
GREENSTONE
H E M A T I T E ORE
HORNSTONE
JASPILITE
bJ
(3.
)F-
LIMESTONE
L)
O
0C
MARLSTONE
MAGNETITE
MARBLE
ONE VALUE AT 29.0

ONE VALUE AT 20.5
DO
ORE
?"
t~"
META-RHYOLITE
MONZONITE
PEGMATITE
PHYLLITE
PORPHYRY
PYROXENITE
QUARTZ
QUARTZITE
SALT
9 o 9
""
O0
SANDSTONE
~ o ~ " 8.
o STATICALLY DETERMINED
~o
ooo
9 DYNAMrCALLY DETERMINED
SCHIST
SHALE
SHONKINITE
SILTSTONE
SKARN
SLATE
SYENITE
TACTITE
TUFF
Q o
99
. I
5
YOUNG'S
I0
MODULUS
15
20
( p s i x 10 6)
FIG. 2--Relationship between rock type and Young's modulus. From Ref 1.1 psi
6895 MPa.
1 0 6 -~
Less work has been done on the tensile modulus of elasticity. Johnston
[3,24] gives values of tensile modulus of elasticity of rock. The values are
somewhat lower than the corresponding compression values. Houghton [17]
discusses the concept of predicting the tensile strain capacity of mass concrete so that tensile strains caused by thermal volume changes can be
accommodated by the concrete without cracking. Relationships with age
MEININGER ON PHYSICAL PROPERTIES 665
for tensile strength and modulus of elasticity enter into the determination
of strain capacity for rapidly loaded concrete. Creep is a third factor which
enters into the estimate when stress is applied slowly.
Iyer et al [8] measured E and Poisson's ratio on aggregate rock material
both wet and dry and before and after 100 cycles of freezing and thawing.
Before freezing E was about 80 000 MPa (11 to 12 106 psi) for both the
quartzite and limestone rock. Poisson's ratio averaged about 0.14 for the
quartzite and 0.30 for the limestone. After freezing only the wet limestone
showed a significant reduction in E to about 62 000 MPa (9 X 106 psi).
Brandon [6], along with compressive and tensile strength of rock, gives
values for Poisson's ratio at stress levels well below half of ultimate strength.
For the strong rocks, over 70 MPa (10 000 psi) compressive strength,
Poisson's ratio ranged from about 0.06 and 0.08 for several sandstones and
quartzites to values over 0.20 for a variety of rock types. Weaker rocks of
the sandstone, graywacke, basalt, and granite varieties ranged from 0.04
to 0.13. Counter to the general trend, one 90 MPa (13 000 psi) conglomerate
had a Poisson's ratio of 0.05 and one 32 MPa (4600 psi) sandstone had a
Poisson's ratio of 0.21. Hansen [10] gives measurements of the complete
range of mechanical properties of three aggregates and discusses the procedures used in testing very small specimens.
The stiffness of an aggregate can have an important effect on the drying
shrinkage and creep of concrete and mortar. Hobbs [25] for concrete and
Chamberlin [26] for mortar show a relationship of decreased drying shrinkage with increasing aggregate modulus of elasticity. Chamberlin estimated
the E of sands from dynamic E measurements on mortar. He was able to
use the relative effective elastic modulus as a quality indicator which correlated with sand equivalent. Lott and Kesler concluded in their studies of
crack propagation in plain concrete [18] that cracks are arrested only when
the E of the aggregate is greater than the E of the paste. Philleo [27] indicates that except for cases in which the creep of the aggregate itself is a
factor, which is usually not the case, the influence of the aggregate on
creep of concrete is confined to the elastic constants and the concentration
of the aggregate in the concrete.
Los Angeles Abrasion

The Los Angeles abrasion test--ASTM Test for Resistance to Abrasion
of Small Size Aggregate by Use of the Los Angeles Machine (C 131) and
ASTM Test for Resistance to Abrasion of Large Size Coarse Aggregate by
Use of the Los Angeles Machine (C 535)--is accepted and used almost
universally in the United States as a specification qualification test for
coarse aggregate for concrete. It is also used widely around the world to
evaluate aggregates. The L. A. Abrasion Test has supplanted the older
Deval Abrasion Tests--ASTM D 2 and D 289--which have been withdrawn
666
as ASTM standards. The L. A. Abrasion Test was developed to overcome

some of the defects found in the Deval test. It is much quicker to prepare
a sample and obtain the resulting losses. Loss values are higher and encompass a larger range than that of the Deval Test. The Deval Test, which
consisted of tumbling cubical pieces of rock (ASTM Methods D 2), or
graded aggregate with an abrasive charge of steel spheres (ASTM Methods
D 289), in a cast iron cylinder for 10 000 revolutions, is discussed in more
detail by Woolf [28]. The Deval Tests are still on the books as AASHTO
standards and are referenced in a few specifications in the United States.
But in view of its declining use in favor of the L. A. Abrasion Test the
Deval Test will not be discussed further here.
The L. A. Abrasion Test, ASTM Test C 131, was first adopted as a tentative ASTM standard in 1937. At first it provided for testing only 38.1 or
19.0 m m (11/2 or aA in.) maximum size aggregates (Gradings A and B).
Later 9.5 m m (3/8 in.) and No. 4 sizes were added (Gradings C and D). In
1947 a revision was prepared to add three more gradings for large aggregate
sizes--75 m m (3 in.), 50 m m (2 in.), and a second narrower 38.1 m m
(11A in.) m a x i m u m size grading (Gradings E, F, and G ) - - w i t h an increased
sample size from 5000 to 10 000 g and a doubling of the number of revolutions from 500 to 1000 for these larger sizes. This was an attempt to obtain
generally corresponding percentages of wear for each grading when a material of uniform quality was tested. Such uniformity of results was usually
obtained for Gradings A, B, C, and D, within a reasonable margin of error.
For the larger gradings, however, agreement was found in some instances
and contrary data in other cases: To clarify the situation in 1964 and 1965
the larger sizes were deleted from ASTM Test C 131 and a new method
was adopted, ASTM Test C 535, for them.
For a detailed description of the equipment and procedures see ASTM
Test C 131. The method was revised in 1976 to add a precision statement.
Briefly the method is a dry abrasion and impact test which involves placing
a sample of aggregate and a charge of steel balls into a hollow steel cylinder
that is 508 m m (20 in.) long and 711 m m (28 in.) in diameter, is closed on
the ends, and has a single rigid steel shelf extending 89 m m (31/2 in.) into
the chamber. The cylinder is rotated at 30 to 33 r p m for 500 revolutions.
Percent loss or wear is obtained as the difference between the original
sample weight and the amount of the final sample coarser than the 1.70 m m
(No. 12) sieve, expressed as a percentage of the original weight. Often the
loss is determined at 100 and 500 revolutions. For a uniform material the
ratio of the loss after 100 revolutions to the loss after 500 revolutions should
not greatly exceed 0.2. A large number of flat or elongated pieces might
also cause this ratio to increase. A nominal 5000-g sample is used for the
nominal gradings given in Table 1. The required charge of nominal 420-g
steel spheres is also indicated in Table 1.
The precision statement recently added to ASTM Method C 131 is based
1250
. . . .
.
.
.
1250
3/4 in.)
1 in.)
A
B
C
Grading
19.0 to
25.0 mm
(1 to
38.1 to
25.0 mm
(U/2 to
1250
2500
.
.
.
1/2 in.)
19.0 to
12.5 m m
(3/4 to
1250
2500
3/8 in.)
12.5 to
9.5 m m
(1/2 t o
. . . .
. . . . .
2500
1/4 in.)
9.5 to
6.3 m m
(3/8 to
W e i g h t in E a c h Sieve F r a c t i o n , g
.
.
2500
No. 4)
6.3 to
4.75 mm
(1/4 to
T A B L E 1 - - S a m p l e grading and steel sphere charge, A S T M Test C 131.
5000
...
No. 8)
4 . 7 5 to
2.36 mm
(No, 4 to
12
11
No. o f Steel
Spheres
F~
.,q
O
"0
m
"In
F"O
"0
-T
.<
IT1
z
z
668
on 19.0 mm (3A in.) maximum size aggregate samples distributed to a

large number of laboratories throughout the United States by the AASHTO
Materials Reference Laboratory. The samples were tested by both ASTM
and AASHTO procedures which differ only in minor respects for this particular test. Multilaboratory coefficient of variation was found to be 4.5
percent and single operator coefficient of variation 2.0 percent. These data
are based on samples ranging in loss from 10 to 45 percent. In another
precision investigation within one state the data show a higher level of
variability on samples having a loss of 13 to 18 percent [29].
The same equipment is used for ASTM Test C 535. Present Gradings
1, 2, and 3 correspond to the E, F, and G gradings deleted from ASTM
Test C 131. They are given in Table 2 with the required charge of steel
spheres. Note that 10 000-g samples are used and that 1000 revolutions of
the drum are used. Loss is again determined using the 1.70 mm (No. 12)
sieve, and an indication of uniformity can be obtained by comparing the
loss after 200 revolutions with that after 1000 revolutions. The ratio should
not be greatly different from 0.2 for uniform material.
Unfortunately no data are available at this time to use in the development of a precision statement. Due to the limited application of large size
aggregate in construction today, this method is not utilized as much as
ASTM Test C 131. Caution should be exercised in comparing values obtained from ASTM Tests C 131 and C 535. They may give different results
for similar materials. For this reason desirable specification limits for
larger sizes may not necessarily be the same as those judged to be needed
for the smaller sizes of coarse aggregate.
Wear or loss in the L. A. Test is the result of impact and surface abrasion. Impact probably causes more loss, at least during early revolutions of
the drum before any fines are created which might tend to cushion the
impact forces. To some extent harder rocks and minerals, even though
they are strong, tend to shatter more than softer materials which can better
absorb the force. On the other hand, softer minerals will be more susceptible to surface wearing; and the product of the degradation action in that
case may be more of a dust rather than the larger angular pieces resulting
from the shattering of hard pieces.
Even though it is widely used in specifications, L. A. abrasion loss cannot be consistently related to the performance of concrete. Judgment may
be needed in establishing limits for different circumstances; and, in some
cases where aggregate performance is acceptable in concrete, the L. A.
abrasion limits have to be ignored.
ASTM Specification for Concrete Aggregates (C 33) now includes a
modular format where many of the required properties of coarse aggregates
are varied depending on severity of exposure to weather and the end use of
the concrete. However, due to the lack of enough significantly explicit data
1
2
3
Grading
2500
. . . .
.
.
.
75 to 63 mra
(3 to 21/z in.)
.
.
.
.
2500
.
63 t o 5 0 m m
(21/2 to 2 in.)
5000
5000
50 to 38.1 m m
(2 to lt/2 in.)
5"000
5000
38.1 to 25 m m
(11/2 to 1 in.)
W e i g h t in E a c h Sieve Fraction, g
T A B L E 2 - - S a m p l e grading and steel sphere charge, A S T M Test C 535.
...
...
5000
25 to 19 m m
(1 to 90 in.)
12
12
12
No. of Steel
Spheres
O~
O~
r
m
co
-11
0
"13
m
-n
p-
09
"1"
.<
0
z
m
m
z
z
G)
m
670
relating concrete performance with L. A. abrasion wear value, the limit has
been set at 50 percent for all categories. In addition, judgment provisions
such as the following in ASTM Specification C 33 have had to be inserted
in many specifications to allow use of known satisfactory materials:
"Coarse aggregates having test results exceeding the limits specified . . .
may be accepted provided that concrete made with similar aggregate from
the same source has given satisfactory service . . . . or in the absence of
a demonstrable service record, provided that the aggregate produces concrete having satisfactory characteristics when tested in the laboratory."
ASTM Specification C 33 specifies requirements for a wide range of aggregate sizes. The determination of which grading of ASTM Tests C 131 or
C 535 is to be used to test an aggregate product is based on the requirement that the L. A. loss or wear is to be determined using the grading
most nearly corresponding to the grading to be used in the concrete.
Crushed air-cooled blast-furnace slag is excluded from the L. A. abrasion
requirement. In its place a minimum compact unit weight is required.
Smith [30] used coarse aggregates in concrete with L. A. abrasion losses
ranging from 13 to 39 percent. He found no significant correlation between
L. A. loss and the abrasion of concrete using three different test procedures: (1) Davis steel ball, (2) dressing wheel, and (3) the Ruemelin shot
blast apparatus. The most significant correlations with concrete abrasion
resistance were with concrete strength and water/cement (w/c) ratio. The
soft limestone did decrease the concrete abrasion resistance at strengths
below 55 MPa (8000 psi). Above that strength no coarse aggregate effect
could be detected.
The loss in the L. A. abrasion test has been correlated by Woolf [31]
with the strength of concrete prepared with a wide variety of aggregates.
However, laboratory studies reported by BIoem and Gaynor [32] failed to
show such a relationship with strength. Jumper [33] and Walker and
Bloem [34], in researches where aggregates of high and low L. A. abrasion
loss were blended to produce a range of L. A. values, did show some correlation of reductions in L. A. abrasion values with increased strength.
Jumper blended San Fernando Valley aggregate to obtain L. A. values
from 42 to 58 percent. Over that range the decrease in compressive strength
for constant cement factor concrete was from about 27 to 23 MPa (3900 to
3300 psi). Walker and Bloem did not detect any dependence of concrete
compressive streng~th on L. A. abrasion loss, but they did find that high
L. A. abrasion losses did lower flexural strength with the aggregates
studied. Bartel and Walker [35] and Walker [36] showed a slight drop-off
of compressive and flexural strength with increased L. A. abrasion losses.
With respect to the freezing and thawing durability of aggregates tested
in concrete specimens, Gaynor and Meininger [37] did not find a relationship between concrete durability and L. A. abrasion loss at 500 revolutions
or 100 revolutions. This study included 56 coarse aggregates.
Nowlen [38] in research into the effectiveness of aggregate interlock in

concrete pavement joints compared the performance of three gravels with
17, 28, and 46 percent L. A. loss, respectively. The lowest L. A. loss material proved to be most effective in transferring load from one slab to
another in repetitive loading tests. The highest loss gravel was the least
effective in transferring load.
Hobeda [39] shows a good relationship between the British crushing
value and the L. A. abrasion loss in two figures apparently originally appearing in a 1948 British paper by Shergold. Hobeda also reports correlations with the Swedish impact test. Houston [40] experimented with the use
of the L. A. abrasion test for evaluating the quality of lightweight synthetic
aggregates. He developed a sandblast test which was able to discriminate
quality better. He found a relationship of loss in the sandblast test with
porosity of the lightweight aggregate. The Los Angeles abrasion test has
not been used widely for lightweight aggregates.
In comprehensive research into aggregate degradation, West et al [41]
studied the mechanism of abrasion in the L. A. Machine. Fine and coarse
grained limestone aggregates were tested in Grading A with the loss and
gradation of the sample determined after each 50-revolution increment.
Also particles in the larger sizes were marked and identified. More breakage of the coarse grained material occurred in the larger sizes, and the
particles held their angularity longer than did the fine grained limestone
which suffered more surface loss by wear. Percent loss increased in a
straight line relationship with number of revolutions. Figure 3 shows percent minus the 180/zm (No. 80) sieve plus A, which is the amount of dust
lost in the test, versus percent loss determined in the traditional manner on
the 1.70 mm (No. 12) sieve. Lines have been added to show the average
trends for various rock classifications. These data represent tests of aggregates from a number of states.
The granite and gneiss category had high loss through the 1.70 mm
(No. 12) sieve but comparatively low values of minus 180/zm (No. 80) dust.
The carbonates and basalts were intermediate, on the average; but the
spread of the data was large. Sandstones gave a comparatively high amount
of dust for the corresponding loss percentages through the 1.70 mm (No.
12) sieve. West and his co-authors concluded that for carbonate rocks
there was a good relationship between degradation value, an empirical
number derived from field service, and L. A. abrasion loss, and a good
relationship between L. A. abrasion and unconfined freeze-thaw deterioration. For igneous and metamorphic rocks L. A. abrasion correlated with
the empirical degradation value and grain size. Larger grain size rock
showed higher L. A. losses. No good correlations with field performance
were found for basalts in terms of L. A. abrasion loss. Woolf presents some
data [28] showing that L. A. abrasion losses of rounded and crushed
particles from the same parent rock do not vary much.
672
&
70
60
/
/
/
/
50
Oj 40
;if
. .. /
to
20
.~,'~
;~;
- -
GRANITE a GNEISS
BASALT
---- 9
10
0
CARBONATES
SANDSTONES
~ - 0
W
U
~"
FIG.
;,,:
o.
OIABASE
MARBLE
I0
20
30
MINUS 80 + A AFTER TEST (%)
40
3--Production of fines versus L. A. abrasion loss, all states. From R e f 41.
Impact Resistance and Toughness

Other than the L. A. Abrasion Test there is no impact or toughness
test for concrete aggregate in regular use in the United States. There are
however tests of this type used in Europe, Britain, and Scandinavian countries for aggregate evaluation. With respect to concrete aggregate, toughness is usually considered the resistance of the material to failure by impact. Woolf [28] discusses the now withdrawn ASTM Method D-3 for
toughness of rock which involved an empirical test to determine the
distance in centimeters through which a 2-kg h a m m e r falls to cause
failure of a small cylinder drilled from the aggregate material. Toughness
values for a wide variety of rock are tabulated in Ref 2. Woolf [28] discusses the Jackson Impact Test in which an aggregate particle is rested on
a rigid lower steel ball and a guided upper ball is allowed to fall on t h e
particle. Two concepts--whether or not the particle fractures and the
673
height of drop needed to cause fracture--have been used in an attempt to

measure soft particles in aggregate. Woolf [42] indicates that some hard
but brittle particles also broke in the device. Different size balls have been
tried in this test. Woolf also reported other impact methods tried in an
attempt to identify soft particles, including: (a) a rotary machine which
slung individual particles against a steel plate; (b) calibrated blows with a
hand hammer where the type of fracture of the particles was examined
(hard materials yielded sharp edged fragments and soft pieces tended to
crush); (c) blows with a rubber mallet; and (d) a bag abrasion method
where particles were placed in a cloth bag and swung against an anvil or
hit with a rubber mallet. None of the impact type tests did a very consistent job of identifying inferior material in samples of hard quartz gravel to
which soft pieces were added.
Kohler [43] in a comprehensive study reported to the Highway Research
Board in 1972 compared the results of aggregates tested by the German
Impact Test, the L. A. Abrasion Test, the Deval Abrasion Test, the British
Impact Test (British Standard 812) and a modified Marshall Impact Test.
He found that even though the methods use different loads and procedures
the tests characterize "the same or at least similar properties of the aggregates." Good correlation was found between all the tests run in the same
laboratory. High variation of results were found for the L. A. and German
Impact Tests when tests were run in different laboratories, suggesting that
further strict standardization of procedures may be necessary. The eleven
procedures studied in detail were: (1) German Impact (sample in 8 to 12.5
mm range subjected to 10 blows with a 50 kg (110-1b) hammer from a
height of 500 mm (19.7 in.)), (2) L. A. Abrasion (Grading B), (3) L. A.
(Grading C), (4) L. A. (Grading D), (5) L. A. Machine used with German
Grading (8 to 12.5 mm, 5000-g sample, 12 steel spheres), (6) L. A. (same
as 5 except smaller sample), (7) Deval (Grading D of old ASTM Standard
D 289), (8) Deval (Grading E), (9) British Impact (British Standard 812,
12.5 to 9.5 mm (1/2 to 3/8 in.) material subjected to 15 blows by a 14-kg
hammer falling 380 mm), (10) Modified Marshall Impact (8.0 to 12.5 mm
(5/16 to 1/2 in.) aggregate, weighing 260 times the specific gravity in
grams, with 200 blows), and (11) Modified Marshall Impact (same as 10
except sample of 400 g).
Kohler compared L. A. abrasion loss usually specified in the United
States with equivalent L. A. limits comparable to the French and German
requirements. This showed that the limits allowed in the United States
are broad, up to 40 to 50 percent loss, compared to more restrictive
equivalent values of 20 to 30 percent maximum loss for French and
German specifications. These comparisons were made for aggregates used
in surface courses, bituminous mixtures, and base courses.
Hobeda [44,45] found a relationship between the Swedish impact test
and L. A. abrasion, when comparing impact, or brittleness values as they
674
are sometimes called, run on both dry and soaked aggregates. British
crushing values were also determined on particles in the 8 to 11.3 mm size
fraction. In most cases soaking the samples had little effect on the impact
of crushing values. However, some aggregates--those with a large proportion of mica or secondary hygroscopic minerals--did show crushing and
impact values which indicated deterioration. Good correlations are shown
between impact value and crushing value. In the World Road Conference
Report, Mexico, 1975, the Swedish indicate that they have found no correlation between i m p a c t resistance test results of aggregate and wear of
aggregates subjected to studded tires in a road simulator. Aggregate abrasion value by the method contained in British Standard 812 showed much
better correlation with road simulator wear (see Fig. 4).
In other related research Kauranne [47] found a good correlation
between the Swedish impact value and L. A. abrasion loss. Also there was
a tendency for higher abrasion losses in crushed gravels with increasing
percentage of quartz and feldspar and decreasing percentages of mica.
These relationships do not hold up as well for stones. It was found somewhat remarkable that mica increased toughness in laboratory tests in view
of the known poor performance of mica-rich materials in road surfacings.
The aggregate materials included in this study mainly had mica occurring
in small flakes. Mica percentages varied from about 5 to 25 percent. Rad
[48] recently investigated a weight-drop test procedure to measure a coefficient or rock strength based on impacting an aggregate sample with a
2.5-kg weight falling from a height of 56 cm. Pilot experiments showed
that "rounded aggregates are less susceptible to breaking than sharp aggregates, which in turn are less susceptible than flat or long aggregates."
Two crushed stones, a marble-limestone with an L. A. loss of 26 percent
and a granite-gneiss with an L. A. loss of 46 percent were tested. Unconfined compressive strengths were fairly close, 140 versus 165 MPa (20 000
versus 24 000 psi), respectively, and the coefficients of rock strength by the
proposed test were close also. The great disparity in L. A. abrasion loss in
this case is apparently due to the greater brittleness of the granite-gneiss.
Bond [9], in the investigation of tests which might be useful as an indicator of crushing resistance of rock under impact, developed a test which
avoided the traditional drop-weight methods. Pendulum devices were
studied. Usually there is very little rebound using this procedure. The
highest impact resistance was obtained from several granites, a traprock,
and several ores. In the intermediate to high range a wide variety of rock
types and gravels were represented: limestone, traprock, sandstone, granite,
rhyolite, and gabbro diorite. Low values were obtained from a number of
limestones, granite pebbles, and a sandstone.
Another interesting research on impact resistance of rock was conducted
by Singh [49] in which pellets were fired at rock at high velocity. The
crater depth and volume was measured and found to correlate well with
675
5"0
40
q
oQ
30
20
I0
.<
E
E
7O0
j
6~
.2
:g
m
(5
L
t~
<
-tcJ
~oo
;j
400
300
L
200
o'n
I00 --
,?-
W e a r ~n R o a d
3
S~mulator
5"
~ mm
FIG. 4 - - R e l a t i o n s h i p o f aggregate i m p a c t a n d abrasion values to wear caused by s t u d d e d

tires. F r o m R e f 45.
total energy input and impact pressure. "Among the rock properties to
which rock damage is correlated are shore scleroscope hardness and tensile
strength." Rock tested ranged in compressive strength from 28 to 393 MPa
(4 000 to 57 000 psi) with corresponding ranges in tensile strength and
modulus of elasticity.
For concrete aggregate at least minimal impact resistance is required to
prevent undesirable breakdown in aggregate handling and hatching
676
operations. Beyond that, there are few applications of concrete where surface aggregates may be exposed to severe impact loadings. Neville [50] in
discussing the impact strength of concrete discusses pile driving and accidental impacts, such as those which might be imposed on an industrial
floor or from sediment loads in large hydraulic structures. Sandhu [51]
reports that impact resistance of concrete is mainly affected by concrete
compressive and tensile strength although there is some indication that at
the same concrete strength level angular aggregates and those with rougher
surface texture may perform better in impact. Sandhu's data on the number of drop hammer blows to failure of concrete cubes with various aggregates show that a crushed limestone and a crushed brick aggregate gave
better relative impact resistance, apparently due to better bond between
m o l a r and aggregate. Cracks were through the coarse aggregate particles.
For gravel, basalt, and granite aggregates, cracks indicated failure in the
bond zone.
Hardness, Wear, Skid Resistance

"Hardness is the single most important characteristic that controls
aggregate wear" according to Stiffler [52]. Mineral and rock materials
when subjected to rubbing, scratching, and gouging actions, particularly
with the aid of an abrasive, are worn down due to the scratching and
pitting that occurs. Stiffler has observed that, unlike metals, which just
scratch when subjected to the movement of an abrasive on the surface
under a load, minerals are pitted as well. This is apparently due to mineral
grains or particles being pulled from the matrix. There are many potential
ways of measuring the hardness [53].
1. Indentation Hardness--Vicker's hardness number which is the load
divided by the projected area of the indentation; Rockwell hardness [54].
2. Rebound Hardness--Shore scleroscope, Swiss rebound hammer.
3. Scratch Hardness--Moh's hardness scale, scratching with brass
(ASTM Recommended Practice for Estimating Scratch Hardness of Coarse
Aggregate Particles (C 851)) or steel [42].
4. Wear or Abrasion Hardness--Dorry hardness where a core of rock is
subjected to wear with an abrasive on a revolving horizontal wheel [42]
and British aggregate abrasion value (British Standard 812).
5. Rate of cutting or drilling.
Hardness is an often used term in describing desirable aggregate properties, but there has been little agreement as to how measurements should
be made and whether hardness should be a property which is specified
for concrete aggregates. There is a history of ASTM Specification for
Concrete Aggregates (C 33) having a limit on "soft particles" [55]. First
the Douglass particle strength test was used, then no test was specified for
a time, and more recently the scratch hardness test developed by Woolf
677
[42] has been used. Because of the difficulty of applying this procedure in
a meaningful and consistent way to the specification of aggregates, ASTM
Test for Scratch Hardness of Coarse Aggregate Particles (C 235) has been
withdrawn. The procedure is now covered in ASTM Recommended Practice C 851 which is not intended for specification purposes. ASTM Specification C 33 now has no limits on soft particles and no other provisions
referring to hardness.
Michalopoulos et al [4] studied the effect of water on the strength and
hardness of rocks, using the shore scleroscope and the Schmidt hammer as
measures of hardness. Table 3 presents some of the data. In most, but
TABLE 3--Effect of water on rock strength and hardness.
Hardness
Unconfined Compressive
Strength, Avg MPa (ksi)
Dry
Sandstone
Granite
Tonalite
Granite
Limestone
Limestone
Granite
Granite
31 (4.5)
19b (28.4)
61 (8.8)
253 (36.7)
57 (8.2)
77 (ll.1)
178 (25.8)
308 (44.7)
Saturated
24 (3.5)
168 (24.3)
20 (2.9)
217 (31.5)
47 (6.8)
75 (10.9)
157 (22.7)
242 (35.1)
Schmidt
Dry
13
30
15
24
13
20
27
32
Saturated
...
27
...
26
"2'2"
25
31
Shore
Dry
Saturated
18
85
47
90
23
39
88
101
10
84
26
84
16
43
80
93
not all cases, strength and hardness decreased going from dry to saturated
rock. He also established correlations for strength and modulus of elasticity
versus the product of rock hardness times unit weight. Harvey et al [54]
was able to relate hardness of limestone by the Rockwell procedure to
sulfate soundness.
As a general rule concrete strength has been found to be the most significant factor in rate of wear of concrete [56]; but fine and coarse aggregate may be an important factor in some instances for concrete surfaces
subjected to heavy traffic or abrasive forces, particularly lower strength
concrete. Initially the hardness of the fine aggregate is important. The
coarse aggregate will become involved only if there is enough loss of elevation to expose a significant amount of the coarse particles. Polishing is a
special form of wear where abrasive size is quite small, such as typical
road grit at 10 to 40 #m; and the action is such that any texture such as
existing pits, gouges, or scratches are smoothed and polished gradually.
This can happen particularly to exposed cement paste and the top surfaces
of fine or coarse aggregate particles. Wear is "waste or diminish by continual attrition, scraping, percussion or the like" [57]. It is any removal of
678
particles that produce shape changes on a microscopic or macroscopic

scale [52].
The mechanism of concrete wear can vary [30]. For example, in concrete
abrasion tests using the steel balls or dressing wheel, harder aggregates
retard abrasion loss and stand out from the surface. A softer limestone
aggregate, on the other hand, wears along with the paste in these tests.
The reverse is true for a shot blast test where the limestone cushioned the
abrasive shot and decreased the wear compared to more brittle aggregate.
Wear of highway pavements is due in large measure to the fine mineral
abrasive present on the pavements. Studded tires have also been an important factor in road wear. With the phasing out of the studs in many
areas this factor should diminish.
In a review of pavement wear testing Stiffler reports that sliding movements between rubber and pavement of up to 6.4 m m (1A in.) occur in
normal rolling tires. Therefore, many pavement wear researchers utilize a
wearing mode of rubber on pavement specimens, or rubber on mineral
aggregate samples, with an abrasive supplied to simulate road grit. Polishing occurs when a fine abrasive is used and the surface is made up of
materials of similar hardness. Regeneration of the surface can occur when
two components of differing hardness are present and particles of the
weaker softer material wear faster causing the harder material to protrude
and eventually be undercut and torn out leaving an unpolished surface.
Friction forces of tire on pavement is made up of two components: (1)
adhesion at contact and (2) plowing displacement of softer material by
harder. Figure 5 shows the scratching and pitting action caused by wear
of minerals. In wear studies of minerals with three different fine abrasives
of varying hardness it was found that wear rate corresponded fairly well
with Vicker's hardness for calcium carbonate (both impure (i) and crystalline (c)), slag, mullite, and silicon dioxide (both crystalline (c) and fused
( f ) ) (see Fig. 6).
During the first decades of this century, road wear was substantially
due to horses' hooves, iron wheels, and solid rubber tires. Road wear then
decreased to only a minimal problem with the use of pneumatic tires. In
the 1960s, first in the Scandinavian countries and later elsewhere, the
widespread use of studs in tires for winter traction caused a sharp rise in
wear of road surfaces. Hobeda's [45] abrasion by wet and dry methods,
using a modified British abrasion test on a horizontal turntable with an
abrasive, showed increased abrasion loss with moisture for most materials.
Using higher pressure on the specimen caused intensive microcracking of
the mineral specimens which allows moisture to penetrate. In the studded
tire road simulator it was determined that wear of pavements was not
related to aggregate impact resistance but to the wet abrasion test value
with a load of 12.5 kg on the samples (Fig. 4). The least studded tire wear
FIG. S--Microphotographs of several worn minerals. From R e f S2.
679
680
FIG. 6--Wear inversely proportional to hardnessfor Si02 abrasive. From Ref 52.
was found for fine grained quartzites and diabase, and the most for micaceous gneiss and limestone.
In Canada and the United States [58-61] studded tire wear was studied
intensively. Smith found that in an area where almost a third of the
passenger vehicles were equipped with studs, over the winter of 1969-70,
wear on both portland cement concrete and bituminous concrete approached 5 to 8 mm in one season. Pavements with aggregates of similar
hardness to the matrix showed uniform wear. With harder aggregates the
matrix was preferentially worn down around them until the particles were
dislodged by the studs because of lack of embedment. It was determined
that studded tires did not change the skid resistance much. In concrete
containing a sand with a substantial amount of soft minerals, pavements
with traprock and limestone coarse aggregates wore at about the same
overall rate even though the mechanism was somewhat different. When a
100-percent silica sand was used with the harder traprock, wear of both
the matrix and the overall pavement was reduced. Keyser found that the
age of the mortar and whether or not limestone coarse aggregate was used
were significant factors in studded tire wear studies. Other aggregate types,
size of coarse aggregate, and concrete compressive strength was not
significant for the materials used. Preus [60] in reporting test track studies
showed that studs produced more than 100 times as much wear as regular
tires even when sand and salt are applied to the surface. Rosenthal et al
[61] show photos of stud wear marks. They confirmed that studs tend to
skid over hard aggregate and leave grooves in softer material. They discussed some German wear data which showed 2.5 to 7.5 mm (0.1 to 0.3
in.) of pavement wear after only 27 000 passes of studded tire cars. Deceleration areas showed the most wear.
Skid resistance of pavement surfaces in wet weather depends on microtexture and, also, on macrotexture if significant speeds are involved. Macrotexture is controlled by finishing operations and it is important in removing excess water from between the tire and pavement. Microtexture is
controlled by the polishing tendency of the exposed cement paste or aggregate surfaces. Meyer [62] in using a number of concrete finishing textures,
silica gravel and limestone coarse aggregates, and silica or lightweight fine
aggregate, found good skid resistance in all cases. The lightweight fines
did wear faster. In other studies where calcareous fine aggregates were
used in concrete, low skid resistances have been found [63]. For concrete
pavements exposed to normal traffic wear which does not expose much
coarse aggregate during the life of the pavement, it is the fine aggregate
which tends to control the polishing rate and microtexture [46]. Colley et al
[63] in tests at the Portland Cement Association showed, after polishing,
that the energy needed to turn a rubber tire against a fixed concrete specimen decreased markedly as the siliceous particle count in the fine aggregate was decreased below 25 percent. MuUen and Dahir [64] studied the
wearing and polishing characteristics of a number of aggregate sources.
They found no general correlation of properties for all aggregates. For a
granite aggregate there was an inverse relationship between wear resistance
and skid resistance; and, for sandstone, synthetic aggregate, and one
mountain gravel, skid resistance increased as absorption and surface
capacity of the particles increased. For carbonate aggregates, a direct
relationship between sand-size, acid-insoluble, residue, and skid resistance
was implied. Some highway agencies have advocated the use of ASTM Test
for Insoluble Residue in Carbonate Aggregates (D 3042), as a tool in
selecting fine or coarse carbonate aggregates for use in surface courses.
A higher insoluble residue indicates a higher percentage of harder and
perhaps more polish resistant minerals. ASTM Specification C 33 contains
the warning that "Certain manufactured sands produce slippery pavement
surfaces and should be investigated for acceptance before use." ASTM
Recommended Practice for Accelerated Polishing of Aggregates Using the
British Wheel (D 3319), is similar to the polishing value determination in
British Standard 812 for coarse aggregate. The procedure involves polishing oriented coarse aggregate particles held by an epoxy backing using a
rotating rubber tire running against the specimens which are around the
perimeter of a second wheel. Abrasive and water are fed onto the tire
specimen interface at a constant rate. Degree of polish is measured with
the British portable tester (ASTM Test for Measuring Pavement Surface
Frictional Properties Using the British Portable Tester (E 303)). A number
of other laboratory procedures for evaluating pavement materials and
mixtures for polishing are currently in use or under investigation. Franklin
682
and Calder [46] report results of skid resistance research using several
types of fine aggregate which are in decreasing order of performance: (1)
calcined bauxite fines with both good polishing and abrasion resistance,
(2) a gritty sandstone material with good polish resistance but poor abrasion resistance, (3) a flint sand with poor polish resistance but good
abrasion resistance, and (4) a carbonate fine aggregate with both poor
polishing and abrasion performance.
Coarse aggregate characteristics may become more important relative to
concrete skid resistance if practices now in use in Europe are adopted more
widely. One includes sprinkling a skid resistance wearing surface aggregate
on the top of fresh concrete just prior to finishing. The other is brushing
away the mortar to expose coarse aggregate on the wearing surface to give
more macro texture. Weller and Maynard [65] at the British Road Laboratory have developed an accelerated wear machine for pavement samples.
It is more severe than the polish stone value device, and somewhat more
severe than typical road action. They use a dry wearing cycle of 50 h with a
small flint gravel as an abrasive followed by 5 h of wet polishing with a
fine emery abrasive. Polish is evaluated with the British portable skid tester
and texture wear by the sand patch test. The most important characteristic
of the sands tested is hardness. Harder sands stood out from the surface
and showed higher skid resistance after dry wearing than after wet polishing. Conversely for soft dolomite and limestone fine aggregates the particles were worn flush with the cement paste during the dry wearing cycle;
and, since portland cement paste polishes more during the dry cycle and
gains skid resistance during wet polishing, the skid numbers of the pavement surfaces containing the soft fine aggregate were lowest after dry
wearing and improved during wet polishing.
Figure 7 shows how the British polished stone value (British Standard
812) of coarse aggregate affected the skid resistance of pavement surfaces
after the 55-h wear and polish treatment. For concrete where little coarse
aggregate was exposed the effect was slight, both tined and brushed mortar
surfaces were included. The effect was much greater when the coarse
aggregate was exposed. Studies of 31 concrete pavements in the United
States [66] showed an average 20 percent reduction in skid resistance over
5 years, although, current values are reported as still being adequate for
most of these pavements. Highest losses in skid resistance consistently
occurred with limestone coarse aggregate.
Wet Durability and Attrition Tests

In recent years there has been increased interest in abrasion and attrition procedures run in a wet environment. Woolf [67] in 1966 briefly described the then fairly new California durability index test for fine and
coarse aggregate which measured the tendency of an aggregate to break
683
60-
"o
50
/-
u
0
40
t"
30
c
cl
20
Exposed stone
Tined concrete
i.
Brushed concrete
-o
~
v
u~
10-
020-
I
30
I
~.0
[
50
[
60
I
70
I
80
Polished stone value

FIG. 7--The effect o f polished stone value on the skid resistance value after wear and
polish. From R e f 65.
down into plastic clay-like fines. It has been known that some aggregates
degrade in a different manner in the presence of water than that exhibited
in dry test methods [68]. For concrete aggregates the principal value of
such tests are identification of aggregates which may create undesirable
fines during handling, batching, and mixing which could affect concrete
properties such as strength and water requirement. These tests perhaps
have more significance for base courses and aggregate surfaced roads
where the possibility of interparticle movement in the presence of water
may occur over the period of service of the material.
Highway Research Board Circular 144 [69] indicates use by state transportation agencies of various test methods relating to aggregate durability
(other than freezing and thawing). The L. A. abrasion (a dry test) was used
by the great majority; however, a number also indicated use of some type
of wet abrasion or attrition procedure including: California Durability
684
Index (AASHTO T 210, an ASTM Standard is in preparation by Committee D-4 on Road and Paving Materials), Washington Degradation Procedure, Idaho Degradation, Nevada Air Test, the Oregon Air Test, and
Modified L. A. abrasion tests using water. A few relied on petrographic
analysis to identify mineral constitutents which are associated with rapid
weathering. Almost all agencies reported some degradation problems; in
most cases the problems were termed moderate.
In Sweden a wet degradation test was performed in a rotary device [44].
Some materials showed marked differences compared to dry abrasion. In
many cases minus 75 #m (No. 200) amounts were more than double compared to dry attrition, and sediment heights in a sedimentation column
were much greater after the wet attrition. A modified wet L. A. abrasion
procedure was used in a National Cooperative Highway Research Project
in the United States [71]. A crushed graywacke aggregate with moderate
dry L. A. loss, and good skid resistance, showed a somewhat high sediment
height in the wet L. A. test and in the California and Washington durability tests. Others [41, 72, 73] have tried modified L. A. abrasion tests with
water. Larson et al proposed a modified L. A. test consisting of 250 dry
revolutions followed by 250 wet revolutions with performance being evaluated after both the wet and dry periods. Percent loss after 250 dry revolutions correlated well with the standard S00 revolution value, and the sediment height after the second 250 revolutions with water present correlated
favorably with sediment height for the same material tested by the Washington Degradation Test. The Larson proposal was to develop specification
limits for loss after 250 dry revolutions and for sediment height after an
additional 250 revolutions with water present.
Breese [73] gives a good history of the development of wet abrasion tests
in the Western United States. He experimented with different procedures
designed to detect comminution degradation and alteration degradation.
He concluded that Nevada had a number of degradation susceptible aggregates and that the California durability index test is not as harsh as some
of the other procedures.
Ekse et al [72] compared properties of fines produced from a standard
L. A. abrasion test and an extended 4-h test without the steel balls. The
aggregates were crushed basalt, a pit-run river gravel, and a crushed river
gravel all from sources in the State of Washington. Plasticity index of the
fines produced from the extended test were respectively 6.8, 2.0, and 1.0.
Fines from the standard L. A. test were found to be nonplastic. A wet test
was also run which tended to accelerate the production of plastic fines,
but it was found messy to clean out the drum of the L. A. machine. Helm
in Idaho [74] found little production of plastic fines in a modified dry L. A.
procedure. He describes the Idaho degradation test which involves: wet
tumbling in a glass jar, with initial and final gradations, and the sand
685
equivalent test to evaluate performance. Amount of increase in aggregate

surface area was judged to be a key characteristic.
Minor [75] describes the development of the Washington degradation
test which was prompted by the poor performance of several basalt crushed
stones and volcanic gravels which gave losses in the standard L. A. abrasion
test ranging from 12 to 50. These aggregates contained altered minerals
which contributed to rapid weathering. In studies in Pennsylvania and
Indiana [76, 77] to help identify the properties of shales which might help
determine suitability for its use in embankment or as a granular material,
the Washington Degradation Test was one of the tests which interested
these researchers. In addition to the Washington Test, good indications of
performance were the ultrasonic disaggregation test and the ethylene
glycol soak test used by the Corps of Engineers. At Purdue the Swiss Hammer was found to be a potentially useful toel in identifying hard shales.
High rebound readings were obtained for two shales which showed less
than 40 percent loss in the L. A. abrasion test and which gave good results in the Washington Test.
The California durability index for fine and coarse aggregate [78] was
developed to measure breakdown during construction and normal use
under traffic conditions. It is intended to complement the sand equivalent,
cleanness value, L. A. abrasion and sulfate soundness procedures. There is
little correlation between durability index and L. A. abrasion which, as an
impact test, tends to pulverize the rock and produces fines which in many
cases are unlike those resulting from normal construction and service degradation. The durability index test starts with a washed sample and then
measures the quality of the fines generated from interparticle abrasion during a wet agitation period. It is stated that it is not presently anticipated
that the test will be specified for concrete or asphalt surfacing aggregate.
Figure 8 shows California durability index values versus L. A. abrasion.
The precision of the fine and coarse California durability tests [29] is fairly
good with the standard deviation increasing with decreasing index values.
The single-operator standard deviation is about 1.5 to 3.5 and the multilaboratory standard deviation is 2 to 5.
Goonewardane [79] in recent work with the Washington degradation test
looked into sources of variation. He concluded that both surface attrition
and leaching of clays from the aggregate pores were involved in the test.
The value obtained is very dependent on the surface area of the charge and
finer samples will indicate more degradation. Aggregate quality improved
with repeated agitation cycles as found by Hveem [78]. Weathered basalt
was found to produce lower durability results than fresh basalt.
Degradation or attrition of aggregates during mixing and agitating of
concrete has been the subject of a number of investigations [80-83]. In
studies by Slater and more recently by Gaynor and by McKisson it was
686
TESTS
AND
PROPERTIES
OF
Inn
CONCRETE
GEND
AGGREGATES
i
II
aGGREGATE ( ~ x #4)i
IREGATE (~14 x 2 0 0 ) ~ - - - f
I
I
1
I
x
bJ
o
Z
).J
m
(]c
tl
"-IE~........
II--I Z'~ "-"
,, ,r ,!T
i ~
I
olo
BY'I
I
DURABILITY
Io
20
30
LART-% Loss at 5 0 0
rll t
TEST
40
60
Revolulions
Less Durable -.
FIG. 8--L.
',,
50
i-
A. abrasion loss versus California durability index. From R e f 78.
identified that during prolonged mixing the fine aggregate with its much
greater surface area tended to break down by attrition much more readily
than the coarse aggregate. In 2 h of mixing in a laboratory mixer, in one
case, and 2 h and 40 min, in another case, little change was noted in the
coarse aggregate grading. The amount of minus 150 #m (No. 100) material
generated from the fine aggregate depended on aggregate hardness. A fine
aggregate known to break down, caused an increase in minus 150/~m (No.
100) material of almost 12 percent in 2 h in one study. In the other study a
soft limestone sand caused an increase in minus 150/~m (No. 100) sizes of
about 6 percent in 160 min. Hard quartz sand only caused an increase
of 1 percent in the same 160-min mixing period.
Several tests have been developed to measure the propensity of fine aggregates to degrade due to attrition during concrete mixing [82, 84,85]. Davis
et al, and Tuthill and Adams, in the discussion of that paper, report data
using a vane type attrition chamber, mounted in a drill press and turned
at 800 rpm, to evaluate fine aggregate attrition. They used a 1000-g sample
of aggregate. Meininger [85] reported on research conducted in the NSGANRMCA Joint Research Laboratory using the same attrition chamber but
reducing the sample size to 500 g for increased convenience and also to
spread the results out more. Table 4 shows results of percent minus 75 #m
(No. 200) sizes generated by the attrition test and the decrease in fineness
modulus for a number of aggregates. Values for 1000-g samples and 500-g
samples are both reported; the 500-g results have been found to be roughly
double the loss for 1000-g samples in terms of generation of minus 75 /~m
(No. 200) when a coarser grading is tested. Since grading affects the results it has been found that the use of several standard gradings is the best
practice.
Higgs reports another approach to the same problem [84] using a paint
shaker to attrition a 1000-g regraded sample of fine aggregate. He investigated a number of basalt fine aggregates and found that those causing
rapid slump loss in concrete yielded more than 8 or 9 percent minus 75 #m
(No. 200) fines in the shaker test. These sands also had high magnesium
sulfate soundness losses. One sand which produced 6 percent fines in the
test caused some problems on a construction project with increased aggregate fines.
Impact on Specifications
Tests which measure directly abrasion, strength, or toughness characteristics of an aggregate are not generally referenced in specifications for
concrete aggregates in the United States. The one exception is the L. A.
Abrasion Test for coarse aggregate which is used in ASTM Specification
C 33 and ASTM Specification for Aggregates for Radiation-Shielding Concrete (C 637) as well as in a large number of agency specifications. ASTM
Test for Clay Lumps and Friable Particles in Aggregates (C 142), which
detects particles which can be broken between the fingers after soaking, is
used in a number of ASTM aggregate specifications to detect extremely
weak particles. The predecessor of the L. A. Abrasion Test, the Deval
Abrasion Test, has been withdrawn as an ASTM standard although it is
still an AASHTO Standard. At least one state still uses the Deval Test.
The Deval and L. A. Abrasion are also used widely in other countries as
indicators of aggregate quality. The scratch hardness test method, ASTM
Test C 235, has been withdrawn and is no longer referenced in ASTM
concrete aggregate specifications. It has been reissued as a recommended
practice, however.
There appears to be a difference in the approach taken in the United
States and Canada with respect to using aggregate strength and impact
tests in specifications as compared to European and British practices
which do utilize tests such as the crushing value, 10 percent fines value,
0.9
1.5
1.4
5154
4957A
4957C
5160A
4954
5076
4621
5160B
5044A
4957B
4989A
4989B
5044D
5075
5044C
5044B
5019B
4933
5154
4965
5160A
5065B
5065A
"5168B
5016B
1.1
2.2
2.7
3.7
3.4
2.0
4.4
3.3
3.9
0.9
1.4
2.3
1.4
1.2
Absorption,
%
Fine Aggregate
Source
Decrease in
Fineness Modulus
Minus 75 tzm
(No. 200)
Generated
Coarse-No. 1
Coarse-No. l
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Coarse-No. 1
Medium-No. 2
Medium-No. 2
Medium-No. 2
Medium-No. 2
Medium-No. 2
Medium-No. 2
Medium-No. 2
0.20
0.25
0.22
0.40
0.41
0.35
1.07
0.76
1.01
0.55
0.78
0.81
0.89
0.56
0.67
0.77
0.95
0.94
0.11
0.16
0.25
0.27
0.20
0.19
0.18
5.3
5.4
5.8
5.9
6.6
8.9
10.3
11.3
12.6
12.8
14.1
14.6
15.2
16.0
16.2
22.2
23.6
24.6
3.3
3.4
3.5
5.1
5.7
5.9
6.0
500-g Sample a. NSGA-NRMCA Joint Research Laborato~
Test Fine Aggregate

Grading b
Attrition Test Results
T A B L E 4--Tests of fine aggregate in vane-type attrition chamber.
silica sand
silica sand
silica sand
silica sand
silica sand
river sand
manufactured limestone
some mica
manufactured sand
occasional strength problems
some mica
some mica
manufactured s a n d
commercial sand
angular river sand
angular river sand
strength problems
slump loss problems
silica sand
silica sand
silica sand
commercial sand
commercial sand
commercial sand
small amount shale
Remarks
-I
m
6o
G~
I"11
-I
ill
0
0
z
0
0
"13
60
m
"11
-I
"U
--t
ITI
60
"--4
60
O~
t~o
Oo
0.8
2.5
2.2
1.8
1.7
1.6
1.3
1.3
2.3
2.2
2.2
0.8
4.7
J
B
C
D
E
F
G
H
1
L
K
Good
Poor
0.22
0.22
0.20
0.36
0.44
0.64
0.10
0.18
0.22
0.31
2
2
2
2
2
2
2
2
2
2
2
0.1 l
0.20
0.14
0.18
0.19
0.19
0.21
0.16
0.16
0.24
0.27
3
6
6
6
6
6
6
6
6
8
9
6.4
6.90
7.1
7.3
8.0
9.0
2.7
4.0
4.9
9.2
Medium-No. 2 approximate
0.20
0.52
6
18
lO00-g S a m p l e a, Davis, Mielenz, and Polivka Data [82]
Medium-No.
Medium-No.
Medium-No.
Medium-No.
Medium-No.
Medium-No.
Medium-No.
Medium-No.
Medium-No.
Medium-No.
Medium-No.
lO00-g Sample a, Tuthill and A d a m s Data [82]
Medium-No. 2
Medium-No. 2
Medium-No. 2
Medium-No. 2
Medium-No. 2
Medium-No. 2
Fine-No. 2A3
Fine-No. 2A3
Fine-No. 2A3
Fine-No. 3
acceptable sand
slump loss problems
satisfactory sand
satisfactory sand
satisfactory sand
satisfactory sand
satisfactory sand
satisfactory sand
satisfactory sand
satisfactory sand
satisfactory sand
sticky concrete, lower strength
sticky concrete, lower strength
small amount shale

commercial sand
small amount shale
commercial sand
some mica
commercial sand
commercial sand
commercial sand
100
100
100
100
100
These are the sand gradings prior to the attrition test.
Coarse-No. 1
Medium-No. 2
Fine-No. 2A3
Fine-No. 3
80
88
100
100
87
50
74
80
85
70
25
49
50
60
52
10
20
20
30
15
2
5
0
10
2
0
0
0
0
1
OThe 500-g sample has been shown to yield attrition results at least 50 percent higher than results obtained with the 1000-g s a m p l e - - b o t h in
terms of decrease in fineness modulus and minus No. 200 material generated in the test.
b
Gradings, % Passing
No. 4
No. 8
No. 16
No. 30
No. 50
No. 100
No. 200
i14
...
iii
0.9
1.4
...
...
...
5016C
5117
5016A
5065C
5160B
4621
5065A
5065B
5065C
4621
O~
O0
r
"o
212
0
"o
m
~o
---I
t'3
w-
-o
-r
..<
0
Z
ITI
z
7
fi3
m
2o
690
and impact value as potential specification tests. In North America the

practice is to, first, rely more on past service record and judgment for
those aggregates which have performed satisfactorily; and, second, to test
the properties of concrete made with the aggregate in question if special
abrasion, impact, or strength properties are required. For example, ASTM
Test for Abrasion Resistance of Concrete by Sandblasting (C 418) and
ASTM Test for Abrasion Resistance of Horizontal Concrete Surfaces
(C 779) cover the abrasion testing of concretes by four procedures depending on anticipated service condition--sandblasting, revolving disks, dressing
wheel, and ball bearings. Quite often these methods might be used to
evaluate the performance of alternative aggregate sources as well as various
mix proportions. With respect to aggregate strength, ASTM Specification
for Lightweight Aggregate for Structural Concrete (C 330) and ASTM
Specification for Lightweight Aggregates for Concrete Masonry Units
(C 331), require the aggregate to be of sufficient strength to yield certain
minimum compressive and tensile splitting values when used in concrete.
Concluding Remarks
There is a good deal of history and data available concerning the measurement of physical properties of aggregates and attempts to relate these
properties with concrete properties. However, the evaluation and testing
of concrete containing alternative materials remains the best approach to
assuring performance when special abrasion, strength, or toughness properties are needed.
References
[1] Udd, J. E., "Physical Properties of Rocks and Probable Applications," A S T M STP 373.
American Society for Testing and Materials, 1965, p. 114.
[2] Woolf, D. O., "Results of Physical Tests of Road-Building Aggregate," Bureau of Public
Roads, 1953.
[3] Johnston, C. D., "Strength and Deformation of Concrete in Uniaxial Tension and Compression," Magazine of Concrete Research, March 1970, p. S.
[4] Micbalopoulos, A. P. and Triandafilidis, G. E., "Influence of Water on Hardness
Strength, and Compressibility of Rock," Bulletin, Association of Engineering Geologists,
Vol. 13, No. 1, Winter 1976, p. 1.
[5] Kaplan, M. F., "Flexural and Compressive Strength of Concrete as Affected by the
Properties of Coarse Aggregate," Journal, American Concrete Institute, May 1959,
p. 1193.
[6] Brandon, J. R., "Rock Mechanics Properties of Typical Foundation Rocks," Report
REC-ERC-74-10, Bureau of Reclamation, 1974.
[7] Collins, J. J. and Hsu, T. T. C., "Properties of Florida-Oolite Concrete," Journal of
Testing and Evaluation. American Society for Testing and Materials, March 1977, p. 141.
[8] Iyer, Rahn, and Remarkrishnam, "Durability Tests on Some Aggregate for Concrete,"
Journal, Construction Division, American Society of Civil Engineers, Sept. 1975,
p. 593.
[9] Bond, F. C., "Crushing Tests by Pressure and Impact," Technical Publication No. 1895,
American Institute of Mining and Metallurgical Engineers, 1946.
[10] Hansen, M. R., "Determination of the Mechanical Properties of Three Concrete Aggregates Using Small Scale Test Specimens," Manuscript 4035, American Concrete Institute,
1977.
[1l] Ross, A. W. and Kienow, K. K., "Arizona Volcanic Cinder Concrete--A Comparative
Study," Bulletin No. 9, University of Arizona, Engineering Experiment Station, 1959.
[12] Lorman, W. R., "Coral and Coral Concrete," Technical Report No. 068, U.S. Naval
Civil Engineering Laboratory, 1960.
[13] Tanigawa, Y., "Model Analysis of Fracture and Failure of Concrete as a Composite
Material," Cement and Concrete Research, Pergamon Press, Sept. 1976, p. 679.
[14] Hosking, J. R. and Tubey, L. W., "Research on Low-Grade and Unsound Aggregates,"
Report LR 293. Road Research Laboratory, 1969.
[15] Clark, G. B., "Deformation Moduli of Rock," Testing Techniques for Rock Mechanics,
A S T M STP 402. 1966, p. 133.
[16] Hartley, A., "A Review of the Geological Factors Influencing the Mechanical Properties of Road Surface Aggregates," Quarterly Journal of Engineering Geology, Vol. 7,
No. 1, Great Britain, 1974, p. 69.
[17] Houghton, D. L., "Determining the Tensile Strain Capacity of Mass Concrete," Journal,
American Concrete Institute, Dec. 1976, p. 691.
[18] Lott, J. E. and Kesler, C. E., "Crack Propagation in Plain Concrete," HRB Special
Report 90. Highway Research Board, 1966, p. 204.
[19] Ko, K. C. and Haas, C. J., "The Effective Modulus of Rock as a Composite Material,"
International Journal of Rock Mechanics and Mineral Sciences, Pergamon Press, 1972,
p. 531.
[20] Hobbs, D. W., "The Dependence of the Bulk Modulus, Young's Modulus, Shrinkage
and Thermal Expansion of Concrete Upon Aggregate Volume Concentration," Technical
Report TRA 43Z Cement and Concrete Association, Dec. 1969.
[21] Hansen, T. C., "Influence of Aggregate and Voids on Modulus of Elasticity of Concrete,
Cement Mortar, and Cement Paste," Journal, American Concrete Institute, Feb. 1965,
p. 193.
[22] Hirsch, T. J., "Modulus of Elasticity of Concrete Affected by Elastic Moduli of Cement
Paste Matrix and Aggregate," Journal, American Concrete Institute, March 1962, p. 427.
[23] Shah, S. P. and Winter, G., "Inelastic Behavior and Fracture of Concrete," SP-20,
American Concrete Institute, 1968, p. 5.
[24] Johnston, C. D., "Deformation of Concrete and Its Constituent Materials in Uniaxial
Tension," HRB Record No. 324. Highway Research Board, 1970, p. 66.
[25] Hobbs, D. W., "Influence of Aggregate Restraint on Shrinkage of Concrete," Journal,
American Concrete Institute, Sept. 1974, p. 445.
Hardened Concrete Mortar," HRB Record No. 124, Highway Research Board, 1966, p. 18.
[27] Philleo, R. E., "Can Research in Volume Change Assist the Designer," TRB Record
613. Transportation Research Board, 1976, p. 10.
[28] Woolf, D. O., "Toughness, Hardness, Abrasion, Strength, and Elastic Properties,"
A S T M STP 169 and 169A. American Society for Testing and Materials, 1955 and 1966.
[29] Benson, P. E. and Ames, W. H., "The Precision of Selected Aggregate Test Methods,"
TRB Record No. 539, Transportation Research Board, 1975, p. 85.
[30] Smith, F. L., "Effect of Aggregate Quality on Resistance of Concrete to Abrasion,"
Cement and Concrete. A S T M STP 205. American Society for Testing and Materials,
1958, p. 91.
[31] Woolf, D. O., "The Relation Between Los Angeles Abrasion Test Results and the Service
Records of Coarse Aggregates," Proceedings, Highway Research Board, 1937, p. 350.
[32] Bloem, D. L. and Gaynor, R. D., "Effects of Aggregate Properties on Strength of Concrete," Journal, American Concrete Institute, Oct. 1963.
[33] Jumper, E. A., Herbert, J. D., and Beardsley, C. W., "Rattler Losses Correlated with
Compressive Strength of Concrete," Journal, American Concrete Institute, Jan. 1956,
p. 563.
[34] Walker, S. and Bloem, D. L., "Effect of Soft Sandstone in Coarse Aggregate on Properties of Concrete," Series 88 and 107, National Sand and Gravel Association, distributed
with NSGA T.I.L. No. 72, 1949.
692
[35] Bartel, F. F. and Walker, S., "Concrete-Making Properties of Gravels from Southwestern
United States," Series 62, National Sand and Gravel Association, distributed with NSGA
T.I.L. No. 48, 1946.
[36] Walker, S., "The Flexural Strength of Concrete," 29th Annual Meeting of the National
Sand and Gravel Association, 1944.
[37] Gaynor, R. D. and Meininger, R. C., "Investigation of Aggregate Durability in Concrete," HRB Record No. 196. Highway Research Board, 1967, p, 25.
[38] Nowlen, W. J., "Influence of Aggregate Properties on Effectiveness of Interlock Joints
in Concrete Pavements," Journal, Portland Cement Association Research and Development Laboratories, May 1968, p. 2.
[39] Hobeda, Peet, "Bergmaterial Till Vagbyggnad," Special Report No. 84. Statens Vaginstitut, Stockholm, 1969.
[40] Houston, J. T., "A Sandblast Abrasion Test for Synthetic Aggregate Evaluation," ASTM
Materials Research and Standards. American Society for Testing and Materials, Aug.
1970, p, 17.
[41] West, T. R., Johnson, R. B,, Smith, N. M., and Aughenbaugh, N. B., "Degradation of
Aggregates," Final Report NCHRP Project 4-2, Purdue University School of Engineering,
Nov. 1966.
[42] Woolf, D. O., "Methods for the Determination of Soft Pieces in Aggregate," Crushed
Stone Journal, Sept. 1951, p. 13.
[43] Kohler, Guntram, "Comparative Investigations of International Test Methods for SmallSized Coarse Aggregates," HRB Record No. 412, Highway Research Board, 1972.
[44] Hobeda, Peet, "Influence of Moisture on the Strength and Wearing Properties of Road
Aggregate," Report No. 42, National Swedish Road and Traffic Institute, Stockholm,
1974.
[45] Hobeda, Peet, comments at World Road Conference, Mexico, 1975.

[46] Franklin, R. E. and Calder, A. J. J., "The Skidding Resistance of Concrete--The Effect
of Materials Under Site Conditions," Laboratory Report 640, Transportation and Road
Research, United Kingdom, 1974.
[47] Kauranne, Kalevi, "On the Abrasion and Impact Strength of Gravel and Rocks in Finland," Bulletin No. 243, Commission Geologique de Finlande, 1970.
[48] Rad, P. F., "A Simple Technique for Determining Relative Strength of Aggregates,"
Journal of Testing and Evaluation. Nov. 1975, p. 477.
[49] Singh, M., "Rock Breakage by Pellet Impact," No. FRA-RT-70-29, ITT Research Institute, 1969, (PB 190 965).
[50] Neville, A. M., "Hardened Concrete--Physical and Mechanical Aspects," Monograph
No. 6, American Concrete Institute, and the Iowa State University Press, 1971.
[51] Sandhu, R. S., "Resistance of Concrete to Impact Loading," Indian Concrete Journal,
May 1963, p. 169.
[52] Stiffler, K., "Mineral Wear in Relation to Pavement Slipperiness," Report J-15, Pennsylvania State University Joint Road Friction Program and Automotive Safety Research
Program, 1967. Also, Stiffler, K., "Relation Between Wear and Physical Properties of
Roadstones," HRB Special Report 101, Highway Research Board, 1969.
[53] Davis, H. E., Troxell, G. E., and Wiskocil, C. T., The Testing and Inspection of Engineering Materials, McGraw-Hill, N.Y., 1964.
[54] Harvey, R. D., Fraser, G. S., and Baxter, J. W., "Properties of Carbonate Rocks Affecting Soundness of Aggregate," Illinois Minerals Note 54, Illinois State Geological Survey,
Feb. 1954.
[55] Gray, J. and Renninger, F., "The Concept of Hardness as Applied to Mineral Aggregates," HRB Record No. 196, Highway Research Board.
[56] Scripture, E. W., Jr., Benedict, S. W., and Bryant, D. E., "Floor Aggregates," Journal,
American Concrete Institute, Dec. 1953, p. 305.
[57] Prior, M. E., "Abrasion Resistance," Significance of Tests and Properties of Concrete
and Concrete-Making Materials, A S T M STP 169,4, American Society for Testing and
Materials, 1966.
[58] Smith, P. and Schonfeld, R., "Studies of Studded-Tire Damage and Performance in
Ontario During the Winter of 1969-70," HRB Record No. 352, Highway Research
Board, 1971, p. 1.
693
[59] Keyser, J. H., "Resistance of Various Types of Bituminous Concrete and Cement Concrete
to Wear by Studded Tires," HRB Record No. 352. Highway Research Board, 1971, p. 16.
[60] Preus, C. K., Discussion of Ref 59, HRB Record No. 352, Highway Research Board,
1971, p. 31.
[61] Rosenthal, P. et al, "Evaluation of Studded Tires," HRB-NCHRP Report No. 61, 1969.
[62] Meyer, A. H., "Wearability of P C Concrete Pavement Finishes," Transportation Engineering, Journal, American Society of Civil Engineers, Aug. 1974, p. 719.
[63] Colley, B. E., Christensen, A. P., and Nowlen, W. J., "Factors Affecting Skid Resistance and Safety of Concrete Pavements," HRB Special Report No. 101, Highway Research Board, 1969, p. 80.
[64] Mullen, W. G. and Dahir, S. H. M., "Skid Resistance and Wear Properties of Aggregates for Paving Mixtures," North Carolina State University Highway Research Program
Interim Report, June 1970, (PB 197 625).
[65] Weller, D. E. and Maynard, D. P., "Influence of Materials and Mix Design on the
Skid Resistance Value and Texture Depth of Concrete," Report No. LR 334, British
Road Research Laboratory, 1970.
[66] "Surface Wear of Portland Cement Concrete Pavements," Circular No. 170, Transportation Research Board, Sept. 1975.
[67] Woolf, D. O., "Toughness, Hardness, Abrasion, Strength, and Elastic Properties,"
Significance of Tests and Properties of Concrete and Concrete-Making Materials, A S T M
STP 169,4, American Society for Testing and Materials, 1966.
[68] Melville, P. L., "Weathering Study of Some Aggregates," Proceedings, Highway Research Board, Vol. 28, 1948, p. 238.
[69] Hendrickson, L. G. and Shumway, R. D,, "Analysis of Questionnaire on Aggregate
Degradation," Circular No. 144, Highway Research Board, July 1973.
[70] Vantil, C. J., Cart, B. J., Vallerga, B. A., and Hilliard, J. M., "Requirements for
Wear-Resistant and Skid Resistant Highway Pavement Surfaces," reviewed in TRBNCHRP Research, Results Digest No. 89. Nov. 1976.
[71] Larson, L. J., Mathiowetz, R. P., and Smith, J. H., "Modification of the Standard Los
Angeles Abrasion Test," Record No. 353. Highway Research Board, 1971.
[72] Ekse, M. and Morris, H, C., "A Test for Production of Plastic Fines in the Process of
Degradation of Mineral Aggregates," Symposium on Road and Paving Materials, A S T M
STP No. 277, American Society for Testing and Materials, 1959, p. 122.
[73] Breese, C. R., "Degradation Characteristics of Selected Nevada Mineral Aggregates,"
Engineering Report No. 4, Civil Engineering Department, University of Nevada, 1966.
[74] Helm, R. L., "An Investigation of the Variation of Degradation-Susceptibility Within
An Aggregate Source," Proceedings, 3rd Annual Engineering Geology and Soils Engineering Symposium, Boise, Idaho, April 1965, p. 1.
[75] Minor, C. E., "Degradation of Mineral Aggregate," Symposium on Road and Paving
Materials, A S T M STP 277, American Society for Testing and Materials, 1959, p. 109.
[76] Reidenouer, D. R., Geiger, E. G., Jr., and Howe, R. H., "Shale Suitability--Phase II-Final Report," Pennsylvania Department of Transportation, Bureau of Materials, Testing, and Research, April 1974.
[77] Chapman, D. R., "Shale Classification Tests and Systems--A Comparative Study,"
Report No. 75-11, Joint Highway Research Project, June 1975.
[78] Hveem, F. N. and Smith, T. W., "A Durability Test for Aggregates," HRB Record No.
62, Highway Research Board, 1964, p. 119.
[79] Goonewardane, K., "Behavior of Aggregate in the Washington Degradation Test,"
Journal of Testing and Evaluation, American Society for Testing and Materials, Vol. 5,
No. 1, Jan. 1977, p. 16.
[80] McKisson, R. L,, "Degrading of Aggregate by Attrition in a Concrete Mixer," Report
No. C-1308, Bureau of Reclamation, 1969.
[81] Gaynor, R. D., "Effect of Prolonged Mixing on the Properties of Concrete," Publication
No. 111, National Ready Mixed Concrete Association, 1963.
[82] Davis, R. E., Mielenz, R. C., and Polivka, Milos, "Importance of Petrographic Analysis
and Special Tests Not Usually Required in Judging Quality of Concrete Sand," Journal
of Materials, American Society for Testing and Materials, Vol. 2, No. 3, Sept. 1967,
p. 461.
694
[83] Slater, W. A., "Tests of Concrete Conveyed from A Central Mixing Plant," Proceedings.
American Society for Testing and Materials, Part 2, 1931, p. 510.
[84] Higgs, N, B., "Montmorillonite in Concrete Aggregate Sands," Bulletin, Association of
Engineering Geologists, Vol. 12, No. l, Winter 1975, p. 57.
[85] Meininger,R. C., "Aggregate Degradation During Mixing," National Ready Mixed
Concrete Association Convention, 1977.
Bibliography
ACI Committee 201, "Durability of Concrete in Service," Report 201, American Concrete
Institute, 1962.
ACI Committee 210, "Erosion Resistance of Concrete in Hydraulic Structures," Report 210,
American Concrete Institute.
Byrne, F. E. and Simms, J. V., "Rock Penetration as a Test of Physical Properties," Proceedings. 3rd Annual Engineering Geology and Soils Engineering Symposium, Boise,
Idaho, 1965, p. 173.
Howard, T. R,, Haye, R. D., and Hoover, L. W., "Investigation of Test Methods to Determine Basalt Aggregate Breakdown," Proceedings. Fourth Annual Engineering Geology
and Soils Engineering Symposium, Moscow, Idaho, 1966, p. 301,
Olsen, R. E., "A Review of European Practice for Laboratory Evaluation of Aggregate Polishing Characteristics," H R B Record No. 341. Highway Research Board, 1969, p. 15.
Price, W. H., "Erosion of Concrete by Cavitation and Solids in Flowing Water," Proceedings.
American Concrete Institute, Vol. 43, 1947, p. 1009.
',herwood, W. C. and Mahone, D. C., "Predetermining the Polish Resistance of Limestone
Aggregates," H R B Record No. 341. Highway Research Board, 1970, p, I.
H. K . C o o k 1
Chapter 39--Thermal Properties
Introduction
The thermal properties of concrete aggregates are of greater or lesser concern to the user, depending upon the nature of the concrete structure and the
exposure to which it is subjected. Examples of where knowledge of thermal
properties are of greater concern are in the design of massive structures,
such as dams, where thermal and volume stability are important and in
lightweight concrete structures where insulating value is a primary factor.
Except for thermal coefficient of expansion, the thermal properties of the aggregates used in normal concrete structures, such as office buildings,
warehouses, pavements, and the like, seldom are considered. An understanding of the effects of the thermal properties of aggregates on the properties of
concrete can have an effect on improving concrete quality and in some instances may well be mandatory.
The attention that has been given to the thermal properties of concrete aggregates is not as great as the amount of research and the volume of testing
performed in connection with other properties of aggregate. This may be
because the significance of thermal properties is not as apparent, and in
many instances the effects do not appear to be as important as the effects of
other properties of aggregates.
A discussion of the significance of the thermal characteristics of concrete
aggregates is further complicated by the effect of the relationship between
the thermal properties of the concrete as a whole and the thermal properties
of the component materials. Insofar as possible, the discussion in this paper
is confined to the significance of tests of thermal properties of the aggregates;
however, in some instances it has been necessary to discuss briefly the thermal properties of the concrete as affected by the aggregates. The significance
of tests of thermal properties of concrete is discussed in another paper in this
publication, z
The need for a better understanding of the effects of the thermal
1Consulting engineer, Concrete, Cleveland, Ohio 44114.
2See p. 242.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
696
characteristics of aggregates, particularly the thermal coefficient of expansion on the durability of concrete, has been expressed by Allen [l],a Woolf
[2], Scholer [3], and others.
The significance of other thermal properties has been given much less attention than has thermal expansion, as evidenced by the fact that the great
majority of available references deal with thermal expansion, Undoubtedly
the reason for this is that the thermal expansion of aggregate and hence of
concrete is important from the standpoint of structural design in all concrete
structures, whereas other thermal properties are usually of concern only in
special instances, particularly in massive structures.
The thermal properties of aggregate referenced in the literature are thermal coefficient of expansion, specific heat, thermal conductivity, and thermal diffusivity, and they are discussed here in that order.
Thermal Coefficient of Expansion

The earliest reference to work on this subject, included in this paper, is by
Hallock [4] and is entitled, appropriately enough, "Preliminary Notes on the
Coefficients of Thermal Expansion of Certain Rocks." Some aspects of this
property, particularly with respect to the volume change of concrete as affected by the aggregates, are discussed in a paper on volume change in this
publication .4
Numerical Values for Thermal Coefficients of Expansion

Although values for the thermal coefficients of expansion of aggregates
from specific locations are contained in nearly all of the references listed and
are available from many other sources, Refs 5 to 15 list more values than
most of the others. It should be noted that Ref 15 is concerned completely
with the thermal coefficient of expansion of portland cement rather than aggregates. It is included because a discussion of the thermal properties of concrete aggregates would not be complete without some information on the
thermal properties of the cements with which they are used.
An average value for the linear thermal coefficient of expansion of concrete
may be taken as 9.9 10-6/~ (5.5 X 10-6/~
but the range may be from
about5.8 10-6/~ (3.2 1 0 - 6 / ~
14.0 10-6/~ (7.8 10-~/~
depending upon the type and quantities of the aggregates, the mixture proportions, and other factors. Hydrated cement pastes may range from 10.8
10-~/~ (6 10-6/~ to 16.2 10-6/~ (9 10-6/~ and mortars from
about 7.9 10-6/~ (4.4 10-6/~ to 12.6 10-6/~ (7.0 X 10-6/~
Mitchell [5] has found that at early ages or at certain critical saturations the
4See p. 226.
COOK ON THERMAL PROPERTIES 697
linear thermal coefficient of expansion of cement pastes may be somewhat

higher than those reported above. He reports values as high as 22.3
10-6/~ (12.4 10-6/~ for some samples of near cement specimens. He
indicates that variation of thermal expansion of neat cement is related to the
amount and characteristics of the gel and to an optimum (or pessimum)
moisture condition producing maximum thermal expansion.
The linear thermal coefficient of expansion of common rocks ranges from
about 0.9 10-6/~ (0.5 10-6/~
about 16.0 10-6/~ (8.9
10-6/~
[13]. It should be kept in mind that some minerals exhibit
anisotropic characteristics or the property of expanding more in one direction or parallel to one crystall6graphic axis than another. The most notable
example is calcite, which has a linear thermal coefficient of expansion of 25.8
X 10-~/~ (14.3 10-6/~ parallel to its C axis and --4.7 10-6/~
(--2.6 10-6/~
perpendicular to this direction. Potash feldspars are
another groups of minerals exhibiting anisotropy. It is for this reason that the
cubical expansion of rocks and minerals is not always related directly to the
linear expansion, and this possibility should be kept in mind when investigating the thermal properties of aggregates.
Effect of Thermal Expansion

There seems to be fairly general agreement that the thermal expansion of
the aggregate has an effect on the durability of concrete, particularly under
severe exposure conditions or under rapid temperature changes. The agreement as to whether a high or a low coefficient of expansion is most desirable
is not so general. In one of the earliest papers on the subject, Pearson [16] attributes a concrete failure to the use of an aggregate of low-thermal coefficient subject to severe frost action. Subsequent laboratory investigations
described in the paper and in a subsequent paper [17] included freezing-andthawing tests of concrete containing aggregates of both low and high thermal
coefficients. The indications were that the concretes containing the aggregates of low thermal coefficient failed much more rapidly under the freezing cycle employed. Walker et al [18], on the other hand, report the results of
heating and cooling concrete specimens over the temperature range of 4.4 to
60~ (40 to 140~
at various rates. They found that "changes in
temperature were destructive to the concrete with sudden changes in
temperature being much more severe than slower ones; and concretes having
higher coefficients of expansion were less resistant to temperature changes
than concretes with lower coefficients." It also was determined that the thermal coefficients of expansion of concrete and mortar containing different aggregates varied approximately in proportion to the thermal coefficient and
quantity of aggregate in the mixture. The anomalous results obtained by
Pearson and by Walker et al perhaps may be explained by the fact that the
concrete in Pearson's studies was subjected to temperatures below the freezCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:32 EDT 2014
698
ing point, whereas that used by the other workers was not. The above is conjecture but is somewhat supported by Koenitzer [19], who obtained coefficients of linear expansion of several concretes and the aggregates used in
them over temperature ranges of -12.8 to 26.7~ (9 to 80~
29.4 to 88~
(85 to 190~
and -128 to 88~ (9 to 190~
in both moist and dry conditions. One of Koenitzer's conclusions was that the elastic and thermal expansion properties for any one material vary with the conditions of test, the
greatest variation being caused by freezing.
The U.S. Bureau of Reclamation's experience with aggregate at Grand
Coulee Dam Showed an extremely high resistance to freezing and thawing.
Tl~is aggregate was predominately basalt with a thermal expansion of about
7.2 10-6/~ C (4 10-6/~ s The Bureau of Reclamation also found [6]
in the Kansas-Nebraska area that replacement of part of the aggregate with
limestone, which had a thermal expansion averaging 4.5 10-6/~ (2.5
10-6/~
greatly increased the durability of the concrete in which it was
used. It should be noted that the addition of limestone was for the purpose of
inhibiting cement-aggregate reaction, and the improvement in durability
probably was primarily because of the inhibition of this reaction, but the low
thermal expansion apparently introduced no adverse effects.
Callan [20,21] has statistically analyzed 78 combinations of aggregate in
concrete with respect to durability in freezing and thawing and differences in
thermal expansion between the coarse aggregate and the mortar. He coneludes that, where the difference between coefficients of expansion of coarse
aggregate and mortar is large, the durability of the concrete may be considerably lower than would be predicted from the results of the usual acceptance tests. Where the differences between these coefficients exceeds 5.4
10-6/~ (3.0 10-6/~
caution should be used in the selection of the
aggregate combination for highly durable concrete. Walker et al [18] found
no relationship between resistance of concrete to temperature changes and
differences between thermal coefficients of aggregates and mortar. Again
Callan's work is based on studies over the freezing-and-thawing range,
whereas the results of the other workers were based on experiments above
the freezing temperature. Swenson and Chaly [22] include some discussion
of the effect of thermal coefficient of expansion of the aggregates on concrete durability and caution against the possible deleterious effects if large
differences are observed. Smith [23] presents calculations to indicate the
potential magnitude of physical incompatibility of the matrix and aggregate in concrete based on differences in thermal expansion. Kennedy and
Mather [24], in attempting to correlate laboratory-accelerated freezing and
thawing with natural weathering at Treat Island, Me., state among other
conclusions that, while there appears to be a correlation between the
resistance of concrete to freezing and thawing and differences in thermal exs L. J. Mitchell, private communication.
COOK ON THERMAL PROPERTIES 699
pansion between the coarse aggregate and the mortar, the correlation is probably usually of lesser importance than other characteristics of the concrete.
The above discussion of potential imcompatibility between the matrix and
the aggregate of concrete has serious implications, although, as yet, the proven correlation in actual field experience is not strong. Seemingly, there is
the possibility of difficulty under unusual situations but sound engineering
judgment would not dictate unusual precautions as being necessary with normal concrete and normal aggregate.
The preceding discussion has dealt with the effects of thermal expansion of
aggregates over termperature ranges that can occur under natural exposure
conditions. Much less is known about these effects under more extreme
ranges as, for example, under the conditions that may exist in burning
buildings. The Bureau of Reclamation reports [14] that at a temperature of
572.7~ (1063 ~ quartz changes state and suddenly expands 0.85 percent,
usually producing a disruptive effect at the suface of concrete in which it is
used. Endell [25] has reported the results of experiments to determine the
structural and expansion changes of concrete aggregates with temperatures
up to 1200~ (2192~
These are considered to be highly specialized conditions and are not discussed further here, but are discussed further in another
paper in this publication. 6
Methods of Determining Thermal Expansion of Aggregates

Several ingenious methods have been developed for determining the thermal coefficient of expansion of coarse aggregate. The majority of the test
methods are based on the measurement of linear expansion over a
temperature range. This range is usually 55 ~ (100~ or more because the
change in unit length per degree is extremely small, and the multiplication of
the change over a substantial temperature range greatly increases the facility
and precision of the measurements. Likewise, the longer the specimen the
greater is the accuracy and precision of the determination. However, the size
of a representative specimen that can be obtained from a coarse aggregate
rapidly approaches a practical maximum. Except in the case of a crushed aggregate of sufficient uniformity to permit obtaining a larger specimen that
will be representative of the sand sizes, this means of multiplying the change
in unit length is not available for determining the linear expansion of fine aggregate. Another device that has been used to obtain additional multiplication of the length change is the optical lever.
The method described by Willis and DeReus [26] is an example of
measurements made over a considerable temperature range with the use of
an optical lever. The specimens were 25.4-mm (1-in.) diameter cores, 50 mm
(2 in.) long, drilled from the aggregate specimens to be tested and placed in a
6See p. 74.
700
controlled-temperature oil bath with a range of 2.78 4- 1.7~ (37 --2 3~ to

60 + 2.8~ (140 + 5~
The vertical movement of the specimen as the
temperature was varied was measured by reading the image reflected by the
mirror of the optical lever, having a 25.4-mm (1-in.) lever arm, on a vertical
scale placed 6.1 m (20 ft) from the mirror, by means of a precise level. It is
reported that consideration of the possible errors involved in the
measurements indicates that the calculated coefficients are probably accurate to 4- 3.6 10-V~ (2.0 X 10-7/~
Another method is the interferometer method described by Merritt [27]
and modified by Saunders [28]. Detailed descriptions of the apparatus, the
preparation of specimens, and the test procedure are also given by Johnson
and Parsons [8]. A third method for determining the thermal coefficient of
expansion of coarse aggregate is that developed by the Corps of Engineers
[21,29] in which an SR-4 strain gage is bonded to a prepared piece of aggregate and readings taken over a temperature range of 1.7 to 57.2 ~ (35 to
135~
This method requires that the piece of aggregate be sliced in three
mutually perpendicular directions, two of these directions to lie in the major
structural plane of the rock, if such plane can be located. The strain gages
are then mounted so as to measure strain in each of the three directions. The
purpose of this requirement is to determine if anisotropy or preferred crystal
orientation exist.
Also the current method used by the Bureau of Reclamation and described
by Mitchell [5] is avaiable. This method employs specimens from 25.4 to 76.2
mm (1 to 3 in.) in size coated with wax and held in fulcrum-type extensometer frames. Measurements are made with electromagnetic strain gages
with electronic indicators while the specimen is immersed in a circulating
ethylene glycol solution held at the desired temperature. The details of the
method are well described in the reference.
Because of the size and usually heterogeneous nature of fine aggregate,
none of the above methods are readily adaptable to the determination of the
thermal coefficient of expansion of this type of material. The usual approach
has been to determine the linear expansion of mortar bars containing the fine
aggregate. Of course, the results obtained include the effect of the length
change contributed by the cement. Verbeck and Hass [30] have developed a
dilatometer method for determining the thermal coefficient of expansion
which is particularly adaptable for use with fine aggregate. The method
determines the cubical thermal coefficient of expansion from which the
linear expansion may be calculated. The apparatus consists of a 1-1itre
dilatometer flask to which is attached a capillary-bulb arrangement containing electrical contacts spaced over a calibrated volume. The flask is filled
with aggregate and water and the apparatus allowed to come to equilibrium
at one of the controlling electrical contacts. The equilibrium temperature is
noted and the procedure repeated at the other electrical contact. After proper calibration, the only measurements required are the weight of the water
COOK ON THERMAL PROPERTIES
701
placed in the flask and the temperature needed to produce an expansion

equivalent to the volume between the electrical contacts. This method offers
a tool for determining directly the average thermal expansion properties of
the smaller aggregate sizes that have not been available heretofore.
Thermal Conductivity, Thermal Diffusivity, and Specific Heat

Thermal conductivity, thermal diffusivity, and specific heat are largely interrelated, and all three normally are determined only for concrete as used in
massive structures. Reference 31 indicates the application of these data in
connection with computing concrete placement temperatures and designing
cooling systems, and in other thermal calculations aimed at reducing thermal
volume change and thus cracking in large dams. The same type of
measurements and calculations would apply equally to other massive structures. Thermal conductivity is also of importance in lightweight concrete for
insulation purposes. It has been indicated by some [32-34] that thermal diffusivity may have an important effect on concrete durability.
Thermal conductivity, measured as the rate of heat flow through a body of
unit thickness and unit area with a unit temperature difference between two
surfaces, normally is expressed in calculations for concrete in the SI system
as watts per metre kelvin or the U.S. customary units as Btu/ft-h-~
Thermal diffusivity is defined as the thermal conductivity divided by the
specific heat and density and is a physical property of the material which
determines the time rate of change of temperature of any point within a
body. Its SI units are square metres per second 7 and U.S. customary units
are square feet per hour.
Specific heat is the amount of heat required to raise the temperature of a
unit mass of material one unit of temperature. Its SI units are joules per
kilogram kelvin 7and U.S. customary units are Btu per lb per ~
Thermal Conductivity of Lightweight Aggregates

One of the more useful properties of lightweight concrete is its low thermal
conductivity. Since thermal conductivity usually varies directly with density,
aggregates of low density produce concrete of lower conductivity. For the
same reason, as indicated by Tyner [35], moisture has a tremendous effect on
thermal conductivity. He reports that in a 1:5 mixture of Florida limerock
concrete an increase of moisture from 0 to 5 percent increases the thermal
conductivity by 23 percent and an increase f r o m 0 to 10 percent increases the
conductivity by 46 percent. Davis and Kelly [36] also state that "the presence
7Users of the SI systemshould be aware that there are several "metric" units which havebeen
used alternatively, for example, there are the calorie gram, the calorie kilogram, the calorie
gram (IT), and others. In the interpretation or comparison of thermal values, the user should
be sure of the particular units used in calculations.
702
of a small amount of moisture in the interior of a lightweight concrete greatly

increases its thermal conductivity; hence, under conditions of continuous or
intermittent exposure to moisture, if a high degree of insulation is desired, an
aggregate (and concrete) of relatively low absorption should be used." Kluge
et al [37] and Price and Cordon [38] also have found pronounced reductions
in the thermal conductivity of concrete containing lightweight aggregate, but
they indicate that the reduction seems to be influenced more by the reduction
in the density of the concrete than by the characteristics of the aggregate.
While the relationship of reduced thermal conductivity with reduced density holds generally for normal and lightweight aggregates, investigations of
barite for use in concrete for radiation shielding by Witte and Backstrom [39]
indicate its thermal conductivity to be somewhat lower than normal concrete
aggregate.
Thermal Diffusivity
Investigations have indicated that the normal diffusivity of the aggregate
may have an influence on the durability of the concrete in which it is used.
Thomson [32] states that for a given body with specified boundary conditions
the thermal stresses depend on certain physical properties of the materials.
In a homogenous body such physical properties as thermal conductivity,
specific heat, and the density of the material influence the temperature
distribution and the thermal stresses during the transient period only in a
certain combination known as the thermal diffusivity. If in a mixture such as
concrete the thermal diffusivities and conductivities are the same for each
material, the body can be thought of as being thermally homogeneous. Since
a difference in diffusivities would result in different rates of diffusion of heat
through the aggregate and cement, it is believed that such a combination
would result in higher thermal stresses than those existing in homogeneous
bodies. Nothstine [33] and Weiner [34] have reported the results of their
approach to the problem. Weiner's work was instigated by the failure of a
gravel concrete, exposed to natural freezing and thawing accompanied by
thermal shock and characterized by bond failure and internal expansion. He
attributes the failure to the relatively high thermal coefficient of expansion of
]the concrete, which is responsible for surface stress, and to the diffusivity of
the gravel which, being higher than the mortar, responds more quickly to
temperature changes, resulting in differential volume change. Fox and Dolch
[40] in an investigation of four limestones found a large change in the thermal diffusivity with a relatively small degree of saturation. The increases in
diffusivity ranged from 20 to 59 percent for saturations of less than 5 percent.
The authors of the references cited essentially agreed that the thermal diffusivity of the aggregates apparently has an effect on the durability of concrete but that further work is needed to determine the significance of the
effect and to find a practical means for using this knowledge to improve
concrete durability.
703
Specific Heat of Aggregates

Specific heat is of considerable importance in connection with the calculations involved in the control of placement temperatures and the limiting of
thermal volume change of mass concrete. The specific heat of the aggregate
contributes materially to the specific heat of the concrete [31].
Methods of Test for Conductivity, Diffusivity, and Specific Heat

Methods are available for the direct determination of thermal conductivity, thermal diffusivity, and specific heat. However, as a matter of practical
convenience, it is customary to determine diffusivity and specific heat and
calculate conductivity or to determine conductivity and specific heat and
calculate diffusivity. This is possible because the formula includes all three
values, and knowing the values for any two, and the density, permits solving the equation for the unknown property. The formula is
k =- hcp
where
k
h
c
p
= thermal conductivity in w/re. k (Btu/ft-h-~

= thermal diffusivity in mE/s (ft2/h),
---- specific heat in J/kg. k (Btu/lb-~
and
= density in kg/m 3 (lb/ft3).
Whether conductivity is determined directly and diffusivity calculated, or

vice versa, is largely a matter of the most convenient equipment setup
available and the preference of the laboratory doing the work. Since the conductivity of a mixture is not an additive function of the constituents [41], the
conductivity and diffusivity of aggregate cannot be obtained indirectly by
determining these properties of the concrete and mortar. Some of the test
methods included in the references to this paper are based on determinations
made on concrete specimens, but in most cases they may be used with aggregates if properly modified with respect to specimen size and shape.
The Corps of Engineers [29], Thomson [32], and Fox and Dolch [40]
describe methods for the direct determination of diffusivity of stone and concrete. All of the methods depend basically on obtaining time-temperature
differential curves between the temperatures at the center and the surface of
a specimen by starting at essentially equilibrium temperature, then changing
the surface temperature, and plotting the time-temperature curve until
equilibrium conditions are obtained at the new surface temperature. Depending upon the degree of accuracy desired, refinements can be made
by grinding the specimen to a sphere, for example, and by refining the
instrumentation. The determination is by no means routine in nature
and requires careful experimentation, precise equipment, and a capable
operator.
704
Thermal conductivity when measured directly is determined by ASTM

Test for Steady-State Thermal Transmission Properties by Means of the
Guarded Hot Plate (C 177) or similar methods [35,42]. Since the equipment
is designed for use with a flat specimen either square or round, it is not
adapted easily to precise determinations of the conductivity of such materials
as concrete aggregates. For this reason it is generally considered better to
determine diffusivity and calculate conductivity for such materials. The
Bureau of Reclamation [43] has developed a method for the determination of
the thermal conductivity of concrete by the use of a 203 by 406-ram (8 by
16-in.) hollow cylindrical concrete specimen. This procedure probably could
be used for coarse aggregate provided a specimen of the required size and
shape could be fabricated from a large rock specimen.
The specific heat of aggregate usually is determined by a procedure known
as the method of mixtures [29]. It is a calorimetric procedure wherein the net
heat required to raise the temperature of a specimen of known weight a given
amount is measUred.
References 44-52 provide additional background information on the
theory and mathematics of thermal tests. Other references may be found appended to many of the references cited here.
Conclusions
Test methods are available which when used with proper attention to procedure, specimen size and shape, instrumentation, and technique are entirely adequate for the determination of the thermal properties of aggregates.
There appears to be no doubt that the normal properties of the aggregates,
particularly thermal expansion, have an effect on the durability and other
qualities of concrete. Investigations reported to date do not present a clearcut picture of the effects that might be expected, and some aspects of the
problem are controversial. The ultimate solution must be based on the performance of aggregates of known thermal properties in concrete, and, as is
normal in this field of investigation, the major difficulty is to separate the effects of the thermal properties of the aggregates from the numerous other
variables existing in the concrete. There is a real need for additional research
work on the subject, both to resolve existing controversy and to improve concrete further as a construction material.
References
[1] Allen, C. H., "Influence of Mineral Aggregates on the Strength and Durability of Concrete," Mineral Aggregates, A S T M STP 83, American Society for Testing and Materials,
1948, p. 153.
[2] Woolf, D. O., "Needed Research," Mineral Aggregates, ASTM STP 83. American Society
for Testing and Materials, 1948, p. 221,
[3] Scholer, C. H., "Durability of Concrete," Report of Significance of Tests of Concrete and
705
Concrete Aggregates, A S T M STP 22,4, American Society for Testing and Materials, 1943,
p. 29.
[4] Hallock, W., "Preliminary Notes on the Coefficients of Thermal Expansion of Certain
Rocks," Bulletin No. 78, U.S. Geological Survey, 1891.
[5] Mitchell, L. J., "Thermal Expansion Tests on Aggregates, Neat Cements and Concrete,"
[6] "A Study of Cement-Aggregate-Incompatibility in the Kansas-Nebraska Area," Report
C-964, U.S. Bureau of Reclamation.
[7] Griffith, J. H., "Thermal Expansion of Typical American Rocks," Bulletin No. 128. Iowa
Engineering Experiment Station, 1936.
[8] Johnson, W. H, and Parsons, W. H., "Thermal Expansion of Concrete Aggregate Material," Journal of Research, RP 1578, National Bureau of Standards, Vol. 32, 1944, p. 101.
[9] Parsons, W. H. and Johnson, W. H., "Factors Affecting the Thermal Expansion of Concrete Aggregate Materials," Journal, American Concrete Institute, April 1944; Proceedings, Vol. 40, p. 457.
[10] "Laboratory Investigations of Certain Limestone Aggregates for Concrete," Technical
Memorandum No. 6-371, Waterways Experiment Station, U.S. Army Corps of Engineers,
Vicksburg, Miss., Oct. 1953.
[11] "Test Data on Concrete Aggregates in Continental United States," Technical Memorandum No. 6-370, Vols. 1-5, Waterways Experiment Station, U.S. Army Corps of Engineers, Vicksburg, Miss., Sept. 1953.
[12] Rhoades, R. and Mielenz, R. C., "Petrography of Concrete Aggregate," Journal, American Concrete Institute, June 1946; Proceedings, Vol. 42, p. 581.
[13] Rhoades, R. and Mielenz, R. C., "Petrographic and Mineralogic Characteristics of
Aggregates," MineralAggregates, A S T M STP 83. American Society for Testing and Materials, 1948, p. 20.
[14] Concrete Manual, U.S. Bureau of Reclamation, 7th ed., 1963, p. 60.
[15] Meyers, S. L., "Thermal Coefficient of Expansion of Portland Cement," Industrial and
Engineering Chemistry, Vol. 32, Aug. 1940, p. 1107.
Journal, American Concrete Institute, June 1942; Proceedings, Vol. 38, p. 36-41.
[17] Pearson, J. C., "Supplementary Data on the Effect of Concrete Aggregate Having a Low
Thermal Coefficient of Expansion," Journal, American Concrete Institute, Sept. 1943;
Proceedings, Vol. 40, p. 33.
[18] Walker, S., Bloem, D. L., and Mullen, W. G., "Effects of Temperature Changes on
Concrete as Influenced by Aggregates," Journal, American Concrete Institute, April 1952;
Proceedings, Vol. 48, p. 661,
[19] Koenitzer, L. H., "Elastic and Thermal Expansion Properties of Concrete as Affected
by Similar Properties of the Aggregate," Proceedings, American Society tbr Testing and
Materials, Vol. 36, Part 2, 1936, p. 393.
[20] Callan, E( J., "Thermal Expansion of Aggregates and Concrete Durability," Journal,
American Concrete Institute, Feb. 1952; Proceedings, Vol. 48, p. 485; discussion, Journal,
American Concrete Institute, Dec. 1952; Proceedings, Vol. 48, p. 504-511.
[21] Callan, E. J., "The Relation of Thermal Expansion of Aggregates to the Durability of
Concrete," Bulletin No. 34, Waterways Experiment Station, U.S. Army Corps of Engineers, Vicksburg, Miss., Feb. 1950.
[22] Swenson, E. G. and Chaly, V., "Basis for Classifying Deleterious Characteristics of Concrete Aggregate Materials," Journal, American Concrete Institute, May 1956; Proceedings. Vol. 52, p. 987.
[23] Smith, G. M., "Physical Incompatibility of Matrix and Aggregate in Concrete," Journal,
American Concrete Institute, March 1956; Proceedings, Vol. 52, p. 791.
[24] Kennedy, T. B., and Mather, K., "Correlation Between Laboratory Accelerated Freezing
and Thawing and Weathering at Treat Island, Maine," Journal, American Concrete
Institute, Oct. 1953, Proceedings, Vol. 50, p. 141.
[25] Endell, K., "Experiments on Concrete Aggregate Materials to Determine Structural and
Linear Expansion Changes with Temperatures up to 1200~ '' Ernst and Sohn, Berlin;
(revised) Tonindustrie-Zeitung, Vol. 53, 1929, p. 1472; Ceramic Abstracts, Vol. 9, 1930,
p. 159; Chemical Abstracts, Vol. 24, p. 2265.
706
[26] Willis, T. F. and DeReus, M. E., "Thermal Volume Change and Elasticity of Aggregates
and Their Effect on Concrete," Proceedings, American Society for Testing and Materials,
Vol. 39, 1939, p. 919.
[27] Merritt, G. E., "The Interference Method of Measuring Thermal Expansion," Journal of
Research, RP 515, National Bureau of Standards, Vol. 10, 1933, p. 591.
[28] Saunders, J. B., "Improved lnterferometric Procedure with Application to Expansion
Measurements," Journal of Research. RP 1227, National Bureau of Standards, Vol. 23,
1939, p. 179.
[29] Handbook for Concrete and Cement. Waterways Experiment Station, U.S. Army Corps of
Engineers, Vicksburg, Miss., 1949.
[30] Verbeck, G. J. and Haas, W. E., "Dilatometer Method for Determination of Thermal
Coefficient of Expansion of Fine and Coarse Aggregate," Proceedings. Highway Research
Board, Vol. 30, 1950, p. 187.
[31] "Cement and Concrete Investigations," Boulder Canyon Project Final Reports, Part 7;
"Thermal Properties of Concrete," Bulletin No. l, Bureau of Reclamation, 1940.
[32] Thomson, W. T., "A Method of Measuring Thermal Diffusivity and Conductivity of Stone
and Concrete," Proceedings. American Society for Testing and Materials, Vol. 40, 1940,
p. 1073; discussion, p. 1081.
[33] Nothstine, L. V., "Thermal Diffusivity and Modulus of Elasticity in Relation to the Durability of Concrete," Kansas State College, Manhattan, Kan., 1940.
[34] Weiner, A., "A Study of the Influence of Thermal Properties of the Durability of Concrete," Journal, American Concrete Institute, May 1947; Proceedings, Vol. 43, p. 997.
[35] Tyner, M., "Effect of Moisture on Thermal Conductivity of Limerock Concrete," Journal,
American Concrete Institute, Sept. 1946; Proceedings, Vol. 43, p. 9.
[36] Davis, R. E. and Kelly, J. W., "Lightweight Aggregates," Mineral Aggregates, A S T M
[37] Kluge, R. W., Sparks, M. M., and Tuma, E. C., "Lightweight-Aggregate Concrete,"
Journal, American Concrete Institute, May 1949; Proceedings, Vol. 45, p. 625.
[38] Price, W. H. and Cordon, W. A., "Tests of Lightweight-Aggregate Concrete Designed for
Monolithic Construction," Journal, American Concrete Institute, April 1949; Proceedings.
Vol. 45, p. 581.
[39] Witte, L. P. and Backstrom, J. E., "Properties of Heavy Concrete Made with Barite
Aggregates," Journal, American Concrete Institute, Sept. 1954; Proceedings. Vol. 51,
p. 65.
[40] Fox, R. G., Jr. and Dolch, W. L., "A Technique for the Determination of a Thermal
Characteristic of Stone," Proceedings, Highway Research Board, Vol. 30, 1950, p. 180.
[41] Austin, J. B., "Factors Influencing the Thermal Conductivity of Non-Metallic Materials,"
Thermal Insulation Materials, A S T M ST,~ 39, American Society for Testing and Materials, 1939, p. 17.
[42] "Projected Testing Method for Conductivity of Materials," Bulletin No. I9, International
Association of Testing and Research Laboratories for Materials and Structures, RILEM,
Nov. 1954, p. 3.
[43] Materials Laboratory Procedures Manual. U.S. Bureau of Reclamation, 1951.
[44] Shack, Alfred, translated from the German by Hans Goldschmidt and E. P. Partridge,
Industrial Heat Transfer, Wiley, New York, 1933.
[45] McAdams, W. H., Heat Transmission, McGraw-Hill, New York, 1942.
[46] Carslaw, H. S. and Jaeger, J. C., Conduction of Heat in Solids, Oxford University Press,
London, 1947.
[47] Clark, H., "The Effects of Simple Compressing and Wetting on the Thermal Conductivity
of Rocks," Transactions. American Geophysical Union, Part 3, 1941, p. 543; Chemical
Abstracts, Vol. 36, 1942, p. 2820.
[481 Clark, H. and Birch, F., "Thermal Conductivities of Rocks and Its Dependence Upon
Temperature and Composition," American Journal of Science, Vol. 238, 1940, p. 529;
ChemicalAbstracts. Vol. 34, 1940, p. 7796.
[49] Ingersoll, L. R., Zobel, O. J., and Ingersoll, A. C., Heat Conduction with Engineering and
Geological Applications, McGraw-Hill, New York, 1948.
]50] Jakob, M., Heat Transfer. Wiley, New York, 1949,
707
[51] Niven, C. D., "Thermal Conductivities of Some Sedimentary Rocks," Canadian Journal
of Research, Vol. 18, 1940, p. 132; ChemicalAbstracts, Vol. 28, 1934, p. 7455.
[52] Kingery, W. D. and McQuarrie, M. C., "Concepts of Measurements'and Factors Affecting
Thermal Conductivity of Ceramic Materials, Journal, American Ceramic Society, Vol. 37,
No. 2, Part 2, Feb. 1954, p. 67.
Sidney D i a m o n d 1
Chapter 40, Chemical Reactions

Other than Carbonate Reactions
Introduction
The proper role of aggregates in concrete is usually perceived as mechanical, rather than chemical. Aggregates are selected on the basis of their
favorable physical properties, and unless something unusual occurs, they
are supposed to remain inert and unreactive chemically.
However, a growing body of experimental evidence indicates that even in
"normal" concrete a certain amount of chemical reaction must be occurring
between the aggregate grains and the surrounding cement paste, if only at
the interface. Direct microscopical evidence of interfacial chemical responses
[1-3] 2 is complemented by reports of differences in behavior of otherwise
identical concretes made with aggregates of different chemical character,
for example, carbonate aggregates as compared to siliceous aggregates
[4, 5]. Thus the usual assumption about the chemical inertness of aggregates
under normal circumstances is not correct.
This situation is perhaps less surprising when considered in terms of the
chemical environment that exists within concrete. Recent research in which
the pore solution has been expressed from hardened cement pastes [6, 7]
and mortars [8, 9] made from even moderate alkali-content cements reveals
that the environment surrounding the aggregates is one in which the pH is
considerably higher than often realized. Values of pH as high as 13.5 or more
are not unusual; further, the alkali concentration often approaches 1 N,
and may exceed it.
It is becoming apparent that aggregates in concrete are exposed over most
of their service life to small amounts of highly concentrated alkali hydroxide
solutions containing little dissolved calcium, even though the solutions are
saturated with respect to calcium hydroxide. Such solutions constitute
lProfessor of engineering materials, Purdue University, West Lafayette, Inc. 47907.
Copyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:41 EDT 2014 7O8
Universidade
do Gois pursuant
Agreement.
Copyright9
1978 tobyLicense
ASTM
International
www.astm.org
DIAMOND ON CHEMICAL REACTIONS OTHER THAN CARBONATE
709
a much more aggressive environment than the conditions with which

aggregate rocks have come to equilibrium during their geochemical history.
Under these circumstances the idea that most aggregates in concrete undergo
at least surface chemical reactions is not surprising.
However, we are concerned here not with interfacial reactions, but with
the effects of bulk chemical reactions in which significant portions of at
least some aggregate grains are transformed chemically to reaction products,
which may be deleterious and cause distress in concrete.
Consideration of these matters would be facilitated, in the writer's opinion, if a clear distinction is recogized between the aggregate reactions per se
and any expansion, cracking, strength loss, or other deleterious effect that
may occur as a secondary consequence of such reaction. The same reaction
may produce a deleterious secondary effect in some circumstances and not
in others.
Most known aggregate reactions in concrete are brought about in response
to the alkaline character of the pore solution and are properly designated as
"alkali-aggregate" reactions. Such reactions are known to occur with at
least two distinct kinds of aggregates--those that are largely composed of
silica (or siliceous minerals or glasses) and those that are largely composed
of carbonates, particularly dolomitic carbonates. Considerations relating to
the latter, the "alkali-carbonate" reactions, are specifically excluded from
this paper, and are covered elsewhere in this volume.
With respect to noncarbonate alkali reactions, Gillott [10] has suggested
that a distinction be drawn between reactions involving the various forms of
silica (opal, chalcedony, chert, flint, tridymite, cristobalite, microcrystalline,
or strained quartz, etc.) for which the designation "alkali-silica" reactions
would be reserved, and another class of reactions involving complex rock
types having as active components various layer-lattice and other silicate
minerals, such reactions being called "alkali-silicate" reactions. The rocks in
question include graywackes, phyllites, siltstones, argillites, arenites, and
others. The distinction is not clear-cut, because many of these rocks include
fine quartz as well as the silicate minerals, and some contain dolomitic
grains. The utility of setting up a separate category of "alkali-silicate" reactions is not universally accepted [11].
In addition to these reactions specifically attributed to the alkaline character of the pore solutions in concrete, other reactions can occur with specific
components of certain aggregates that can produce deleterious consequences.
These have been discussed in some detail by Hansen [12]. The specific reactions involved include delayed hydration of periclase (MgO), of various iron
oxide constituents, and even of calcium oxide (CaO), which is an occasional
constituent of slags and expanded shales [13]. These cause distress resulting
from the enlargement of the individual grains. Other sources of distress
include various ferrous sulfides that can oxidize and expand [14]; dehydrated
clays and some zeolites that can undergo hydration and expand; and anCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:41 EDT 2014
710
hydrite, formed in some expanded shale and expanded clay aggregates,

which can hydrate to gypsum and cause problems both by direct expansion
pressure and by subsequent formation of secondary ettringite.
Discussion of these reactions also might lead to the consideration of problems associated with contamination of aggregate materials by various soluble
salts, a related problem and one of much current concern in Middle Eastern
countries. However, such problems are outside of the intended scope of the
present paper. In fact, the available information on occurrences of nonalkaline reactions in general is so meager that these will not be discussed
further, and the .remainder of this paper will be concerned exclusively with
the alkali reactions.
Mechanisms Responsible for Distress Associated with Alkali-Silica Attack

The writer has recently reviewed certain aspects of this subject [15,16].
Most of the available information on noncarbonate alkali-aggregate attack
has been gathered in terms of reaction and distress with silica-bearing aggregate. It is presumed that the slower and perhaps more complicated reactions that occur with the silicate rocks are at least generally similar to what
occurs with the silica aggregates.
The practical consequences of alkali-silica attack are that the affected
structure or structural members: (1) are placed under expansive stresses,
(2) expand, (3) crack, (4) lose strength, and (S) by virtue of cracks reaching
the exterior, become exposed to further deterioration from external sources.
An additional consequence of importance in some cases is that the external
appearance of the affected structure is degraded by deposits of alkali-silica
sols which bleed to the outside, carbonate, and leave patchy white deposits.
The problems generated by alkali-aggregate attack range from the purely
cosmetic to the complete failure of the structure and the necessity for its
replacement.
The time scale involved may be rapid (of the order of several months to
only a few years), or it may be slow (of the order of 30 years or more). In the
latter case the distress may be evident at a fairly early age but progress only
slowly, or it may appear suddenly without warning many years after the
structure has been placed in service.
Alkali-silica reactions and distress have been observed in many countries,
and probably exist unnoticed or unremarked in many others. While a typical
map cracking pattern often occurs at the exposed surface of the concrete,
other causes of distress in concrete may generate similar signs. In some
occurrences it is difficult or impossible to decide whether alkali reactions
are the primary cause of the problem, since distress due to shrinkage, sea
water attack, freezing and thawing, or sulfate attack from other sources may
also be present. Significant contributions to this difficult problem of field
recognition have been made by Idorn [17] in his systematic survey and analyCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:41 EDT 2014
711
sis of the causes of distress in Danish concrete structures, and by Mielenz

[18] in his detailed discussion of the microscopic aspects of the problem.
The broad outlines of the mechanism responsible for the distress are
generally agreed on by most investigators, but a number of the details remain
to be settled. The affected (silica) aggregates react in situ with the alkaline
solutions in such a manner that they are converted to alkali silica glasses or
relatively dry gels. If free water is available, it is absorbed by these gels,
which then swell and generate the expansive forces that do the damage. Both
coarse and fine aggregate grains may be affected. In some cases it is clear
that only part of a grain is affected, the remainder being resistant to the
reaction. Furthermore, a full gamut of different responses have been observed when distressed concrete is examined petrographically. Some affected
aggregates retain their apparent rigidity and most of their original microscopic features, and merely expand. Other aggregates seem to convert completely to gelatinous products in situ--they exert enough expansive stress
on their surroundings to have caused cracks in the surrounding cement paste
matrix, but if dried specimens are examined the altered residual "grain"
is seen to have shrunk away from the original contact with the paste, and
may display shrinkage cracks.
If water is easily available, the reaction products may grow progressively
more hydrated and convert gradually to fluid sols and even to solutions.
These may run through previously generated cracks and sometimes accumulate in void spaces. Hansen [19] and other workers believe that such isolated
pockets of reaction product sol will exert further expansive stresses by drawing in additional water through the "membrane" of cement paste surrounding them, that is, by osmotic action. In any event, the fundamental action
of the initial mechanically competent reaction products in attracting water
and swelling is osmotic in character, the reaction product gel itself constituting in effect a "virtual" membrane. Osmotic action with self-constrained gels does not require an intact membrane to surround the gel, and is
continued even after cracks develop in the concrete.
The expansion and cracking that constitute the deleterious features of the
attack are not invariable consequences of the occurrence of the alkali-silica
reaction. Under many circumstances reaction does not lead to distress. This
may come about if the concrete is sufficiently porous to provide local accommodation for the expansive effect; if it is under a sufficiently strong
compressive loading so that the expansive stresses produced are overbalanced; if the available water is too limited in amount to produce the requisite swelling of the reaction product; if the character of the reaction
product produced is such that the gel converts rapidly to a sol at relatively
low water content and hence ceases to exert expansive stresses; perhaps if
the reaction product is soon converted to a calcium alkali-silicate gel of
limited swelling ability; or for various other reasons not fully understood.
Furthermore, it is well established that although considerable reaction may
712
occur, expansion and distress may not be severe if the fraction of aggregate
present that is sensitive to alkaline attack is either much smaller than or
much larger than some "pessimum" proportion, or conversely, if the alkali
content of the pore fluid is much less or much greater than some "pessimum" amount.
A reliable and simple means of detecting reaction independently of the
distress it may cause would be helpful in distinguishing cases where no alkali
reaction has occurred from those where reaction has occurred but has not
led to visible distress in the particular concrete. In the latter instance,
changes in such things as the alkali content of the cement, the cement
content of the mix, or the conditions of exposure may convert the aggregate
from a seemingly innocuous one to a deleterious one.
Methods of Avoiding Alkali Aggregate Reaction Problems

Avoiding the deleterious consequences of alkali-aggregate reactions in
concrete is not particularly complicated in principle. The entire gamut of
techniques ordinarily practiced can be summarized as follows
1. Do not use reactive aggregates.
2. Do not use cement with a high content of alkalies.
3. When bringing possibly reactive aggregate and high alkali cement
together for economic or other reasons, use an admixture (pozzolan) which
experience or specific tests indicate will mitigate the deleterious consequences of any reaction that may develop.
ASTM standardization activities have been addressed to each of these
aspects. The resulting set of practices, with a few exceptions, has been successful in preventing large scale damage to concrete in the past. Whether it
will continue to do so in the future is open to question.
Methods of Avoiding the Use of Reactive Aggregates

To avoid the use of reactive aggregate it is necessary to know which
aggregates are reactive.
The first line of defense, and the one technique practiced to the exclusion
of all others by probably the great majority of concrete producers, is simply
to use aggregate of known satisfactory service record. This usually works well
if there is such aggregate available and economically competitive in the area,
assuming that the currently available aggregate is really the same as the
aggregate on which the service record is based (new faces are constantly
opened up in any active quarry), and assuming that relevant parameters of
the concrete being produced are similar to those for the concrete on which
the service record is based. These parameters include the alkali content of
the cement to be used, the richness of the mix (that is, the amount of cement
added per unit volume), and the possibility of additional alkali being brought
713
in from sources external to the concrete during its exposure. It should be

apparent that satisfactory service from a quarry under one set of conditions
is in itself no guarantee of continued satisfactory service if either the actual
rock or the conditions of use have changed materially.
The second way in which the use of reactive aggregate can be avoided is
simply to "inspect" the aggregate and reject any that inspection discloses
as being reactive. Unfortunately, "inspection" for the presence of reactive
components in aggregate cannot be carried out satisfactorily by means
available to most aggregate or concrete producers. In order to have much
chance of screening out deleterious aggregate "inspection" really entails one
or more of the following: (a) a complete petrographic examination by a
petrographer familiar with the problem, (b) a chemical test requiring analytical chemical facilities and some laboratory expertise, and (c) a "proof test"
where one mixes high alkali cement with the aggregate and waits to see if
expansion will take place.
All these approaches have been the subjects of ASTM standardization
activity.
The ASTM Recommended Practice for Petrographic Examination of
Aggregate for Concrete (C 295) provides specific and helpful information
with reference to such important matters as sampling techniques and details
of what should be determined in examination of natural sands and gravels,
of drilled cores, of ledge rock, etc. No attempt is made to provide criteria
for the recognition of individual minerals or rock constituents, this being
assumed to be part of the professional equipment of the petrographer.
Unfortunately, in applying this recommended practice to the recognition
of reactive components in concrete aggregates, the petrographer has certain
practical problems to overcome. Some components, for example opal or
chalcedony, are clearly deleterious and can be recognized readily. Other
components, for example cherts, may be deleterious in some cases and not
in others. Quartz, close to being present universally, may be deleterious in
microcrystalline or cryptocrystalline sizes, especially when strained, and
reports are accumulating that even strained megascopic quartz may sometimes be reactive, but not always. The petrographer has a considerable
problem with silicate rocks that sometimes exhibit alkali-silicate responses,
again, not with any great degree of predictability. Dolar-Mantuani [20] has
provided guidance in such decisions.
A related difficulty has to do with proportions of possibly reactive components that might be present. One of the peculiar features of the alkaliaggregate problem is the existence of the "pessimum proportion" phenomenon. Aggregates with a proportion of reactive component much higher than
the pessimum, and with the reactive species thus readily detectable by the
petrographer, would in fact be less deleterious than aggregate containing a
much smaller content of the same species, which might escape detection.
The ultimate in this unusual state of affairs must surely be the aggregate
714
responsible for the cracking and distress observed in the Vale de la Mare
Dam on the Island of Jersey. The reactive component (beekite, an opalchalcedony complex) is ascertained reliably to make up perhaps 0.02 percent
of the total aggregate [21].
The second line of defense in screening out reactive aggregate is the chemical test, often designated the "quick chemical test" for its most desirable
attribute. This test, the ASTM Test for Potential Reactivity of Aggregates
(Chemical Method) (C 289), is based originally on research carried out by
Mielenz et al at the Bureau of Reclamation [22]. The idea is that one distinguishes between a "deleteriously reactive" rock and an "innocuous" rock
by crushing the rock to standard size (150 to 300 #m), immersing it in a
1 N sodium hydroxide solution, and "cooking" it in a sealed container at
80~ for 24 h. The resulting suspension is then filtered, and the filtrate is
analyzed for two parameters: the amount of silica dissolved in it, and the
reduction in the original hydroxide ion concentration due to reaction with
the granular rock. Roughly speaking, deleterious rocks are those with which
unit reduction in alkalinity is accompanied by at least unit production of
soluble silica; conversely, innocuous rocks are considered to be those with
which either little or no reduction in hydroxide concentration occurs (that
is, little reaction takes place), or else rocks with which significant reduction
in hydroxide ion concentration is observed, but accompanied by a less-thanequivalent production of soluble silica.
As currently specified, the reduction in OH-ion concentration, Re, is
plotted on a linear ordinate against the concentration of soluble silica, Sc
on a logarithmic abscissa, and a standard line, modified from the original
1:1 correspondence on the basis of experience with mortar bar testing and
field service reports, is used to separate "innocuous" aggregates from those
not considered innocuous. Such a chart is shown in Fig. 1.
In the current version of the test method, the aggregates not considered
innocuous are further subdivided into those considered "potentially deleterious" and those considered "deleterious." Somewhat paradoxically, the
"potentially deleterious" aggregates are those with which a large reduction
of OH-concentration is accompanied by a correspondingly large (or larger)
production of soluble silica, that is, a great amount of reaction of the type
with which we are concerned has taken place. The aggregates considered
"deleterious" on the other hand have shown only a modest lowering of the
OH-ion concentration accompanying a considerable production of silica.
The rationale is based generally on the idea that aggregates that show extensive reaction in the test are likely to contain greater than the pessimum proportion of reactive constituents, and so are capable of causing serious distress
only if diluted with additional unreactive material [23]. In any event, the
subdivision seems to be backed up by mortar-bar results attesting to its
validity.
The results of the quick chemical test may be anomalous if certain conCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:41 EDT 2014
715
stituents that react with the alkaline solution differently from the susceptible
silica or siliceous components of concern are present. The presence of magnesium or ferrous carbonates (dolomite, magnesite, or siderite) or of serpentine in the aggregate is known to result in unusually large reductions in OHconcentration and also restrictions on the amount of soluble silica that would
otherwise be produced, thus perturbing the interpretation of the quick chemical test. The presence of calcite has been said to cause difficulties in some
circumstances, and there may be other unknown aggregate constituents
that also interfere. Thus it is often desirable to combine the quick chemical
test with at least some petrographic examination to be sure that known or
suspected components that may result in anomalous interpretations are
absent.
It should be understood that the quick chemical test involves reaction
conditions (temperature and pressure) different from those at which concrete
undergoes alkali-aggregate reaction in the field, and is in no sense a predictor of the extent of expansion which a given"deleterious" aggregate might
produce in a given concrete. Nevertheless it appears to serve its designated
purpose well, at least with respect to the alkali-silica reaction. It is appreciated particularly where quick decision is required, the results being
obtainable in a few working days.
Perhaps the screening method most closely related to the actual behavior
of field concrete is the method whereby mortar bars consisting of the aggregate and a highly alkaline cement are prepared and, exposed to controlled
temperature and humidity, and the resulting expansion, if any, is measured.
Such a scheme is the basis of ASTM Test for Potential Alkali Reactivity
of Cement-Aggregate Combinations (Mortar-Bar Method) (C 227). In this
method a procedure is prescribed for preparing mortar bars using a specified
gradation of fine aggregate and a reference cement of high alkali content (or
alternatively, the contemplated job cement). The mortar bars are stored in
special containers designed to maintain a 100 percent relative humidity
atmosphere at a slightly elevated temperature (37.8~
and their lengths
are measured at intervals to monitor the rate and extent of any expansion
occurring. Coarse aggregate to be tested for alkali reactivity is first crushed
to the specified fine aggregate gradation. Cement-aggregate combinations
that show expansions greater than 0.10 percent in 6 months under this
regime "usually should be considered capable of harmful reactivity."
Combinations which show expansions of 0.05 percent at 3 months are considered "potentially capable of harmful reactivity."
A somewhat different set of exposure conditions is prescribed for mortar
bars in an alternative mortar-bar test, ASTM Test for Potential Volume
Change of Cement-Aggregate Combinations (C 342). This test is said to have
particular applicability to certain cement aggregate combinations common
in Oklahoma, Kansas, Nebraska, and Iowa. A variant form of reaction has
been reported from this region, which is characterized by a hot, dry environCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:41 EDT 2014
716
700
9
o
600
--.
Aggregates causing m o r t a r expane~on more than 0.1 percent in o year

* h e n used with a cement c o n t a i n i n g 1.38 per cent a l k a l i e s ,
A g g r e g a t e s causing rttortar exDan$1on tess t h a n 0.1 p e r c e n t in o year
under some condifions,

Aggregates for which mortar expansion data are not a v a i l a b l e but
which ate indicoter to be d e l e t e r l o u ~ by petrographic examination.
Boundary
which are indicated to be innocuous by pe, r o g r o p h i c e x o m i n a t l o n

llne b e t w e e n innocuous and d e l e t e r i o u s a~gregotes.
500
"400
?
.g.
~ 300
C
g
o
a~
i
~1oo
a.5
5.o
Quantity
FIG.
7.5 ~o
Sc-Dissolved
z5
5o
Silica
75~oo
(millimoles
z5o
soo rso ,ooo
25oo
per l i t e r )
1--Decision chart for interpretation of the results of A S T M Test C 289,
ment much of the year, and where much of the available aggregate is in the
form of relatively fine gravel, necessitating concrete mix designs of higher
than usual cement content. The reaction under these circumstances appears
to be characteristically slow and unspectacular in early years, but to result
eventually in severe map cracking, expansion, and degradation. ASTM Test
C 342 (based originally on the work of Conrow [24]) prescribes curing the
mortar bars in water at room temperature (23~ up to 28 days, followed
by a week of curing at an elevated temperature (55~ with the humidity
717
maintained at 100 percent relative humidity, then a week of drying in a

drying oven at the same temperature, and finally a return to the original
23 ~ water-curing conditions. Subsequently the expansion is monitored at
periodic intervals. Criteria are provided in the Appendix to ASTM Specification for Concrete Aggregates (C 33), indicating that aggregate-cement combinations producing expansion equaling or exceeding 0.200 percent at 1
year may be considered unsatisfactory.
The ASTM mortar-bar tests described above have a long history of use
in the United States, and most concrete technologists feel that they are
successful and reliable as indicators of deleteriously expansive aggregates.
However, other researchers, particularly many technologists from outside
the United States, are concerned that mortar bars may lack a sufficiently
close correspondence to field concrete to be entirely reliable, and have
developed concrete prism tests for what are considered more realistic assays
of possibly deleterious aggregates. Current German practice uses concrete
prisms (cubes) 30 cm on a side, and Dahms [25] reported that not only are
the cube expansions different from those of mortar bars, but the shape
and dimensions of the concrete prisms themselves significantly affect the
results, "for example, no expansions or only very small ones could be
found in (concrete) beams, yet the 30-cm sized cubes showed quite substantial cracks and precipitations of alkali silicate gel."
Difficulties with ASTM Test C 227 mortar bars and a preference for
concrete prisms have also been expressed by Canadian researchers [26];
this is true especially when considering rocks susceptible to the less-wellunderstood "alkali-silicate" reactions. Further, these researchers suggest
that the expansion "schedule" specified in ASTM Test C 227 (0.1 percent
at 6 months or 0.05 percent at 3 months) is not appropriate for the alkalisilicate reacting rocks, which take longer to expand.
Limitation on the Alkali Content o f the Cement
The second general way of staying out of trouble with alkali-silica reactions is to avoid the use of high-alkali cement, thus limiting the amount
of alkali that would be expected to be brought into the pore solution of
the resulting concrete. The ASTM Specification for Portland Cement
(C 150) suggests that a limit of 0.60 percent on the total alkali content
(percent Na20 plus 0.658 percent K20) may be specified when the cement
is to be used in concrete with aggregates that may be deleteriously reactive.
The value of 0.60 percent was arrived at empirically by correlation with
field indications of distress, and seems to be a reasonable figure.
However, the specification does not take into account the proportion
of cement to be used in a particular concrete. Clearly a 446 kg/m 3
(750 lb/yd 3) mix design will result in incorporating more than half again
as much alkali in a unit volume of concrete as will a 307 k g / m 3 (420 lb/yd 3)
718
mix design, and will have a proportionately higher potential for chemical
attack. Further, limiting the alkali content of the cement does not guard
against the potential introduction of additional alkalies from other
sources--sea water, deicing salts, etc. Finally, the value of the alkali
content limitation specified ceases when it becomes impossible or uneconomical for the cement producer to provide cement that meets the limit.
With the growth of energy constraints, dry process kilns, often with suspension preheaters, are replacing older designs, and these energy-efficient
kilns provide only limited ability to reduce excess alkali in the raw mix
by venting. This development, combined with the increasing tendency to
recycle alkali-rich waste cement dust to the raw mix has resulted in a
progressive increase in alkali contents of most cements in recent years.
It seems likely that only those few suppliers blessed with low alkali raw
materials will be able to offer cement that complies with the 0.60 percent
alkali limitation for the indefinite future.
Use of Pozzolanic or Other Admixtures to Mitigate the Effects of AlkaliAggregate Reactions

Experience has indicated that introduction of significant proportions of
certain pozzolanic materials, as a substitute reactant that combines with
the alkali solution more rapidly than the deleterious aggregate and which
produces reaction products without deleterious consequences, can prevent
distress in concrete. Deciding what constitutes an appropriate pozzolan for
this purpose is the function of a special mortar-bar testing procedure
described in ASTM Test for the Effectiveness of Mineral Admixtures in
Preventing Excess Expansion of Concrete Due to Alkali Aggregate Reaction (C 441). The procedure is based on the mortar-bar test of ASTM Test
C 227, except that the aggregate used is a specified gradation of highly
reactive borosilicate glass rather than an aggregate to be tested for its
potential reactivity. The glass is combined in a mortar with a high alkali
reference cement (alkali content of at least 1.0 percent). Control mortar
bars are prepared without admixture; test mortar bars prepared at the
same time contain 25 percent less cement than the control mix, and
incorporate an amount of the test pozzolan of the same volume as that
of the cement removed. Control and test bars are maintained at 37.8~
and 100 percent relative humidity, and are measured at 14 days. The
beneficial effect of the admixture (if any) is displayed in terms of the
reduction in expansion of the test mortar bars over that of the control,
reduction of 75 percent or more being deemed satisfactory.
In a variant of the test method, Section 4.3 of ASTM Test C 227 prescribes that the contemplated job cement be used instead of a reference
high-alkali cement, and that the contemplated pozzolan be used in the
amount proposed for use on the job. For this "job mix" mortar, a limitation in expansion of 0.02 percent over the 14 day period is specified.
719
ASTM Specification for Fly Ash and Raw or Calcined Natural Pozzolan
for Use as a Mineral Admixture in Portland Cement Concrete (C 618)
lists the 75 percent reduction in expansion of the test mortar as compared
with the control mortar, and the 0.02 percent maximum expansion of
the "job mix" mortar as optional requirements for pozzolans, to be applied
at the purchaser's request. ASTM Specification C 618 indicates that the
test for mortar-bar expansion of the "job mix" mortar is to be preferred
over that for reduction of expansion with a reference cement, if the portland cement to be used in the work is known and available.
The use of other than pozzolanic admixtures to mitigate the effects of
alkali-aggregate reactions is not addressed by an ASTM specification,
but evidence in the literature suggests that small quantities of lithium
salts added as admixtures reduce expansions of reactive combinations
significantly [27]. The possible use of barium salts also has been suggested, but the latter will precipitate a highly insoluble barium sulfate
instantly on addition to suspensions of ordinary gypsum-containing portland cement. The experimental work in which the beneficial effects of
barium salts were described [28] was carried out with gypsum-free ground
clinker specially prepared for the purpose.
Another significant possibility for the mitigation of the deleterious
effects of alkali reaction is the possible use of pozzolanic cement or of
blast-furnace slag cement. Present specifications for such blended hydraulic cements (such as ASTM Specification for Hydraulic Cements
(C 595)) make only passing reference to alkali reactions, but considerable
European experience attests to the potential value of at least some such
blended cement products in controlling the deleterious consequences of
alkali reactivity.
Concluding Remarks
All sources of available information about trends in the cement industry
suggest that the drift to higher alkali contents in most portland cements
will continue in the foreseeable future. This comes at a time when high
quality aggregates with proven service records are becoming progressively
more difficult to obtain in many areas due to various factors, but particularly to urban restrictions on continued operation of aggregate sources.
While the trend does not seem to have caused a noteworthy upsurge in
alkali-related problems in the recent past, scattered reports of such
problems continue. It is probably too soon for occurrences of distress
associated with the recent upsurge in alkali contents of portland cements
to have manifested themselves, since a lag time of a number of years before
such distress is noted would be expected. Nevertheless, it would not be
surprising if a number of recently completed structures eventually show
such distress, especially where adherence to the optional limitation of 0.6
percent total alkali content is not traditionally practiced.
720
T h e r e has b e e n an i n c r e a s e of r e s e a r c h in d e f i n i n g t h e d e t a i l s o f a l k a l i
attack under various circumstances, and the summary of one and the
c o m p l e t e p u b l i s h e d p r o c e e d i n g s o f two o t h e r r e c e n t i n t e r n a t i o n a l conf e r e n c e s on t h e q u e s t i o n are now a v a i l a b l e [29,30]. It is h o p e d t h a t cont i n u e d r e s e a r c h on t h e s u b j e c t will b e f r u i t f u l in d e v e l o p i n g e n h a n c e d
m e t h o d s of r e c o g n i z i n g a n d c o p i n g with t h e p r o b l e m .
Acknowledgments
T h e w r i t e r is p l e a s e d to a c k n o w l e d g e his d e b t to G u n n a r I d o r n , K a t h a r i n e
a n d B r y a n t M a t h e r , a n d Bill D o l c h - - f r i e n d s a n d c o l l e a g u e s w h o h a v e prov i d e d i n f o r m a t i o n , c o n v e r s a t i o n , a n d g u i d a n c e with r e s p e c t to alkalia g g r e g a t e p r o b l e m s o v e r a n u m b e r of years.
References
[1] Barnes, B. D., "Morphology of the Paste-Aggregate Interface," Ph.D. thesis, School
of Civil Engineering, Purdue University, West Lafayette, Ind., Jan. 1976.
[2] Farran, J. et al, "Etude de l'Aureole de Transition Existant entre les Granulats d'un
Mortier et la Masse de la pate de Ciment Hydrate," Liaisons de Contact dans les
Materiaux Composites Utilises en Genie Civil, RILEM International Colloquium, Vol. 1,
Toulouse, France, 1972, pp. 60-76.
[3] Javelas, R., Maso, J. C., and Ollivier, J. P., Cement and Concrete Research, Vol. 4,
No. 2, March 1974, pp. 167-175.
[4] Alexander, K. M., Wardlaw, J., and Gilbert, D. J., "Aggregate Cement Bond,
Cement Paste Strength, and the Strength of Concrete," International Conference on
the Structure of Concrete, London, 1965, pp. 59-81.
[5] Scholer, C. F., RecordNo. 210, Highway Research Board, 1967, pp. 108-117.
[6] Longuet, P., Burglen, L,, and Zelwer, A., Revue des Materiaux, No. 676, 1973, pp.
35-41.
17] Abbassi, Gh., Revue des Materiaux. No. 691, Nov.-Dec. 1974, pp. 315-322.
[8] Barneyback, R. S. and Diamond, S., "Preliminary Studies on Pore Solutions Expressed from Alkali Silica Reactive Mortars," Annual Meeting, American Ceramic
Society, May 1976. Abstract, Bulletin, American Ceramic Society, Vol. 55, No. 4,
April 1976, p. 407.
[9] Diamond, S. and Barneyback, R. S., Proceedings, 1976 Alkali Symposium, The
Effect of Alkalies on the Properties of Concrete, Cement and Concrete Association,
London, 1977, in press.
[10] Gillott, J. E., "Practical Implications of the Mechanisms of Alkali-Aggregate Reactions," Symposium on Alkali-Aggregate Reaction, Preventative Measures, Reykjavik,
Iceland, Aug. 1975, pp. 213-230.
[11] Mather, B., Cement and Concrete Research, Vol. 6, No. 6, Nov. 1976, pp. 813-814.
[12] Hansen, W. C., "Chemical Reactions," Significance of Tests and Properties of Concrete and Concrete-Making Materials, A S T M STP 169 A. American Society for
Testing and Materials, 1966, pp. 487-496.
[13] Hansen, W. C., "Anhydrous Minerals and Organic Materials as Sources of Distress
in Concrete," Record No. 43, Highway Research Board, 1964, pp. 1-7.
[14] Mielenz, R. C., "Reactions of Aggregates Involving Solubility Oxidation, Sulfates,
or Sulfides," Record No. 43, Highway Research Board, 1964, pp. 8-18.
[15] Diamond, S., Cement and Concrete Research, Vol. 5, No. 4, July 1975, pp. 329-346.
[16] Diamond, S., Cement and Concrete Research, Vol. 6, No. 4, July 1976, pp. 549-560.
[17] Idorn, G. M., "Durability of Concrete Structures in Denmark," Technical University
of Denmark, Copenhagen, Jan. 1967.
[18] Mielenz, R. C., "Cracking of Concrete Related to the Alkali-Silica Reaction," DisCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:45:41 EDT 2014
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
721
eussion presented at the U.S.-Japan Joint Seminar on Research on Basic Properties of

Various Concretes, Tokyo, Jan. 1968.
Hansen, W. C., Cement and Concrete Research, Vol. 6, No. 2, March 1976, pp.
323-325.
Dolar-Mantuani, L., "Petrographic Investigation of Alkali-Reactive Silicate Rocks
in Several Structures," Proceedings, 1976 Alkali Symposium, London, A. B. Poole, ed.,
Cement and Concrete Association, 1977, in press.
Midgley, H. G., "The Identification of Opal and Chalcedony in Rocks, and Methods
of Estimating the Quantity Present," Proceedings, 1976 Alkali Symposium, London,
A. B. Poole, Ed., Cement and Concrete Association, 1977, in press.
Mielenz, R. C., Greene, K. T., and Benton, E. J., Journal, American Concrete Institute, Vol. 19, No. 3, Nov. 1947, pp. 193-224.
Mielenz, R. C. and Witte, L. P., Proceedings, American Society for Testing and
Materials, Vol. 48, 1948, pp. 1071-1103.
Conrow, A. D., Proceedings, American Society for Testing and Materials, Vol. 52,
1952, pp. t205-1227.
Dahms, J., "Influences on the Alkali Aggregate Reaction Under Field Conditions,"
Proceedings. 1976 Alkali Symposium, London, A. B. Poole, Ed., Cement and Concrete
Association, 1977, in press.
Grattan-Bellew, P. and Litvan, G. G., "Testing Canadian Aggregates for Alkali
Expansivity," Proceedings. 1976 Alkali Symposium, London, A. B. Poole, Ed., Cement
and Concrete Association, 1977, in press.
McCoy, W. J. and Caldwell, A. G., Journal, American Concrete Institute, Proceedings.
Vol. 47, 1951, pp. 693-706.
Hansen, W. C., Journal, American Concrete Institute; Proceedings, Vol. 31, 1960,
pp. 881-883.
Asgeirrsson, H., Ed., Symposium on Alkali Aggregate Reaction--Preventative Measures, Icelandic Building Research Institute, Reykjavik, Iceland, 1975.
Poole, A. B., Ed., Proceedings. 1976 Alkali Symposium, London, Cement and Concrete
Association, 1977, in press.
H. N. W a l k e r ~
Chapter 41--Chemical Reactions of

Carbonate Aggregates in Cement
Paste
Introduction
It has long been recognized that no aggregate is completely inert in the
environment of portland cement paste. This environment is always at
least as alkaline as a saturated solution of calcium hydroxide and may for
certain sources of cement exceed this alkalinity [1]. 2 We only need to consider those reactions which take place at a rate sufficient to affect the concrete during its normal life span. Carbonate aggregate reactions may be
beneficial, innocuous, or deleterious.
The most common beneficial reactions are those of enhancement of the
bond of the aggregate to the paste by the epitactic overgrowth of the calcium hydroxide of the cement paste on the calcium carbonate crystals of
limestone and dolostone aggregates [2,3].
The production of rims of discoloration within the aggregate particle
has been associated with concretes showing distress as well as with good
durable concretes. Usually these rims will show a different solubility in
acid than does the center of the aggregate particle. It is now considered
that such rims are generally innocuous and occasionally may indicate a
reaction which strengthens the bond.
The most deleterious of the chemical reactions occurring between carbonate aggregates and cement paste is the expansive dedolomitization
reaction. Dedolomitization can be an innocuous reaction but in specific
kinds of carbonate rocks it can result in expansion of the aggregate particles and destructive expansion of the concrete in which it is used, The
expansive dedolomitization reaction, its occurrence, probable cause, and
suggested means of control will be the major subject of this p a p e r .
IMaterials research petrographer, Virginia Highway Research Council, Charlottesville,
Va. 22903.
Universidade
do Gois pursuant
Agreement.
authorized.
Www.astIII.OI'g
Copyright9
1978tobyLicense
ASTM
International
WALKER ON ALKALI-CARBONATE REACTIONS
723
It is generally considered that kinds of carbonate aggregates which

have been associated with D-cracking [4], distress due to weathered materials [5], and pavement blow-ups [6-8], are not basically chemically
reactive, but rather that their deleterious nature is due to excess porosity,
lack of physical stability, lack of cohesiveness on wetting and drying, lack
of freeze thaw durability, or the presence of expanding clay minerals.
However, as Axon and Lind [9] have stated, "Factors both physical and
chemical which cause deterioration of concrete tend to work concurrently
and not independently. Consequently any factor contributing to concrete
deterioration cannot be ignored. For example a slight amount of alkali
carbonate or alkali silica reaction may, with certain materials greatly increase the rate of deterioration of concrete by frost action."
Expansive Dedolomitization Reaction
Evidences
The expansive dedolomitizafion reaction or expansive alkali-carbonate
reaction, as it is frequently called, is first manifested in the field by many
of the same evidences of expansion that have been found to be associated
with the expansive alkali-silica aggregate reactions [10,11]. These evidences
include the closing of joints and the squeezing out of joint filler materials,
the buckling of adjacent members, the crushing of weaker concretes, and
the presence of the distinctive "pattern" or " m a p " cracking on surfaces.
The pattern cracking is pseudo-hexagonal and to the geologist is reminiscent of the sort of cracking usually associated with mud cracks and columnar jointing, which are shrinkage phenomena. The typical cracks occur
in the surface of the concrete that has not been as affected by the expansion as has the portion of the concrete in contact with the soil or other
moisture source. The cracks may be severe and may extend more than
two thirds of the way through the slab, or they may be visible only when
the surface is moistened with water. The concrete in the surface areas
between the cracks generally retains its integrity and apparently is unaffected by the reaction. Examination of the expanded portions of the concrete yields little positive evidence of the reaction. There are no characteristic reaction products discernible by visual inspections or microscopical examination. There is no clear cut evidence such as the silica
gel in the alkali-silica reaction. Occasionally certain surface cracks may
be traced to a typical carbonate coarse aggregate particle. Expansive carbonate reactions produced in laboratory fabricated mortar bars and concrete bars have shown instances of peripheral cracking and other cracking
within the aggregate particles [12,13]. The cracking of the paste is not
distinctive but is common and does generally connect aggregate particles
to each other or to the surface, or both.
724
The expansion of the aggregate particles in cement paste can be illustrated by the results obtained by treatment of samples of carbonate rocks
in solutions of sodium hydroxide over a period of 1 year as shown in Fig.
1 from Newlon, Ozol, and Sherwood's study of test methods.
Characteristics of the Expansive Carbonates

The expansive alkali-carbonate rocks reaction was reported first by
Swenson in 1957 [10]. He discovered that the severe distress found in the
concrete of the Kingston, Ont., area could be attributed to a specific
argillaceous dolomitic limestone, with low porosity, good physical properties [14], and of Middle Ordivician age [10,15]. The acid insoluble residue
and carbonate analyses of this limestone are shown in Table 1. The deleterious nature of this aggregate could not be detected by any of the ASTM
tests or procedures available at that time. Testing in specimens of concrete (as detailed below) showed that the expansion was greatest with
large size aggregate, in the presence of moisture, at high temperatures,
and, most importantly, with high alkali cements.
The microstructure of the alkali-expansive carbonate rocks is made up
of small rhombs of dolomite, generally under 25 #m in diameter set in
matrixes of calcite micrite, grain size 2 to 6/~m (mostly aphanocrystalline),
with finely disseminated clay material. This texture is illustrated in Fig. 2.
The matrixes of these rocks generally appear dark and murky in standard
thickness (30-#m) petrographic sections. Hand specimens are noted for
fine grain, conchoidal fracture and frequently for dark color.
6.0 ~
. . . . . . . . . . . . .~ . . . ~ ~
,,f~- NOTE C H A N G E
" ~---~
~ IN S C A L E ~
1
~.0 ~
~ 12-9
--UNRESTRAINED
. . . . . . RESTRAINED
/
O~
.
'
9
J
~
0.4
12~9
0.2 ~" /
0
'.
~i
15,9
16
24
32
40
48
WEEKS
FIG. 1--Length changeof typical rock prisms [12].

725
Table l--Composition of the alkali-reactive expansive carbonate rocks.

Acid Insoluble
Residue, %
Dolomite %
of Total
Carbonate
S to 15
10 to 20
about 50
40 to 60
13 to 29
21 to 49
33
46 to 73
75 to 87
over 90
Kingston Early Expanders [10,14,15]

Iowa, Illinois & Indiana Early
Expanders [20,16]
Virginia Early Expanders [17,18]
Gull River Ontario Lake Expanders [15]
Virginia Late Expander [12]
FIG. 2--Typical texture of alkali-reactive carbonate rock [12].
Hadley [16] reported on studies of similar aggregates from Iowa, Illinois, and Indiana. Newlon and Sherwood [17,18] found expansive alkalireactive rocks in Virginia. The typical compositions of these aggregates
are shown in Table 1. It can be shown that the majority of the expansive
alkali-reactive carbonate rocks, whenever they occur, have in common,
at least small areas of the texture described above [10,14,16-18], and shown
in Fig. 3, that is rhombs of dolomite "floating" in a silty and clayey mi-'
crite sized calcite matrix and that those areas have a composition that
includes high amounts of acid-insoluble residue and approximately equal
amounts of calcite and dolomite.
It was noted that the particular composition of the carbonate rock that
was found to be most reactive was uncommon in nature. Swenson and
Gillott [14] theorized that the rare occurrence of carbonate rocks with
nearly equal proportions of calcite and dolomite might indicate a "constitutional instability" which made it possible for these rocks to react with the
726
FIG. 3 - - A vein of reactive texture surrounded by innocuous texture. The rock is alkalireactive [12].
cement alkalies. Steidtmann [19], speaking of the conditions of carbonate

deposition states, " . . . the margin between dolomite and limestone bearing
conditions is a narrow one and hence near the critical stage slight changes
in chemical conditions may completely change the nature of d e p o s i t i o n . . . "
Thus, we can assume that this intermediate composition cannot be expected to occur in uniform lithologies of wide extent but will occur only
within isolated areas of a time zone or facies. Most carbonate rocks perform very well in portland cement concrete. There is only a very limited
type of narrow distribution that will react expansively with the alkalies
which may be found in cements.
Reactions of the Alkali-Expansive Carbonate Rocks

Hadley [16,20] proposed that the following reaction had taken place in
the expanding carbonate rocks exposed to alkali solutions
CaMg(CO3)2 + 2 M O H = Mg(OH)2 + CaCO3 + M2CO3
in which M represents potassium, sodium, or lithium. In concrete, the
727
alkali carbonate produced by this reaction will react with hydration products
of portland cement; for example
Na2CO3 "1- Ca(OH)2 = 2NaOH + CaCO3

This last reaction indicates that the sodium hydroxide could have been
regenerated until all the dolomite had reacted or all the alkali had been
used up in secondary reactions. Seligman [21] has pointed out that the
reaction producing the carbo-aluminate end-product, detected by Hadley
[16], would also regenerate the alkalies and thus further the dedolomitization.
Hansen [22] has theorized that the dolomite is more expansive with
sodium hydroxide than with the other common alkalies because the sodium
ions penetrating the dolomite lattice form sodium carbonate hydrated
with ten molecules of water before the crystal structure disintegrates into
solution; potassium carbonate requires less space because it hydrates with
only two water molecules; lithium carbonate requires still less space because it does not form a hydrate. Hadley found that a large dolomite rhomb
mounted in a strain gage and immersed in 3 M NaOH expanded about
0.15 percent in 100 days.
It was found that the rate of dedolomitization was highest when the
calcite to dolomite ratio was nearly equal to one [11,16,17,22]. All limestones whose composition approached this ratio reacted with alkalies when
the rock was finely ground; only those rocks with the typical fine-grained
texture including dolomite rhombs in an argillaceous calcite micrite reacted to the alkali when in an unground state [23]. However, Lemish and
Moore showed that the same dedolomitization reactions occurred in all
dolomitic rocks when in the presence of alkalies. The dolomitic rocks
which exhibited stable size and those that shrunk also dedolomitized [24].
Buck and Mather [25] found that dedolomitization occurred only with the
expansive rocks and never with the nonexpansive ones.
The reactions and the end products, namely alkali-carbonates, brucite, and calcite, are well documented by X-ray diffraction analyses of
reacted limestones. However, the reactions are a volume by volume replacement and cannot explain the observed final expansion found in the limestones,
in the concrete, or in Hadley's dolomite crystal [16]. Feldman and Sereda
[26] postulated that an expanding water absorbing substance formed in
conjunction with the dedolomitization reaction. Gillott [27] found evidence of water uptake by dedolomitized rock. Later work by Walker [13]
and Buck [28] indicated the formation of layer minerals of the hydrotalcite or sjogrenite groups, or both. In nature, these mineral groups are
related closely to and often occur with brucite. The hydrotalcite-sjogrenite
minerals possibly may help promote the water absorption and expansion.
728
The clay minerals, most particularly illite, are always found in alkaliexpansive limestones. It has been postulated that the clays weaken the
strueture of the rock to the point that the expansion of the included dolomite crystals is not restrained and becomes expressed externally to the
extent that even the surrounding concrete is expanded [11,29]. This theory
must be considered in the light of Hilton's work [30] in which it was shown
that certain porous argillaceous limestones, known to expand in alkali
solutions, did not have sufficient internal textural rigidity to expand under
a restraining force of about the same magnitude as would be found in
concrete paste. His data suggest that only certain limited combinations
of textural properties and dolomite versus calcite content can produce
the requirements for detrimentally expansive carbonate rocks.
Swenson and Gillott noted that the dolomite crystals in the rocks causing
expansion at Kingston are zoned by the inclusion of growth rings containing a cloudy material [31,32]. They postulated that the cloudy material
is a clay. Theorizing that the dolomite crystals were formed under conditions of great pressure, the authors thought that the included clay would
be free of water and in an active state. Once the dolomite is removed or
disrupted by dedolomitization, the clay absorbs water and causes expansion. This process may cause a portion of the expansion in those rocks
in which the cloudy zones appear. Such zones have not been observed in
the dolomite rhombs of the expansive rocks of Virginia.
Sherwood and Newlon [23] showed that powdered samples of the reactive
rocks react with strong concentrations of NaOH and form brucite (MgOH)
and gaylussite (CaNa2(CO3)2 9 5H20). Brucite and gaylussite have been
found at the surface of rock samples which have reacted with NaOH in
certain studies [24]. Other studies [13] have shown that, whereas the brucite can occur within the reacted rocks, the gaylussite can be precipitated
from the solutions in which prisms of reactive carbonate have been stored
for expansion testing.
Swenson and Gillott [27,31] investigated the expansion of powdered
samples of the reactive rocks in alkaline solutions by means of special
compression cells and dilatometers. It was determined that expansions of
more than 7 percent were obtained in some instances, that rate and expansion varied with the particle size, and that the rock material absorbed
water while expanding. The influence of texture was confirmed by the
fact that expansions were greater with the more coarsely crushed samples.
Despite detailed work performed by numerous investigators the exact
mechanism causing the expansion of these dolomitic rocks when exposed
to highly alkaline environment is still to be determined. Any proposed
mechanism must account for the roles of the required texture, the calcitedolomite ratio, the presence of illite, and the results observed with the
various alkalies. It is thought by many that the alteration of the dolomite
729
crystals by the dedolomitization reaction must in carbonate rocks of the

proper texture and composition allow the alkalies greater access to the
interstitial illite as well as to any other included clays.
Test Methods
Older Methods--ASTM Test for Potential Reactivity of Aggregates
(Chemical Method) (C 289) is not appropriate for reactive carbonate rocks
because it was specifically designed to detect the amount of silica reacted
with the cement alkalies. Swenson [IO] found that the m o r t a r bar test
method (ASTM Test for Potential Reactivity of Cement-Aggregate Combinations (Mortar-Bar Method) (C 227)), commonly used to detect the
alkali-silica reaction, was too slow and interpreted improperly to detect
the deleteriously expansive alkali-carbonate reactive rocks. He found that
aggregate which produced definite deterioration in the field produced
only slight expansions in the mortar bars; that is, it produced insufficient
expansion to be classed as reactive by the criteria in use for alkali-silica
reactive aggregates.
Determination of Length Change of Concretes--The practice of using
the length change bars of concrete as a method of test for alkali-carbonate
reactive aggregate was begun by Swenson [10] and later reported by Smith
[33,34], Ryell, et al [35], and Newlon et al [12]. Basically the method tests
aggregate in small, 75 by 75 by 280-mm (3 by 3 by 11-in.) bars of concrete,
using various cements and storing the specimens in a very moist environment. The bars are measured for length change at regular intervals and the
percentage change is calculated. The method is in draft form under the
jurisdiction of ASTM Subcommittee C09.02.02 on Chemical Reactions of
Portland Cement Concrete. The method will be titled "Test for Length
Change of Concrete Due to Alkali Carbonate Rock Reaction." The advantage of this proposed method is that it can use the ingredients in the proportions proposed for structural and for paving concrete to be used in constructions. Aggregates less than 19.0 mm (3A in.) in size can be used in the
same grading as will be used in the final mixture. The method accelerates
the reaction by the type of storage conditions only. The present state of the
draft indicates that the specimens shall be classified as potentially reactive
if the expansion equals or exceeds 0.015 percent at 84 days, 0.025 percent at
180 days, and 0.03 percent at 1 year. The data for the later ages are preferred.
The method used by Smith [33,34] was closer to the method of the Canadian standard CSA A 23.2.24, Test for Alkali Aggregate Reaction. The
Canadian test uses aggregate specifically proportioned into three sizes between 25.0 mm (1 in.) and 4.75 mm (No. 4 sieve) and a standard cement.
The alkali content is augmented by the addition of reagent grade sodium
730
hydroxide so that the alkali content of the mixture shall be 1 percent by

weight of the cement with the Na20 equivalent calculated as Na20 +
0.658 K20. Smith [34] indicated that beams which had expanded more
than 0.02 percent at 84 days were considered to contain deleteriously expansive reactive aggregate. This test is commented on by Dolar-Mantuani
[36] as follows: " T h e great advantage of the test is that the aggregate is
tested in the coarse aggregate size embedded in concrete paste, thus is very
similar to the real field conditions of the mix. The use of high alkali cement had the intention to speed up the reaction but the addition of sodium
hydroxide solution may change the reaction because of the greater availability
of alkalies from the hydroxide than would occur in field concrete. From
this point of view, the added alkali is a possible disadvantage."
Despite the possibility that the chemical reaction may be accelerated
when the aggregate is of small size and has more surface area, the expansion has bee'n shown to be accentuated by larger size aggregate by both
Swenson an(: Gillott [14] and Newlon et al [12]. Thus the concrete bar
method of de~ecting expansive alkali-carbonate aggregate reactivity avoids
the possible diminution of the expansive effect of crushing the aggregate to
sand size as would be required for the testing of the aggregate in mortar
bars of 25 by 25 mm (1 by 1 in.) cross section.
The collection of large size aggregate samples, the preparation of a large
number of concrete mixtures, and the long storage required for the bulky
specimens may make this test method prohibitively expensive for use in any
survey of a large number of carbonate aggregate quarries.
Determination of Length Change of Aggregate Rock--Hadley [20] developed a test for the expansion of the aggregate itself which has become
ASTM Test for Potential Alkali Reactivity of Carbonate Rocks for Concrete Aggregates (Rock Cylinder Method) (C 586). In essence, this method
measures the change in length of small (approximately 35 mm long) right
circular cylinders (originally four-sided prisms) of aggregate rock as they
are altered by soaking in individual containers of 1 N sodium hydroxide
solution. Measurements of the length of each specimen of aggregate are
made with a special length comparator at intervals of 1 week for the first
month and every 4 weeks thereafter. The test method states: "Appreciable
expansion should indicate the need for further testing. In the light of current knowledge, it appears that expansions in excess of 0.1 percent are
indicative of chemical reactions and should warrant additional testing preferably in concrete. Usually expansive tendencies are evident after 28 days
of immersion in alkali, however, exceptions to this have been noted.
. . . quantitative predictions of concrete expansion in service solely from
results of this test method are not yet possible (ASTM Test C 586)."
Thus, it is recommended that the rock cylinder method be used only as a
screening method and that quantitative data be obtained from tests in
concrete such as the concrete bar method outlined above or ASTM Test for
731
Length Change of Hardened Cement Mortar and Concrete (C 157) modified by the substitution of moist storage for dry storage.
Petrographic Examination--Newlon et al [12] state that there are no
known exceptions to the following: "(1) All limestones which possess the
characteristic texture and composition as described in ASTM C 294 and
described above and shown in Fig. 18 [reprinted here as Fig. 2] will react
or dedolomitize in an alkaline environment; and (2) all limestones which
dedolomitize in an alkaline environment possess the characteristic texture
and composition." These authors go on to point out that rocks which possess the characteristic texture and composition may be deleteriously expansive when embedded in concrete and that all rocks which have shown
themselves to be deleteriously expansive do possess the characteristic texture and composition over at least small portions of their volume. Thus,
the texture determination can be used as a method of screening out all the
possible potentially deleteriously expansive carbonate aggregates. This
texture is best observed and defined by a trained petrographer examining
standard (or somewhat thinner) petrographic thin sections of the rocks in
question by means of a polarizing microscope with standard equipment.
(See ASTM Recommended Practice for Petrographic Examination of Aggregates for Concrete (C 295).) It has been shown that the most expansive
rocks are those with between 10 and 30 percent acid-insoluble components
and a calcite to dolomite ratio approaching unity. The textural classification of the aggregate is, therefore, often accompanied by an analysis of
the mineral phases by X-ray diffraction, by staining and volumetric analyses of the thin section. Alternatively, various methods of instrumental
analysis of the elemental composition may be used.
The reactive texture frequently may be recognized by a petrographer,
with sufficient experience, by examination of polished or finely ground
surfaces of rock which have been subjected to a light acid etch. Various
dyes may prove useful.
Late Expanders
The exceptions to the ability of the rock cylinder test to indicate alkaliexpansive carbonate rocks within 28 days are the rocks known as late expanders. These rocks were first noted by Dolar-Mantuani [15] during
studies of the Gull River carbonate rocks in which she used the rock cylinder
test. The expansive aggregates studied were grouped into three classes:
1. Minor-expansion group--expansion from 0.10 to 0.39 percent.
2. Early-major-expansion group--expansion of at least 0.40 percent,
showing expansion in 2 to 5 weeks. This group usually reached its maximum expansion in 10 to 40 weeks, after which the length of the cylinders
either remained more or less constant or increased at a very slow rate.
3. Late-major-expansion group--expansion of at least 0.40 percent,
732
starting to expand after 25 weeks (in one case after more than one year)
and continuing to expand for the full testing period [15].
The late expanding rocks are reported to have continued expanding for
five years [37,38] and some of them have expanded for 11 years (personal
communication). Later work by Dolar-Mantuani [37] defined the late
expansive group as those rocks which, when tested by ASTM Test C 586,
the rock cylinder method, have a shrinkage period of at least 10 weeks, of
at least 0.03 percent and a later minimum expansion to 0.20 percent.
These rocks were characterized as having acid-insoluble residues of from 21
to 49 percent and carbonate contents containing 75 to 87 percent dolomite. The
texture was of rhombs of dolomite floating in a clayey, fine-grained ground
mass, but the ground mass was not as fine-grained as in the early expanders and was composed of interlocking dolomite as well as calcite,
clay, and silica minerals.
Newlon et al [12] found only one such aggregate (Fig. 4) among the
22 aggregate samples for a comparison study of the various test methods.
These authors state: "Sample 13-1 was a delayed expander whose expansion was not reflected in the concrete. While intensive studies have been
made of this type of rock by Dolar-Mantuani (1964), these have been
limited to studies of rock without parallel studies of concrete. The single
FIG. 4--Texture of late expanding carbonate rock from Virginia [12].

733
sample in this project obviously does not permit generalization but

it sugt
gests that the delayed expansive tendency in rocks is not manifested in
concrete, at least up to 5 years in service"
In this regard, it should be noted that Smith found no reason to test
concrete prisms more than 84 days after studying the results obtained with
concrete specimens of nearly 7 years of age [34]. The late expanders may
be prevented from expressing their expansion in concrete by the strength of
the more hydrated concrete of ages greater than 10 weeks. If the expansion is sufficiently slow, the concrete may adjust to the changing aggregate by creep and autogenous healing [39]. However, the possibility of
these rocks contributing to the deterioration of concrete at later ages must
not be disregarded entirely.
Rock Variability
The expansion for the wide variety of compositions and for the range of
exansive behaviors found in the aggregate from one lithologic unit, from
one quarry is accounted for by the fact that rocks are heterogeneous in
nature. The conditions of deposition often undergo a continuous and
variable set of changes during the entire history of deposition of the sediments which make up one small piece of rock. Submarine slumping of unconsolidated sediments can mix these deposits. More and greater changes
must be considered as the sample size increases.
Dolomitization is an uneven process. Entire beds of strata may be dolomitized evenly or the changes may take place only along certain cracks,
fissures, bedding surfaces, or stylolites. One portion of a rock may be a
pure aphanoerystalline micrite with no insolubles and another portion may
contain veins and stylolites high in acid insolubles with included dolomite rhombs in a typical reactive texture.
Dolar-Mantuani [15] describes the expansive behavior of a rock cylinder
of variable texture which was tested under ASTM Test C 586, soaked in
1 N sodium hydroxide solution. The cylinder represented a limestone which
contained only 4.1 percent dolomite and only 4.7 percent insoluble residue. The pointed end of the little cylinder contained a stylolite with reactive texture. At 4 weeks the cylinder had expanded close to 0.5 percent.
At this time the stylolitic end broke off. The shortened cylinder was repointed, and testing was resumed. The specimen showed no further expansion. The remainder of the rock was pure lithographic limestone.
The expansion had been produced entirely by the small amount of material
in the narrow stylolite. Thus, the small amount of dolomite and the insoluble residue were arranged in a texture that caused the rock to expand
significantly, even though only a small portion of the length was involved.
Figure 3 shows a similar seam with reactive texture in a thin section of the
Virginia aggregate.
734
The sampling problem is dealt with in detail by Newlon et al [40]. In

part they conclude: "The sampling plan should have two stages. The first,
or preliminary, screening (Stage I) would furnish sufficient information
to establish either the absence or presence of reactive rock within a quarry.
A second, more detailed sampling procedure (Stage II) would be used
where warranted for quarries shown in the Stage I sampling to contain
significant amounts of reactive rock . . . . Sampling is the most complex
aspect of the evaluation of prospective quarry sites or producing quarries
for potential reactivity. This complexity results from the cohabitation and
interbedding of reactive and non reactive rocks in what are designated
as single lithologic units by conventional geologic criteria... Experience
has shown that a minimum of 5 samples and an average of approximately
10 samples per quarry will give a preliminary indication of the amount
of reactive materials present. (Stage I sampling.)" The authors further
explain that the Stage II sampling must be arranged in accordance with
the specific sort of variability that the rocks exhibit in the particular quarry
sampled. In general, they believe that the sampling instructions contained
in ASTM Recommended Practice C 295 closely describe the requirements.
Use and Control of Expansive Carbonate Aggregates

The precautions necessary once an aggregate has been identified as
being potentially deleteriously expansive have been reported by Smith
[33,34], Mather [41], Swenson [42], and Newlon et al [40,43]. In
general these writers are in agreement and they have enumerated the
following procedures.
1. Selective Quarrying--Avoid using a rock determined to be potentially
reactive in the fabrication of portland cement concrete. (Such aggregate
will normally be acceptable for use as a base material or in bituminous
concrete.)
2. Dilution--If it is not economically feasible to avoid the use of the
potentially reactive material, it should be diluted either naturally or by
artificial means until the coarse aggregate contains less than 20 percent
potentially reactive material, or, if both the coarse and the fine are reactive, then less than 15 percent of the the total aggregate. Extremely reactive
materials should be present in less quantity and slightly reactive materials
may somewhat exceed this amount.
3. Smaller Aggregate Size--Although the size of the aggregate does not
slow the chemical reactions involved, the expansion appears to be in direct
proportion to the diameter of the aggregate particles and the smallest
feasible size of the reactive aggregate should be used.
4. Low Alkali Cement--The expansion of these rocks has been shown
to increase with the amount of alkali in the cement. It is often not economically feasible to obtain a cement with such a low alkali content that
735
the most reactive of carbonate aggregate will show no expansion. With the
most reactive rock 0.4 percent alkaline can be sufficient to create expansion and should be used only in conjunction with other control measures.
Low alkali cement may become harder to obtain under environmental control procedures of cement production.
5. Protection from Moisture--Because the reaction takes place only in
the presence of moisture, the potenially reactive rock used with a moderately
low alkali cement should be protected from ground moisture with a vapor
barrier.
6. Design Considerations--If known potentially reactive aggregate is to
be used, the structure or pavement should be designed to accommodate
small amounts of expansion, that is expansion joints, etc., should be used.
Consideration should be given to unsightly cracking that may occur on surfaces exposed to public view.
7. Prohibit Use--If none of control measures 1 through 6 can be applied
in an economically feasible regime that controls the expansion, the use of
the potentially reactive carbonate rock under consideration must be prohibited.
Rim Forming Aggregates

Rims Within the Aggregate
Rims occur within certain carbonate aggregate particles which have
reacted with the surrounding cement paste. They may be observed as dark
discolorations at the very edges of the particles or very slightly inside the
edges. Figure 5 shows this dark discoloration on particles of a reactive
Virginia aggregate after moist storage of a concrete bar fabricated with
high alkali cement containing 0.95 percent Na20 (equivalent). The concrete surface shown has been lapped smooth to enhance detail and contrast. Although rims are thought of generally as a product of a chemical reaction of the carbonate aggregate with the alkalies of the cement, it
sometimes happens that they are more pronounced visually in low alkali
concrete than in high alkali [12,44]. Figure 6 shows the same aggregate
after use in a concrete bar made with cement containing 0.43 percent
Na20 (equivalent), surface finely lapped.
Determination of whether such interior aggregate rims are positive or
negative is accomplished by subjecting a smoothly sawn or smoothly
lapped surface of the concrete to a hydrochloric acid etch. If the discolored rim is wholly or partially less soluble than the interior of the aggregate, it is considered to be a positive rim. Rims which are more soluble
than the interior of the aggregate are classified as being negative. Evidence
of the presence of a rim is destroyed completely on some aggregates by the
action of the acid.
736
FIG. S - - D a r k rims f o r m e d on a Virginia reactive aggregate embedded in a concrete

bar fabricated with high alkali cement and subjected to moist storage [12].
Positive Rims--The first investigation of positive rims was conducted

by Lemish et al [50] and by Bisque and Lemish [29,45,46] who found
dark rims on aggregate particles from the Glory Quarry (Rapid member,
Cedar Valley formation, Devonian age, Iowa) which had been used in concretes showing premature distress. When samples of the concretes were
sawn and etched, it was found that the dark rim was less soluble than
either the interior of the aggregate or a very narrow layer of the aggregate
at the contact with the cement paste. The dark, acid-resistant rims had
developed on the aggregate while it had been embedded in concrete. These
changes led the authors to conclude that a chemical reaction and accompanying transfer of material had occurred. The rims were duplicated
on Glory aggregate in the laboratory in a silica solution at pH 12. Various
chemical experiments and analyses suggested that the rims may have become enriched in silica. It is clear from the work of Bisque and Lemish
[29,46] and Lemish [47] that dedolomitization accompanied the rimforming process. Illitic clay was found to be a necessary mineralogical constituent of the rim-forming carbonate rocks [45]. The source of silica for
the enrichment of the rim was shown not to be the quartz of the fine aggregate [46], but was variously thought to be either the paste itself [29,46],
737
FIG. 6 - - W i d e rims f o r m e d on a Virginia Reactive aggregate (same aggregate as in Fig. 5)

embedded in a concrete bar fabricated with low alkali cement and subjected to moist storage [12]
or the well distributed particles of silica minerals in the carbonate coarse

aggregate [48 ].
It was shown that the growth of the positive reaction shells was not accompanied by excessive expansion of test bars of concrete [45], but testing
did indicate that the Glory aggregate [45] and Jeffersonville, Ind., aggregate [48] were associated with the retardation of strength development which normally accompanies the continuing hydration of concrete as
it matures. It is this lack of strength development which has been thought
to be the cause of the deterioration specifically associated with the Glory
aggregate, but which has not been thought to be associated with all positive
rim-forming aggregates. Some of the rim-forming aggregates were earlier
found by Sweet [49] to be highly susceptible to frost damage. It is not
known if there is any correlation between the susceptibility to frost and the
rim-forming process.
The lithology of the Rapid member typically possessed high, very fine
(0.1 mm) porosity, a texture of dolomite crystals set in a matrix of medium
to coarsely crystalline (1.0 to 0.1 mm) calcite, insoluble residue of 10 to 14
percent, and a high loss under freeze-thaw testing [50,51].
Negative RimsmNegative .reaction rims were first shown by Mather
738
et al [52]. They sometimes occurred in acid etched concrete samples

as negative areas surrounding less soluble aggregate centers. Sherwood and
Newlon tested 224 samples from 43 quarries for rim formation by embedding cubes of the test rock in mortar bars which were then stored over
water for 8 weeks [53]. (See also Hadley [21].) Sherwood and Newlon [53]
reported: "Evidence of rims was found in 25 of the 224 cubes. All but two
of these were negative rims. The composition of the rim-forming rocks and
the type of rims formed can be seen in Fig. 9. [reprinted here as Fig. 7].
Figure 10 [reprinted here as Fig. 8] shows a negative and a positive rim
from this group . . . . (a) the negative rims were virtually limited to high
calcite rocks, (b) the insoluble residue contents are generally low, and
(c) the limited evidence in the case of positive rims points to compositions
high in dolomite. A check for rim vs expansion showed that six of the
25 rimmed rocks also expanded, and that two of these reached expansions
in excess of 0.1 percent after six months. Because the percentage of the rim
growing rocks that expanded is about the same as the percentage of all the
samples that expanded, no relationship between rim growth and expansion
appears to exist. It should be noted also that no effects, either deleterious
or beneficial, have been proven attributable to rim growth at this time."
Further investigations by Buck [44] of the aggregates observed to have
negative reaction rims in the Milford D a m investigations [52] confirmed
the fact that this type of reaction occurs in high calcite rocks. The reaction
does not seem to cause any kind of deterioration. In fact, Buck thought
that the reaction might take place in a wide variety of carbonate aggregates
and might be responsible for the observed increase with time of the strength
of the bond between carbonate aggregate and concrete paste.
Exterior Rims
Discolorations of the mortar surrounding aggregate particles occur in
some cases [50,52,54]. These may take the form of lobes of carbonation
or areas low in calcium hydroxide extending out into the paste. They seem
to be associated generally with aggregates which form positive rims. These
mortar rims or lobes are not known to be indicative of any particular
deleterious reaction with carbonate aggregates.
Synopsis
The chemical reactions which take place between carbonate aggregates
and portland cement paste are many and varied. To date there is only one
type of chemical reaction that is known generally to cause premature
deterioration of concrete pavements and structures. This is the expansive
alkali-dedolomitization reaction. The reaction is most active in high alkali
cements and occurs expansively with only a very restricted type of rock.
LEGEND
INSOLUBLE
RESIDUE
/
k
o ClearlyDefined
/ ~
Negative Rim
L~ Passible
/
/
Negative Rim
* Clearly Defined
Positive Rim
9 Passible
739
\
\
Po
DOLOMITE
CALCITE
FIG. 7--Composition of rim-forming rocks [18].
The rocks contain nearly equal amounts of calcite and dolomite, a substantial amount of illite, and possess a texture of dolomite rhombs in
"dirty" micrite. Rocks of this composition can occur in any of the carbonate sequences. They have most often been found in Ordovician formations. The expansive alkali-reactive carbonate rocks occur in relatively
small quantities in limited areas or time zones. Thus, selective quarrying is
often a preferred method of dealing with them when they occur.
P e t r o g r a p h i c examination (ASTM Recommended Practice C 295) and
the rock cylinder method (ASTM Test C 586) are valuable screening procedures to detect aggregate materials that are potentially susceptible to
the alkali-carbonate rock reaction. A test method for length change of concrete prisms being developed by ASTM Subcommittee C09.02.02 appears to offer the most dependable means of evaluating the susceptibility of
cement-aggregate combinations to expansive effects of this reaction.
References
[1] Hansen, W. C., "Chemical Reactions," Significance of Tests and Properties of Concrete
and Concrete Making Materials, A S T M STP 169A, American Society for Testing and
Materials, 1960.
[2] Lerch, William, "Chemical Reactions," Significance of Tests and Properties of Concrete
and Concrete Making Materials, A S T M STP 169, American Society for Testing and
Materials, 1956.
[3] Mather, Katharine and Mielenz, R. C., "Cooperative Examination of Cores from the
McPherson Test Road," Proceedings, Highway Research Board, Vol. 39, 1960,
pp. 205-216.
740
FIG. 8--Close-up photos o f sawed cubes comparing acid etched and non-etched halves.
Part (A ) shows a negative rim and part (B) a positive rim [18].
741
[4] Bukovatz, J. E., Crumpton, C. F., and Worley, H. E., "Kansas Concrete Pavement
Performance as Related to D-Cracking," Cement Aggregate Reactions, Record No. 525,
Transportation Research Board, 1974, pp. 1-8.
[5] Roy, C. J., Thomas, L. A., Weissman, R. C., and Schneider, R. C., "Geologic Factors
Related to Quality of Limestone Aggregates," Proceedings, Highway Research Board,
Vol. 34, 1955, pp. 400-411.
[6] Moyer, R. A., "An Eleven-Year Study of the Expansion and Contraction of a Section
of Concrete Pavement," Proceedings, Highway Research Board, Vol. 25, 1945, pp. 71-81.
[7] Woods, K. B., Sweet, H. S., and Shelburne, T. E., "Pavement Blowups Correlated with
Source of Coarse Aggregate," Proceedings, Highway Research Board, Vol. 25, 1945,
pp. 147-168.
[8] Sweet, H. S. and Woods, K. B., "Evaluation of Aggregate Preformance in Pavement
Concrete," Journal, American Concrete Institute, Vol. 19, No. 10, June 1948, pp. 10331040.
[9] Axon, E. O. and Lind, Junior, "Alkali-Carbonate ReactivitymAn Academic or a
Practical Problem," Symposium on Alkali-Carbonate Rock Reactions, Record No. 45,
Highway Research Board, 1964, pp. 114-125,
[10] Swenson, E. G., "A Reactive Aggregate Undetected by ASTM Tests," Bulletin No. 57,
[1l] Hadley, D. W., "Alkali Reactivity of Dolomitic Carbonate Rocks," Symposium on
Alkali-Carbonate Rock Reactions, Record No. 45, Highway Research Board, 1964,
pp. 1-19.
[12] Newlon, H. H., Ozol, Michael, and Sherwood, W. C., "Potentially Reactive Carbonate
Rocks," An Evaluation of Several Methods for Detecting Alkali-Carbonate Reaction,
Progress Report No. 5, Virginia Highway Research Council, 71-R33, May 1972.
[13] Walker, H. N., "Reaction Products in Expansion Test Specimens of Carbonate Aggregates," Cement-Aggregate Reactions, Record No. 525, Transportation Research
Board, 1974, pp. 28-37.
[14] Swenson, E. G. and Gillott, J. E., "Characteristics of Kingston Carbonate Rock Reaction," Concrete Quality Control, Aggregate Characteristics and the Cement Aggregate
Reaction, Bulletin No. 275, Highway Research Board, 1960, pp. 18-31.
[15] Dolar-Mantuani, L. M. M., "Expansion of Gull River Carbonate Rocks in Sodium Hydroxide," Symposium on Alkali-Carbonate Rock Reactions, Record No. 45, Highway
[16] Hadley, D. W., "Alkali Reactivity of Carbonate Rocks--Expansion and Dedolomitization," Proceedings, Highway Research Board, Vol. 40, 1961, pp. 462-474.
[17] Newlon, H. H. and Sherwood, W. C., "An Occurrence of Alkali-Reactive Carbonate
Rock in Virginia," Carbonate Aggregate and Steam Curing of Concrete, Bulletin
No. 355, Highway Research Board, 1962, pp. 27-44.
[18] Sherwood, W. C. and Newlon, H. H., Jr., "A Survey for Reactive Carbonate Aggregates
in Virginia," Record No. 45, Highway Research Board, 1964, p. 222.
[19] Steidtmann, Edward, "The Origin of Dolomite as Disclosed by Stains and Other
Methods," Bulletin, American Geological Society, Vol. 28, June 1917, pp. 431-450.
[20] Hadley, D. W., "Alkali-Reactive Carbonate Rocks in Indiana--A Pilot Regional Investigation," Symposium on Alkali-Carbonate Rock Reactions, Record No. 45, Highway
[21] Seligmann, Paul, "General Discussion," Symposium on Alkali-Carbonate Rock Reactions, Record No. 45, Highway Research Board, 1964, p. 111.
[22] Hansen, W. C., "General Discussion," Symposium on Alkali-Carbonate Rock Reactions,
[23] Sherwood, W. C. and Newlon, H. H., Jr., "Studies on the Mechanisms of AlkaliCarbonate Reactions--Part I, Chemical Reactions," Symposium on Alkali-Carbonate
Rock Reactions, Record No. 45, Highway Research Board, 1964, pp. 41-56.
[24] Lemish, John and Moore, W. J., "Carbonate Aggregate Reactions--Recent Study and
an Approach to the Problem," Symposium on Alkali-Carbonate Rock Reactions, Record
[25] Buck, A. D. and Matber, K., "Brucite Formation in Carbonate Rock Prisms," Miscellaneous Paper No. 6-867, U.S. Army Engineer Waterways Experiment Station,
Corps of Engineers, Vicksburg, Miss., Jan. 1967.
742
[26] Feldman, R. F. and Sereda, P. J., "Characteristics of Sorption and Expansion Isotherms
of Reaction Limestone Aggregate," Journal, Vol. 58, No. 2, Aug. 1961, pp. 203-214.
[27] Gillott, J. E., "Mechanism and Kinetics of Expansion in the Alkali-Carbonate Rock
Reaction," Canadian Journal of Earth Sciences, Vol. 1, No. 2, 1964, pp. 121-145.
[28] Buck, A. D., "Control of Reactive Carbonate Rocks in Concrete," Technical Report
C-75-3, U.S. Army Engineer Waterways Experiment Station, Corps of Engineers, 1975.
[29] Bisque, R. E. and Lemish, John, "Chemical Characteristics of Some Carbonate
Aggregate as Related to Durability of Concrete," Air Voids in Concrete and Characteristics of Aggregate, Bulletin No. 196, Highway Research Board, 1958, pp. 29-45.
[30] Hilton, M. H., "Expansion of Reactive Carbonate Rocks Under Restraint," CementAggregate Reactions, Record No. 525, Transportation Research Board, 1974, pp. 9-22.
[31] Swenson, E. G. and Gillott, J, E., "Alkali-Carbonate Rock Reaction," Cement Aggregate Reactionsl Record No. 525, Transportation Research Board, 1974, pp. 21-40.
[32] Swenson, E. G. and Gillott, J. E., "Alkali Reactivity of Dolomitic Limestone Aggregate," Magazine of Concrete Research, Vol. 19, No. 59, June 1967, pp. 95-104.
[33] Smith, P., "Learning to Live with a Reactive Carbonate Rock," Symposium on AlkaliCarbonate Rock Reactions, Record No. 45, Highway Research Board, 1964, pp. 23-27.
[34] Smith, P., "15 Years of Living at Kingston with a Reactive Carbonate Rock," Cement
Aggregate Reactions, Record No. 525, Transportation Research Board, 1974, pp. 23-27.
[35] Ryell, J., Chojnacki, B., Woda, G., and Koniuszy, Z. D., "The Uhthoff Quarry AlkaliCarbonate Rock Reaction--A Laboratory and Field Performance Study," Cement
Aggregate Reactions, Record No. 525, Transportation Research Board, 1974, pp. 43-54.
[36] Dolar-Mantuani, L. M. M., Draft of "Manual of Petrographic Analysis of Concrete
Aggregates," unpublished.
[37] Dolar-Mantuani, L. M. M., "Late-Expansion Alkali-Reactive Carbonate Rocks,"
Mineral Aggregates, Record No. 353, Highway Research Board, 1971, pp. 1-14.
[38] Dolar-Mantuani, L. M. M., "Alkali-Silica-Reactive Rocks in the Canadian Shield,"
Portland Cement Concrete, Record No. 268, Highway Research Board, 1969, pp. 99-117.
[39] Gillott, J. D. and Swenson, E. G., "The Mechanism of the Alkali-Carbonate Rocks
Reaction," Quarterly Journal of Geology, Vol. 2, pp. 7-23.
[40] Newlon, H. H., Sherwood, W. C. and Ozol, M. A., "Potentially Reactive Carbonate
Rocks," A Strategy for Use and Control of Potentially Reactive Carbonate Rocks,
including an Annotated Bibliography of Virginia Research, Progress Report No. 8,
Virginia Highway Research Council, 71-R41, 1971.
[41] Mather, Bryant, "Development in Specification and Control," Cement Aggregate Reactions, Record No, 525, Transportation Research Board, 1974, pp. 38-42.
[42] Swenson, E. G., "Interaction of Concrete Aggregates and Portland Cement--Situation
in Canada," Engineering Journal, Vol. 55, No. 5, 1972, pp. 34-39,
[43] Newlon, H. H. and Sherwood, W. C., "Methods for Reducing Expansion of Concrete
Caused by Alkali-Carbonate Rock Reaction," Symposium on Alkali-Carbonate Rock
Reactions, Record No. 45, Highway Research Board, 1964, pp. 134-150.
[44] Buck, A. D., "Investigation of a Reaction Involving Non-Dolomitic Limestone Aggregate in Concrete," Miscellaneous Paper No. 6-724, U.S. Army Engineer Waterways
Experiment Station Corps of Engineers, 1965.
[45] Bisque, R. E. and Lemish, John, "Effect of Illitic Clay on Chemical Stability of Carbonate Aggregates," Concrete Qualiy Control, Aggregate Characteristics and the Cement
Aggregate Reaction, Bulletin No. 275, Highway Research Board, 1960, pp. 32-38.
[46] Bisque, R. E. and Lemish, John, "Silification of Carbonate Aggregates in Concrete,"
Physical and Chemical Properties of Cement and Aggregate in Concrete, Bulletin
[47] Lemish, John, "Research on Carbonate Aggregate Reactions in Concrete," Transactions, American Institute of Mining, Metallurgical, and Petroleum Engineers,
Vol. 220, 1961, pp. 195-198.
[48] Lemish, John, Harwood, R. J., Hiltrop, C. L., and Werner, M, A., "Compositional
Variations Associated with Carbonate Aggregate Reactions," Properties of Concrete,
Record No. 3, Highway Research Board, 1963, pp. 1-8.
[49] Sweet, H. S., "Research on Concrete Durability as Affected by Coarse Aggregate," Proceedings, American Society for Testing and Materials, Vol. 48, 1948, pp. 988-1017.
743
[50] Lemish, John, Rush, F. E., and Hiltrop, C. L., "Relationship of Physical Properties of
Some Iowa Aggregates to Durability of Concrete," Air Voids in. Concrete and Characteristics of Aggregate, Bulletin No. 196, Highway Research Board, 1958, pp. 1-16.
[51] Hiltrop, C. L. and Lemish, John, "Relationship of Pore-Size Distribution and Other
Rock Properties to Serviceability of Some Carbonate Aggregates," Physical and Chemical
Properties of Cement and Aggregate in Concrete, Bulletin No. 239, Highway Research
Board, 1960, pp. 1-23,
[52] Mather, K., Luke, W. I., and Mather, B., "Aggregate Investigations, Milford Dam,
Kansas--Examination of Cores from Concrete Structures," Technical Report No. 6-629,
U.S. Army Engineer Waterways Experiment Station Corps of Engineers, June 1963.
[53] Sherwood, W. C. and Newlon, H. H., "A Survey for Reactive Carbonate Aggregates
in Virginia," Symposium on Alkali-Carbonate Rock Reactions, Record No. 45, Highway
[54] Mather, Katherine, Buck, A. D., and Luke, W. I., "Alkali-Silica and Alkali-Carbonate
Reactivity of Some Aggregates from South Dakota, Kansas, and Missouri," Symposium
on Alkali-Carbonate Rock Reactions, Record No. 45, Highway Research Board, 1964,
pp. 72-109.
L. D o l a r - M a n t u a n i
Chapter 42--Soundness and

Deleterious Substances
Introduction
During the past 100 or more years many tests have been developed and
applied to assess the quality of concrete aggregates. This paper is a discussion of the significance of five aggregate quality tests currently in use.
One of them, the sulfate soundness test (ASTM Test for Soundness of
Aggregates by Use of Sodium Sulfate or Magnesium Sulfate (C 88)), is
perhaps the most widely used of all methods for determining the overall
quality of aggregate. The other four tests are used to determine the presence
of specific harmful particles or substances which influence the mix proportions of fresh concrete or its early stage hardening, or which damage the
concrete surface under specific circumstances.
Nomenclature
A terminology has grown up over the years for describing the assessment
of aggregate quality. As is common with most terminologies that grow with
time, not all authors use the same terms with the same meaning. The two
terms used in association with the materials and tests discussed here are
"soundness" and "deleterious" and different authors have assigned different meaning to these terms.
The sulfate soundness test is designed to predict the overall performance
of an aggregate in concrete. Bloem [112 used "soundness" as a general
term to describe an aggregate as a whole and uses "deleterious substances"
to describe individual particles or contaminants that are harmful to concrete. The use of the terms "soundness" and "deleterious" in this way causes
some difficulties. In some literature on concrete technology "sound" and
"unsound" are used to describe chemical properties only. Consideration
IConsultant petrologist, Toronto, Ontario, Canada M6M 4Z2.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
DOLAR-MANTUANI ON SOUNDNESS AND DELETERIOUS SUBSTANCES
745
should be given to omitting the term soundness from the ASTM C 88 sulfate
test which is identified adequately by the word sulfate. Actually, the test
measures the unsound and not the sound aggregate constituents.
It is not easy to find agreement on the meaning of "deleterious." Bloem
used it to mean specific, mainly physically undesirable, characteristics of
an aggregate. Petrographers usually use deleterious, and its opposite innocuous, to refer to chemical properties of individual particles. (See Rhoades
and Mielenz [2] and for wider definitions see Mielenz [3,4] and in the
present publication.) A detailed classification of harmful substances, based
on published literature, was proposed by Swenson and Chaly [5]. They use
"deleterious" to mean any quality of an aggregate that is harmful to concrete. In this paper, the term "harmful" will be used to mean detrimental
to concrete in the broad sense.
Sulfate Test
Although individual particles behave differently in concrete, it is desirable
to have general criteria for the acceptance of an aggregate. It is convenient
to test representative aggregate samples to obtain data that show a meaningful relationship between the test results and the quality of concrete
made from the aggregate. No satisfactory single test accomplishes this.
The most frequently used method for measuring overall aggregate quality
and its performance in concrete is ASTM Test C 88. This test was first
published in 1931 as a tentative ASTM method and was approved as a
standard method in 1963. The scope is given as follows: "This method
covers the testing of aggregates to determine their resistance to disintegration
by saturated solutions of sodium sulfate or magnesium sulfate. It furnishes
information helpful in judging the soundness of aggregates subject to
weathering action, particularly when adequate information is not available
from service records of the material exposed to actual weathering conditions."
The sulfate test is carried out by soaking a sample of each fraction in
saturated sodium or magnesium sulfate solution and then oven drying.
Magnesium sulfate generally is more destructive than the sodium salt. The
growth of salt crystals in the pores of the particles was thought originally
to produce disruptive internal forces similar to the action of freezing in
water. After 5 or 10 cycles of soaking and drying, the amount of the material that has been broken or disintegrated is determined by weighing
the coarse aggregate fractions retained on sieves with openings five sixths
of the original ones. The fine aggregate is sieved on the original sieves.
The sulfate loss is expressed as the weighted sum of the difference between
the first and second weighing of each fraction.
T h e physical damage produced during the test includes crumbling,
flaking, splitting, cracking, and granular disintegration of the aggregate
746
particles. To investigate the damage in more detail, after each immersion,

the fraction retained on a 19.0-mm (3A-in.) sieve is examined visually with
or without a lens. After the test the damaged particles are counted and the
smaller size fractions of coarse aggregate are examined to determine if
there is excessive particle splitting. Shaly and other weak, flat particles
may split, producing thin flat particles of approximately the same sieve
size as the original particles before testing. The weight of such physically
damaged particles probably should be added to the loss. Since the aggregate
is washed before testing, the fines and soluble materials are not included
in the loss (see ASTM Test for Materials Finer than No. 200 (75-/~m) Sieve
in Mineral Aggregates by Washing (C 117)).
A petrographic examination of the coarse aggregate before and after the
test gives information about the weakness and strength of the rock varieties
in the aggregate. An examination of the "loss" passing the after-test sieves
and retained in the pan of each fraction may help to characterize the objectionable aggregate, especially the shales and other argillaceous rocks [6].
Interpretation of Sulfate Tests

Although the scope of the method refers to weathering, with no mention
of frost action, the test results usually are interpreted in terms of the behavior of aggregates exposed to frost action.
Explanations of the mechanism by which the sulfate test disrupts rock
particles have been given by many authors but there is no general agreement [7-11]. Some research workers [8] have published data on the properties of the sulfate solutions and observations of the behavior of the sodium
and magnesium salts as they pass through various stages of crystallization
and dehydration. It is generally agreed that alternative soaking in sulfate
solution and oven drying causes salt to accumulate in the pores and cracks
in the rocks. At some number of alternations, the quantity of salt reaches
an amount which on rehydration during the next soaking cannot be accommodated in the pores of some particles and exerts disruptive pressures.
Bloem discussed the properties of the two salt solutions and pointed out
that the magnesium solution is more aggressive than the sodium solution
because there is a greater weight and volume of magnesium salts than
sodium salts in solution.
There is evidence that destructive forces other than the pressure produced
by crystal growth must be involved in the sulfate test. Wuerpel [I0] found
the loss to be approximately proportional to the number of cycles. Appreciable losses have occurred when the test was run with distilled water instead
of sulfate solution, indicating that simple wetting and drying or heating
and cooling disrupts the particles [8]. Hudec and Rogers [12] found that
destruction in the sulfate test was related to the quantity of water adsorbed
by carbonate rocks. These results indicate that wetting and drying is also a
747
destructive force in the sulfate test. Finally, the severity of the test increases
when the drying time is extended beyond that needed to dehydrate the
crystals, indicating damage from prolonged heating.
During the early development of the sulfate test, it was suggested that
the loss depends more on the number of pores than on their size and distribution [13-19]. The damage to concrete due to freezing depends on the
pore structure, elastic properties, size, and water saturation of the aggregate
particles [12-22] as well as the properties of the other ingredients, their
proportions, and the severity of the exposure, to mention a few factors.
Details of frost action are discussed elsewhere in this publication. Various
authors have emphasized that there is little or no theoretical or experimental
support for the assumption that the sulfate test simulates exposure to
freezing and thawing in concrete or gives a reliable indication of field performance [17,13,14,23]. On the other hand, statistical analysis by Vollick
and SkiUman [24] based on 70 samples did indicate a correlation between
the freeze-thaw durability of air-entrained concrete and the soundness of
coarse aggregate as measured by the sodium sulfate test. However, the
wide scatter obtained shows that the durability is influenced by other factors
that are not measured by the sulfate test. Quoting selected papers, starting
with the first description of sulfate solution used for aggregates by Brard
(1829), Mather [25] concludes "that many fundamental aspects of the
physical and chemical effects of the phenomena associated with the formation of hydrated sulfates in aggregates are still poorly understood." This is
still true today.
Chamberlin a has pointed out repeatedly that the fine aggregate particle
size is smaller than the critical size for frost damage in most materials and
therefore sulfate test results on fine aggregates require an entirely different
interpretation than the results on coarse aggregates. From examples of his
and other's investigations it can be concluded that high magnesium sulfate
losses for fine aggregates correlate with properties such as low density, high
absorption and attrition, low resistance to abrasion, and low modulus of
elasticity [17,26-29]. These test results agree with the general petrographic
findings that deeply altered and in situ weathered rocks consisting of detrimental constituents having a physically poor texture, break down on further
weathering into sand fractions which are compositionally or texturally
detrimental, or both, to concrete. This usually is reflected in the coarser
sand fractions by the presence of poor rock particles and in the finer fractions by concentrations of harmful mineral grains.
After a critical review of the literature treating methods of identifying
aggregates subject to destructive volume change when frozen in concrete,
Larsen et al [30] conclude that the sulfate test is too sensitive to test variables and does not measure or reflect the susceptibility of aggregates to
aChamberlin, W. P., personal communication.
748
damage by freezing and thawing. They recommend therefore that "no

further research effort should be devoted to sodium or magnesium sulfate
soundness tests and these should be deleted from specifications as soon as
suitable replacements are available." However, it seems rather unwise to
eliminate these tests, which are probably the most widely used for concrete
aggregates and which have been used for most major constructions on this
continent.
It is the author's opinion that the test does indicate weaknesses in aggregates. If the specification limits are used carefully together with sound
engineering judgment, the test is certainly useful because of the relatively
short time needed to perform it. The value of the test may be enhanced
by petrographic examination of the aggregate after testing to identify better
the weaknesses in the particles.
Reproducibility of Sulfate Tests

After various researchers [8-10,15,31] showed that many details of the
equipment and technique affect the test results, the equipment and procedures were defined more closely to give better reproducibility. Several
suggestions by Wuerpel [11], Paul [31] and Woolf [32,33] for changes were
incorporated in the reviews of the test method; the last was as recently as
1973 in the Annual Book of A S T M Standards. These changes improved
the reproducibility both within and between laboratories. But the failure
in many cases to obtain results of even comparable magnitude probably
stems from three causes: (1) inability to prepare truly representative specimens especially for the coarser fractions of lithologically complex aggregates [34]; (2) differences in following the procedures, such as extra drying
to obtain constant weight; and (3) deviations from the prescribed method,
such as shortening the initial or final drying period before weighing.
A great principal difficulty in obtaining comparable results is the number of different alternative procedures permitted in the method. Either
sodium or magnesium solutions can be used and either 5 or 10 immersion
cycles.
Trends associated with the various optional procedures have been identified [35]. Although they do not permit translation of results obtained with
one type of salt or number of cycles to another, they do indicate approximate degrees of severity in a statistical sense. For fine aggregates, magnesium sulfate is consistently more severe than sodium sulfate, producing
losses about twice as great. For coarse aggregate, magnesium sulfate is
usually the more severe but the difference is smaller, and for many individual aggregates, may be reversed. Losses at 10 cycles average about 50
percent higher than at five cycles for either type of salt and for both coarse
and fine aggregate. The 1976 Annual Book of A S T M Standards seems to
749
favor the magnesium over the sodium solution. This is probably because
the magnesium sulfate solubility is less sensitive to temperature changes
than is the solubility of sodium sulfate [11.31]. Laboratories using sodium
sulfate are hesitant to switch to magnesium because it would reduce the
value of accumulated results, particularly where earlier results from the
same deposit are used for comparison.
It has been suggested that the sulfate test might be used to accept aggregates but not to reject them [33], the assumption being that aggregates
that withstand the test are good while those that break down may or may
not be bad. Unfortunately, the test is not reliable even to that extent.
Certain aggregates with an extremely fine pore structure show almost no
loss in the sulfate test but disrupt concrete when frozen in a saturated condition. Conversely, certain highly porous aggregates disintegrate in the
sulfate test but produce concrete with good resistance to freezing and thawing. Vollick and Skillman [24] report that durable concrete was produced
with aggregates having as much as 32 percent loss in the sodium sulfate
solution. A low sulfate test is usually, but not always, evidence of good
durability, whereas a high loss places the aggregate in a questionable category until performance data become available.
A recommendation that sulfate testing be simplified by testing one
fraction each of the coarse and fine aggregates was not accepted for ASTM
Test C 88. It was rejected because no one size fraction adequately represents
the quality of a graded aggregate. This is evident from petrographic examination of numerous samples [37]. Conversely, the N.Y. State Department
of Transportation tests the 25.4 to 12.7-mm (1 to 1/2-in.) fraction only for
acceptance and as an indication of uniformity of aggregate supplies. This
simplified test (NYSDOT 208 A-76) is considered very helpful in the dayto-day operation of the state transportation department. 4
Specification Limits
The lack of correlation between the sulfate loss and the performance of
aggregates in field or laboratory concrete, militates against setting inflexible
acceptance limits for the sulfate loss and disregarding judgment in interpreting the results [1]. In ASTM Specification for Concrete Aggregates
(C 33), limits are specified for magnesium and sodium sulfate losses for
fine aggregates. For coarse aggregates the sulfate test can be omitted for
certain types of concrete and for concrete placed in certain locations. The
specifications also contains escape clauses for certain cases in which the
aggregate fails to meet the requirements.
4Toung, G. D., personal communication.
750
Precision Statement
Precision indexes, based on the latest data supplied by ASTM Subcommittee C09.03.05 on Methods of Testing and Specifications for Physical
Characteristics of Concrete Aggregates (19 June 1974): for coarse aggregates
having weighted average sulfate losses in a range of 6 to 16 percent for
sodium and 9 to 20 percent for magnesium sulfate are given in Table 1.
Available data thus still discourage any attempt to estimate the laboratoryto-laboratory reproducibility of sulfate test results.
Bloem's paper [1] includes within-laboratory-single-operator data on
5-cycle sodium and magnesium sulfate tests made in triplicate on 36 fine
aggregates and 56 coarse aggregates [38]. The results are given in Table 2.
To the extent that these data apply generally, which assumes meticulous
attention to all testing detail, duplicate determinations by one operator
within one laboratory should be expected to yield numerical losses checking within about 4 for coarse aggregate and 1.5 for fine aggregate 95 percent of the time.
Other Tests for Soundness
Absorption Test
Because of the unsatisfactory correlation between the sulfate test results
and the overall performance of aggregates in concrete, it has been suggested
that the absorption test, which measures pore volume directly, gives better
correlation with the resistance of concrete to laboratory freezing and thawing than the sulfate test [18,19].
Unconfined Freeze- Thaw Test

To replace the sulfate test with the freeze-thaw test on unconfined aggregates seemed to be a logical change to obtain values that more closely
indicate the susceptibility of aggregates to frost action. Over the years,
many attempts have been made to improve the correlation between the
laboratory freeze-thaw test results and the performance of aggregates in
concrete. Variables investigated include sample type and size, pretest treatment, rate and conditions of freezing and thawing, different numbers of
cycles, and the addition of a small volume of alcohol, usually 0.5 percent
[39], to the water. Various investigators have reported a general correlation
between aggregate freeze-thaw resistance and other tests including absorption and sulfate loss. Some have expressed confidence in the ability of these
tests to predict field performance. For instance, the N.Y. Statement Department of Highways successfully uses a laboratory freeze-thaw test
(NYSDOT 208) or the magnesium sulfate test (NYSDOT 207) for acceptance
751
TABLE ] --Precision statement for sulfate losses:a
Multilaboratory
sodium sulfate
magnesium sulfate
Single operator
sodium sulfate
magnesium sulfate
Maximum Coefficient of
Variation (1S %),
max, %
Maximum Difference
between two Tests
(D2S%), b percent of avg
43
35
131
99
23
I1
64
30
~These numbers represent, respectively, the 1S% maximum and D2S% maximum limits as
described in ASTM Recommended Practice for Preparing Precision Statements for Test
Methods for Construction Materials (C 670).
bD2S is the difference 2-sigma limit: difference equalled or exceeded in only S percent of
cases.
TABLE 2--Percentage of coefficient or variation.
Type of Test
Coarse Aggregate"
Fine Aggregateb
14.5
12.4
3.9
6.0
Five-cycle magnesium sulfate

Five-cycle sodium sulfate
"Fixed grading, 25-ram (I-in.) maximum size.

bTypical concrete gradings; loss weighted in accordance with individual grading.
of aggregates. However, a g g r e g a t e rejection m a y require showing, by a
p e t r o g r a p h i c identification, t h a t failed particles have a known unsatisfactory
p e r f o r m a n c e record in concrete structures (see footnote 4). A l t h o u g h the
results from the freeze-thaw test correlate only poorly in the statistical
sense with the results from the New York State or A S T M M e t h o d C 88
m a g n e s i u m sulfate tests, all t h r e e p r o c e d u r e s (with their a c c e p t a n c e criteria)
tend to accept or reject the s a m e m a t e r i a l s (see footnote 3).
Verbeck [20] a n d m a n y others have p o i n t e d out t h a t the destructive
effects of particles freezing in concrete c a n n o t be s i m u l a t e d by freezing
aggregate in an u n c o n f i n e d state. This is relevant for coarse aggregate
fractions but, as m e n t i o n e d above, C h a m b e r l i n points out t h a t these mechanisms generally do not a p p l y to the small particles of fine aggregates. His
thesis a b o u t the different position of fine a g g r e g a t e c o m p a r e d to t h a t of
coarse aggregate suggests t h a t the physical a n d other p r o p e r t i e s of m a n y
n a t u r a l sands have a high degree of m u t u a l correlation [26].
C o n f i n e d Freeze- T h a w Tests
F r e e z e - t h a w tests on aggregates in the confined state in concrete is the

most satisfactory m e t h o d for p r e d i c t i n g the field p e r f o r m a n c e o f concrete
aggregates. These m e t h o d s are discussed elsewhere in this p u b l i c a t i o n .
752
Copper Nitrate Test

The copper nitrate test is performed by first immersing the aggregate
in one normal copper nitrate solution for 16 h (overnight) and then in
strong ammonia for about 2 min and finally washing carefully in tap
water. The one normal solution is prepared by dissolving 250 g of hydrated
copper nitrate, (Cu(NO3)2.3H20) in 1 litre of distilled water. This test was
used to obtain data on harmful aggregates consisting of varieties of argillaceous rocks and shales [40]. The prolonged copper nitrate test proved
to be a good indicator of the presence of harmful argillaceous impurities in
carbonate rocks and sandstones, and of the presence of harmful shale
particles. The damage to the aggregate particles is similar to that produced
by the sodium sulfate test. The quality assessment is based on scaling,
flaking, splitting, cracking, exfoliation, and disintegration of argillaceous
concentrations. There is no acceptance limits for this test because the complex copper nitrate coating on the carbonates prevents evaluation of the
damage by weight [6].
Correlation of Losses with Petrographic Characteristics

Fourteen samples of pure and argillaceous dolostones from a quarry
were subjected to the magnesium sulfate test after being petrographically
examined by Woda. s The number of particles in the 16.0 to 12.5-mm (1/2
to 5/8-in.) and 12.5 to 9.5-mm (5/8 to 3/8-in.) size fractions were counted
before and after testing and the losses were calculated by weight, and
particle count. The correlation among values of sulfate losses, acid insoluble
residues, and absorption, all of which were low, was good for the group of
pure dolostones. For the group of argillaceous dolostones the correlation
was also good for the average values of these parameters, all of which were
high, but it was mostly poor for each of the single samples in the argillaceous
group. The loss values by particle count averaged 13 percent (based on
weight loss) lower than those by weight.
Lithology of Harmful Substances

Harmful substances can be classified in four groups according to their
composition or physical properties.
1. "Clay and silt" refers to incoherent finely disseminated materials finer
than 75 #m (No. 200 sieve). This includes the fine materials that occur in
lenses and layers in gravel and sand deposits and the fracture fines produced by crushing gravel and rocks. Fracture fines include crusher fines
consisting of grains smaller than 75 #m and crusher dust which, before
washing, usually adheres to the surfaces of the crushed particles. Large
SWoda, G., personal communication.
753
amounts of dust are produced when shales, argillaceous limestones, dolostones, some sandstones and similar rocks are crushed.
2. The term "clay lumps" means lumps of clay in the aggregate which
remain cohesive during processing and are not disseminated throughout
the aggregate. These lumps would have to be dispersed mechanically to
produce fines. The lumps contain silt or very fine sand and often are held
together by clay. They do not disintegrate easily during the mixing and
placing of concrete.
Clay lumps and lumps of clayey silt held together by ice may survive the
processing of a concrete mix. Clay lumps lose some material during processand handling, as do friable and intensely physically weathered particles.
This increases the amount of fines in the concrete mix.
3. "Friable" and "soft" particles, and "lightweight" pieces, including
porous chert, are usually distinctly or deeply weathered rocks or mineral
agglomerations. Friable and soft particles are characterized by a poor bond
between the grains. The difference between friable and soft particles is that
individual mineral grains are detached easily from friable particles of any
rock or mineral aggregation while soft particles consist of relatively soft
mineral constituents that are structurally weak and have a low compressive
and tensile strengths, and may be partly friable. Examples are micaceous
minerals and gypsum [41]. Soft particles usually lose material in bits and
pieces but may also lose mineral grains. Deeply weathered rocks, such as
some granites and gneisses, ochers, earthy varieties of weathered iron oxides,
and incompletely cemented sand- or pebble-sized particles, are designated
as friable material, whereas weathered shales, soft limestones, dolostones,
and sandstones with an argillaceous or limonitic matrix lacking proper
cementitious properties are designated as soft or rather weak particles.
Lightweight pieces include highly porous particles of rocks and minerals
which float on a liquid with a density of 2.0. These particles lack frost resistance because they are saturated with water easily. The flotation test
also separates pieces of coal and lignite, both of which have very low densities.
Varieties of very porous weathered chert with a relatively low density are
objectionable as constituents of concrete aggregates. A density of 2.40 is
used to separate the usually harmful chert varieties from the harmless ones.
Some aggregate particles may consist of weathered, porous, and frost
susceptible chert together with its host rock limestone or dolostone. These
may have a higher density and still be harmful [42].
4. "Organic impurities" may consist of coal, lignite, plant roots, twigs,
and other vegetable and animal materials. A distinction is made between
discrete pieces of impurities, such as coal and lignite, and minute particles
of coal and lignite finely disseminated throughout the aggregate. Organic
matter, mainly tannins and compounds derived from the decay of vegetables
or animal matter, may occur as contaminants in particles or as particle
754
coatings. Organic matter which occurs in argillaceous concentrations in

sedimentary rocks usually is not determined by the test for organic impurities but it does influence the quality of an aggregate.
Effect of Harmful Substances on Concrete
Large amounts of fines (material finer than 75 /~m) in an aggregate

increase the water requirement of concrete resulting in excessive drying
shrinkage and low strength of the hardened concrete [43]. The adverse
effect on concrete is greater when the fines contain clay minerals, particularly those of the swelling group such as montmorillonite, smectite, and
nontronite. These minerals increase the amount of volume change in the
hardened concrete and cause the formation of microcracks [44,45]. Fines
in a sand that contained an unusually large proportion of nontronite, in
addition to increasing the water requirement, caused very rapid early stiff.
ening of the concrete mix [27]. Finely divided, dispersed clayey material in
the small amounts frequently found in aggregates usually is not harmful.
When clay lumps survive mixing and placing, and occur at the surface
of hardened concrete, they may breakdown from freezing and thawing,
wetting and drying, or wear and cause unsightly surface pits and popouts.
Friable and soft particles break down easily into numerous smaller pieces.
Depending upon the number of very friable particles, concrete deterioration
may be general but is more often localized in the form of surface pitting or
scaling [15]. Large quantities of soft particles cause a reduction in strength.
Smaller quantities significantly lower the abrasion resistance of concrete
surfaces exposed to severe attrition and impact.
Very low density particles usually are considered unsound [46-50]. Very
porous rocks, such as intensely weathered chert, friable siltstones, and
dolostones, are easily saturated with water but disruption on freezing depends on many other factors [20].
Organic impurities consisting principally of tannins retard the hardening
and reduce concrete strength, particularly at early ages [23,51,52]. Minute
grains of finely divided coal or lignite, in sufficient quantities, will retard
hardening [51], but coal and lignite are usually present as small particles
which may cause localized pitting, popouts, and staining of concrete surfaces marring its appearance [15]. Coal may be sensitive to frost. Anthracite is chemically inert, but could cause surface blemishes. Brown lignite
coal, fossilized wood, tall oil residues, and pine oil contain materials which
readily disperse in the alkaline solution present in concrete causing dark
brown local stains and weak zones. They complicate the painting of concrete surfaces. There is no simple way of separating such compounds from
aggregates [53].
755
Tests for Harmful Substances

The tests discussed below are designed to determine if potentially harmful substances occur in sufficient amounts to be harmful in a particular
construction. Generally accepted limits for these substances are given in
ASTM Specification C 33.
Clay, Silt, and Crusher Fines

ASTM Test C 117 determines the amount of fines in concrete aggregates.
Particles dispersed by the wash water and water soluble materials are removed from the sample during the test. It is assumed that particles which
disintegrate during washing will disintegrate during mixing of the concrete.
The test does not indicate the presence of the detrimental swelling clay
minerals. This information must be obtained by X-ray diffractometry or
other suitable methods. ASTM Specification C 33 recognizes that the dust
from nonargillaceous rocks is less detrimental than clay. The limits are increased when the fines consist "of the dust of fracture, especially free from
clay or shale."
Various other tests have been developed which are reported to indicate
reliably the potential harmfulness of clay constituents in concrete aggregate [44]. Sand-equivalent and sedimentation tests evaluate the clay in
terms of the extent to which a water suspension of the aggregate will settle
in a prescribed time period. According to Gaynor [54], based on examinations of 130 samples, there is almost no correlation between the sand
equivalent value and the amount of minus 75 #m material in the sand.
Others have reported that the tests correlate well with the mixing water
requirement of concrete which, in turn, affects shrinkage and strength. It
might be surmised that, in view of the many factors which affect mixing
water requirements, aggregates should be evaluated for this effect by direct
comparisons in concrete mixes [1].
Experiments to determine the precision of ASTM Test C 117 have shown
that the results of two properly conducted tests from two different laboratories should not differ by' more than 0.6 percent and those conducted by
the same operator on the same coarse aggregate by not more than 0.3 percent.
Clay Lumps and Friable Particles

ASTM Test C 142 has changed over the past ten years from "Test for
Clay Lumps in Natural Aggregates" (C 142-64T), to "Test for Friable
Particles in Aggregates" (C 142-67), and finally to "Test for Clay Lumps
and Friable Particles (C 142-71). It applies only to agglomerations and
weathered particles which may be expected to remain at least partially
756
intact during the concrete mixing. Therefore, the test is performed after
the sample has been tested foi" materials finer than 75 #m and the original
fines have been washed from the sample. The remainder, which is broken
down with the fingers and removed from the sample by wet-screening on
sieves whose openings are slightly smaller than those of the original sieves,
is classified as clay lumps and friable particles which are less detrimental
than fines. The maximum amounts permitted in concrete mixtures is 1
percent for fines and 2 to 10 percent for clay lumps and friable particles
(ASTM Specification C 33). The friable particles that break down during
the wet-screening in ASTM Test C 117 are also expected to break down
during concrete mixing. For this reason, the total amount allowed of material finer than 75 #m, in an aggregate is specified.
Lightweight Pieces
ASTM Test for Lightweight Pieces in Aggregate (C 123) uses heavy
liquids of suitable density to separate low density particles. The samples
are immersed in the heavy liquids in the saturated surface dry condition, a
condition which is not obtained easily, neither for very fine-grained fractions
nor for very large coarse fractions. Coal and lignite particles are separated
by a liquid having a density of 2.0. They are distinguished from other low
density, mostly deep weathered rock particles, by their brownish-black or
black color, following the direction given in ASTM Specification C 33.
Very finely divided coal or lignite is detected by the test for organic impurities.
Potentially physically harmful chert is very porous and lighter than
harmless dense chert. Potentially detrimental chert is distinguished by its
saturated surface dry density, which is specified as less than 2.40 (ASTM
Specification C 33). Formerly, to distinguish harmful chert, an option was
given to use a saturated surface-dry density of less than 2.35, or 5 cycles of
the sulfate test or S0 cycles of the freeze-thaw test at --17.8~ (0~ and
4.4~ (40~
emphasizing the frost susceptibility of such chert particles.
Whereas lignite and coal may be separated by the less hazardous zinc
chloride solution, the requirement of a heavy liquid density of 2.40 for
chert means that toxic liquids must be used.
Experiments have shown that sand-sized particles of chert which were
detrimental in coarse aggregate size, did not damage concrete subjected to
the freeze-thaw test (ASTM Test for Resistance of Concrete to Rapid
Freezing and Thawing (C 666, Procedure B)) rapid freezing in air and
thawing in water [55]. The vulnerability of chert particles depends on the
pore structure which relates approximately to density. Porous chert is
limited to from 3 to 8 percent (ASTM Specification C 33) for aggregates in
which chert appears as a minor constituent. The test for lightweight pieces
757
is repeatable within about 10 percent (95 percent probability level) on

duplicate portions of the same sample [32].
Soft Particles
It is difficult to distinguish clearly between soft and friable particles.
Since 1949, soft particles have been identified by ASTM Test for Scratch
Hardness of Coarse Aggregate Particles (C 235). The test has been discontinued as a standard but retained as ASTM Recommended Practice
C 851. The reason for discontinuing it as a test method was the lack of
consistency in results and the poor correlation of the results with concrete
strength, even when the aggregate contained a high percentage of "soft"
particles [15]. Soft particles sometimes were determined by visual examination but the results were poor because of the different impression of softness
by each observer. Mobs hardness scale for mineral grains also has been
applied for determining the hardness of rock particles, but is quite inappropriate for this purpose.
The test for soft particles, using a brass pencil of specified hardness to
scratch the particles is useful for quick cursory checking a sample in the
field or in the laboratory before the more adequate Los Angeles abrasion
test is performed (ASTM Test for Resistance to Abrasion of Small Size
Coarse Aggregate by Use of the Los Angeles Machine (C 131) and ASTM
Test for Resistance to Abrasion of Large Size Coarse Aggregate by Use of
the Los Angeles Machine (C 535)). It is also useful in petrographic examination of physically unsatisfactory aggregate.
Organic Impurities
The colorimetric test for organic matter, ASTM Test for Organic Impurities in Sands for Concrete (C 40), detects harmful organic substances
but unfortunately also reacts with other organic substances, such as bits of
wood, which in limited amounts do not affect strength or setting time adversely. Therefore, a negative test is normally conclusive evidence that the
aggregate is free from harmful organic matter, but a positive reaction (dark
color) may not forebode difficulty. Some sands contain organic coatings
which produce "accidental" air entrainment in a concrete mixture [36].
These volatile products derived from the anoxemic decomposition of organic
matter from underlying clay beds cannot be detected by the sodium hydroxide color test. The colorimetric test is employed as a warning only and
should be checked by mortar strength and setting time tests. ASTM Test
for Effect of Organic Impurities in Fine Aggregate on Strength of Mortar
(C 87) uses as the basis of comparison the suspect sand after it has been
758
treated to remove the organic matter. Presumably, the only difference between the test and reference mortars is the organic matter in the former.
The colorimetric test gives a highly reliable and reproducible indication
of organic matter in sand. The check for effect on mortar strength by
ASTM Test C 87 is subject to the problems of strength testing, which is
discussed elsewhere. ASTM Specification C 33 allows the mortar strength
ratio to be as low as 95 percent. The slight reduction allows for unavoidable
testing variation.
Petrographic Examinations (Addendum)

Harmful substances, which occur in discrete particles, large enough to
be handled, such as coal, lignite, and friable particles, can be identified
most efficiently by petrographic methods. Petrographic examination 6 gives
numerical data for aggregate particles consisting of harmful substances or
containing them in amounts sufficient to influence the concrete quality
adversely. It does not give numbers corresponding to the specified limits
obtained by technological tests.
The petrographic examination cannot be used for determining either the
amounts of organic materials coating aggregate particles or finely disseminated minute particles of coal and lignite, nor can it be used to determine
the exact amounts of clayey material in the fraction finer than 75/~m (No.
200 sieve size). However, it does identify the clay minerals. The importance
of detailed petrographic examination of aggregates from areas without a
service record is emphasized in a paper by Davis et al [27].
Conclusion
When the sulfate test was first developed it was thought that it would
help to separate aggregates susceptible to frost action from those which are
resistant to it. Today the significance of the test is still evaluated by comparing the results with those from laboratory freeze-thaw tests and with the
general field performance of the aggregate. The correlation with field performance and laboratory freeze-thaw results is frequently poor. Many users
claim that the test is unsatisfactory while others say they obtain satisfactory
results which are useful for selecting and assessing concrete aggregates.
Suggestions have been made to replace the sulfate test with the freezethaw test on unconfined aggregate. This suggestion was rejected on the
grounds that the test results would not reflect the performance of aggregates confined in concrete. Some researchers who recommend elimination
of ASTM Test C 88 do not consider the vast number of test results available for which the aggregate behavior in concrete is known. The test is one
6See p. 539.
DOLAR-MANTUANI ON SOUNDNESS AND DELETERIOUS SUBSTANCES 759
of those often used for assessing a d e p o s i t p r i o r to t a k i n g an option for

p u r c h a s e of an aggregate source.
In the tests for h a r m f u l a g g r e g a t e substances clay a n d silt, clay lumps,
a n d friable particles, lightweight pieces a n d o r g a n i c c o n t a m i n a n t s are included. Some of these substances increase the w a t e r r e q u i r e m e n t thereby
r e d u c i n g strength a n d d u r a b i l i t y o f concrete; some m a y influence the setting
time of a concrete mix, aggravate drying s h r i n k a g e (swelling clay minerals),
or cause surface stains, pits a n d popouts. Finely d i s s e m i n a t e d clay a n d silt
in aggregates are more d e t r i m e n t a l t h a n clay l u m p s a n d friable particles.
This is reflected in the specification limits. T h e organic i m p u r i t i e s are
detected readily a n d their influence on concrete strength should be checked
by A S T M Test C 87.
T h e test for soft particles ( A S T M Test C 235) is no longer a s t a n d a r d
a c c e p t a n c e test b e c a u s e of the lesser i m p o r t a n c e of soft particles except
where good a b r a s i o n resistance is required. It gives a quick p r e l i m i n a r y
indication of the presence of particles which m a y be classified as soft.
The general opinion seems to be t h a t the s t a n d a r d tests for h a r m f u l
substances are useful, satisfactory, easy a n d quick to perform. By following
up the test results a n d e l i m i n a t i n g or r e d u c i n g the a m o u n t s o f h a r m f u l
m a t e r i a l s some basic concrete p r o b l e m s can be avoided.
References
[1] Bloem, D. L., "Soundness and Deleterious Substances," Significance of Tests and Properties of Concrete and Concrete Making Materials, ASTM STP 169A. American Society
for Testing and Materials, 1966, pp. 497-512.
[2] Rhoades, R. and Mielenz, R. C., "Petrography of Concrete Aggregates," Proceedings,
[3] Mielenz, R. C., "Petrographic Examination of Concrete Aggregate," Proceedings,
American Society for Testing and Materials, Vol. 54, 1954, pp. 1188-1218.
[4] Mielenz, R. C., "Petrographic Examination," Significance of Tests and Propert&s of
Concrete and Concrete Making Materials, ASTM STP 169A, American Society for
[5] Swenson, E. G. and Chaly, V., "Basis for Classifying Deleterious Characteristics of
Concrete Aggregate Materials," Proceedings, American Concrete Institute, Vol. 52,
pp. 987-1002; Journal, May 1956.
[6] Dolar-Mantuani, L., "A Guide to Petrographic Evaluation of Natural Concrete Aggregates," in manuscript prepared for print.
[7] McCown, V., "The Significance of Sodium Sulfate and Freezing and Thawing Tests on
Mineral Aggregates," Proceedings, Highway Research Board, Vol. 11, 1931, p. 312.
[8] Garrity, L. V. and Kriege, H. F., "Studies of Accelerated Soundness Tests," Proceedings, Highway Research Board, Vol. 15, 1935, p. 237.
[9] Walker, S. and Proudley, C, E., "Studies of Sodium and Magnesium Sulfate Soundness Tests," Proceedings, American Society for Testing and Materials, Vol. 36, Part 1,
1936, p. 327.
[10] Wuerpel, C. E., "Factors Affecting the Testing of Concrete Aggregate Durability,"
Proceedings, American Society for Testing and Materials, Vol. 38, Part 1, 1938, p. 327.
[11] Wuerpel, C. E., "Modified Procedure for Testing Aggregate Soundness by Use of Magnesium Sulfate," Proceedings, American Society for Testing and Materials, Vol. 39,
1939, p. 882.
760
[12] Hudec, P. P. and Rogers, C. A., "The Influence of Rock Type, Clay Content and Dominant Cation on the Water Sorption Capacity of Carbonate Rocks," Abstract of paper
presented at Geological Society of Canada, Annual Meeting, Edmonton, May 1976.
[13] Woolf, D. O., "Relation Between Sodium Sulfate Soundness Tests and Absorption of
Sedimentary Rock," Public Roads, Dec. 1927, p. 225.
[14] Cantrill, C. and Campbell, L., "Selected Aggregates for Concrete Pavement Based on
Service Records," Proceedings, American Society for Testing and Materials, Vol. 39,
1939, p. 937.
[15] Walker, S. and Bloem, D. L., "The Problem of Deleterious Particles in Aggregates,"
Circular No. 35, National Sand and Gravel Association, 1950.
[16] Adams, A. and Pratt, H. A., "A Comparison of Absorption and Soundness Tests on
Main Sands," Proceedings, American Society for Testing and Materials, Vol. 45, 1945,
p. 771.
[17] Mather, Katharine, "Relation of Absorption and Sulfate Test Results of Concrete Sands,"
Bulletin No. 144, American Society for Testing and Materials, Jan. 1947, p. 26.
[18] Bloem, D. L., "Review of Current and Projected Researches," report to Annual Convention, National Sand and Gravel Association and National Ready Mixed Concrete
Association, Chicago, I11., Feb. 1960.
[19] Sweet, H., "Chert as a Deleterious Constituent in Indiana Aggregates," Proceedings,
[20] Verbeck, G. and Landgren, R., "Influence of Physical Characteristics of Aggregates on
Frost Resistance of Concrete," Proceedings, American Society for Testing and Materials,
Vol. 60, 1960, p. 1063.
[21] Litvan, G. G., "Phase Transitions of Adsorbates, 4. Mechanism of Frost Action in
Hardened Cement Paste," Journal. American Ceramic Society, Vol. 55, No. 1, Jan.
1972, pp. 38-42.
[22] Litvan, G. G., "Frost Action in Cement Paste," Materials and Structures, No. 34, JulyAug. 1973, p. 6.
[23] Lang, F. C., "Deleterious Substances in Concrete Aggregates," Bulletin. National Sand
and Gravel Association, April 1931; Circular No. 10, National Sand and Gravel Association, May 1931.
[24] Vollick, C. A. and Skillman, E. I., "Correlation of Sodium Sulfate Soundness of Coarse
Aggregate with Durability and Compressive Strength of Air-Entrained Concrete," Proceedings, American Society for Testing and Materials, Vol. 52, 1952, p. 1159.
[25] Mather, B., "Sulfate Soundness, Sulfate Attack and Expansive Cement in Concrete,"
International Symposium on Durability of Concrete--1969, RILEM Preliminary Report, Part 2, C209-C220.
Hardened Concrete Mortar," Report No. 24, Highway Research Board, 1966, pp. 18-40.
[27] Davis, R. E., Mielenz, R. C., and Polivka, M., "Importance of Petrographic Analysis
and Special Tests Not Usually Required in Judging Quality of Concrete Sand," Journal
of Materials, American Society for Testing and Materials, Vol. 2, No. 3, 1967, pp.
461-486.
[28] Chamberlin, W. P., "Importance of Petrographic Analysis and Special Tests Not Usually
Required in Judging Quality of Concrete Sand: A Discussion," Journal of Materials,
American Society for Testing and Materials, Vol. 2, No. 3, 1967, pp. 482-483.
[29] Tuthill, L. H. and Adams, R. F., "Importance of Petrographic Analysis and Special
Tests Not Usually Required in Judging Quality of Concrete Sand: A Discussion," Journal
of Materials. American Society for Testing and Materials, Vol. 2, No. 3, 1967, pp.
480-482.
[30] Larsen, T., Cady, P., Franzen, M., and Reed, J., "A Critical Review of Literature Treating Methods of Identifying Aggregates Subject to Destructive Volume Change When
Frozen in Concrete and a Proposed Program of Research," Research Special Report
No. 80, Highway Research Board, 1964, p. 81.
[31] Paul, I., "Magnesium Sulfate Accelerated Soundness Test on Concrete Aggregates,"
Proceedings, Highway Research Board, Vol. 12, 1932, p. 319.
[32] Woolf, D. O., "Improvement in the Uniformity of the Accelerated Soundness Test of
Coarse Aggregate," Bulletin No. 213, American Society for Testing and Materials, Jan.
1953, p. 42.
761
[33] Woolf, D. O., "An Improved Sulfate Soundness Test for Aggregates," Bulletin No. 213,
American Society for Testing and Materials, April 1956, p. 77.
[34] Alexander, M. L., "An Evaluation of the Effect of Sampling on the Repeatability of the
Sodium Sulfate Soundness Test," California Department of Transportation, Dec. 1975,
p. 20.
[35] Bloem, D. L., "Sulfate Soundness Test Relationship," report to Subcommittee III-e of
ASTM Committee C-9, 23 Sept. 1958.
[36] MacNaughton, M. F. and Herbich, J. B., "Accidental Air in Concrete," Proceedings,
American Concrete Institute, Vol. 51, 1955-1956, pp. 273-284.
[37] Dolar-Mantuani, L., "Petrographic Examination of Natural Concrete Aggregates,"
Record No. 120, Highway Research Board, 1966, pp. 7-17.
[38] Bloem, D. L. and Gaynor, R. D., "Effect of Aggregate Properties on Strength of Concrete," Journal, American Concrete Institute, Oct. 1963; Proceedings. Vol. 60, p. 1429.
[39] Brink, H. R., "Rapid Freezing and Thawing Test for Aggregate," Bulletin No. 201,
[40] Dolar-Mantuani, L., "Concrete Aggregate Examination by Prolonged Copper Nitrate
Staining Method," Ontario Hydro Research News, Vol. 14, No. 2, Second Quarter,
1962, pp. 2-14.
[4l] Lang, F. C., "Deleterious Substances," Significance of Tests of Concrete and Concrete
Aggregates, A S T M STP 22 A, American Society for Testing and Materials, 2nd ed.,
1943, pp. 138-144.
[42] Dolar-Mantuani, L., "Harmful Constituents in Natural Concrete Aggregates in Ontario,"
Proceedings, 24th International Geological Congress, Montreal, Section 13, 1972,
pp. 227-234.
[43] Powers, T. C., The Properties of Fresh Concrete, Wiley, New York, 1968.
[44] Temper, B. and Haskel, W. E., "The Effect of Clay on the Quality of Concrete Aggregates," California Highways and Public Works, Vol. 34, Nov.-Dec. 1955; Highway Research Abstracts, Vol. 26, No. 2, Feb. 1956, p. 30.
[45] Lyse, I., "Tests Indicate Effect of Fine Clay in Concrete," Engineering News Record,
Vol. 113, 23 Aug. 1934, p. 233.
[46] Lang, F. C., "Summary of Tests on Effect of Shale in Gravel on Compressive Strength of
Concrete," Proceedings. American Concrete Institute, Vol. 23, 1927, p. 592.
[47] Walker, S. and Proudley, C. E., "Shale in Concrete Aggregates," Proceedings, Highway
[48] Litehiser, R. R., "The Effect of Deleterious Materials in Aggregates for Concrete,"
Circular No. 16. National Sand and Gravel Association, 1938.
[49] Walker, Stanton and Bloem, D. L., "Effect on Heavy-Media Processing on Quality of
Gravel," Circular No. 55, National Sand and Gravel Association, 1953.
[50] Higginson, E. C. and Wallace, G. B., "Control Testing for Separation of Lightweight
Material from Aggregate," Bulletin No. 243, American Society for Testing and Materials,
Jan. 1960, p. 60.
[51] Freeman, P. J., "Effect of Coal and Lignite in Sand for Concrete," Proceedings. American
Society for Testing and Materials, Vol. 29, Part I, 1929, p. 328.
[52] Mather, B., "Testing of Fine Aggregate for Organic Impurities and Compressive Strength
in Mortars," Bulletin No. 178, American Society for Testing and Materials, Dec. 1951,
p. 35.
[53] "Impurities in Aggregates for Concrete," Advisory Note No. 18, Cement and Concrete
Association, London, 1969.
[54] Gaynor, R. D., "Investigation of Concrete Sands," Technical In)brmation Letter No.
266, National Sand and Gravel Association, 13 May 1968, p. 24.
[55] Bloem, D. L., "Factors Affecting Freezing-and-Thawing Resistance of Chert Gravel
Concrete," Research Report No. 18, Highway Research Board, 1963, pp. 48-60.
W. J. M c C o y 1
Chapter 43--Mixing and Curing

Water for Concrete
Introduction
This paper is concerned primarily with the significance of tests of various
types of waters for mixing and curing concrete and makes no attempt to
include the effect of quantity of mixing water. Almost all natural waters,
fresh waters, and waters treated for municipal use are satisfactory as mixing
water for concrete if they have no pronounced odor or taste. Because of
this very little attention is usually given to the water used in concrete, a
practice that is in contrast to the frequent checking of the cement and
aggregate components of the concrete mix.
Mixing Water
A popular criterion as to the suitability of water for mixing concrete is
the classical expression, "If water is fit to drink it is all right for making
concrete." This does not appear to be the best basis for evaluation, since
some waters containing small amounts of sugars or citrate flavoring would
be suitable for drinking but not mixing concrete [1]$ and, conversely,
water siutable for making concrete may not necessarily be fit for drinking.
In an attempt to be more realistic some concrete specification writers attempt to insure that water used in making concrete is suitable by requiring
that it be clean and free from deleterious materials. Some specifications
require that if the water is not obtained from a source that has proved
satisfactory, the strength of concrete or mortar made with the questionable
water should be compared with similar concrete or mortar made with water
known to be suitable. The Corps of Engineers, U.S. Army Specifications
[2], in addition to a general description of acceptable water also states that
if the pH of water lies between 6.0 and 8.0 and the water is free from orIDirector, Cement Technology, Master Builders, Cleveland, Ohio 44118.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
766
ganic matter, it may be regarded as safe for use in mixing concrete. (An
exception to this is the case where sodium sulfates or other natural salts
are present in excessive amounts.) These specifications also state that if
water is of questionable quality it should be tested in mortar cubes for
which the 7 and 28 day compressive strengths should equal at least 90 percent of that of companion test specimens in which distilled water is used
before it is judged to be acceptable [3]. Other than comparative tests of
this type, no special test has been developed for determining the quality of
mixing water, and, hence it is difficult for the man in the field to judge
the fitness of water for use in concrete [4].
The two principal questions regarding mixing water appear to be how
do impurities in the water affect the concrete and what degree of impurity
is permissible? The following discussion is a resum6 of available information on these two items.
Effect of Impurities in Mixing Water

The most extensive series of tests on this subject was conducted by
Abrams [5]. Approximately 6000 mortar and concrete specimens representing 68 different water samples were tested in this investigation. Among
the waters tested were sea and alkali waters, bog waters, mine and mineral
waters, and waters containing sewage and solutions of salt. Tests with
fresh waters and distilled water were included for comparative purposes.
Setting time tests on cement and concrete strength tests from 3 days to
2.33 years were conducted for each of the various water samples. Some
of the more significant conclusions based on these data are as follows:
1. The time of setting of portland cement mixes containing impure mixing waters was about the same as those observed with the use of clean
fresh waters with only a few exceptions. In most instances the waters giving
low relative compressive strength of concrete caused slow setting, but,
generally speaking, the tests showed that time of setting is not a satisfactory
test for suitability of a water for mixing concrete.
2. None of the waters caused unsoundness of the neat portland cement
pat when tested over boiling water.
3. In spite of the wide variation in the origin and type of waters used,
most of the samples gave good results in concrete due to the fact that the
quantities of injurious impurities present were quite small.
4. The quality of mixing water is best measured by the ratio of the 28
day concrete or mortar strength to that of similar mixes with pure water.
Waters giving strength-ratios which are below 85 percent in general should
be considered unsatisfactory.
5. Neither odor nor color is an indication of quality of water for mixing
concrete. Waters which were most unpromising in appearance gave good
MCCOY ON MIXING AND CURING WATER FOR CONCRETE
767
results. Distilled waters gave concrete strengths essentially the same as

other fresh waters.
6. Based on a minimum strength-ratio of 85 percent as compared to that
observed with pure water, the following samples were found to be unsuitable for mixing concrete: acid water, lime soak water from tannery waste,
carbonated mineral water discharge from galvanizing plants, water containing over 3 percent of sodium chloride or 3.5 percent of sulfates, and
water containing sugar or similar compounds. The concentration of total
dissolved solids in these waters was over 6000 ppm except for the highly
carbonated water which contained 2140 ppm total solids. These data support one of the principal reasons for the general suitability of drinking water,
for few municipal waters contain as much as 2000 ppm of dissolved solids
and most contain far less than 1000 ppm. For drinking waters, 1000 ppm
is the maximum permissible in some specifications [6]. Very few natural
waters other than sea water contain more than 5000 ppm of dissolved solids
[7,8].
7. Based on the minimum strength-ratio of 85 percent, the following
waters were found to be suitable for mixing concrete: (a) bog and marsh
water; (b) waters with a maximum concentration of 1 percent SO4; (c) sea
water, but not for reinforced concrete; (d) alkali water with a maximum of
0.15 percent Na2SO4 or NaCI; (e) pumpage water from coal and gypsum
mines; and (j) waste water from slaughterhouses, breweries, gas plants,
and paint and soap factories.
Many of the specifications for water for mixing concrete, especially those
which require that it be potable, would have excluded nearly all of the above
waters, but contrary to this rather general opinion, the test data show that
the use of many of the polluted types of water did not result in any appreciable detrimental effect to the concrete. The important question is not
whether impurities are present, but do impurities occur in injurious quantities? It should be noted that the conclusions in Paragraph 7 are based
entirely on tests of specific samples from the indicated sources, and it should
not be assumed that all waters of the type described would be innocuous
when used as mixing water.
Typical analyses of natural fresh water as reported by the U.S. Geological Survey are given in Table 1. Collins [9] reports that these analyses
represent public water supplies used by about 45 percent of the cities of
the United States that have a population of more than 20 000.
A concrete manual [10] published in Denmark in 1944 points out that
humic acid and other organic acids should be avoided because their presence
means a danger to the stability of concrete.
An article appearing in a 1947 issue of the British publication Concrete
and Constructional Engineering [11] discusses the harmful effects of using
acid waters in concrete and claims that the harmful effects of organic acid
768
TABLE 1--Typical analyses of city watersupplies (ppm)a

Analysis Number
Silica (SiO2)
Iron (Fe)
Calcium (Ca)
Magnesium (Mg)
Sodium (Na)
Potassium (K)
Bicarbonate (HCO3)
Sulfate (SO4)
Chloride (CI)
Nitrate (NO3)
Total dissolved soilds
2.4
0.14
5.8
1.4
1.7
0.7
14.0
9.7
2.0
0.54
31.0
12.0
0.02
36.0
8.1
6.5
1.2
119.0
22.0
13.0
0.1
165.0
10.0
0.09
92.0
34.0
8.2
1.4
339.0
84.0
9.6
13.0
434.0
6
9.4
0.2
96.0
27.0
183.0
18.0
334.0
121.0
280.0
0.2
983.0
7
22.0
0.08
3.0
2.4
215.0
9.8
549.0
11.0
22.0
0.52
564.0
aTaken from Collins [9].

are not evident as soon as those of mineral acids, while deleterious salts
have a greater effect on strength at an early age than at later ages.
Kleinlogel [12] in his book Influences on Concrete states that mixing
water should not contain humus, peat fiber, coal particles, sulfur, or industrial wastes containing fat or acid.
A recent article in the Indian Concrete Journal [13] contains a tabulation
of maximum limits for impurities in mixing water which is summarized in
Table 2.
The limits for suspended particles in Table 2 agree with the requirements
of the U.S. Bureau of Reclamation [14] which has a turbidity limit of 2000
ppm for mixing water.
Sugar is probably the organic contaminant that causes the most concern.
Steinour [15] explains that, although sugar as a contaminant in the field
has gained a bad reputation as a retarder and strength reducer, these
judgments need qualification. Laboratory tests have shown that, although
the smaller amounts retard the setting, they increase the strength development. With larger amounts the setting is further retarded, and early
strengths such as those for two and three days (or even seven days) are
severely reduced. The later strengths, however, are increased, or at least
not affected adversely, provided proper curing is maintained as required.
With still larger amounts the cement becomes quick setting and strengths
are reduced for markedly for 28 days and probably permanently. The
amount of sugar that can cause these different effects varies with the other
factors involved, such as the composition of the cement, cement content,
and ambient conditions.
Use of Sea Water in Mixing Concrete

In addition to the supporting reference previously mentioned in Abrams'
paper [5], the English article [11] also states that seawater with a maximum
769
TABLE2--Tolerable concentrations of impurities in mixing water.

Impurity
1. Sodium and potassium carbonates and bicarbonates
2. Sodiumchloride
3. Sodiumsulfate
4. Calciumand magnesium bicarbonates
5. Calciumchloride
6.
7.
8.
9.
10.
Iron salts
Sodium iodate, phosphate arsenate and borate
Sodiumsulfide
Hydrochloricand sulfuric acids
Sodiumhydroxide
11. Salt and suspended particles
Maximum Tolerance Concentration

1 000 ppm
20 000 ppm
10 000 ppm
400 ppm of bicarbonate ion
2% by weight of cement in plain
concrete
40 000 ppm
500 ppm
even 100 ppm warrants testing
10 000 ppm
0.5% by weight of cement is set not
affected
2 000 ppm
concentration of salts of the order of 3.5 percent does not appreciably reduce the strength, of concrete although it may lead to corrosion of reinforcement.
A paper by Liebs in the German publication Bautechnik [16] contains
the results of comparative 7, 28, and 90 day compressive strength tests of
concrete mixed with fresh water and with seawater. The data show that the
seawater concrete had about 6 to 8 percent lower strength than the fresh
water concrete. No efflorescence was observed. The Indian Concrete Journal
article [13] pointed out that concrete made with seawater may have higher
early strength than normal concrete, but strengths at later ages (after 28
days) may be lower. Steinour [15] states the use of seawater may cause a
moderate reduction in ultimate strength, an effect that can be avoided by
the use of a higher cement content. He also notes that concrete in which
seawater is used as mixing water is sound but its use may cause efflorescence or dampness, and, in reinforced concrete, the risk of corrosion of the
steel is increased. Seawater definitely should not be used for making prestressed concrete.
Hadley [17] points out that the seawater was used in the concrete mix
for the foundation of the lighthouse at the extremity of the Los Angeles
breakwater which was built by the U.S. Engineer Corps in 1910 and that
25 years later it was examined and found to be in good condition with sharp
edged corners and no disintegration. There are several references in the
literature which indicate that salt water has been used in mixing plain
concrete without incurring trouble at later periods. Much of the concrete
in the Florida East Coast Railway was mixed with sea water with no detrimental effect due to its use [18]. Most engineers are of the opinion that
seawater should not be used for mixing reinforced concrete; however,
Dempsey [19] describes construction of military bases in Bermuda using
770
coral aggregate and concludes that seawater seems to be satisfactory for

making reinforced concrete and causes no problem beyond an acceleration
in stiffening of the mix. No harmful effect on the durability of reinforced
concrete had occurred at the end of four years.
Effect of Algae in Mixing Water on Air Content and Strength of Concrete

A rather extensive series of laboratory tests reported by Doell [20] has
shown that the use of water containing algae had the unusual effect of entraining considerable quantities of air in concrete mixes with an accompanying decrease in strength. The data in Table 3 were extracted from
Doell's paper and are based on tests with 19.0 mm ( 90
maximum size
aggregate concrete having a water/cement ratio of 0.5 and a slump of 38.1
to 76.2 mm (1.5 to 3 in.), with the ratio of coarse to fine aggregate constant.
In addition to the detrimental effect on strength, one of the important
aspects of these data is that considerable quantities of air can be entrained
in concrete by the use of mix water containing algae.
Effect of Hardness of Mixing Water on Air Content of Concrete

Wuerpel [21] reports data on a series of air determination tests with
waters of various degrees of hardness which show that the air content was
not affected by the amount of hardness in the water.
Curing Water
There are two primary considerations with regard to the suitability of water
for curing concrete. One is the possibility that it might contain impurities
that would cause staining and the other is that it might contain aggressive
impurities that would be capable of attacking or causing deterioration of
the concrete. The latter possibility is unlikely, especially if water satisfactory
for use in mixing concrete is employed. In some instances the staining or
discoloration of the surface of concrete from curing water would not be
TABLE 3--Effect of algae in mixing water on air content and strength of concrete.
Mix
Number
10
8
7
5
9
Algae in
Mix Water,
%
Air in
Concrete,
%
Compressive
Strength, 28 day,
MPa (psi)
none
(control)
0.03
0.09
0.15
0.23
2.2
33.30 (4830)
2.6
6.0
7,9
10.6
33.37
27.86
22.75
17.03
(4840)
(4040)
(3320)
(2470)
771
objectionable. The most common cause of staining is usually a relatively

high concentration of iron or organic matter in the water; however, relatively
low concentrations of these impurities may cause staining, especially if
the concrete is subjected to prolonged wetting by runoff of curing water
from other portions of the structure [2].
Test data included in an unpublished Waterways Experiment Station
Miscellaneous Paper by the Corps of Engineers, U.S. Army [22], show
that there is not a consistent relation between dissolved iron content
and degree of staining. In some cases 0.08 ppm of iron resulted in only
a slight discoloration, and, in other cases, waters with 0.06 ppm of iron
gave a moderate rust colored stain, while 0.04 ppm produced considerable brownish black stain. Generally speaking, the conditions of these
tests were such as to accentuate the staining properties of the water,
since considerably more water was evaporated over a unit area than
would be the case in most instances in the field.
With respect to organic impurities in water, it is virtually impossible
to determine from a chemical analysis if the water would cause objectionable staining when used for curing concrete. It is advisable to
use a performance type test procedure, an example of which is Test
Method CRD-C 401-57 of the Corps of Engineers, U.S. Army [23].
This method of test outlines three procedures for evaluating the staining
properties of water proposed for use in curing concrete. The Preliminary
Method is intended for use in selecting sources that are worthy of more
complete investigation and consists of evaporating 3000 ml of the test water
in the concave area formed by the impression of a 101.6 mm (4-in.) watch
glass in the surface of a neat white cement or plaster of Paris specimen. The
Complete Method can be used to evaluate those sources that the Preliminary
Method indicates to be promising. In the Complete Method, 11.35 litres
(3 gal) of test water drips on to a mortar specimen exposed to heat lamps
and forced circulation of air. The Field Method is intended as a means of
evaluating the water finally selected for use and involves the curing of a
1.858 m 2 (20-ft 2) slab of concrete with the test water under maximum
exposure to the sun with the test slab placed at a slight angle to keep it in
a wet condition with a minimum runoff. The test results by each of these
three methods are evaluated by visual observation.
The latest Corps of Engineers Standard Practice Manual for Concrete
[24] notes that the Standard Guide Specifications do not require that water
be nonstaining. However, the manual does clearly state that there must be
no permanent staining of surfaces where appearance is important. For
these surfaces the contractor has the option of using nonstraining water or
of cleaning the surface after completion of moist curing. No cleaning is
required of surfaces which will subsequently be stained when the structure
is in service.
772
Summary
Mixing Water
The significance of the foregoing information presented indicates that
any naturally occurring or municipal supply waters suitable for drinking
purposes can be used as mixing water for concrete and that most naturally
occurring waters ordinarily used for industrial purposes are satisfactory.
Many waters, which upon casual examination would be judged to be unsuitable because of color, odor, or contamination with impurities as in the
case of marsh water, alkaline sulfate waters, and water containing industrial
wastes, would be found to be satisfactory when tested in mortar or concrete
since in many instances the strength would be greater than 85 to 90 percent
of the strength of comparative specimens made with pure waters. In the
case of seawater, a strength reduction ranging from 8 to 15 percent can be
expected depending on job conditions; however, seawater ordinarily is not
recommended for use as mixing water in reinforced concrete. The hardness
of water does not affect air content of concrete, but algae can entrain air
and significantly reduce strength.
Curing Water
It is improbable that a water used for curing would attack concrete if
it were of the type suitable for use as mixing water. Organic matter or iron
in the curing water can cause staining or discoloration of concrete, but this
is rather uncommon especially where a relatively small volume of water is
used; however, the suggested performance tests [21] will determine if a
water possesses any potential staining qualities.
References
[1] Clair, M. N., "Effect of Sugar of Concrete in Large Scale Trial," Engineering NewsRecord, March 1929, p. 473.
[2] "Requirements for Water for Use in Mixing or Curing Concrete," CRD-C 400, Handbook
for Concrete and Cement, Corps of Engineers, U.S. Army.
[3] "Method of Test for Compressive Strength of Mortar for Use in Evaluating Water for
Mixing Concrete," CRD-C 406, Handbook for Concrete and Cement, Corps of Engineers,
U.S. Army.
[4] Bauer, E. E., Plain Concrete, 3rd ed., McGraw-Hill, New York, 1949, p. 82.
[5] Abrams, D. A., "Tests of Impure Waters for Mixing Concrete," Proceedings. American
[6] Public Health Service Drinking Water Standard, U.S. Department of Health Service,
Washington, D. C., 1946.
[7] Proudley, C. E., "Effect of Alkalies on Strength of Mortar," Public Roads, Vol. 5, 1924,
pp. 25-27.
[8] Giesecke, F. E. and Parkinson, G. A., "Effects of Various Salts in the Mixing on the
Compressive Strength of Mortar," Bulletin No. 2730, University of Texas, Engineering
Research Series, Aug. 1927.
773
[9] Collins, W. D., "Typical Water Analyses for Classification with Reference to Industrial
Use," Proceedings, American Society for Testing and Materials, Vol. 44, 1944, pp.
1057-1061.
[10] Plum, N. H., Christiani, and Nielsen, "Concrete Manual," Bulletin No. 39, Copenhagen,
Denmark, 1944.
[11] "Water for Making Concrete," Title No. 44-197, Proceedings, American Concrete
Institute, Vol. 44, 1948, pp. 416-16; reprinted from an editorial "Water for Making
Concrete," Concrete and Constructional Engineering , London, Vol. 42, No. 10, 1947, p. 25.
[12] Kleinlogel, A., Influences on Concrete, Frederick Ungar Publishing Co., New York,
p. 158.
[13] "Requirements of Mixing Water for Concrete," Indian Concrete Journal, March 1963,
pp. 95, 98, 113.
[14] Concrete Manual, U.S. Department of Interior, Bureau of Reclamation, 1963, p. 74.
[15] Steinour, H. H., "Concrete Mix Water-How Impure Can It Be?" Journal, Portland
Cement Association, Research and Development Laboratories, Vol. 2, No. 3, 1960,
pp. 32-50.
[16] Liebs, W., "The Change of Strength of Concrete by Using Sea Water for Mixing and
Making Additions to Concrete," Bautechnik, 1949, pp. 315-316.
[17] Hadley, Homer, "Letter to Editor," Engineering News-Record, May 1935, pp. 716-717.
[18] "Job Problems and Practice," Proceedings, American Concrete Institute, Vol. 36, 1940,
pp. 313-314.
[19] Dempsey, J. G., "Coral and Salt Water as Concrete Materials," Proceedings, American
[20] Doell, B. C., "The Effect of Algae Infested Water on the Strength of Concrete," Proceedings, American Concrete Institute, 1954, pp. 333-342.
[21] Wuerpel, C. E., "Influence of Mixing Water Hardness on Air Entrainment," Proceedings,
[22] "Water for Use in Mixing or Curing Concrete," Waterways Experiment Station Miscellaneous Paper. Corps of Engineers, U.S. Army, unpublished, 1953.
[23] "Method of Test for the Staining Properties of Water," CRD-C 401-57, Handbook for
Concrete and Cement, Corps of Engineers, U.S. Army.
[24] Engineer Manual, Standard Practice for Concrete, EM 1110-2-2000, Corps of Engineers,
U.S. Army, Dec. 1963, p. 38.
R. E. Carrier 1
Chapter 44--Curing Materials
Introduction
Environment adjacent to and within a freshly placed mass of portland cement concrete has a profound influence on the ultimate strength and
durability which the concrete will exhibit. Temperature and moisture are the
two most significant environmental properties. Generally, moist environments having moderately low temperatures provide optimum curing
conditions. These conditions allow the cement to hydrate to the fullest extent
possible within the physical-chemical limitations which always constrain
complete hydration. For a given concrete--one having a fixed water/cement
(w/c) ratio, cement factor, and other properties--it is usually desirable to advance hydration to the maximum extent possible through maintenance of a
favorable curing environment.
Inadequate or improper curing may result in the physical disruption of the
hydration cement gel products or in a low extent of cement gel formation
(hydration), or both. Numerous studies have shown that improper curing
results in a variety of undesirable effects, such as lower strength, scaling,
dusting, high permeability, low resistance to abrasion, several types of cracking, and low resistance to weathering.
Curing requirements may differ substantially depending on the ultimate
use intended for the concrete. Hence, "use" must be considered in the definition. As a minimum, "curing is the process of maintaining a satisfactory
moisture content and a favorable temperature in concrete during the period
immediately following placement so that hydration of the cement may continue until the desired properties are developed to a sufficient degree to meet
the requirements of service" [1].2 Examples of uses which differ significantly
from the standpoint of curing requirements are mass concrete as used in
dams (a relatively small surface area per unit volume of concrete), structural
concrete as used in beams, columns, etc. (generally moderate to high surface
tVice president, Ackenheiland Associates, Pittsburgh, Pa. 15216.

Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
lntcrnational
www.astm.org
CARRIER ON CURING MATERIALS
775
area per unit volume of concrete), and highways or floor slabs (high surface area per unit volume of concrete). The desired service properties of
these three types of concrete are also widely different. For mass concrete,
minimal strength is required, but low heat of hydration is important. Adequate curing of interior concrete in a mass concrete structure is virtually ensured since it is surrounded by additional concrete. In structural concrete,
uniform strength is the principal requirement. Curing must be such that
hydration proceeds relatively quickly and uniformly throughout the entire
cross section of beams and columns. Adequate curing is particularly important for the cement which surrounds the reinforcing steel. In highway or floor
slab concrete, strength is perhaps the most important and desired property.
High surface strength resists abrasion, cracking, and weathering. In short,
curing requirements as specified in construction practice may differ markedly due to the differing intended uses of concrete. Similarly, a wide range of
acceptable curing methods and materials presently are being used.
There are a number of physical factors which influence the selection and
specification of the curing process and materials. Apart from the uses of concrete described above, several of the more important physical properties
which affect and are affected by the curing process are dealt with below.
Temperature Effects
Hydration of cement is an exothermic process. Heat liberation, commonly
referred to as the heat of hydration, normally amounts to about 100 cal/g of
cement. In massive construction, this amount of heat can be reduced to 60 to
70 cal/g by reducing the cement factor and percentage of high-heatproducing cement compounds, tricalcium silicate and tricalcium aluminate.
Frequently, refrigerant systems consisting of coolant pipes are placed in the
fresh concrete during placement of massive pours. In the Hoover Dam, for
example, loops of 2.5 cm (1 in.) 14 gage tubing were spaced 1.75 m (5.75 ft)
on center for each 1.5 m (5 ft) left of concrete. Cooling water at 10~ (50~
and lower was circulated through the tubing as soon as placement was
complete.
The rate of hydration reaction is controlled to a large extent by
temperature. At temperatures near or slightly below 0~ (32 ~ the rate of
hydration and thus early strength gain is nil. Thus, nonmassive placements
having large exposure areas (for example, thin floor slabs) may remain
plastic for much longer than normal periods at these temperatures. It is important to ensure that freezing does not occur to fresh concrete.
Two conditions are necessary in order for damage to occur. First, the concrete temperature, not the air temperature, must drop below 0~ (32~
Second, the concrete must be at an early enough stage of curing to still contain large quantities of bulk or uncombined water. If this water freezes, the
expanding ice in the plastic concrete causes a permanent increase in the
776
porosity of the concrete, thus reducing the strength and durability of the concrete. Prolongation of freezing conditions can result in excessive heaving in
the concrete. However, the ultimate temperature below freezing is not related
to the magnitude of damage except with respect to the extent of concrete
affected.
Early frost damage in concrete can result in the loss of about half of the
compressive strength. The modulus of rupture may also be reduced by as
much as half the normal value. Durability, as measured by the number of
cycles to a 25 percent weight loss, may be reduced by more than half.
Freezing after a minimum curing period of 24 h at 20 ~ (68 ~ causes only
slight reductions in the compressive strength, provided that no additional
water is supplied to the curing concrete. For virtually no change in flexural
strength, a minimum initial curing time of at least 36 h at 20~ (68~ is
necessary.
A common misconception relative to winter concreting is that calcium
chloride (dissolved in the mix water at the rate of not more than two percent
by weight of cement) acts as an antifreeze. Rather, it acts as an accelerator,
hastening the hydration reaction if and only/fthe temperature of the mixture
is well above freezing.
Another common misconception relative to winter concreting is that there
is no danger of freezing of fresh concrete which has cured for several days. As
hydration of freshly mixed concrete progresses, some free water becomes
combined chemically with the cement and some is drawn into smaller
capillary spaces. The temperature at which this capillary water freezes
becomes progressively lower as the capillary diameter decreases. If the
temperature of fresh concrete of typical mix proportions has been maintained at or above about 13 ~ (55~ for several days prior to freezing, there
is little danger of detectible strength loss due to freezing. However, if fresh
concrete has been maintained for several days at lower temperatures prior to
freezing, damage may result. In this case, the hydration reaction, being
retarded by low temperatures, has consumed very little of the original mix
water. Should this chemically free water then be subjected to freezing, the ice
crystals which form may disrupt the hydration products which have formed,
thereby diminishing the ultimate strength. The mechanisms, water quantities, and curing periods associated with early freezing problems have been
the subject of numerous research projects [2].
Where ambient temperatures fall below 5~ (40~
insulation is considered desirable by many specifying agencies. Generally the insulation
blankets currently in use serve two purposes, to maintain adequate
temperatures in curing concrete and to minimize evaporative losses. Insulation in winter concreting of certain members such as bridge piers, foundations, and the like can cause overheating damage to concrete. Probably more
damage occurs due to curing at too high a temperature than too low, even in
winter concreting. This may occur when the heat of hydration for a relatively
777
massive placement is contained too tightly within the insulation blankets.

This is not to say that it is not desirable to both insulate and heat concrete
placed in the winter. The important criterion is that concrete temperature be
maintained in the range of 5 to 71~ (40 to 160~
In addition to insulation for winter concreting, mix water and sometimes
aggregates are heated prior to mixing with cement. This allows a more
uniform initiation of the hydration reaction and may be particularly valuable
for structural concretes. These techniques do not replace the need for
insulation.
Ordinarily the temperature of curing concrete should not exceed about
71~ (160~
High temperatures may cause rapid evaporation of the mix
water (appears as drying of exposed edges of the member), thus curtailing
the hydration reaction in those areas. Also high temperatures, if not uniform
throughout the curing mass, may cause physical disruption. If large variations occur before the concrete has developed sufficient strength to resist the
resulting thermal stresses, cracking occurs. Flash set, excessive shrinkage,
and random cracking are also related to high early temperatures. For hot
weather concretes exposed to solar radiation, mix water frequently is cooled
to avoid problems associated with rapid set. Evaporative sprays over aggregate stockpiles or ice melted in the mix water are used where the combined effects of solar exposure and heat of hydration raise concrete
temperatures to about 40~ (100~ or higher.
The optimum curing temperature is usually considered about 32~
(73 ~
Curing at lower temperatures may give slightly higher 28-day compressive strengths whereas curing at higher temperatures may increase the
early strength at the expense of 28-day and ultimate strength. Figure 1 shows
some general relationships between strength at various ages for different curing temperatures.
Moisture Effects
The precise mechanism by which hydrating cement gains strength is still
incompletely understood. It is known that cement gel is the precipitated product formed during the chemical combination of cement and water. Cement
gel precipitates only into water-filled space. This is demonstrated most obviously by the inability of cement gel to fill air cavities such as entrained air
bubbles in concrete. Hence, curing practices should be such as to keep concrete as nearly saturated as possible until the originally water-filled space has
become filled with hydration products.
It is also important that there be no interruption in the availability of water
to hydrating particles during the early curing period. As shown in Fig. 2, if
curing water is denied for a period sufficient to allow curing concrete to dry,
it never again regains the strength of continuously moist cured concrete, even
after long periods of subsequent moist curing. Apparently, once the capillary
778
TESTS A N D PROPERTIES OF OTHER MATERIALS
>.
I10
C3
~00
- -s-s ~
9O
~z
tu
u
.+#~~//:"""+P
uJ
j-
= ==
.~f
70
'8
"1-
60
"z ,., 5 0
li:
~or) o
40
~f[,,
m ~
30
i/
(n
ILl
a.
n.
20
>
80
/ ,'
,
'
"
'
//
/
I0
/
. ~ -~--- - -
'"
,oso___ _ _ . / - ~_...~--
~
_
. a~ ~
Concrete Mixed,Ploced
And Cured M o i s t At
Tempero lures Indicoted
'
0
I
28
AGE
OF T E S T
, DAYS
FIG. 1--Relationships between strength and curing age for concrete cured at various
temperatures (from Portland Cement Association Manual, Design and Control of Concrete
Mixtures).
continuity to interior areas of hydrating cement is broken, it is never again

fully regained by the addition of moisture.
Powers [3] and others [4-6] have shown that portland cement virtually
stops hydrating when concrete dries to the point where the internal relative
vapor pressure falls below about 0.80 of the saturation pressure (or about 80
percent internal relative humidity). The hydration rate for relative vapor
pressures between 0.30 and 0.80 is relatively low (see Fig. 3) and is negligible
below 0.30. For maximum hydration, concrete must be kept saturated or as
nearly so as is practical.
Generally, the volume of mix water added to concrete is sufficient to allow
adequate curing. Maximum hydration for an average cement (which is usually distinct from complete hydration) requires about 0.44 g of water per
gram of cement, plus the curing water which must be added to keep the cement saturated. If no curing water is added, self-desiccation which is caused
by the consumption of water during hydration occurs.
In some respects, self-desiccation is desirable. Air dried concrete cylinders
are approximately 15 to 30 percent stronger than the same concrete when
saturated (either before drying or after being soaked for 15 to 30 hours).
779
Q
1%1
tlL
0
MOIST C JRED
ENTIRE T I M E ' ~
160
I.Z
hi
U
~ 140
12:
W
Q.
0
U
~'"
~Z 1 2 0
./
//<
100
3:
(.9
Z
I,I
e:
60
]
::~.....~- -
f ]
/
,,,
[/
-.
IN AIR
MOIS" CURED, FTER

I YEA IN AIR., ~. J
~..
IST CURI!D AFTER

IONTHS N AIR
MOI!|T CURl O
_________-
NTIRE T ME
AFT ',R 2B DJYS

IN AIR
~- 4O
o
0
W 9
20
0
O
50
IOO
150
200
250
AGE AT TEST-DAYS
3OO
350
400
0
S,J
FIG. 2--Strength of concrete increases while moisture and unhydrated cement are still
available. Note delayed strength rise when moist curing is resumed (from Portland Cement
Association Manual, Design and Control of Concrete Mixtures).
Another phenomenon, known as autogenous healing, is enhanced by selfdesiccation of curing concrete. Autogenous "healing" across fine cracks or in
weakly bonded areas within the matrix (between cement paste and aggregate, for example) may occur late in the life of the concrete member due
to hydrating of previously unhydrated cement. If substantial quantities of
unhydrated cement are available because of early cessation of the hydration
reaction, autogenous healing is enhanced. One final advantage of selfdesiccation is the improvement in freeze-thaw resistance. Water in larger
capillaries is consumed by self-desiccation. The remaining water is drawn
preferentially into the smaller capillaries where it is much less susceptible to
freezing.
Efforts to prohibit the evaporative losses of large quantities of mix water
during the early stages of curing by the application of membrane coatings are
desirable, especially where concrete members are in contact with soil. Membrane curing does not assure as full an extent of hydration as water or moist
curing, but results achieved over the years have proven satisfactory. Membrane curing has also become much more economical than the former wet
covering techniques for many continuous-production concrete uses such as
highways.
780
.40
.36
IZ
bJ
,,,
u
u.
0
.32
,,~
28
"~ O T A L
WATER
N O N - EVAPORATED
WAT ER
(.9
tr
"Q_
'
.2 4
f,/.)
n,.
c9
20
Z
w
U
>m
Q.
D .12
Z
ILl
Fn,-
.08
.04
/
0
0
FIG.
20
40
60
RE LATIVE HUMIDITY
80
I00
3--Amounts of water taken up by dry cement exposed to water vapor 6 months (from
Ref 4).
Moisture losses other than the physical and chemical uptake of water by
the hydration process occur by evaporation, bleeding, absorption by certain
forming materials, and by gravity-induced flow. Control of each of these
losses may be achieved by a different technique. For example, excessive
bleeding may render a freshly applied membrane-forming compound ineffective by disrupting the continuity of the membrane. Absorbent forming
materials may be sprayed with a hydrophobic material or oiled to retard
moisture loss. Evaporation may be controlled by reducing exposed surface
781
area (for example, with sheet coverings) or by increasing the relative humidity in the air immediately adjacent to the concrete member (for example, fog
sprays used on bridge decks). Gravity losses may be partially controlled by
placing an underlayment barrier (for example, sheet material) beneath the
fresh concrete. However, gravity and bleeding losses may be controlled most
effectively by altering mixture proportions.
The most significant single factor affecting the magnitude of volume
change and associated cracking in concrete is moisture content. Early
volume changes due to bleeding are usually not serious (excluding prestressed concrete) since significant stresses cannot develop in plastic concrete. However, high early evaporative losses, particularly from high unit
water content mixes can cause drying shrinkage and plastic shrinkage
cracks. Plastic shrinkage cracks are relatively shallow, have relatively wide
openings at the exposed concrete surface, and are relatively short. They are
usually good indicators of inadequate early curing, especially where a set
retarder has been used in the concrete. Drying shrinkage cracks are a random pattern though they frequently reflect edge drying. As their name implies, they are caused by rapid drying of fresh concrete surfaces. Research
has shown that cracks associated with early volume change can be greatly
reduced through adequate curing.
Not all curing concretes require additional effort by man. Concrete curing
behind forms, for example, typically does not require additional moisture.
Studies [8] show that even for moderately severe exposures (high
temperatures, high wind velocity, and low relative humidity) the internal
relative humidity 5 cm (2 in.) below exposed concrete surfaces remained sufficiently high to allow curing to proceed for at least 28 days (see Fig. 4). Conversely, the upper 0.65 cm (1A in.) of concrete dried to the point where curing
activity virtually ceased in slightly more than 1 day. This experiment points
out the importance of adequate curing for surface concrete which, in service,
is exposed to abrasion, weathering, or other deteriorative mechanisms.
Methods of Curing
Three general methods for curing are in common use in the field: (a) low
permeability sheet materials, usually of polymer materials, (b) low
permeability membrane-forming films, sprayed or otherwise applied like
paints, and (c) wet coverings such as burlap or wet sprays. A fourth method,
used extensively at concrete products manufacturing plants, consists of immersion in lagoons or placement in enclosures maintained at high relative
humidity. These facilities usually include temperature control intended to accelerate the hydration process. Temperatures are raised at a uniform rate of
not more than about 20~
(40~
for larger units and about 3S~
(6S~
for smaller sizes until a maximum of 71~ (160~ is reached.
Low permeability sheet materials and membrane-forming compounds are
782
I00
so
,oo ~ " > < " ~ S ~ ' - -
a ~
~ -
,-V,/
40
~
j
' IX/ ~
. . . . .
~'0o10
11
NOTE:
100% = 200sq. rt.lgal,
20
o/
'/,
V,
,V~
DEPTH
BELOW
SURFACE
I
2
(INCHES)
FIG. 4--Moisture distribution with depth at 28 days for specimens sprayed with various applications of white pigmented curing compound and subjected to a severe exposure (from
Ref8).
used principally where the supply of water is limited or where an excessive investment in piping or water trucking is required to maintain wet coverings.
Wet coverings, soaker hoses, or fog sprays frequently are used where the
strength of the surface concrete may be critical to the life of the entire structure, such as for concrete bridge decks.
Water resistant paper, reinforced with fibers and treated to reduce
shrinkage, is effective in retarding moisture loss. It also may provide some
temperature control. It is applied either immediately after casting fresh concrete, or it may be applied subsequent to a preliminary wet curing by fog
spray or wet burlap. Some of these sheet materials have a burlap-like texture
on their underside. Plastic films, such as polyethylene or combinations of
plastic with fibers or fabric for strength, are used in the same manner as
water resistant paper products.
Liquid membrane-forming compounds typically are sprayed on exposed
surfaces of freshly placed concrete after the water sheen has disappeared.
They dry rapidly resulting in a continuous film which retards moisture egress
from the curing concrete. After several months, depending on the use and exposure of the sprayed surfaces, most curing compounds flake off (with the
possible exception of asphalt emulsion or similar flexible coatings).
Temperature control for concrete placed in the summer months or in
warm climates is achieved commonly through the use of white or reflecting
curing materials. White sheet coverings and white pigmented curing compounds usually are specified for concrete exposed to solar radiation.
783
Effectiveness of Curing Methods

Of the three curing methods cited, the wet curing methods, if properly carried out, produce concrete having superior strengths. Sheet materials and
membrane-forming compounds are not as effective principally because less
moisture is available to promote hydration. Sheet materials and membrane
films each have finite permeability, allowing small quantities of mix water to
escape from the hydrating concrete. The moisture-retaining effectiveness of
sprayed membranes is not as heavily dependent on the application rate as it
is on the constituent ingredients of the film and the continuity of coverage the
film exhibits on a microscopic scale.
Experiments [8] show that for lightly textured surfaces, applications of
from one half to two times the normal application rate of white pigmented
curing compound are effective in maintaining the 28-day internal relative
humidity of curing concrete at or above the 80 percent level, at least for
depths of 2.5 cm (1 in.) or greater. Experience has shown that the time of
application may affect moisture retention properties more significantly than
the thickness of the film. If applied too early--before the water sheen has
disappeared--the membrane becomes scaly and exhibits discontinuities
upon drying. Excessive bleeding can also disrupt the continuity of an otherwise appropriately applied membrane, thus rendering it ineffective in
moisture retention.
While solar reflectance is desirable in some instances, excessive pigmenting of the membrane vehicle may reduce moisture retention efficiency.
Hence, there is a trade off between reflectance and moisture retention
properties.
Two types of compounds are used widely, white pigmented and clear or
translucent. It is most difficult to determine the field application rate for
sprayed compounds. No widely accepted standard test is available. At least
one state highway department has attempted to measure the field application
rate by laying an absorbent fiat-lying paper material of known weight on the
pavement just prior to passage of the curing compound spraying machine.
After spraying the surface, the sprayed paper is weighed again to determine
the application rate. Still another method suggests visually comparing the
whiteness of field sprayed concrete with that of a laboratory prepared mortar
plaque sprayed at the specified rate. In practice, however, this method is
used rarely.
It is generally fairly easy to detect nonuniform applications of white
pigmented curing compounds by the alternating white and gray colors of
freshly sprayed surfaces. To aid in determining uniformity of coverage of
clear or translucent compounds--for that matter to aid the person doing the
spraying in differentiating what has been sprayed and what has not--a
fugitive dye is frequently added. The dye is readily visible immediately after
spraying, but becomes inconspicuous after a week's exposure or less in
sunlight.
784
Other properties which may affect a membrane's effectiveness are its

viscosity, curing period, and sag characteristics. This is particularly true for
concrete members which do not lie horizontally or for those with deep textures such as is now done in pavement construction. Curing compounds may
flow from the ridges of the texture to the sag areas.
Tests and Their Significance

Water retention capability is perhaps the most significant property of a
curing material. ASTM Test for Water Retention by Concrete Curing
Materials (C 156) provides a laboratory technique for estimating the efficiency of membrane-forming compounds and sheet materials. In this test,
moisture loss is measured by weight for a freshly mixed mortar plaque
specimen subjected to 72 h of exposure in a cabinet held at 37.8 ___ 1.1~
(100 ___ 2~ and a relative humidity of 32 +_ 2 percent. The mortar plaques,
made with standardized mix proportions and mixing procedures, are finished so that their surfaces have a very light texture. Curing compounds are
sprayed or brushed on the surface at the rate of 0.2 dm3/m2 (1 gal/200 ft2).
The same plaque specimens are used for testing the moisture retention efficiency of sheet materials. Sheet materials are laid over the surface of the
plaque and sealed at the edges with a sealing compound.
ASTM Specification for Sheet Materials for Curing Compounds (C 171),
and ASTM Specification for Liquid Membrane Forming Compounds for
Curing Concrete (C 309), each allow a maximum moisture loss of 0.055
g/cm 2 through the membrane during the aforementioned 72-h test. ASTM
Specification C 171 identifies three types of sheet materials: waterproof
paper, polyethylene film, and white burlap-polyethylene. Each of these are in
common use. Minimum tensile strength properties and minimum percent
elongation are specified. These properties are tested according to ASTM
Tests for Wet Tensile Breaking Strength for Paper and Paper Products (D
829), and ASTM Tests for Tensile Properties of Thin Plastic Sheeting (D
882). A minimum percentage reflectance is specified for white or reflecting
sheets and compounds as tested by ASTM Test for 45-deg, 0-deg Directional
Reflectance of Opaque Specimens by Filter Photometry (E 97).
In addition to the above described water retention and reflectance criteria,
membrane curing compounds must dry to the touch after a 4-h exposure at
23 + 1.7~ (73.4 ___ 3~
50 percent relative humidity, and an air velocity
of approximately 183 m/min (600 ft/min) passing horizontally across the
specimens. Further, compounds must not be tacky or track off concrete
when walked upon after 12 h. Acceptable compounds must possess a certain
long term pigment settling resistance and must adhere to but not be deleterious to concrete surfaces. Both resin and petroleum or natural (the so-called
"wax bases") vehicle solids are used. Membrane compounds must have a
consistency which allows them to be sprayed readily at 4 ~ (40 ~
These test methods are intended principally to assess the characteristics of
785
concrete curing materials. They do not provide information on whether curing of field constructed members is adequate. The tests are significant only
as base or minimum criteria for curing materials. The fact that a particular
curing material is acceptable by ASTM standards does not ensure that its
performance in the field will be adequate. For example, a membraneforming compound may perform adequately on a pavement having its surface prepared with a burlap drag but inadequately on rougher pavement surfaces prepared with a rake of textured float. Evidence suggests that even the
application of a second coat of curing compound on heavily textured surfaces
may not assure adequate moisture retention,
No widely accepted standard test is available for testing the moisture retention performance of concrete constructed in the field. A small surface
mounted device for visually estimating relative humidity through color
change at the surface of concrete members has been proposed [8], but has
not gained wide acceptance. Abrasion tests and impact or penetration
resistance may provide valuable performance information as to the adequacy
of curing at the surface.
Many variables affect the field performance of curing materials. Among
the most important are the type of water used with wet curing methods, the
condition and age of the surface concrete at the time of application of
membrane-forming compounds, the air flow under or around sheet materials
(particularly at edges or corners), the heat transmission/absorption properties of sheet material or insulation/curing materials, and the texture and exposure conditions of the concrete surface to which curing materials are being
applied. Sound judgment by knowledgeable and experienced concrete
specialists must be exercised to achieve strong durable concrete, particularly
where surface properties are important. Figure 5 profoundly demonstrates
the importance of curing for surface concrete.
FIG. 5--Two specimens made from same concrete mix showing the effects of curing on abrasion resistance of concrete (from Ref 9). (a) Membrane curing compound. (b) No curing.
786
References
[1] "Curing of Concrete Pavement," Current Road Problems, No. 1-R Highway Research
Board, revised edition, March 1962.
[2] Cady, P. D. and Carrier, R. C., "Methods For Analysis of Hardened Concrete: Early Frost
Damage," Pennsylvania State University Special Report No. 1, Clearinghouse, Springfield,
Va., PB 183-602.
[3] Powers, T. C., "A Discussion of Cement Hydration in Relation to Curing of Concrete,"
Proceedings, Highway Research Board, Vol. 27, 1947, pp. 178-188.
[4] Jesser, L. and Berg, U., Hi~ttenmiinn Jahrb. rnontan. Hochschule Leoben, Vol. 75, No. 2,
1927, p. 69.
[5] Rodt, V., Zement, Vol. 14, 1925, p. 520.
[6] Gause, G. R. and Tucker, J., "Method for Determining the Moisture Condition in Hardened
Concrete," Journal of Research, National Bureau of Standards, Vol. 25, 1940, pp. 403-416.
[71 "Design and Control of Concrete Mixtures," Bulletin No. T-12, Portland Cement Association, 1952, p. 7.
[8] Carrier, R. E. and Cady, P. D., "Evaluating Effectiveness of Concrete Curing Compounds,"
Journal of Materials, American Society for Testing and Materials, Vol. 5, No. 2, June 1970,
pp. 294-302.
[9] Kennedy, H. L. and Prior, M. E., "Abrasion of Resistance," Significance of Tests and
Properties of Concrete and Concrete Aggregates A S T M STP 169, American Society for
Paul KEeger 1
Chapter 45--Air-Entraining
Admixtures
Introduction
The use of intentional air entrainment in concrete is a well-established
means of greatly enhancing the ability of concrete to resist the potentially
destructive effect of freezing and thawing. Its use should be mandatory
when concrete is to be exposed to such an environment, particularly when
chemical deicers are being used, as on pavements and bridge decks.
A thorough survey of the early development of air entrainment is presented by Gonnerman [1]. 2 The following paragraph from Gonnerman's
report is of particular significance.
"These projects (test roads constructed in 1935-1937) revealed no relationship between surface scaling and composition of the cement, but
they did show clearly that portland cement that inadvertently contained
"crusher oil" reduced surface scaling as did many of the blends of portland
and natural cement that contained tallow added during grinding of the
natural cement. Laboratory tests disclosed that the beneficial effect of the
crusher oil and tallow was due entirely to the additional air entrapped in
the concrete by these air-entraining agents."
Other investigators [2-6] came to similar conclusions during the late
thirties and early forties. In these early instances, the air entrainment was
not intentional but resulted from the pre~ence of the crusher oil or the use
of the tallow as a grinding aid during the production of the cement. These
were the forerunners of the materials now used to produce air-entraining
cements and called additions. Materials similar to presently used additions
are called air-entraining admixtures when added with the other concrete
ingredients at the time of mixing. This report is concerned with the class
of materials. What are these materials; how do they function, both as to
the process of entraining air and enhancing durability; how can they be
1Director, Concrete Materials Research Department, Research and Development Laboratories, Portland Cement Association, Skokie, Ill. 60076.
2The italic number in brackets refer to the list of references appended to this paper.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
788
specified and tested to ensure adequate performance? These are some of

the questions which we will attempt to answer in this chapter.
Definitions
It will be helpful to define certain of the terms used in discussing air
entrainment. This listing is by no means complete, but the following definitions appear appropriate.
Air Entrainment--The introduction of air in the form of discrete air
voids or bubbles dispersed throughout the mixture as a result of the use
of air-entraining agents.
Entrained Air--The air, made up of discrete air voids, which becomes
part. of a mixture during the process of air entrainment.
Entrapped Air Voids--Air voids not resulting from intentional air entrainment. Such voids are larger than those resulting from intentional air
entrainment and are at times referred to as natural air voids.
Entrained Air Voids--Air voids resulting from the use of intentional air
entrainment. Such voids are generally spherical in shape and considerably
smaller than the natural air voids.
Air-Entraining Admixture--A material added to cementitious mixtures at
the time materials are batched for mixing, the use of which results in intentional air entrainment. (See ASTM Definition of Terms Relating to
Concrete and Concrete Aggregates (C 125) for the definition of an admixture.)
Air-Entraining Addition--Air-entraining material interground with, or
to be interground with, hydraulic cement. (See ASTM Definition of Terms
Relating to Hydraulic Cement (C 219) for a definition of an addition.)
Air-Entraining Agent--A material the use of which results in intentional
air entrainment when included in a mixture; a term that should be used
only when it is intended to refer to materials that can be used both as airentraining additions or admixtures.
Materials Used as Air-Entraining Admixtures
T h e r e are many materials capable of functioning as air-entraining admixtures. In an extensive evaluation program, the Bureau of Public
Roads [7] separated 27 commercial air-entraining admixtures submitted
for test into the following classifications: (1) salts of wood resins (pine wood
stumps), (2) synthetic detergents (petroleum fractions), (3) salts of sulfonated
lignin (paper pulp industry), (4) salts of petroleum acids (petroleum refining), (5) salts of proteinaceous materials (processing of animal hides),
(6) fatty and resinous acids and their salts (paper pulp and animal hide
processing), and (7) organic salts of sulfonated hydrocarbons (petroleum
KLIEGER ON AIR-ENTRAINING MIXTURES
789
refining). Of these 27 materials, 17 were liquids, S were powders, 4 were in

flake form, and 1 was semisolid.
Some materials, such as hydrogen peroxide and powdered aluminum
metal, can be used to entrain gas bubbles in cementitious mixtures but
are not considered as acceptable air-entraining admixtures, since they do
not produce an air void system which will enhance resistance to freezing
and thawing.
Generally, only small quantities of air-entraining admixtures are required to entrain the desired amount of air. These are of the order of 0.05
percent of active ingredient by weight of the cement.
Function of Entrained Air
The major reason for the use of intentionally entrained air is to provide
concrete with a high degree of resistance to freezing and thawing and the
use of chemical deicers. (The discussion to follow will also be applicable to
the use of air-entraining cements, in which the air-entraining agent or
material is used as an addition during the grinding of the cement clinker.)
There are numerous other advantages, also, to the use of intentionally entrained air. For example, plasticity and workability are increased, enabling
a reduction in water content. Uniformity of placement and consolidation
can be achieved more readily, thus reducing segregation, and bleeding is
reduced. These and other advantages are discussed in detail in a number
of the appended references, for example [8, 9].
To achieve the improvement in frost resistance, the intentionally entrained
air must have certain characteristics. Not only is the total volume of importance, but more importantly the size and distribution of the air voids
must be such as to provide efficient protection to the cement paste. Powers ~
[10-12] contributions to the understanding of how entrained air functions
in providing increased frost resistance have been outstanding.
This work developed the concept of internal hydraulic pressure created
by the resistance to flow or movement of excess water volume produced
during the freezing process as being the mechanism responsible for distress.
To keep this internal pressure below the tensile or rupture strength of the
paste, Powers showed that the air voids must be well distributed throughout the matrix (cement-water paste component) and sufficient in number
so that each void provides protection to the cement paste surrounding it,
and the protected volumes overlap to leave no unprotected paste. Each
void provides space into which the excess water can move during freezing,
and the voids should be sufficiently close to one another so as to avoid high
internal hydraulic pressure in the paste during movement of water to the
air void.
Later work by Powers and Helmuth [13] indicated that, in addition to
790
the generation of hydraulic pressure during freezing, another important

factor may be the diffusion of gel water to capillary cavities contributing to
the growth of ice bodies in these cavities resulting in the development of
expansive forces. This factor may be of significance in nonair-entrained
concrete. The presence of an adequate air-void system provides spaces containing ice which competes with the ice in the capillary cavities for this unfrozen gel water, and thus avoids the development of these expansive
forces.
Powers [11] developed the concept of void spacing factor to characterize
an air-void system and analyzed laboratory freezing and thawing data
available at that time to show that the void spacing factor for frost resistance should be about 0.01 in. or less. The void spacing factor is defined by Powers as the average maximum distance from a point in the
cement paste to the nearest air void. This is an indication of the distance
water would have to travel, during the freezing process, to reach a protective air void. More recent work by Mielenz et al about 0.152 to 0.203 mm
(0.006 to 0.008 in.) is required for extreme exposures. Extensive freezing
and thawing tests by Klieger [15] provided further substantiation of the
void spacing factor concept. These and other tests [16] called attention to
the need for different volumetric air-content requirements for concretes made with different maximum sizes of coarse aggregate. Table 1
shows this effect of maximum size of aggregate on the optimum air content, along with void spacing factors. Although the total air contents of
the mixtures shown in Table 1 vary through a wide range, the air content of the mortar fraction is essentially constant at about 9 percent. As
the maximum size of coarse aggregate increases and consistency and cement content held constant, less mortar is required in the mixture [15];
therefore, change in total concrete air content with change in maximum
size of coarse aggregate is to be expected.
Most of the early field and laboratory work on air-entrained Concretes
dealt with paving-type concrete in which the coarse aggregate was generally about 38.1 mm (1189 in.) maximum size. In such concretes without
intentional air entrainment, the air content (composed of entrapped air
voids which are too large to be effective with respect to improving frost resistance or workability but which are included necessarily when air content
is expressed in volumetric terms) may range up to as high as 1.0 or 2.0 percent. Field and laboratory tests indicated that when such concretes were
provided with about 3 percent additional entrained air the resistance to
freezing and thawing and deicer chemicals was enhanced greatly. This led to
the specification that such concrete should contain 4.5 ___ 1.5 percent air,
based on the premise that the coarse aggregate would have a maximum
size of 38.1 mm (11/2 in.) and that the average amount of entrapped air
would be about 1.5 percent and composed of relatively large and ineffective air voids. For these concretes, the provision of an additional - 3 . 0
791
TABLE 1--Characteristics of concretes at aptimum a a& content (Tables 17 and 18, of R e f 15).
Max
Size of
Aggregate,
mm (in.)
Total
Air
Content,
%
Mortar
Air
Content,
%b
Paste
Air
Content,
%b
BubbleSpacing
Factor,
mm (in.) c
Net
w/c
Ratio, by
Weight
Cement Content: (376 lbs/yd 3)

63.5
38.1
19.1
9.5
No. 4
(21/2)
(11/2)
(3/4)
(3/8)
(No. 4)
4.5
4.5
5.5
8.5
12.5
8.8
8.3
8.9
11.1
12.5
18.5
18.2
19.8
26.2
31.6
0.305
0.330
0.330
0.229
0.152
(0.012)
(0.013)
(0.013)
(0.009)
(0.006)
0.56
0.58
0.67
0.75
0.91
0.178
0.203
0.229
0.279
0.305
(0.007)
(0.008)
(0.009)
(0.011)
(0.012)
0.42
0.42
0.47
0.53
0.67
0.178
0.203
0.178
0.203
0.203
(0.007)
(0.008)
(0.007)
(0.008)
(0.008)
0.36
0.36
0.38
0.42
0.5J

63.5
38.1
19.1
9.5
No. 4
(21/2)
(11/2)
(3/4)
(3/8)
(No. 4)
4.5
4.5
5.0
6.5
9.0
9.1
8.5
8.3
8.7
9.0
16.7
16.4
16.9
19.7
23.0

63.5
38.1
19.1
9.5
No. 4
(21/2)
(11/2)
(3/4)
(3/8)
(No. 4)
4.5
4.5
5.5
7.0
10.0
9.2
8.4
9.2
9.6
10.0
14.7
14.3
16.8
19.4
23.4
a Optimum determined from relation between expansion during 300 cycles of freezing and
thawing and air content of concrete.
b Calculated as follows
Mortar air content,
% =
Paste air content,
% =
C+W+S+A
A
C+W+A
100
100
where
C = absolute volume of cement,
W = volume of net mixing water,
S = absolute volume of fine aggregate (saturated surface dry), and
A = volume of air.
CCalculated from linear traverse data using method outlined in Bulletin No. 33, Portland
Cement Association Ill].
percent of intentionally entrained air by the use of an air-entraining adm i x t u r e p r o v i d e d a n a i r - v o i d s y s t e m well d i s t r i b u t e d t h r o u g h o u t t h e m a t r i x

and containing a sufficient number of air voids to meet the void spacing
f a c t o r r e q u i r e m e n t s w h i c h l a t e r s t u d i e s d e m o n s t r a t e d as n e c e s s a r y .
S p e c i f i c a t i o n s b a s e d o n v o l u m e o f e n t r a i n e d a i r still r e m a i n t h e o n l y
792
practical way of specifying intentionally entrained air, despite the fact

that other parameters are of more importance than volume alone. Examples
of such use are found in American Concrete Institute (ACI) Standard
Recommended Practice for Selecting Proportions for Concrete (ACI 211.1)
and specifications of agencies such as the Corps of Engineers, U.S. Army,
and the Bureau of Reclamation. For example, the Corps of Engineers,
U.S. Army, Guide Specifications for Concrete, CE 1401.01, October 1953,
Section -08c, Entrained Air Content, states:
"The total calculated air content of that portion of the concrete containing aggregate smaller than the 38.1-mm (ll/5-in.) square mesh sieve
shall be between 4 and 7 percent of the volume of the concrete based on
measurements made on concrete immediately after discharge from the
mixer."
Although this was directed primarily to mass concrete, it was a recognition of the need for considering the effect of the maximum size of coarse
aggregate on the required air content of concrete.
Specifications and control tests will continue to be based on the volume
of air entrained in the concrete, rather than on the size and distribution
of the air voids in the cement-paste matrix, until a means is developed
for readily determining other air void parameters directly on the freshly
mixed concrete in the field. It would be highly desirable to have a test
method available that could provide a measure of size and distribution of
air voids within a few minutes after completion of mixing. Not having such
a test method available, it is indeed fortunate that concretes having total
air contents in the range of the optimum air contents shown in Table 1
will have the proper size and distribution of air voids when the air-entraining admixture used meets the requirements of ASTM Specification
for Air-Entraining Admixtures for Concrete (C 260).
The importance of size and distribution of air voids, as contrasted with
total volume of voids alone, can be seen in the results of a study made in
the laboratories of the Portland Cement Association. Air-entrained concretes were prepared using an acceptable proprietary air-entraining admixture and four nonproprietary materials which exhibited a potential
for entraining air. A nonair-entrained concrete was also included in these
tests. In addition to the determination of air content of the freshly mixed
concrete as de cribed in ASTM Test for Air Content of Freshly Mixed
Concrete by the Pressure Method (C 231), the air content, void spacing
factor, specific surface, and number of voids per lineal inch of traverse
were determined on the hardened concretes as described in ASTM Recommended Practice for Microscopical Determination of Air-Void Content,
Specific Surface, and Spacing Factor of the Air-Void System in Hardened
Concrete (C 457). Additional refinements of the technique enabled the
determination of the total number of air voids per unit volume. The results of these measurements and the performance of the concretes when
frozen and thawed while immersed in water are shown in Table 2.
1.8
6.0
6.0
5.0
5.8
5.2
None
A
B
D
E
Fa
1.1
4.1
4.1
3.2
3.5
3.9
Air, %
0.031
0.157
0.193
0.130
0.201
0.378
(0.8)
(4.0)
(4.9)
(3.3)
(5.1)
(9.6)
Number
of Voids/
mm (in.)
0.787
0.330
0.254
0.330
0.220
0.152
(0.031)
(0.013)
(0.010)
(0.013)
(0.009)
(0.006)
Void
Spacing
Factor, mm (in.)
11 889
15 236
18 897
16 377
22 716
38 976
(302)
(387)
(480)
(416)
(577)
(990)
Specific
Surface, m2/m 3
(in.2/in. 3) of air
a A commercial air-entraining admixture meeting the requirements of ASTM Specification C 260.
Air Content,
(ASTM C 231),
%
AirEntraining
Admixture
4 881
6 712
13 425
15 866
47 598
231 279
(0.08)
(0.11)
(0.22)
(0.26)
(0.78)
(3.79)
No. of Voids/m 3
(in. 3) Concrete,
millions
Air-Void Characteristics (ASTM Recommended Practice C 457)
TABLE 2--Influence of air-void characteristics on resistance of concrete to freezing and thawing.
19
29
39
82
t00
550
Freezing and
Thawing
Cycles for
0.10%
Expansion
-,4
t,D
r,.o
x
--I
r"
-n
rn
Go
z
z
63
--4
-n
rn
63
m
I--
7~
794
It is apparent from these data that an air-entraining admixture must

not only be capable of entraining some volume of air but also that the
air void system must be characterized by a large number of small, well
distributed air voids in order to provide a high degree of frost resistance.
Factors Influencing Amount and Character of Entrained Air
The works of Mielenz et al [14] and Bruere [9,17-19] have made significant contributions to the understanding of the mechanism by which airentraining admixtures function and the influence of a number of different
variables. Mielenz and his co-workers dealt extensively with the origin,
evolution, and effects of the air-void system in concrete. They showed that
in the concentrations normally used in concrete, air-entraining admixtures
are positively adsorbed at air-water interfaces and that the surface tension
of the water is decreased about 25 percent. This adsorption at air-water
interfaces produces a "film" of air-entraining admixture which influences
the air-retention properties of discrete bubbles formed during mixing.
For some air-entraining admixtures, the calcium salt of the active constituent in the admixture may be only slightly soluble in water. In such instances, the film at the air-water interface may include a precipitated
solid or gelatinous film enclosing each air bubble. Admixtures which produce a relatively insoluble precipitate in portland cement concrete include
sodium soaps of wood resin, such as neutralized Vinsol resin and sodium
abietate; sodium soaps of lignin derivatives, rosin, or fatty acid; or triethanolamine salts of sulfonic acid. Many calcium salts of sulfonic acids
are soluble in water and many air-entraining admixtures in which the
surface-active constituent is a sulfonate would not form such precipitated
films around the air bubbles.
The amount and character of the air entrained in concrete is influenced
by numerous factors, some of which are: (1) concentration of the air-entraining admixture and its influence on surface tension; (2) the presence
of other chemical admixtures, such as water reducers and retarders; (3)
time of mixing; (4) speed of mixing, that is, rate of shearing action; (5)
consistency of mixture; (6) temperature of the mixture; (7) water cement
(w/c) ratio and water content of the mixture; and (8) the gradation of the
solids in the mixture, including the cement. Mielenz et al [14] have concluded from theoretical considerations of pressure in air voids due to hydrostatic pressure of overlying concrete and the curvature of the air-water
interface (that is, bubble size) and the solubility of air bubbles in water
that both the total volume of air and the size distribution of the air
bubbles can change in the unhardened concrete due to interchange of air
between bubbles and dissolution of air. Their tests on air-water foams and
dilute cement pastes tended to corroborate these conclusions. However,
Bruere [19] has shown that such changes do not take place to any signifiCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:47:39 EDT 2014
795
cant degree in air-entrained pastes after cessation of mixing, although

such interchange is of significance during the mixing process and may be
for a few minutes thereafter.
Further information on the influence of numerous variables on the
characteristics of air entrained in concrete is of interest and is presented
in the following paragraphs.
Type and Amount of Air-Entraining Admixture

Mielenz and his co-workers theorize that the type of organic material
which constitutes the active ingredient in the air-entraining admixture
influences the amount and character of entrained air voids by its effect
on: (1) surface tension; (2) the elasticity of the film at the air-water interface, (3) transmission of air across the air-water interface, and (4) adhesion
of the air bubbles to particles of cement or aggregate. All of these factors
will be operative during the mixing operation.
At the same volumetric air content, different air-entraining admixtures
will produce air-void systems having different specific surfaces, number
of air voids per unit volume, and spacing factors. Part of the differences
occur during the mixing operation and part during handling, placing,
and compaction. The potential importance of these differences with respect
to freezing and thawing durability was illustrated by the data in Table 2.
Increasing the amount of acceptable air-entraining admixtures will
increase the volume of air-entrained, increase the specific surface of the
air-void system, and decrease the spacing factor.
The presence of other chemical admixtures, such as water reducers or
retarders, may result in a synergistic effect which operates to increase the
amount of air entrained significantly.
Fine Aggregate--Changes in grading of sand may alter the volume and
nature of air in the mortar [20]. Appreciable increase in the quantity of
very fine particles of sand will decrease the amount of entrained air and
may reduce the maximum and median size of the individual air bubbles
[21]. Appreciable increase in the quantity of the middle sizes of sand will
tend to increase the air in the mortar. Sand gradation is of more importance
in leaner mixes. In the richer mixes, the influence of gradation on entrained air is not as marked.
Sands have been reported as containing organic matter difficult to remove by ordinary measures which cause entrainment of large quantities
of air in concrete. Such sands are relatively rare, but concretes made with
them illustrate the importance of the distribution of the air as compared
with simple volume. Macnaughton suggests [22] this type of entrained air
be termed "accidental air" and since "it may be unstable in nature and
vary in size... [it] should not be confused w i t h . . . 'entrained air'."
Cement--As the cement content increases, the air-entraining potential
796
of an admixture will tend to diminish, and an increase in the fineness

of cement will result in a decrease of the air entrained in the mortar [2325]. Some regular (nonair-entraining) cements naturally entrain more air
than others, and these require less air-entraining admixture to develop a
given mortar air content.
Water--Increase in w/c ratio is likely to result in an increase in air
content. Although the volume of air entrained may increase, the specific
surface of the air voids generally decreases and the spacing factor increases [14]. Nondegradable detergents present in water can result in excessively high and variable air contents.
Consistency--Within the normally used range, increase in initial slump
is accompanied by an increase in air content in concrete mixtures [8,23,25].
Work by Klieger [16] indicates that the optimum mortar air content remains at about 9.0 percent.
Mixing--The effect of mixing action on the amount of air entrained
varies with the type and the condition of the mixer [8,26]. The amount
of air entrained by any given mixer will decrease appreciably as the blades
become worn, or as the mixing action is impaired if hardened mortar is
allowed to accumulate in the drum and on the blades. An increase in entrained air will occur if the mixer is loaded to less than rated capacity,
and a decrease will result from overloading the mixer. A stationary mixer,
a paving mixer, and a transit mixer may develop significant differences
in the volume of air entrained in a given concrete mixture. The air content will increase with increased time of mixing up to about 2 min in stationary [8] or paving mixers (and to about 15 min in most transit mixers),
after which the air content may remain approximately constant for a considerable period before it begins to drop off. The reduction in air may
result from an increase in very fine particles in the mixture with prolonged
mixing action or simply from an increase in the ratio of air-escape to foamgeneration in the latter portion of the mixing period. Portland Cement
Association tests of transit-mixed concrete indicate that the air void system,
as characterized by specific surface and spacing factor, is not harmed by
prolonged agitation.
Temperature--For a constant amount of air-entraining admixture, less
air will be entrained at 310.9 K (100~ than at 294.3 K (70~ and more
will be entrained at 277.6 K (40~
In other words, everything else being
equal, air entrainment varies inversely with temperature [8,24]
Vibration--Intensive internal vibration applied to concrete will cause
air bubbles to rise to the surface and be expelled. The larger natural voids
are expelled most readily [14,27-29]. Moderately small bubbles may tend
to work upward if the vibration is intense and prolonged. There is increasing evidence, however, that the critically important spacing of small
entrained-air bubbles in the matrix is disturbed very little, even by intense
vibration. If vibration is applied as it should be, with just enough intensity
797
and duration to effect consolidation, and if the mixture is designed properly,

removal of the effective portion of the entrained air will not occur. In some
instances externally applied vigorous vibration may cause an increase in
air content. However, in this case the added air is in the form of relatively
large natural voids.
Admixtures--Solid, finely divided admixtures such as fly ash or bentonite will reduce the air content. The effect becomes marked with fly ash
as the carbon content increases. Calcium chloride can be used successfully with air-entraining admixtures, but if added in direct contact with
certain common admixtures, a rapid chemical reaction may occur between the two, largely inhibiting air entrainment. On the other hand, some
tests have indicated that when the inhibiting reaction does not take place,
the air content may be increased about 1 percent. Especial care must be
taken when other chemical admixtures are used in conjunction with airentraining admixtures, as noted earlier.
One of the most frequent and pronounced causes for variations in air
content results from variation in the amount, type, or condition of the
air-entraining admixture itself. This last cause, as with many of the other
variations in concrete, is a function in turn of the alertness and adequacy
of the control and inspection given the work.
Any influence which would maintain or even actually improve the distribution of the bubbles within the matrix or increase the ratio of airboundary surface to air volume would be desirable. It is certain that an
improved surface-to-volume ratio would reduce considerably the total
volume of air needed to be entrained compared with what is now considered
normal for optimum results.
Work under way in the research laboratories of the Portland Cement
Association [30] and elsewhere points toward advances in the comprehension of the mechanics of air entrainment which should further establish
this major advance in the technology of concrete. For the time being,
however, evaluation must continue for the most part along the indirect
lines of volume of air in the total mixture.
Methods for Determining Air-Void Characteristics
Freshly Mixed Concrete

Since parameters such as specific surface and spacing factor of entrained
air are more reliable criteria of effectiveness than the volume of entrained
air, the ideal test would be one which could measure these characteristics
in the freshly mixed concrete. Unfortunately, no such method has yet been
developed. For field control purposes, presently available test techniques
can determine with reasonable accuracy only the volume of entrained air.
The gravimetric method described in ASTM Test for Unit Weight, Yield,
798
and Air Content (Gravimetric) of Concrete (C 138), is a simple test to

make. It is based on the relation of actual to calculated air-free unit weight
of the concrete and produces reasonably accurate results when aggregates
of relatively uniform specific gravity are used. However, errors attributable
to inadeate sampling may be introduced where the specific gravities of
the fine and coarse aggregate differ materially or where either aggregate
is itself composed of particles of materially differing densities.
The most widely used method is that described in ASTM Test C 231
which is a procedure for determining the air content of freshly mixed concrete by application of pressure. Based on Boyle's law, that at a constant
temperature the volume of a given quantity of any gas varies inversely
with the pressure to which the gas is subjected, the method is generally
adequate for use with all ordinary types of mortar or concrete containing
reasonably dense aggregate. Large errors will be introduced where highly
vesicular or porous aggregates are used due to the impracticability of differentiating between the air in the aggregate particles and the entrained air
in the paste. There are a number of pressure meters on the market which,
by this method, produce satisfactory results with concrete mixtures containing aggregate up to about 50.8 mm (2 in.) in maximum size. When
larger aggregate is used, the larger particles may be removed by hand
and the effect of such a removal calculated in arriving at the total air content [31]. Passing samples of freshly mixed concrete over a sieve to exclude
large particles may lead to important errors resulting from loss of air
during the screening operation.
The volumetric method described in ASTM Test for Air Content of
Freshly Mixed Concrete by the Volumetric Method (C 173) is most useful
for determining the air contents of concretes made with lightweight aggregates [32]. This technique eliminates the possibility of significant errors
in differentiating between air in the aggregate particles and air in the
paste.
Hardened Concrete
The important characteristics of the entrained-air voids can be determined most readily in hardened concrete by microscopical examination
of sawed and ground surfaces of the hardened concrete. ASTM Recommended Practice C 457 is a procedure for determining the total air volume,
specific surface, and spacing factor of air voids by either a linear traverse
method or a modified point-count method. Additional refinements of the
linear traverse equipment enable the measurement of chord size distribution
of air voids from which the total number of air voids per unit volume can
be calculated, as shown in Table 2.
Mielenz and his co-workers [14] show the results of measurements of air-
799
void characteristics of cores taken from a wide variety of structures. Although such measurements are time consuming, they can provide reassuring evidence of the effectiveness of the air-entraining admixture in providing the desired air-void system.
The volume of air in hardened concrete can also be determined by a
high pressure method developed by the Illinois Highway Department [33].
Current Specifications for Air-Entraining Admixtures

It was realized rather early in the development of air entrainment that
the size and distribution of the air voids could be expected to be of major
importance with respect to the effectiveness of the entrained air in enhancing durability. Since there was no ready and reliable means for determining these air-void characteristics and since there was a need to
evaluate the influence of these air-entraining materials on other concrete
properties, a performance type specification was developed in 1950 by
ASTM.
ASTM Specification C 260 evaluates the effects which any given airentraining admixture under test may exert on the bleeding, compressive
and flexural strength, resistance to freezing and thawing, strength of bond
to steel, and the length change on drying of a concrete mixture in comparison with a similar concrete mixture containing a reference air-entraini n g admixture. The methods by which these effects may be tested are given
in ASTM Testing Air-Entraining Admixtures for Concrete (C 233). The
criteria of ASTM Specification C 260 are assumed to afford assurance
that if, under the particular conditions of the specified mixtures and conditions in ASTM Method C 233, the particular sample of the admixture
under test exerts satisfactory influence on certain properties of the laboratory concrete, it will be reasonable to expect that the quantity of the airentraining admixture represented by the sample will develop satisfactory
entrainment of air in field concrete without deleterious effects. However,
the Scope of ASTM Method C 233 acknowledges that the "tests are based
on arbitrary stipulations permitting highly standardized testing in the
laboratory and are not intended to simulate actual job conditions."
All of the elaborate testing required by ASTM Specification C 260
reduces simply to a very indirect method of determining whether: (1) the
particular admixture under test will produce relatively stable air voids
which will become widely dispersed throughout the matrix of f e l d mortar
or concrete so as to produce an air-void system having the proper characteristics for enhancing durability, and (2) that the admixture contains
nothing which will have a deleterious chemical effect on such mixtures.
This specification, with its attendant test methods, while rather elaborate,
has provided a means for evaluating air-entraining admixtures on a per-
800
formance basis. The wide variet~ ch~.nically of materials which can function as satisfactory air-entraining admixtures precludes the inclusion of
chemical requirements.
In recognition of the fact that the tests required are elaborate and timeconsuming and that consequently they may not be performed as often as
desirable to ensure conformance, ASTM Specification C 260 contains a
section on Optional Uniformity Requirements which can be invoked to
ensure that subsequent shipments are identical to the sample tested for
performance.
A further consideration which is receiving attention is the possibility
that the freezing and thawing tests, which are probably the most costly and
time-consuming part of the testing procedure, can be supplanted by an
examination of the characteristics of the air-void system produced by the
admixture under test and a comparison with the system produced by the
reference admixture. ASTM Recommended Practice C 457 provides the
means for such a comparison; however, further information is needed
on the accuracy and reproducibility of the method. Such information is
being developed by ASTM Committee C-9 on Concrete and Concrete Aggregates, which is responsible for the various ASTM specifications and test
methods to which this report has referred. Until that time, however, the
job performance of an air-entraining admixture must still be based on the
direct measurement of "total air content" and on its effect on other readily
measured properties, such as slump, bleeding, and strength. In this respect, it is fortunate that the simple air-entraining admixtures available so
economically in the United States appear to produce, in almost all cases,
about the optimum air-void system in the matrix when the total air contents shown in Table 1 are present in the concretes.
References
[1] Gonnerman, H. F., "Air-Entrained Concrete, A Look at the Record," Consulting Engineer, Oct. 1954, pp. 52-61.
[2] Swayze, M. A., "More Durable Concrete With Treated Cement," Engineering NewsRecord, Vol. 126, 19 June 1941, pp. 946-949.
[3l Anderson, A. A., "Experimental Test Data in Connection with the Development of
Chloride Resisting Concrete by the Use of Treated Portland Cements and Blends with
Natural Cements," 17th Annual Proceedings, Association of Highway Officials of the
North Atlantic States, Feb. 1941, pp. 67-89.
[4] Kellerman, W. F. and Runner, D. G., "The Effect of Using a Blend of Portland and
Natural Cement on Physical Properties of Mortar and Concrete," Proceedings, American
Society for Testing and Materials, Vol. 38, 1938, pp. 329-350; discussion, pp. 351-353.
[5] Lawton, E. C., "Durability of Concrete Pavement-Experience in New York State," Proceedings, American Concrete Institute, Vol. 35, 1939, pp. 561-578; discussion,
pp. 580-1 to 580-5.
[6] Hansen, W. C., "History of Air-Entraining Cements," Proceedings, American Concrete
Institute, Vol. 58, pp. 243-245; Concrete Briefs, Aug. 1961.
[7] Halstead, W. J. and Chaiken, B., "Chemical Analysis and Sources of Air-Entraining
Admixtures for Concrete," Public Roads, Vol. 27, No. 12, Feb. 1954, pp. 268-278.
KLIEGER ON AIR-ENTRAINING MIXTURES 801
[8] Lerch, William, "Basic Principles of Air-Entrained Concrete," Research Laboratories

of the Portland Cement Association, 1950. Contains extensive bibliography to 1950.
[9] Bruere, G. M., "Relative Importance of Various Physical and Chemical Factors on
Bubble Characteristics in Cement Paste," Australian Journal of Applied Science,
Vol. 12, No. 1, March 1961, pp. 78-86.
[10] Powers, T. C., "A Working Hypothesis for Further Studies of Frost Resistance of
Concrete," Journal, American Concrete Institute, Feb. 1945; Proceedings, Vol. 41,
pp. 245-272. Discussion by R. D. Terzaghi, Douglas McHenry, H. W. Brewer, A. Collins, and T. C. Powers, Journal, American Concrete Institute, Nov. 1945, Supplement;
Proceedings, Vol. 41, pp. 272-1 to 272-20; Bulletin No. 5.4, Research Department, Portland Cement Association.
[11] Powers, T. C., "The Air Requirement of Frost-Resistant Concrete," Proceedings, Highway Research Board, Vol. 29, 1949, pp. 184-211; Bulletin No. 33, Research Department, Portland Cement Association.
[12] Powers, T. C., "Void Spacing as a Basis for Producing Air-Entrained Concrete," Journal,
American Concrete Institute, May 1954; Proceedings, Vol. 50, pp. 741-760. Bulletin No.
49, Research Department, Portland Cement Association.
[13] Powers, T. C. and Helmuth, R. A., "Theory of Volume Changes in Hardened Portland
Cement Paste During Freezing," Proceedings, Highway Research Board, Vol. 32, 1953;
Bulletin No. 46, Research and Development Laboratories, Portland Cement Association.
[14] Mielenz, R. C., Wokodoff, V. E., Backstrom, J. E., and Flack, H. L., "Origin, Evolution, and Effects of the Air Void System in Concrete. Part 1--Entrained Air in Unhardened Concrete," Journal, American Concrete Institute, July 1958; Proceedings,
Vol. 55; "Part 2--Influence of Type and Amount of Air-Entraining Agent," Journal,
Aug. 1958; Proceedings, Vol. 55; "Part 3--Influence of Water-Cement Ratio and
Compaction," Journal, American Concrete Institute, Sept. 1958; Proceedings, Vol. 55;
and "Part 4--The Air Void System in Job Concrete," Journal, American Concrete Institute, Oct. 1958; Proceedings, Vol. 55.
[15] KIieger, Paul, "Effect of Entrained Air on Strength and Durability of Concrete Made
with Various Maximum Sizes of Aggregate," Proceedings, Highway Research Board,
Vol. 31, 1952, pp. 177-201; Bulletin No. 40, Research and Development Laboratory,
Portland Cement Association.
[16] Klieger, Paul, "Further Studies on the Effect of Entrained Air on the Strength and
Durability of Concrete Made With Various Maximum Sizes of Aggregates," Proceedings,
Highway Research Board, Vol. 34, 1955; Bulletin No. 77, Research and Development
Laboratory, Portland Cement Association.
[17] Bruere, G. M., "Effect of Type of Surface-Active Agent on Spacing Factors and
Surface Areas of Entrained Bubbles in Cement Pastes," Australian Journal of Applied
Science, Vol. 11, No. 2, 1960, pp. 289-294.
[18] Bruere, G. M., "Air-Entrainment in Concrete," Australian Journal of Applied Science,
Vol. 11, No. 3, 1960, pp. 399-401.
[19] Bruere, G. M., "Rearrangement of Bubble Sizes in Air-Entrained Cement Pastes During
Setting," Australian Journal of Applied Science, Vol. 13, No. 3, 1962, pp. 222-227.
[20] Scripture, E. W., Jr., Hornihrook, F. B., and Bryant, D. E., "Influence of Size Grading
of Sand on Air Entrainment," Journal, American Concrete Institute, Nov. 1948; Proceedings, Vol. 45, pp. 217-228.
[21] Mather, Katharine, "Crushed Limestone Aggregates for Concrete," Mining Engineering,
Oct. 1953, pp. 1022-1028; TechnicaIPublicationNo. 36164, American Institute of Mining,
Metallurgical, and Petroleum Engineers.
[22] Macnaughton, M. F. and Herbich, J. B., "'Accidental' Air in Concrete," Journal.
American Concrete Institute, No. 1954; Proceedings. Vol. 51, pp. 273-284.
[23] Scripture, E. W., Jr. and Litwinowicz, F. J., "Some Factors Affecting Air Entrainment,"
Journal, American Concrete Institute, Feb. 1949; Proceedings, Vol. 45, pp. 433-442.
Data on effects on air entrainment of variations in slump, cement factor, and sand to
total aggregate ratio of the concrete mixture.
[24] Scripture, E. W., Jr., Benedict, S. W., and Litwinowicz, F. J., "Effect of Temperature
and Surface Area of the Cement on Air Entrainment," Journal, American Concrete
Institute, Nov. 1951; Proceedings, Vol. 48, pp. 205-210.
802
[25] Wright, P. J. F., "Entrained Air in Concrete," Proceedings, Institute of Civil Engineers,
London, Vol. 2, Part 1, May 1953, pp. 337-358.
[26] Scripture, E. W. and Litwinowicz, F. J., "Effects of Mixing Time, Size of Batch, and
Brand of Cement on Air Entrainment," Journal, American Concrete Institute, May
1949; Proceedings, Vol. 45, pp. 653-662.
[27] Higginson, E. C., "Some Effects of Vibration and Handling on Concrete Containing
Entrained Air," Journal, American Concrete Institute, Sept. 1952; Proceedings, Vol. 49,
pp. 1-12.
[28] Crawley, W. O., "Effect of Vibration on Air Content of Mass Concrete," Journal,
American Concrete Institute, June, 1953; Proceedings, Vol. 49, pp. 909-920; discussion
by C. E. Wuerpel, pp. 920-1 to 920-2.
[29] Tuthill, L. H., "Vibration of Mass Concrete," Journal, American Concrete Institute,
June 1953; Proceedings, Vol. 49, pp. 921-931, (especially p. 929 for "Effect on Air
Entrainment").
[30] Powers, T. C., "Topics in Concrete Technology and Mixtures Containing Intentionally
Entrained Air," Journal, Portland Cement Association Research and Development
Laboratories, Vol. 6, No. 3, Sept. 1964.
[31] "Handbook for Concrete and Cement," Method CRD-C 41-52, Waterways Experiment
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
Station, 1952. See also "Investigation of Field Methods for Determining Air Content of
Mass Concrete," Technical Memo No. 6-352, Waterways Experiment Station, Nov.
1952.
Symposium on Measurement of Entrained Air in Concrete, Proceedings, American
Lindsay, J. D., "Illinois Develops High Pressure Air Meter for Determining Air Content of Hardened Concrete," Proceedings, Highway Research Board, Vol. 35, 1956,
pp. 424-435.
Blanks, R. F. and Cordon, W. A., "Practices, Experiences, and Tests with AirEntraining Agents in Making Durable Concrete," Journal, American Concrete Institute,
Feb. 1949; Proceedings, Vol. 45, pp. 469-488.4.
Brickett, E. W. and Kennedy, H. L., "The Application of Air-Entraining Agents in
Concrete and Products," Pit and Quarry, Vol. 38, No. 9, March 1946, p. 144.
Brown, L. S. and Pierson, C. U., "Linear Traverse Technique for Measurement of
Air in Hardened Concrete," Journal, American Concrete Institute, October, 1950;
Proceedings, Vol. 47, pp. 117-124; Bulletin No. 35, Research Department, Portland
Cement Association.
Halstead, W. J. and Chaiken, Bernard, "Chemical Analysis and Sources of AirEntraining Admixtures for Concrete," Public Roads, Vol. 27, No. 12, Feb. 1954,
pp. 268-278.
Jackson, F. H. and Timms, A. G., "Evaluation of Air-Entraining Admixtures for Concrete," Public Roads, Vol. 27, No. 12, Feb. 1954, pp. 259-267.
Kennedy, T. B., "Air Entrainment and Its Effect on the Design of Concrete Mixtures,"
Air Entrainment in Concrete Design, Bulletin No. 30, Waterways Experiment Station,
Nov. 1947, pp. 1-18.
Klieger, Paul, "Effect of Entrained Air on Concrete Made with So-Called 'Sand Gravel'
Aggregates," Journal, American Concrete Institute, Oct. 1948; Proceedings, Vol. 45,
pp. 149-163; Bulletin No. 22, Research Department, Research and Development
Laboratories, Portland Cement Association.
Mardulier, F. J., "Particle Size of Dispersed Carbon Black Affects Entrainment of
Air in Concrete," CivilEngineering, Vol. 18, No. 3, March 1948, p. 43.
Mather, Bryant, "The Testing of Aggregates in Air-Entrained Concrete," Air Entrainment in Concrete Design, Bulletin No. 30, Waterways Experiment Station, Nov. 1947,
pp. 41-56.
Menzel, C. A., "Development and Study of Apparatus and Methods for the Determination of the Air Content of Fresh Concrete," Journal, American Concrete Institute, May
1947; Proceedings, Vol. 43, pp. 1053-1072.5; Bulletin No. 16, Research Department,
Portland Cement Association.
Scripture, E. W., Jr., Benedict, S. W., and Litwinowicz, F. J., "Air Entrainment and
Resistance to Freezing and Thawing," Journal, American Concrete Institute, Dec. 1951;
803
[45] Symposium on Entrained Air in Concrete, Journal, American Concrete Institute, June
1946; Proceedings, Vol. 42, p. 601.
[46] Taylor, T. G., "Determination of the Air Content of Mortars by the Pressure Method,"
A S T M Bulletin No. 155, American Society for Testing and Materials, Dec. 1948,
p. 44 (TP2S2); Bulletin No. 27, Research Department, Portland Cement Association.
[47] Taylor, T. G., "Effect of Carbon Black and Black Iron Oxide on the Air Content and
Durability of Concrete," Journal, American Concrete Institute, April 1947; Proceedings,
Vol. 44, p. 613; Bulletin No. 23, Research Department, Portland Cement Association.
[48] Wuerpel, C. E., "Addition of Air-Entraining Agent at Concrete Mixer Advocated," Civil
Engineering, Vol. 16, No. 11, Nov. 1946, pp. 496-498.
[49] Wuerpel, C. E., "Field Use of Cement Containing Vinsol Resin," Journal, American
Concrete Institute, Sept. 1945; Proceedings, Vol. 42, pp. 49-82; Concrete Research
No. 84, Second Interim Report, Part 2, Central Concrete Laboratory.
[50] Wuerpel, C. E., "Laboratory Studies of Concrete Containing Air-Entraining Agents,"
Journal, American Concrete Institute, Feb. 1946, Proceedings, Vol. 42, p. 305. Contains
extensive bibliography up to 1945.
[51] Wuerpel, C. E., "Some Effects of Air-Entrainment and Coarse Aggregate Type on the
Properties of Concrete," National Crushed Stone Association, Washington, D.C., 1945.
[52] Verbeck, G. J., "The Camera Lucida Method for Measuring Air Voids in Hardened
Concrete," Journal, American Concrete Institute, May 1947; Proceedings, Vol. 43,
pp. 1025-1040.5; Bulletin No. 15, Research Department, Portland Cement Association.
L. H. T u t h i l l I
Chapter 46--Mineral Admixtures
Introduction
Admixtures generally are used to provide an economical means of improving one or more properties of fresh or hardened concrete. Among the
effects sought are: reduction in bleeding, increased workability, acceleration
or retardation of hydration or setting, increased strength, reduction in heat
of hydration, added resistance to freezing and thawing, increased impermeability, improved resistance to aggressive waters and soils, and reduction
of expansion caused by reactive aggregate and alkalies in cement. It should
be noted that many of these improvements may be obtained without the
use of admixtures if proper steps are taken in the design of concrete for
the desired uses. However, admixtures designed for specific uses may effect
the desired improvements in properties at a lower cost. Economy is, therefore, one of the matters of prime consideration.
An admixture is defined as a material other than water, aggregates, and
hydraulic cement that is used as an ingredient of concrete or mortar and
is added to the batch immediately before or during its mixing, according to
ASTM Definition of Terms Relating to Concrete and Concrete Aggregates
(C 125). A wide variety of materials are so classified. Among these are
organic compounds, such as triethanolamine, sulfonated lignins, oils, fats,
resins, and carbohydrates; inorganic compounds, such as calcium chloride,
borax, sodium carbonate, and sodium silicate; and finely divided minerals.
Only mineral admixtures are discussed in this paper.
Mineral admixtures include any essentially insoluble material other than
cement and aggregate, which is used as an ingredient for concrete, and is
added to the batch immediately before or during mixing. Mineral admixtures include natural materials, processed natural materials, and
artificial materials. They are finely divided and so form pastes to supplement
portland cement paste, in contrast to soluble substances which act as
chemical accelerants or retardants during the hydration of portland cement
or otherwise modify the properties of the mixture.
1Concrete engineeringconsultant, Sacramento, Calif. 95822.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
TUTHILL ON MINERAL ADMIXTURES 805
For years, powdered mineral admixtures have been used in concrete to

improve workability and to alleviate bleeding. Materials .for this purpose
included cementitious materials, such as natural cement, hydrated lime,
hydraulic lime, and blast-furnace slag; pozzolanic materials, such as fly
ash, diatomaceous earth, calcined shale, and volcanic ash; and inert
materials, such as clay, talc, and stone dust.
Although pozzolans are still used in some areas to supplement sand that
is deficient in fines, air entrainment notably aids workability and reduces
bleeding more uniformly and economically than do most of the mineral
admixtures, but it does somewhat reduce strength in other than lean
concretes. Air entrainment also has a significant effect on increasing resistance to deterioration of concrete caused by freezing and thawing and
has led to an almost complete elimination of the use of chemically inert
mineral admixtures, since, except for pozzolans, they do not generally
impart any benefits to concrete other than improved workability and reduced
segregation.
The types of mineral admixtures used at the present time generally are
limited to those having pozzolanic properties. Although the major portion
of this paper is devoted to a discussion of properties, methods of testing,
and significance of specifications requirements for mineral admixtures
having pozzolanic properties, a brief summary of the history and use of
other mineral admixtures is presented.
Admixtures with Low Reactivity

This type of admixture includes such materials as ordinary clay, ground
quartz, ground limestone, bentonite, hydrated lime, and talc. These
materials were used primarily to improve workability prior to the advent of
the use of air-entraining agents. Their value is based upon their ability to
form soft, plastic pastes which separate and lubricate the aggregate particles in the fresh mortar or concrete mixture. Increasing the ratio of surface
area of solids to volume of water in the paste provides an effective method
of reducing the degree of bleeding. This generally increases the stiffness
of the paste and at a given slump effects a wider separation of the aggregate
particles in the concrete, thereby increasing the workability. The ratio of
surface area of solids to volume of water may be increased by increasing
the cement content or by adding a suitable mineral admixture.
The "inert" mineral admixtures vary in their ability to increase workability and reduce bleeding. Good admixtures will form pastes which contain a maximum proportion of solid matter and a minimum proportion
of water. This requires that the mineral particles do not have unfavorable
shapes and that the surface area is not too great. Williams [1] 2 compared
806
the workability afforded by several admixtures and attempted to gage the

effect of each in resisting segregation in concrete. He concluded that the
relative efficiency of admixtures was indicated by the volume of paste produced by a given weight of each.
Cementitious Materials
Cementitious materials used as admixtures include natural cements,

hydraulic limes, slag cements, and pulverized granulated blast-furnace
slag. Some of these materials, although not used as admixtures to a great
extent at the present time, may provide economical means of improving
specific properties of concrete.
Natural Cements
The development of natural cements in a logical manner was started in
Europe during the late 18th century. Natural cement is the product obtained
by finely pulverizing calcined argillaceous limestone (ASTM Specification
for Natural Cement (C 10)). Natural cements are calcined at a temperature
below 1000~ which is sufficient to drive off carbonic acid gas. This
temperature is below the temperature required for the formation of tricalcium silicate and tricalcium aluminate in portland cement. However,
the intimately associated compounds of calcium and silicon found in natural cement rock will react at these lower temperatures to form dicalcium
silicate, the major compound in natural cement, and monocalcium silicate.
The percentages of these two compounds present in natural cements are
dependent upon the quantity of lime present in the raw material. Because
of its compound composition, natural cement produces concrete which
develops strength slower than portland cement concrete. However, concretes
containing natural cement show a greater rate of strength gain at later
ages, greater resistance to sulfate attack, and lower heat of hydration than
do concretes containing Type I portland cement. When natural cement is
used as an admixture in portland cement concrete, there is no significant
interaction between the hydration products of the portland cement and
the natural cement. Natural cement-portland cement combinations simply
alter the percentages of the hydraulic compounds present, resulting in the
changes in properties of concrete mentioned.
Blast-Furnace Slag
Granulated blast-furnace slag is a nonmetallic product consisting essentially of silicates and alumino-silicates of calcium that is developed
simultaneously with iron in a blast furnace and is granulated by quenching
the molten material in water or in steam, and air, according to ASTM
Specification for Blended Hydraulic Cements (C 595). The use of blast

furnace slag as a cementitious material dates at least from 1774 when Loriot
made a mortar using finely ground blast-furnace slag and slaked lime.
Portland blast-furnace slag cements have been used widely in Europe since
about 1855. The use of blast-furnace slags as cementitious materials generally is limited at the present time to the manufacture of portland blastfurnace slag cements and slag cements. Blast-furnace slag, for use in the
manufacture of portland blast-furnace slag cement, was required in the
1960 Federal Specification SS-C-197b [2] to comply with the following
requirement for composition
CaO + MgO + U3A1203
SiO2 + 2/3A1203
1.0
Granulated blast-furnace slag having a chemical composition within

the following ranges generally will meet the requirements of the above
formula:
Silicon dioxide (SiO2)
30 to 40 percent
Aluminum oxide (AI20a )
8 to 18 percent
Ferrous oxide (FeO)
0 to 1 percent
Calcium oxide (CaO)
40 to 50 percent
Magnesium oxide (MgO)
0 to 8 percent
Manganic oxide (Mn2 03 )
0 to 2 percent
Sulfide sulfur (S)
0 to 2 percent
The use of granulated blast furnace slag as a constituent of portland
blast-furnace slag cement does not conform to the definition of an "admixture": that it is a material added to the batch immediately before or
during mixing. Although in practice blast-furnace slags generally are not
used as admixtures, their use as such would give results similar to those
obtained by intergrinding with portland cement. In addition, closer control
of the desired properties of the concrete would result from the regulated
additions at the mixer. Blast furnace slags have both cementitious and
pozzolanic properties. The use of blast furnace slag in combination with
portland cement in concrete results in a reduction in the total percentage
of uncombined calcium hydroxide present and an increase in the quantities
of calcium silicates and aluminates as compared with a corresponding
concrete containing straight portland cement. Such concrete will have a
lower temperature rise, may tend to inhibit expansion due to alkali-aggregate
reaction, likely will exhibit decreased permeability, and may exhibit increased strength at later ages and increased resistance to aggressive attack
of sea water and natural acid waters, as compared to concrete containing a
Type I portland cement. It should be noted that these advantages will be
gained only under suitable curing conditions for the concrete and may vary
808
greatly with the composition and fineness of the slag and the composition
and fineness of the cement. The relative improvements are likely to be
greater for concrete in which the cementitious material contains the largest
proportion of the slag component.
Slag Cement
Slag cement is defined in previously mentioned ASTM Specification C 595
as the "finely divided material consisting essentially of an intimate and
uniform blend of granulated blast-furnace slag and hydrated lime in which
the slag constituent is at least 60 percent of the weight of the slag cement."
Slag cements have lower strengths than portland blast-furnace slag cements.
Because of their low strength, they generally are used in combination with
portland cement in concrete. In the United States, the proportions usually
used are about 50 percent slag cement and 50 percent portland cement.
Economy is the primary advantage to be derived from the use of slag
cement. Concrete containing slag cement in combination with portland
cement gains strength slower and may have less temperature rise than
corresponding straight portland cement concrete. At later ages, however,
the strengths of slag cement-portland cement concrete which has been
moist cured generally are comparable to straight portland cement concrete.
Concrete containing this type of cement usually exhibits good resistance to
sulfate attack and better plasticity than does corresponding straight portland cement concrete. However, most countries have abandoned production
of this cement, because it is relatively unstable in storage and is lower in
strength-producing characteristics.
ASTM Specification C 595 and the 1960 Federal specifications [3] cover
two types of slag cements, Types S and SA. Type S is defined as slag
cement for use as a blend with portland cement in making concrete and
as a blend with hydrated lime in making masonry mortar. Type SA is an
air-entraining slag cement for the same uses as Type S. Chemical requirements for slag cements are the same for both types. These limits require
the composition of the slag to conform to the following
CaO + MgO + 1/3A1203
_> 1.0
SiO2 + V3AI203
The chemical requirements for the slag cement itself are directed toward
limiting the quantities of certain compounds present. The chemical requirements listed in ASTM Specification C 595 are as follows:
Sulfur trioxide (SOa), max
Sulfide sulfur (S), max
Insoluble residue, max
Loss on ignition, max
4.0
2.0
1.0
4.0
percent
percent
percent
percent
TUTHILL ON MINERAL ADMIXTURES
809
Specification C 595 also has limitations on the types and quantities of

processing additions which may be interground with slag cements. These
limits are as designated in ASTM Specification for Processing Additions
for Use in the Manufacture of Hydraulic Cements (C 465).
The physical test limits for slag cements require that the specific surface
by the Blaine air permeability method be an average of not less that 4700
cm2/g with no single value being less than 4200 cm2/g. This fineness
which is approximately 50 percent higher than that found in most portland
cements, helps to improve the strength-producing properties of slag. As
compared with requirements in ASTM Specification C 150 for portland
cement, the ASTM Specification C 595 requirements for slag cement are:
(Vicat) initial set the same at 45 min; final set 1 h less at 7 h; soundness
0.30 percent less at 0.50 percent; and air content of mortar the same at
12 percent maximum for nonair-entraining cement but 3 percent less at
19 percent, with no lesser minimum limit. The compressive strength requirements for Type S slag cement at 7 and 28 days' age are 600 and 1500
psi, respectively.
Hydraulic Lime
Hydraulic lime is obtained from the burning of limestone containing
clay. The alumina and silica of the clay combine in some degree with the
lime produced in the burning, and the resulting compounds are responsible
for the hydraulic properties of the material. The composition of limes may
vary from high quality limes, containing less than 1 to 2 percent silica
and alumina, to hydraulic limes containing up to 25 percent of these constituents. As the content of silica and alumina increases, the rapidity of
slaking and the evolution of heat, which is characteristic of limes high in
calcium oxide decreases until, with the limes having a high degree of hydraulic activity, no appreciable immediate reaction occurs unless they are
finely ground. Hydrated limes, both with and without hydraulic properties,
have been used in the past as admixtures for concrete.
ASTM Specification for Hydraulic Hydrated Lime for Structural Purposes
(C 141) requires that calcium and magnesium oxides comprise between 65
and 75 percent of the hydraulic hydrated lime. The silica content, which,
along with iron and aluminum oxides, combines with the calcium and
magnesium oxides and imparts the hydraulic properties to the lime, is
required to be between 16 and 26 percent. The total of iron and aluminum
oxides is limited to a maximum of 12 percent, and carbon dioxide may not
exceed 8 percent. In addition to the chemical requirements, hydraulic lime
must also conform to physical requirements for fineness, time of setting,
soundness, and compressive strength. Requirements for fineness are that
not more than 0.5 percent may be retained on a No. 30 sieve and not
more than 10 percent may be retained on a No. 200 sieve; time of setting
limitations by the Gilmore needle test, are a maximum of 2 h for initial
810
set, and 24 h for final set. The autoclave test for soundness permits not
more than 1.0 percent expansion; the compressive strength requirements
are a minimum of 1723.6 kPa (250 psi) at 7 days' age and 3347.4 kPa
(S00 psi) at 28 days' age.
At the present time, hydraulic hydrated lime rarely is used as an admixture in concrete, primarily due to the lack of any real advantage being
derived from such use. One of its principal uses is as an ingredient in
masonry mortars.
Pozzolanic Materials
The term "pozzolan" is employed to designate a siliceous or siliceous

and aluminous material, which in itself possesses little or no cementitious
value but will, in finely divided form and in the presence of moisture,
chemically react with calcium hydroxide at ordinary temperatures to form
compounds possessing cementitious properties (ASTM Definition of Terms
Relating to Hydraulic Cement (C 219)). Calcium hydroxide is liberated
during the hydration of portland cement, and pozzolans combine with this
liberated calcium hydroxide to form stable cementitious compounds which
contribute to strength and watertightness. Generally, it is agreed that the
siliceous ingredient of a pozzolan should be in an amorphous or noncrystalline state, such as glass, opal, or thermally altered clay for best
activity. Crystalline siliceous materials, such as quartz even in finely divided
form, combine, but slowly, with lime, except with curing at elevated temperatures, and, therefore, are considered to be quite inactive.
The early Greeks and Romans used pozzolans in combination with lime
to improve the cementing qualities of lime. Such materials were used in a
variety of constructions, including the aqueducts of ancient Rome, some
of which are still in service. The first large scale use of pozzolan in the
United States took place in 1910 when a rhyolite tuff was used in the
construction of the Los Angeles aqueduct. This was followed by the use of
ground granite in Arrowrock Dam and pulverized sandstone in Elephant
Butte Dam, which were built by the Bureau of Reclamation between 1911
and 1916. Although the degree of pozzolanic activity of the Arrowrock
ground granite and Elephant Butte ground sandstone was very low, especially considering modern specifications requirements, they were interground
with cement to produce"sand cements," which produced strengths in
tension tests equal to straight portland cement at one year and gave greater
strengths at five years' age. The use of varying amounts of pozzolan in
concrete, particularly in mass concrete construction, has continued intermittently since that time.
The amounts of pozzolan used in concrete has varied from 5 percent or
less of the total cementitious material, when employed solely for the purpose
of improving workability, to between 15 and 40 percent, when utilized
811
for its pozzolanic properties. In the United States, the amount used generally
ranges between about IS and 35 percent by weight of the total cementitious
material. In Italy, where pozzolans interground with portland cement are
used extensively, an attempt is made to proportion the amounts used to
obtain a balance between the total amount of silica, alumina, and iron
oxide with the total amount of calcium oxide available in the cementpozzolan mixture. In this way, most of the free lime in cements is combined
with silicates and aluminates, the cementitious properties being thereby
more completely utilized. To conform to this principle, Italian specifications
require that the lowest quantity of pozzolan to be used with a cement be
determined by the following composition ratio to which the cement-pozzolan
mixture in total should agree
SiO2 % + A1203% + Fe203 %
>1
CaO%
=
Pozzolans are classified into two major groups: (1) raw or calcined natural and (2) artificial. Natural pozzolans include such materials as some
diatomaceous earths, opaline cherts and shales, tufts, and volcanic ashes
or pumicites. Each may or may not require calcination, depending upon its
clay content or grinding or both in order to be suitable for use as pozzolan.
Artificial pozzolans include the fine fly ash produced from the burning
of powdered coal which is caught in electrostatic precipitators, waterquenched boiler slag, and precipitated silica. Suitable fly ash has a lowcarbon content and a fineness usually about the same as that of portland
cement. Most of the fly ash particles are in the form of tiny glassy spheres;
usually the low carbon fly ashes of the United States require no further
processing. The spherical shape and texture of the low carbon fly ashes
generally result in a reduced water requirement when used in concrete. On
the other hand, concrete containing most natural or calcined natural pozzolans requires more water than does the corresponding portland cement
concrete.
The American Society for Testing and Materials, the federal agencies
such as the Bureau of Reclamation and the Corps of Engineers, as well as
at least one state agency, the California Department of Water Resources,
have established specifications for pozzolans. Chemical limitations, in percent, are given in Table 1 and physical requirements, in maximum and
minimum values, are given in Table 2.
These requirements provide for a minimum of 70 to 75 percent of the
pozzolan to be composed of silicon dioxide, aluminum oxide; and ferric
oxide. The other main compounds present in pozzolans have maximum
limits placed upon the quantities permitted to be present. These compounds are magnesium oxide, sulfur trioxide, loss on ignition, moisture
content, and exchangeable alkalies.
812
TABLE 1--Chemical
requirementsfor pozzolans.
Bureau of Reclamation
and Corps of Engineers
Item
Silicon dioxide (SiO2) + alum i n u m oxide (A1203) + ferrite
oxide (Fe203) not less than
Magnesium oxide (MgO), not
more than
Sulfur trioxide (SO3), not more
than
Loss on ignition, not more than
Moisture content, not more than
Exchangeable alkalies as Na20,
not more than
TABLE
ASTM Specification
C 618
Natural
Pozzolan
Fly
Ash
Natural
Pozzolan
Fly
Ash
70.0
70.0
70.0
70.0
5.0
5.0
5.0
. ..
4.0
8.0
3.0
4.0
6.0
3.0
4.0
10.0
3.0
5.0
12.0
3.0
1.5
1.5
...
1.5
2--Physical requirementsfor pozzolans.

Bureau of Reclamation and
Corps of Engineers [4]
Item
Fineness
Specific surface, Blaine air
permeability apparatus, cm 2/
cm 3, not less than
Material retained on No. 325 mesh
sieve, percent, not more than
than
Compressive strength
With portland cement, percent
of control, 28 days, not less
than
With lime, 7 days, min M P a (psi)
Increase in drying shrinkage of
mortar bar, percent shrinkage
of pozzolan bar minus percent
shrinkage of control bar, not
morethan
Water requirement, percent not
more than
Reduction of reactive expansion at
14 days, percent, not less than
Mortar expansion at 14 days
percent, not more than
Soundness, autoclave expansion,
percent, not more than
ASTM Specification
C 618
Natural
(Type N)
Fly
Ash
Natural
(Type N)
Fly
Ash
12 000
6 500
12 000
6 500
20
34
...
75
(900)
85
(900)
75
(800)
85
(800)
0.03
0.03
0.03
0.03
115
105
115
105
75
...
75
...
0.020
0.020
0.020
0.020
0.50
0.50
0.50
0.50
813
The principal technical benefits generally derived from the use of pozzolan in concrete are increased impermeability (Fig. 1) [5], lower heat of
hydration (Fig. 2) [5], reduced alkali-aggregate expansion (Fig. 3), and
improved workability. One of the disadvantages of using pozzolan is that
concrete containing pozzolan generally develops strength more slowly than
portland cement concrete. However, concretes containing some pozzolans
under moist or mass curing conditions continue to gain in strength for a
long period of time and may produce concretes whose strengths are equal
to or substantially greater than the strength of corresponding straight portland cement concretes. In mass concrete, this type of strength gain is
satisfactory. However, paving and structural concrete generally require
early strength properties comparable to those obtained in straight portland
cement concrete. Therefore, when pozzolan is used for these purposes, it
is necessary to proportion the concrete mixes to provide higher early
strengths.
One method frequently used in certain localities to partially counteract
the lower strength of portland cement-pozzolan concrete at early ages is
the redesign of concrete mixtures to incorporate a greater amount of pozzolan in the mix than the cement it replaces [6,7]. Redesign of mixtures
in this manner can be justified economically only in localities where the
cost of the pozzolan is low enough to permit production of this concrete
without an increase in cost. Pozzolans lower the resistance of concrete to
deterioration caused by freezing and thawing unless longer than usual
moist curing is provided, which in most instances is not always practicable
kg/m 3
40 X 10 -4
125
150
'
'
175
200
225
' I
I
250
275
[ ,
12 X 10 -4
35
'
--~
9 i0
3o
~ 100% portland
cement
-8
25
20
"6
~E
,4
70 to
.~-
75% p o r t l a n d
cement
25 tO 30% fly ash
200
250
300
350
Cementing Materials,
400
450
500
ib/yd 3 concrete
FIG. 1--Permeability of concrete with and without pozzolan.

814
60
If.
30
t)
15
~7
Jo
50
40
0
0
0
n-
2
tO
E
3 Sacks of cemenfing maferial per

cubic yard of concrete :
A - N o Pozzolon replacement
13 - 30*/. Fly Ash replacemenf, by weighf.
C - 30~, Calcined Diafomaceous Shale
replocemenf, by w?.
E
u
;
0
"D
00
tO
Age
15
in Days
20
25
50
FIG. 2--Effect of pozzolan on temperature rise of concrete.
during construction. The magnitude of the reduction is usually not large

and varies with the pozzolan. With recommended air entrainment, any
such effect on durability is likely to be small in comparison with the benefits
of using pozzolan. However, under prolonged moist curing conditions, many
pozzolans have a beneficial effect on the durability of concrete as measured
by freezing and thawing tests. In most regions having enough freezing and
thawing weather to affect durability, exposure ambients are sufficiently
moist so much of the time that full natural curing and maturity of exposed
pozzolan-cement concrete is sooner or later attained; consequently it becomes amply and comparably durable.
Conformance with the specifications limits listed in Table 2 generally
will assure that pozzolan will impart the desired properties to concrete in
which it is used. However, to assure procurement of pozzolan of maximum
effectiveness, one state agency, the California Department of Water Resources, improved the following four physical requirements in their specifications for N-type natural pozzolan as compared with corresponding requirement values shown in Table 2.
Pozzolanic activity index:
815
0,7
0.6
CALCINED
0.5
0.4
QUARTZ (CONTROL)
0.3
0.2
Nol~-reoc//ve
0. I
Aggregole7
0
0.7
0.6
OIL
-=: 0.5
CA LCI N E D
PUMICITE
ALCINED
UE NTE
0.4
~ 0.5
CALCINED
c 0.2
',,0.1
0
0.7
0.6
FL:ASH
0.5
NCALCINED
0.4
X'PUMICITE
0.3
0.2
0.1
0
0
12
18
6
t2
18 0
G
12
POZZOLAN REPLACING HIGH-ALKALI CEMENT,
PERCENT BY W E I G H T .
t8
F I G . 3--Effect of small proportions of pozzolans on chemical reactivity in mortar bars;

expansion after 15 months storage. (Note." 1.'2 mortar in 25 by 25 by 278.mm (1 by 1 by 11 88
in. ) bars sealed moist at IO0~ )
With portland cement, at 28 days, min, percentage of control

With lime, at 7 days, min MPa (psi)
Water requirement, max, percentage of control
Reactivity with cement alkalies:
Reduction of mortar expansion at 14 days, min percent
90
(1000)
110
85
The fineness requirement is important to assure that a uniform and sufficient surface area per unit quantity of material is present to react with the
calcium hydroxide which is liberated during the hydration of cement. The
816
fineness of a pozzolan will affect its other properties, primarily its strength
producing ability. The more finely ground a material is, the more rapid is
the rate of chemical reaction and the greater is the proportion of the pozzolan which reacts. The workability of concrete is decreased as the fineness
is decreased. Coarse pozzolans of poor particle shape, particularly volcanic
glasses, require a larger amount of water in concrete for a given slump.
The water requirement of concrete may also be increased by pozzolans of
very high fineness, such as some diatomites, when used in large amounts.
The greater the water requirement of a pozzolan, the greater will be the
drying skrinkage of the concrete in which it is used. Coarse pozzolans of
poor particle shape, such as some of the coarse volcanic ashes, also contribute to bleeding in freshly mixed concrete.
Methods of test in use for determination of fineness of pozzolans include
specific surface (cm2/g or cm2/cm3 of solid volume) by Blaine air permeability apparatus, per cent of material retained on a (No. 325) mesh sieve,
and calculated mean particle diameter.
The compressive strength of portland cement-pozzolan mortar is a property of considerable importance, since the value of a pozzolan when employed in combination with portland cement depends primarily upon its
effect upon strength of concrete. The contribution of a pozzolan to strength
of concrete arises in the cohesion of the hydration products of cement pozzolan combinations and in the adhesion of the matrix to the grains of sand
or other aggregate. Concretes containing pozzolan, when fog-cured 28 days,
usually produce strengths which are lower than similarly cured concrete
containing the same weight of portland cement only. However, the difference
in strength becomes progressively less as the age of the concrete increases,
and, at later ages, concrete containing optimum amounts of pozzolan
frequently develops higher strengths (Fig. 4). The slower strength-producing properties of portland cement-pozzolan concretes are considered to be
due to the slow rate of reaction between the calcium hydroxide liberated
as the cement hydrates and amorphous silica or siliceous glass to form
silicates and aluminosilicates of lime that are not found in portland cement
concretes which do not contain pozzolan. An evaluation of the compressive
strength-producing properties of pozzolans is made by comparing the compressive strengths of portland cement-pozzolan mortars, as determined
from tests on S0-mm (2-in.)-cube specimens, with compressive strengths
of a similar mortar containing the same portland cement without pozzolan.
Another test used for determining the strength-producing properties of
pozzolan is the test for compressive strength with lime, which test is considered effective in determining the total activity of pozzolan (ASTM Specification for Fly Ash and Other Pozzolans for Use with Lime (C 593), Sec.
7). Pozzolans will react with hydrated lime in the presence of moisture
developing a "set" by the formation of cementitious products and increasing in strength to an extent that depends on the activity of the pozzolan.
817
6000
40.O
35O
5000
300
4000
250
3000
200 ~
150
2000
8
10.0
50
2|
9!
AGE- DAYS
365
O0
T30
F I G . 4--Effect of some pozzolans on compressive strength of concrete. A a n d B--470 lb

cementitious material per cubic yard of concrete. A--No pozzolan, Cement A, Type II. B - 15 percent N-type pozzolan replacement, by weight. C, D a n d E--282 lb cementitious
material per cubic yard of concrete. C--30 percent N-type pozzolan replacement by weight. D-No pozzolan, Cement B, Type 11. E--25 percent flyash pozzolan replacement by weight.
Therefore, a lime pozzolan sand mortar is used to obtain an evaluation

of the activity of the pozzolan without introducing other variables from
portland cement. Results of this test are generally reported after 7 days'
age and, therefore, provide a more rapid evaluation of the strength-producing properties of a pozzolan than do portland cement-pozzolan strengths
which generally are reported at 28 days' age. It is considered that the portland cement-pozzolan mortar test provides a more reliable indication of the
actual performance to be expected from concrete containing a given pozzolan. However, the lime-pozzolan test for compressive strength is also
useful as a uniformity test for compressive strength after a source has been
established.
As previously mentioned, a direct relationship exists between water requirement and drying shrinkage; the greater the amount of water required,
the greater the drying shrinkage. When pozzolan is added to concrete as a
partial cement replacement, the water requirement for concrete of a given
consistency usually is changed. Natural pozzolans have a tendency to increase
the water requirement, and fly ashes of low carbon content usually lower
it slightly. A high-water requirement will increase drying shrinkage of conCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:31:09 EDT 2014
818
crete, which in turn will increase the amount of cracking which may occur,
and may harmfully affect other properties of concrete, such as strength,
durability as measured by freezing and thawing, impermeability, and sulfate resistance. Pozzolan specifications limit the percent increase in both
water requirement and drying shrinkage of mortar containing portland
cement-pozzolan combinations over that of portland cement mortar without pozzolan, to assure that the shrinkage characteristics of concrete containing pozzolan are not adversely affected to a harmful degree. Figure 5
[5] shows the effect of various pozzolan replacements on drying shrinkage
and autogenous length change of portland cement concrete.
However, these greater laboratory shrinkage values for the N-type natural
pozzolans are not always reflected in a visible additional amount of cracking in field structures. It is not as if there were cracks in the pozzolan
concrete and none in the concrete without it. Both have cracks. This was
noted where concrete with and without such pozzolan was examined particularly for this comparison, in Davis Dam (some parts without and some
with) and in structures of the Delta-Mendota Canal (without) and of the
California Aqueduct (with) which are in identical service conditions running
parallel on the west side of the San Joaquin Valley. If visible cracking is
to be eliminated it will not be accomplished as much by choice of materials
as it can be approached by judicious use of contraction joints and grooves.
.|_="-3
~
zoo
~'~ i'~~
~Drying shrinkage at 50% relative humidity

.~
9o days ,o0 car,no.
c>
3 Sacks of cementi
.J= -400 - cubic yard of concrete :
~ ~
A 8 E - 5 0 % Fly Ash replacement, b y w t . ~
~
|
B & F - NO replacement
I ~
~
"-500
C & G - 50% Pumice replocemenf, b y w t . - ~ '
~
~
O F~ H - 50% Calcined Shale replacement, by wt.
~
-600
14
28
Age in Days ( L o q
90
Scale )
180
E
F
G
H
365
FIG. 5--Drying shrinkage and autogenous length change of lO0 by 100 by 750-ram (4 by 4
by 30-in. ) bars of portland cement concrete with variouspozzolan replacements.
Moreover, if any question of alkali-silica expansive reaction may be involved,

the considerable superior reduction of such expansion imparted by the Ntype, nonfly ash pozzolan is a much more important factor in its favor than
its minor and seldom evident contribution to shrinkage is to the contrary.
One of the important benefits derived from the use of pozzolan is its
effectiveness in controlling potentially disruptive expansion in concrete
resulting from the formation of an alkali-silicate gel, which is the product
of reaction between certain types of mineral aggregates (cherts, opal, and
so on) and the alkalies (sodium and potassium oxide) in cement.
Initially, the only apparent methods of controlling alkali-aggregate reaction in concrete were to use low-alkali cements or to avoid the use of
reactive aggregates. Subsequently, some pozzolans were found to effectively
reduce expansion caused by alkali aggregate reaction. Other pozzolans
have little effect in controlling this expansion. Therefore, it is necessary
to test pozzolans individually in order to evaluate their ability to control
this expansive reaction. An accelerated test (ASTM Test for Potential
Alkali Reactivity of Cement-Aggregate Combinations (Mortar-Bar Method)
(C 227)), which measures the effectiveness of a pozzolan in reducing expansion of mortar made with high-alkali cement and reactive aggregate,
is generally used for this purpose. The amount of pozzolan needed in concrete to control reactive expansion will vary with the individual pozzolan,
aggregates, and with the alkali content of the cement.
In general the N-type pozzolans will be found to be considerably superior
to the S-type and fly ash pozzolans in their ability to reduce expansion due
to the alkali-silica reaction. This is indicated in the absence of a value for
minimum percent reduction in expansion allowed for fly ash by the ASTM
and Federal specifications cited in Table 2. Also it is shown by less reduction in expansion with 15 and 18 percent of the fly ash in the lower left
unit and of the S-type pumicite in the lower right unit of Fig. 3. In this
figure it will be noted that all these tests of different pozzolans were made
with high-alkali cement. All but the two just mentioned were nonfly ash,
natural N-type pozzolans, except for the ground quartz, nonpozzolan control shown in the upper left unit of Fig. 3, which only began to reduce
expansion when used in amounts of 15 and 18 percent. At these amounts,
all the others except the pumicite, which is only slightly better than the
fly ash in this respect, reduced expansion very positively even with the highalkali cement and reactive aggregate, to substantially less than 0.10 percent after 15 months of storage under conditions highly conducive to maximum alkali-silica expansion.
Accordingly, this distinction between N-type and other pozzolans, which
is not clearly made in some specifications, is well worth considering and
making if one of the values expected from use of a pozzolan is control of
alkali-silica expansion. However, replacement of some portion of the cement with any pozzolan will have the effect of diluting the alkalies just
820
that much and to that extent cause some reduction in their reaction with
silica aggregate.
As a safety precaution against the possibility of increased alkali-silica
expansion in concretes containing small (pessimum) amounts of certain
pozzolans less than 10 percent by weight of the cement, ACI Committee
212 [8] advises that pozzolans should not be used in amounts less than
about 15 percent by weight of total cementitious material.
Another substantial benefit from the use of fly ash and N-type pozzolans
meeting requirements of ASTM Specification for Fly Ash and Raw or
Calcined Natural Pozzolan for Use as a Mineral Admixture in Portland
Cement Concrete (C 618), is markedly improved sulfate resistance. Aside
from their ability to combine with and thus deny lime for expansive combination with sulfates that penetrate the concrete, all such good pozzolans
materially improve the impermeability of the concrete so that it is less
vulnerable to penetration of solutions carrying the sulfates.
According to a wide range of long-time testing of sulfate resistance at
the U.S. Bureau of Reclamation, all types of cement are notably improved
in this respect by pozzolans. However, this does not mean that thereby
Type I was made as resistant as Type V. The more severe the exposure,
the greater was the benefit of the pozzolan in resisting it. The order of
resistance reported was (1) Type V with fly ash (greatest resistance), (2)
Type II with fly ash (barely greater than Type V), (3) Type V, (4) Type II,
(5) Type I with fly ash, (6) Type I (least resistance).
With 12 fly ashes tested there were two N-type pozzolans. These did as
well as many of the fly ashes in air-entrained concrete. No lignite and subbituminous fly ashes were included. Later indications are that these do not
reliably increase sulfate resistance. Ideally, potential job cement and pozzolan should be tested for specific combined capability. Unfortunately,
such tests require many months, even several years, of specimen exposure
in strong sulfate solutions to make positive distinctions. Whether or not
such tests are made, indications are clear that greatly improved resistance
to severe exposure to sulfates, will be obtained if 15 to 30 percent of the
cement (preferably Type V) is replaced by weight with fly ash or N-type
pozzolan meeting the requirements of ASTM Specification C 618.
The potential danger of unsoundness occurring in concrete due to the
presence in pozzolan of either the periclase form of magnesia, "dead burnt"
calcium oxide, or an excessive amount of sulfur trioxide, points out the
importance of determining the characteristics of the" pozzolan in this regard.
The expansion of mortar bars determined by the autoclave soundness test,
as required for portland cements, is the procedure used for evaluation of
the effect of a pozzolan on soundness of concrete.
Much has been done in recent years toward the development and refinement of the use of mineral admixtures. Through improved technology and
a thorough knowledge of proper design of concrete mixes to meet specific
needs, the use of m a n y a d m i x t u r e s previously used has b e e n e l i m i n a t e d . I n

a d d i t i o n , this i m p r o v e d technology has r e s u l t e d in m o r e e c o n o m i c a l a n d
a d v a n t a g e o u s uses o f m i n e r a l a d m i x t u r e s , which have p r o v i d e d b e t t e r
concrete a t lower cost. B a s e d u p o n p r e s e n t technological advances, it is
a n t i c i p a t e d t h a t the f u t u r e will c o n t i n u e to show m a r k e d i m p r o v e m e n t s
in concrete technology, which will u n d o u b t e d l y include f u r t h e r r e f i n e m e n t s
in the use of m i n e r a l a d m i x t u r e s .
Acknowledgments
This writer wishes to give full credit to E. C. H i g g i n s o n , the p r e c e d i n g
a u t h o r of this p a p e r in the 1966 edition of A S T M S T P 169A. T h e g r e a t
m a j o r i t y o f its original content r e m a i n s u n c h a n g e d in this edition. P r i n c i p a l
differences are an u p d a t i n g of the various agency specifications a n d A S T M
s t a n d a r d s m e n t i o n e d a n d f u r t h e r i n f o r m a t i o n r e g a r d i n g the i m p r o v e m e n t s
i m p a r t e d to concrete by the various pozzolans, p a r t i c u l a r l y as to sulfate
resistance a n d as to effective control of t h e alkali-silica reaction.
References
[1] Williams, G. M., "Admixtures and Workability of Concrete," Proceedings, American
[2] "Cement, Portland, Blast-Furnace Slag," Federal Specification SS-C-197b, 26 Sept.
1960.
[3] "Cement, Slag," Federal Specification SS-C-218a, 21 Nov. 1960.
[4] "Pozzolan, for Use in Portland Cement Concrete," Federal Specification SS-C-1960/5,
2 Dec. 1975.
[5] Higginson, E. C., "Manual de Control del Concreto," Publicacion Numero 68, Del
Instituto de Ingenieria, Universidad Nacional Autonoma de Mexico, 1962.
[6] Lovewell, C. E. and Washa, G. W., "Proportioning of Concrete Mixtures Using Fly
Ash," Proceedings, American Concrete Institute, Vol. 54, 1958, p. 1093.
[7] Frederick, H. A., "Application of Fly Ash for Lean Concrete Mixes," Proceedings,
[8] ACI Committee 212, "Admixtures for Concrete," Journal, American Concrete Institute,
Sept. 1971, Sec. 6.2.5.
Bibliography
Blanks, R. F. and Kennedy, H. L., "The Technology of Cement and Concrete," Concrete
Materials, Vol. 1, Wiley, New York, 1955.
Bogue, R. H., The Chemistry of Portland Cement, Reinhold Publishing Corp., New York,
N.Y., 1955.
Concrete Manual, U.S. Department of the Interior, Bureau of Reclamation, 7th ed., 1963.
"Cooperative Tests of Fly Ash as an Admixture in Portland Cement Concrete," Report of
Subcommittee III-h on Methods of Testing and Specifications for Admixtures, Proceedings,
Davis, R. E., "Use of Pozzolans in Concrete," Proceedings, American Concrete Institute,
Vol. 46, 1950, p. 377.
Dexheimer, W. A., "Pozzolans in Concrete Dams," Question No. 23, R. 97, 6th International
Congress on Large Dams, New York, N.Y., 1958.
"Investigations of Portland Blast-Furnace Slag Cements," Technical Report No. 6-445,
Waterways Experiment Station, Corps of Engineers, U.S. Army, Vicksburg, Miss., Dec.
1956.
822
Lea, F. M. and Desch, C. H., The Chemistry of Cement and Concrete, St. Martin's Press
Meissner, H. S., "Mineral Admixtures," Significance of Tests and Properties of Concrete and
Concrete Aggregates, A S T M STP 169, American Society for Testing and Materials, 1956,
pp. 375-387.
Mielenz, R. C., Greene, K. T., and Schieltz, N. C., "Natural Pozzolans for Concrete,"
Economic Geology, Vol. 46, NO. 3, 1951, pp. 311-328.
Use of Pozzolanic Materials in Mortars and Concretes, A S T M STP 99, American Society for
Testing and Materials, 1950.
Bryant Mather 1
Chapter 47
Chemical Admixtures
Preface
The original version o f A S T M STP 169 [1] 2 published in 1956 did not have
a chapter on "Chemical Admixtures" since at that time there was no ASTM
standard covering chemical admixtures. The ASTM Specification for
Chemical Admixtures for Concrete (C 494) was originally issued in 1962 and,
hence, in A S T M STP 169-,4 [2], issued in 1966, there was a chapter. It was
prepared by Dr. Bruce Foster [3]. The present chapter is an updating of Dr.
Foster's paper.
This discussion is limited to certain features of chemical admixtures
regarded as most appropriate in the context of this volume. For a more comprehensive review of knowledge in this complex field, reference should be
made to such of the following works as may be relevant. In 1967, RILEM
published "Proceedings, International Symposium on Admixtures for Mortar and Concrete" [4] held in Brussels. In Vol. 6, pp. 167-185, there is a
discussion of "Quality Control of Admixtures." In this discussion there is a
presentation of a series of 11 tests of the admixture itself, tests of the admixture in pastes, mortars, and concretes. The 11 tests mentioned are: (1) dry
solid content, (2) loss on ignition, (3) presence of sugar or chloride, (4) pH of
solution, (5) surface tension of solution, (6) foam stability, (7) specific surface (of powders), (8) color and odor, (9) density, (10) infrared or X-ray spectra, or both, (11) methods of identification indicated by the manufacturer.
In 1971 the American Concrete Institute (ACI) published the fourth report
of its Committee 212 on Admixtures for Concrete, entitled "Guide for Use of
Admixtures in Concrete" [5]. This committee which was organized in 1943
has functioned in close, but informal, liaison with the Transportation
Research Board Committee A2E05 on Chemical Additions and Admixtures
for Concrete and Subcommittee C09.03.08 on Test Methods and Specifications for Admixtures for Concrete of ASTM Committee C-9 on Concrete and
I Acting chief, Structures Laboratory, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, Miss. 39180.
2 The italic numbers in brackets refer to the list of references appended to this paper.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
824
Concrete Aggregates. Symposia, bibliographies, reports of research, and the

state-of-the-art reviews sponsored by all three groups have been studied and
used, with benefit, by all three. Reviews covering significant portions of the
area were distributed in 1970 as Part IV (V.IV) of the Proceedings of the
Fifth International Symposium on the Chemistry of Cement, entitled "Admixtures and Special Cements," [6] and as "The Science of Admixtures,"
proceedings of a symposium organized jointly by the Concrete Society and
the Cement Admixture Association, London, 6 Nov. 1969 [7].
Also in 1971, the Highway Research Board published the report on "Admixtures in Concrete," prepared by its Committee A2E05 [8].
In 1973 a book entitled Admixtures for Concrete of 253 pages with text in
parallel columns in English and French, prepared by Albert Joisel, was
published [9].
In 1975 and 1976 annual reviews were published [10,11] of published
work in the preceding year. Odler [I0] defines "chemical admixtures" as "inorganic or organic compounds which, when added to portland cement in
small amounts (usually less than one percent by weight of cement), modify
the properties of the cement paste of hardened cement stone in a desired
way." He reported that in 1974 there were three review papers: Bonzel and
Krumm, [12] Ivanov et al [13], and Vavrin [14]. Odler [10] noted that 21
papers, for which he gives references, were published in 1974 discussing the
effects of chlorides on hydration of portland cement and on the nature of the
products formed. A review of the mechanism of acceleration by calcium
chloride was given by Skalny and Maycock [15].
Introduction
Although all concrete admixtures are chemicals in a literal sense, by convention, in concrete technology, the term "chemical admixture" is restricted
to water-soluble substances other than those used solely for the purpose of air
entrainment. Most chemical admixtures react chemically with the cement in
concrete. This discussion will be confined to set-controlling and water reducing admixtures. Chemical admixtures of other types are classified and
described in reports of the ACI [5, I6].
A water-reducing admixture is defined as one that reduces the quantity of
mixing water required to produce concrete of a given slump; a retarding admixture as one that retards the setting (that is, increases the time of setting)
of concrete; an accelerating admixture as one that accelerates the setting and
early strength development of concrete; a water-reducing and retarding admixture as one that reduces the quantitiy of mixing water required to produce concrete of a given slump and retards the setting of concrete; and a
water-reducing and accelerating admixture as one that reduces the quantitiy
of mixing water required to produce concrete of a given slump and accelerates the setting and early strength development of concrete. These five
B. MATHER ON CHEMICAL ADMIXTURES 825
types of admixtures are referred to in ASTM Specification C 494 as Types A,

B, C, D, and E, respectively.
Types of Materials and Their Action in Concrete
Many materials, when introduced into mortars or concretes have been
found to modify the setting properties of portland cement [16-18]. Often, use
of such materials also results in a lower water requirement to produce a given
slump.
The materials in widespread use as set-retarding and water-reducing admixtures in the United States in 1959 were classified by Prior and Adams [19]
into four classes: (1) lignosulfonic acids and their salts, (2) modifications and
derivatives of iignosulfonic acids and their salts, (3) hydroxylated carboxylic
acids and their salts, and (4) modifications and derivatives of hydroxylated
carboxylic acids and their salts. In each of these, the primary component has
both water-reducing and set-retarding properties. In the formulation of
products of Classes 2 and 4 these admixtures may be modified by the addition of other components to give various degrees of retardation, no significant change in setting time, or acceleration, while at the same time preserving the water-reducing properties. They also may be modified by the addition
of an air-entraining admixture.
Odler [11] reported that a series of additional compounds was described to
be effective as water-reducing or plasticizing agents, or both, among them:
acrolein; polyglycerol; a combination of anthranylic acid, formaldehyde and
phenol; polyethylene-oxide with a molecular weight from 8 X l0 s to 6.7 X
106; polyacrylic and polymethacrylic acid or their alkali metal or ammonium
salts; admixture of triethanolamine and gallic acid; glutamic acid amide;
and a compound with the formula (NaO2). CH(SOaNa)- CHR(CO2Na), where
R is H or Me.
The accelerating admixture which has found greatest use is calcium
chloride [16,20]. The organic material triethanolamine also is used extensively but usually in combination with other materials, and little published
information is available on its properties [21,23]. Therefore, the discussion of
accelerators will be confined primarily to calcium chloride.
It is to be expected that the extent of any modification of setting time will
depend upon the properties of the cement to which an admixture is added.
The effect of calcium chloride on acceleration of strength development has
been reported to vary with the cement with which it is used [18], while the
behavior of concrete containing water-reducing retarders appears to be
dependent even more upon the properties of the cement. Polivka and Klein
[24] reported that the effectiveness of the latter materials, both from a waterreducing standpoint and from a retarding standpoint, was greater with
cements of low alkali and low C3A content. Tuthill et al [25] and those who
discussed their paper have reported excessive retardation when retarders
826
were used with certain cements, which could be overcome by an increase in

the sulfate content of the cement. Often, the excess retardation was accompanied by stiffening, even though the cement without the water-reducing
retarder showed no early stiffening and had a normal setting time. Seligman
and Greening [26] have reported on work that provides information on the
chemistry of this phenomenon.
Bruere [27] and Dodson and Farkas [28] have noted that the effect of
water-reducing retarders depends in some cases on the time at which the
retarder is added to the mixture. Dodson and Farkas [28], working with
cements little affected by the addition of either lignosulfonate or hydroxylated carboxylic admixtures, found that the efficiency of the admixtures as
retarders, and their capacity to enhance air entrainment and reduce water requirement, was increased greatly by a 2-min delay in their addition after mixing commenced.
The unfavorable behavior of some admixtures with certain cements and
under certain conditions of use which have been reported is counterbalanced
by a record of successful use under controlled conditions in many concreting
operations. However, when experience with specific admixture-cement combinations under similar job conditions is not available, tests with specific
materials should precede a decision for use in construction.
When two or more admixtures are added ~o a concrete mixture, they
should be added separately during the mixing operation unless they have
been shown to be compatible when added to the concrete as a single material.
Effects on Freshly Mixed Concrete
Water Reduction
The water reduction resulting from the use of traditional water-reducing
admixtures ranges from 5 to 15 percent [29]. The water reduction found with
lignosulfonate water-reducers may be contributed to by the air entrained by
these materials. In addition to varying with the particular cement employed,
the amount of water reduction with a given admixture is also influenced by
dosage, cement content, type of aggregate, and the presence of other admixtures, such as air-entraining agents or pozzolans.
Water-reducing admixtures are effective with all types of portland cement,
portland blast-furnace slag cement, portland-pozzolan cement, and high
alumina cement.
The addition of recommended amounts of calcium chloride has been
found to reduce the water requirement by a small, but definite, amount over
that required to produce the same slump, with no chloride added.
At the time or preparation of this paper (December 1976), ASTM Subcommittee C09.03.08 on Admixtures for Concrete had a task group on
chemical admixtures that was working on the modification of ASTM
B. MATHER ON CHEMICAL ADMIXTURES
827
Specification C 494 to deal with a number of products that had recently come
on the market in the United States (principally from Europe and Japan) that
permitted greater degrees of water reduction than referred to above. These
have been used to produce high-strength concrete by taking advantage of the
opportunity they provide to lower the water/cement (w/c) ratio more than
generally has been practical with more traditional water-reducing admixtures or to produce "flowing" concrete, that is, concrete of much greater
slump. It has been reported that some of these products provide the alteration of properties that characterizes their use for a relatively short period of
time. Such admixtures may be added to ready mixed concrete after arrival at
the site for placement. The relevant available data may be located beginning
with the work in the United Kingdom [30], Germany [31,32], or the United
States [33].
Setting Time
The magnitude of the effect on setting time of the addition of calcium
chloride depends not only upon the dosage, but also upon the particular cement, the temperature, and other factors. The recommended maximum
dosage has a substantial effect on setting time at normal temperatures, and
can produce a very rapid set. at high temperatures, as is also the case with
very large dosage at normal temperatures.
Similarly, the retardation of setting time brought about by retarders is
dependent not only upon the particular cement with which they are used, but
also upon the temperature, dosage, and other factors. Overdosage may produce excessive setting times of 24 h or more, but in such cases, if the concrete
finally sets and has been protected from drying, ultimate strengths developed
may be satisfactory if forms are left in place for a sufficient length of time.
With either lignosulfonate or hydroxylated carboxylic admixtures, the degree
of retardation can be controlled by varying the dosage, provided that the
allowable air content with lignosulfonate materials is not exceeded. Classes 2
and 4 water reducers may be formulated to give no retardation or to produce
acceleration, both effects being produced by the incorporation of a catalyst
or an accelerator.
Air Entrainment
Lignosulfonate water-reducing retarders usually entrain 2 to 3 percent of
air when used in normal dosages. The hydroxylated carboxylic admixtures do
not entrain air, nor does calcium chloride. However, all three materials
enhance the effectiveness of air-entraining agents from the standpoint of
volume of air produced, so that less air-entraining agent is required when
added to concrete containing one of these other admixtures. Air-entraining
properties may be incorporated in the modified materials of Classes 2 and 4
828
by the incorporation of a suitable air-entraining agent. When the airentraining agent is not incorporated in another admixture by the manufacturer, it and other admixtures should be added to the concrete separately
unless tests have shown that the two materials are compatible when added to
the concrete as a single material.
Bleeding
Water-reducing admixtures that entrain air reduce bleeding, the reduction
being due to the entrained air and the lower content. Hydroxylated carboxylic water-reducing retarders have been reported to increase the rate and
amount of bleeding. Such bleeding has been suggested as being responsible
for a portion of the strength increase observed with the use of these materials.
Slump Loss
Contrary to expectations, water-reducing retarders usually have not been
found effective in reducing slump loss resulting from substantial delays, after
mixing, in placing the concrete. As pointed out above, the use of retarders
with some cements may actually produce an early stiffening. However, the
addition of a water-reducing admixture will give a high initial slump with the
same w/c ratio and permit more slump loss before concrete becomes unworkable.
Calcium chloride may result in early stiffening and in many cases,
therefore, is not added until after mixing has commenced.
Effects on Hardened Concrete
Strength
Usually compressive strength is increased 10 to 20 percent by use of a
water-reducing admixture based on lignosulfonate or a salt of a hydroxylated
carboxylic acid [16]. The percentage of strength gain at 3 and 7 days is usually higher than that at 28 days, while tests up to 5 years in duration have
shown a continued strength benefit.
As pointed out earlier, lignosulfonate water-reducers normally entrain
some air. The portion of the water reduction attributable to the air content
can, with suitable adjustment of the amount of sand, compensate largely or
wholly for the loss of strength due to the entrained air. The balance of the
water reduction when using a lignosulfonate water reducer, and the reduction produced by a hydroxylated carboxylic water reducer, are effective in increasing the 28-day strength over that which would be produced with similar
concrete without admixtures. Further, this increase in strength is generally
greater than would be predicted from the reduction in w/c ratio.
In common with retarders, calcium chloride also shows maximum strength

gain at early ages [34]. Concrete with 2 percent calcium chloride has often
been found to be stronger at 1 year than similar concrete without the admixture. Accelerators are particularly effective at very early ages and at relatively
low temperatures.
Flexural strength generally is improved by all of these admixtures, but the
effect is less pronounced than that on compressive strength, and calcium
chloride may even cause some reduction in flexural strength at late ages.
Permeability
The permeability of concrete is not changed significantly by any of the
chemical materials, except that the concrete matures more rapidly, and in
the case of accelerators may be affected less by failure to provide effective
curing.
Drying Shrinkage
The available data on the effect of accelerators and retarders on drying
shrinkage are conflicting, probably because of the influence of variations in
test procedures employed. Tremper and Spellman, [23], using concrete
specimens of 3 by 3-in. (approximately 75 by 75-mm) cross section made
from a blend of Type II cements, moist cured for 7 days, and dried at 50 percent relative humidity for 28 days, found drying shrinkages of 8 to 17 percent
greater than similar concrete without admixtures for one hydroxylated and
two lignosulfonates, and shrinkages of 30 percent greater for calcium
chloride. The percent increase in shrinkage over the control concrete was
usually higher for shorter than for longer drying periods, particularly with
accelerators. The relative shrinkage where calcium chloride was added was
found to be dependent on the SO3 content of the cement used. Other investigators [22,34-36], using various procedures and drying times, have
reported drying shrinkage figures which are usually lower than those of
Tremper and Spellman.
In assessing the importance of these figures it should be kept in mind that
they are comparable in magnitude with effects which may be introduced
through choice of cement, choice of maximum aggregate size, contamination
of aggregate, and choice of aggregate source.
Resistance to Freezing and Thawing

While the entrained-air void spacing obtained with Class 1 water-reducing
retarders is slightly greater than that for an equivalent amount of entrained
air produced by typical air-entraining admixtures, the performance of concrete containing water-reducing retarders, as measured by freezing and
830
thawing tests, often has been found to be better than concrete of the same air
content, but without the water-reducing retarders. This increase might be
the result of the reduction in w/c ratio.
Calcium chloride has been found to improve the early resistance of concrete to freezing and thawing, either in the presence or absence of iceremoval salts, but to reduce somewhat the eventual durability of the fully
cured concrete [34,35, 37].
Resistance to Sulfates
Laboratory tests have shown some small improvement in sulfate
resistance, through the use of water-reducing retarders [35], but the use of
calcium chloride has been found to decrease somewhat the resistance of concrete to sulfates [32].
Heat of Hydration
The chemical admixtures have been shown to have little, if any, effect on
the total heat liberated during the hydration of cement [21,35], but they may
have a pronounced effect upon the rate at which the heat is liberated. Concrete containing calcium chloride liberates heat earlier, as would be expected
from the earlier strength development.
Corrosion of Metals
The addition of calcium chloride to reinforced concrete has not been found
to contribute significantly to corrosion of the reinforcing. If the concrete is
proportioned properly, consolidated properly to form a continuous contact
with the steel, and of adequate cover thickness, no corrosion problems normally are encountered. However, chlorides should not be used in prestressed
concrete [38-40]; in concrete where stray currents are present [41]; in concrete in which dissimilar metals are imbedded, such as aluminum conduit
and steel reinforcement [42]; or where galvanized forms are to be left in place
[43]. Also calcium chloride may bring about corrosion where elevated
temperature is employed during curing [44]. By contrast, another accelerator, stannous chloride [44], when properly used, was found not to contribute to corrosion. The usual water-reducing retarders, unless modified by
the addition of chlorides, have not been found to bring about corrosion.
There is some evidence that retarding admixtures, when used in conjunction
with chlorides, reduce the increased electrical conductivity which normally
would result from use of the chlorides, and hence the resulting corrosion
[411.
Applications
Water-Reducing Admixtures
Water-reducing admixtures may be used with no change in cement content
and slump to produce concrete with a lower w/c ratio; with no change in cement content and w/c ratio to produce a higher slump; or with reduced cement content and unchanged w/c ratio and slump. In the first case the usual
benefits accruing from the use of a lower w/c ratio normally will be obtained,
and, in many cases an increase in strength greater than normally produced
by the reduction in water possibly may result. In the second application
easier placing of concrete, or a higher slump with delayed placing of the concrete, may be obtained. In the third application a reduced cost may result.
Retarders
Retarders may be used to delay the setting of concrete during hot weather,
or to extend the vibration limit so that large members can be cast and consolidated without cold joints, and without damage to the freshly placed concrete due to settlement of forms as concreting proceeds.
Accelerators
Accelerators have their primary application in cold weather concreting
where they may be used to permit earlier starting of finishing operations, and
in certain cases, the application of insulation; reduce the time required for
curing; and permit earlier removal of forms or loading of the concrete.
Dosage
In the case of water-reducing and retarding admixtures dosages required
to produce specific results usually are recommended by the manufacturers.
Variation in the dosage often can be made to obtain the desired concrete
properties under particular job conditions. In other cases, the manufacturer
may change the formulation of the admixture to suit the conditions under
which it will be used. Increase in the dosage in a multipurpose material to obtain one particular effect might not be feasible because, as an example, too
much or too little air entrainment might result.
Calcium chloride usually is added in amounts of 1 to 2 percent by weight of
the cement, and the latter figure should not be exceeded.
Specifications
A specification for a concrete admixture should serve several purposes.
1. It should provide test methods and specification limits by which the
832
material to be tested may be judged as to its general ability to perform the

functions for which it is purchased. Tests for strength, water requirement,
and setting time measure properties important in the use of chemical admixtures.
2. Specifications should provide test procedures and test limits against
which the material may be judged from the standpoint of not producing
deleterious properties in the concrete. Excessive drying shrinkage, low
resistance to freezing and thawing, and excessive setting time are examples of
such properties.
3. The specification should provide procedures whereby the performance
of a particular admixture with particular sources of cement and aggregate,
and specific job conditions can be assessed.
4. A specification should provide a ready means of identifying materials in
successive shipments, both from the standpoint of composition and concentration, to give the user some assurance that the material being used is
uniform and is the same as that which was tested.
The current version of the ASTM Specification C 494 recognizes each of
these needs. A test procedure is outlined using a blend of cement, aggregates
of specified grading, and mixture proportions based on ACI Standard 211.1,
Recommended Practice for Selecting Proportions for Normal and
Heavyweight Concrete, and it recommends that, where practicable, tests be
made using the cement, aggregates, and air-entraining admixture proposed
for the specific work. It contains recommendations that can serve to assure
that a particular admixture previously subjected to all the specification tests
and found satisfactory when tested with specific aggregates, air-entraining
agent, and blend of cement would perform satisfactorily with the other concrete ingredients on a particular job. The materials covered by ASTM
Specification C 494 are often found particularly useful at temperatures well
below or above the laboratory temperature of 23~ (73~ required by the
specification, and a few tests at temperatures anticipated in the field might
be profitable.
The specification requires the manufacturer to recommend, upon request,
appropriate test procedures for establishing the equivalence of materials
from different lots. Qualitative, and to some extent quantitative, compositional analysis of water-reducing retarders may be obtained by infrared or
ultraviolet absorption spectroscopy [45, 46]. One alternative to this type of
test consists of watching the rate of hardening, water requirement, and
slump in the job concrete as use of a new lot of admixture is started, while
another involves laboratory concrete tests using the new batch and a reserve
sample from the initial batch [47].
The complete sets of tests prescribed in ASTM Specification C 494 is so extensive, and requires such an extended period of time, that it is unlikely that
it will be performed more than a few times for any given admixture, and
perhaps never for specific lots of materials used in construction work. These
833
tests are desirable and necessary to establish the potential value of an admixture, and the test data should be available to the purchaser, but as a practical
matter, less extensive tests might well be used for assessing the admixture's
performance with the specific materials and conditions of given concrete
work.
Since its introduction, the specification has been undergoing constant improvement and clarification, and further revisions no doubt will be made.
In establishing specifications for concrete admixtures, allowance must be
made not only for the statistical variation of test results on the concrete containing the admixture, but also that of the control concrete without admixture. For example, if it is intended that a water-reducing admixture increase
the 28-day compressive strength by 20 percent, and the limit therefore is set
at 120 percent of the control, an admixture which actually, on the average,
will produce an increase of 20 percent in strength would fail to meet the
specification requirement 50 percent of the time. The limits given in Table 1
of ASTM Specification C 494 take this factor into account by lowering the
strength requirements and the water requirement by approximately 10 percent. These figures were selected originally as a matter of good judgment,
and they are subject to future change in the light of more extensive statistical
considerations of the problem. The specification requirement for Type B
(retarding) admixture that the strength of treated concrete at any age be at
least 90 percent of that of the untreated control has the objective of requiring
that there be no sacrifice in strength due to use of the admixture. However,
under the requirements of the specification, an admixture which actually
caused a 10 percent reduction in strength would be rejected, on the average,
only 50 percent of the time. The rate of rejection for true strengths above or
below this point depends on the standard deviation of the test data and the
number of test specimens. Protection to the consumer against acceptance of
inferior products can be increased by improving test procedures so that the
variance is lowered, increasing the number of specimens, or both.
The balance between probability of accepting an inferior product and rejecting a satisfactory one can be adjusted within limits, but substantial improvement can be achieved only by further reduction in the variance through
better testing and test methods, or the use of more specimens.
References
[1] Significance of Tests and Properties of Concrete and Concrete Aggregates, A S T M STP
169, American Society for Testing and Materials, (1955) 1956.
[2] Significance of Tests and Properties of Concrete and Concrete-Making Materials, A S T M
STP 169-.4, American Society for Testing and Materials, 1966.
[3] Foster, B. E., "Chemical Admixtures," A S T M STP 169-.4, American Society for Testing
and Materials, 1966, pp. 556-564.
[4] Proceedings, International Symposium on Admixtures for Mortar and Concrete, RILEM,
Brussels, Vol. 8, 1967, including report of RILEM Working Group on Admixtures,
"Classification and Definitions of Admixtures."
834
[5] "Guide for Use of Admixtures in Concrete," Journal, American Concrete Institute;
Proceedings, Vol. 68, 1971, pp. 646-676.
[6] "Admixtures and Special Cements," Proceedings, 5th International Symposium on The
Chemistry of Cement, Tokyo, 1968, Part 4, The Organizing Committee, Cement Association of Japan, Tokyo, 1970, including "Use of Surface-Active Agents in Concrete," in
Proceedings, International Symposium on the Chemistry of Cement, Tokyo, Part 4,
pp. 1-29.
[7] "The Science of Admixtures," The Concrete Society, London, 1970.
[8] "Admixtures in Concrete-Accelerators, Air Entrainers, Water Reducers, Retarders,
Pozzolans," Special Report 119, Highway Research Board, 1971.
[9] Joisel, Albert, Admixtures for Cement (Les Adjuvants du Ciment), published by the
author, Paris, 1973.
[10] Odler, Ivan, "Admixtures," Cements Research Progress 1974, Cements Division, American Ceramic Society, Columbus, Ohio, 1975, pp. 79-91.
[11] Odler, Ivan, "Admixtures," Cements Research Progress 1975. Cements Division, American Ceramic Society, Columbus, Ohio, 1976, pp. 83-95.
[12] Bonzel, J. and Krumm, E,, "Concrete Additives," VDZ Zement Taschenbuch, 1974/75,
Bauverlag GmbH, Wiesbaden-Berlin, 1974.
[13] Ivanov, F. M., Batrakov, V. G,, and Lagoida, A. V., "Additives for Concrete and Construction Mortar," Beton Zhelezobeton, Vol. 6, 1974, pp. 2-5.
[14] Vavrin, F., "Effect of Chemical Additions on Hydration Processes and Hardening of
Cement," Preprint 11, International Symposium on the Chemistry of Cement, Moscow,
1974.
[15] Skalny, J. and Maycock, J. M., "Mechanisms of Acceleration by Calcium Chloride: A
Review," Journal of Testing and Evaluation, Vol. 3. No. 4, 1975, pp. 303-331.
[16] "Admixtures for Concrete," Journal, American Concrete Institute; Proceedings, Vol. 60,
No. 11, Nov. 1963, pp. 1481-1524.
[17] Steinour, H. H., "Concrete Mix Water--How Impure Can it Be?" Journal. Research and
Development Laboratories, Portland Cement Association, Vol. 2, No, 3, Sept. 1960,
pp. 32-50.
[18] Lea, F. M., The Chemistry of Cement and Concrete, 3rd ed., Chemical Publishing Co.,
[19[ Prior, M. E. and Adams, A. B., "Introduction to Producers' Papers on Water-Reducing
Admixtures and Set-Retarding Admixtures for Concrete," Effect of Water Reducing and
Set Retarding Admixtures on Properties of Concrete, A S T M STP 266, American Society
for Testing and Materials, 1960, p. 170.
[20] Vollmer, H. C., "Calcium Chloride in Concrete," Bibliography No. 13, Highway Research
Board, 1952 (annotated).
[21] Newman, E. S., Blaine, R. L., Jumper, C. H., and Kalousek, G. L., "Effects of Added
Materials on Some Properties of Hydrating Portland Cement Clinkers," Journal of Research, U.S. National Bureau of Standards, Vol. 30, 1943, p. 281.
[22] Mather, Bryant, "Effects of Three Chemical Admixtures on the Properties of Concrete,"
Miscellaneous Paper No. 6-123, Report No. 3, Corps of Engineers, U.S. Army, May 1956.
[23] Tremper, Bailey and Spellman, D. L., "Shrinkage of Concrete--Comparison of Laboratory and Field Performance," Record No. 3, Highway Research Board, 1963, p. 30.
[24] Polivka, Milos and Klein, Alexander, "Effect of Water-Reducing Admixtures and SetRetarding Admixtures as Influenced by Portland Cement Composition," Effect of WaterReducing Admixtures and Set-Reducing Admixtures on Properties of Concrete, ASTM
[25] Tuthill, L. H., Adams, R. F., Bailey, S. N., and Smith, R. W., "A Case of Abnormally
Slow Hardening Concrete for Tunnel Lining," Journal, American Concrete Institute;
Proceedings, Vol. 57, No. 9, March 1961, p. 1091; discussion by K. E. Palmer,
Proceedings, Vol. 57, p. 1828, and Bailey Tremper, Proceedings, Vol. 57, p. 1831.
[26] Seligman, Paul and Greening, N. R., "Studies of Early Hydration Reactions of Portland
Cement by X-Ray Diffraction," Record No. 62, Highway Research Board, 1964, p. 80,
[27] Bruere, G. M., "Importance of Mixing Sequence when Using Set-Retarding Agents with
Portland Cement," Nature, Vol. 199, 1963, p. 32.
[28] Dodson, V. H. and Farkas, E., "Delayed Addition of Set-Retarding Admixtures to PortCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:31:08 EDT 2014
835
land Cement Concrete," Proceedings, American Society for Testing and Materials, Vol.
64, 1964, pp. 816-829.
[29] Vollick, C. A., "Effect of Water-Reducing Admixtures and Set-Retarding Admixtures on
the Properties of Plastic Concrete," Effect of Water-Reducing and Set-Retarding Admixtures on Properties of Concrete, ASTM STP 266, American Society for Testing and
[30] Hewlett, Peter and Rixom, Roger, "Superplasticised Concrete," Concrete, Sept. 1976,
pp. 39-42.
[31] Sasse, H. R., "Water-Soluble Plastics as Concrete Admixtures," Proceedings, 1st International Congress on Polymer Concretes, Paper No. 2, London, 1975.
[32] "Flieszbeton," pages 327-350, VDZ Zement Taschenbuch 1976/77, Bauverlag GmbH,
Wiesbaden-Berlin.
[33] Tynes, W. O., "Investigation of Proprietary Admixtures," USAEWES TR C-77-1,
April 1977, Vicksburg, Miss.
[34] Shideler, J. J., "Calcium Chloride in Concrete," Journal, American Concrete Institute;
Proceedings, Vol. 48, No. 7, 1952, p. 537.
]35] Wallace, G. B. and Ore, E. L., "Structural and Lean Mass Concrete as Affected by
Water-Reducing, Set-Retarding Agents," Effect of Water-Reducing and Set-Retarding
Admixtures on Properties of Concrete. ASTM STP 266. American Society for Testing and
[36] Grieb, W. E., Werner, G., and Woolf, D. O., "Water-Reducing Retarders for Concrete,"
Public Roads, Vol. 31, 1961, p. 136.
[37] Klieger, Paul, "Curing Requirement for Scale Resistance of Concrete," Bulletin No. 150,
[38] Evans, R. H., "Effects of Calcium Chloride on Prestressing Steel and X-Rays for Anchorage Lengths," Proceedings, World Prestressed Conference, San Francisco, Calif., 1957.
[39] Monfore, G. E. and Verbeck, G. J., "Corrosion of Prestressed Wire in Concrete," Journal,
American Concrete Institute; Proceedings. Vol. 57, No. S, Nov. 1960, pp. 491-516.
[40] Godfrey, H. T., "Corrosion Tests on Prestressed Concrete Wire and Strand," Journal,
Prestressed Concrete Institute, Vol. 5, No. 1, 1960, pp. 45, 48-51.
[41] Kondo, Yasuo, Takeda, Akihilo, and Hideshima, Setsuji, "Effect of Admixtures Electrolytic Corrosion of Steel Bars in Reinforced Concrete," Journal, American Concrete Institute; Proceedings, Vol. 56, No. 4, 1959, pp. 299-312.
[42] Wright, T. E., "An Unusual Case of Corrosion of Aluminum Conduit in Concrete,"
The Engineering Journal, Vol. 38, 1955, p. 1357.
[43] Mange, C. E., "Corrosion of Galvanized Steel in Contact with Concrete Containing
Calcium Chloride," 13th Annual Conference, National Association of Corrosion Engineers, St. Louis, Mo., 1957; abstract, Proceedings, American Concrete Institute, Vol. 54,
Nov. 1957, p. 431.
[44] Vivian, H. E., "Some Chemical Additions and Admixtures in Cement Paste and Concrete," Monograph 43. Proceedings, 4th International Symposium on Chemistry of
Cement, National Bureau of Standards, Vol. 2, 1960, p. 909.
[45] Swenson, E. G. and Thorvaldson, T., "Detection of Lignosulfonate Retarder in Cement
Suspensions and Pastes," Effect of Water-Reducing and Set-Retarding Admixtures on
Properties of Concrete, A S T M STP 266, American Society for Testing and Materials,
1960, p. 159.
[46] Halstead, W. J. and Chaiken, Bernard, "Water-Reducing Retarders for Concrete-Chemical and Spectral Analyses," Bulletin 310, Highway Research Board, 1962, p. 33.
[47] Tuthill, L. H., Adams, R. F., and Hemme, J. M., Jr., "Observations in Testing and Use
of Water-Reducing Retarders," Effect of Water-Reducing and Set-Retarding Admixtures
on Properties of Concrete, A S T M STP 266, American Society for Testing and Materials,
1960, p. 97.
L. M. L e g a t s k i ~
Chapter 48--Cellular Concrete
Introduction
The basic contribution of cellular concrete to the field of concrete
technology is the ability to control the density of concrete over a wide range.
Density control is achieved by adding a calculated amount of a proper foam
to a slurry of water and cement, with or without the addition of sand or other
aggregate. The wet density range for cellular concrete mixes is usually considered to be from about 320 to 1920 kg/m 3 (20 to 120 Ib/ft3).
Reducing the density of concrete by adding foam, of course, is accompanied by reduction in all of the strength properties as well as a reduction in
thermal conductivity. Since density, strength, thermal conductivity, and
member cross section may now be considered as variables, it becomes possible to select a concrete density and size of element that will satisfy strength
requirements and, in addition, provide increased insulating value and reduced weight. The structural elements referred to above usually will be those
which must resist modest loads on a relatively small span. Examples of
nonstructural cellular concretes of broad utility are those used primarily for
insulation, for filling space, or both.
It is the objective, in this chapter, to discuss cellular concretes in general,
their physical properties, the procedures for mix design, and finally the range
of their applications.
General
Cellular concretes are lightweight concretes. They consist of a system of
macroscopic air cells uniformly distributed in either a matrix of cement paste
or of aggregate and cement paste. The cell size varies approximately from
0.10 to 1 mm (0.004 to 0.04 in.). The skin of the air cells must be tough and
persistent in order to withstand the rigors of mixing and placing, during
which periods the cells are separated, coated with cement paste, and the concrete is pumped or transported to the casting position.
lprofessor emeritus of civil engineering, Universityof Michigan, Ann Arbor, Mich. 48104.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
LEGATSKI ON CELLULAR CONCRETE
837
In this discussion it is assumed that the air cells are preformed as foam and
added to the slurry in the mixer. It is also possible to form the air cells in the
slurry by chemical reaction or by vigorous mixing of the slurry with a proper
foam concentrate in a high-speed mixer. The chemical reaction method of
foaming is suited only to precasting and requires a large capital investment in
plant. In the high-speed mixing method the volume of air cells produced
depends on the amount and properties of the foam concentrate and the mixing time as well as the temperature of the water and other materials. Consequently the control of batch density is difficult at best.
In reference to the density of cellular concrete, confusion can be avoided
by always stating the moisture condition of the mix along with the density.
The significant moisture conditions are wet density (density of the fresh concrete), air-dry density (at stated age and curing condition), and the oven-dry
density.
The air-dry density of cellular concrete probably represents, as well as any
other density measure, the condition of the material in its place in the construction. The change in density due to air-drying is a functio~ of
temperature, humidity, duration of the drying period, the wet density of the
concrete, the water-cement (w/c) ratio, and the surface-area ratio of the element. The relation of air-dry density to wet density would therefore seem to
be a complicated one. However, for laboratory conditions: temperature 21
to 24~ (70 to 75~ relative humidity 50 percent and drying period 28 days,
the air-dry density of 152 by 305 mm (6 by 12 in.) cylinders is about 80
kg/m 3(5 lb/ft a) less than the wet density for cellular concretes of wet densities from 640 to 1920 kg/m a (40 to 120 lb/fta). Cylinders made, cured, and
air-dried under job conditions in the west and southwest states have shown
density losses up to about 160 kg/m 3 (10 lb/ft3). These latter cylinders were
76 by 152 mm (3 by 6 in.) of concrete whose wet density ranged from 480 to
575 kg/m 3 (30 to 36 lb/ft3). Oven-dry density of cellular concrete usually is
used only for determination of thermal conductivity. For this purpose the
oven-dry density may be calculated with sufficient accuracy from the mix
data by assuming that the water required for hydration of the cement is 20
percent of the weight of the cement. The oven-dry density, D, is calculated
as follows
D = 1.2c +
Akg/m3( 1"2c + A lb/ft3~

\
27
(1)
where
c = weight of cement kg/m 3 (lb/yd 3) of concrete, and
A = weight of aggregate kg/m a (lb/yd a) of concrete.
There are no definite upper and lower limits of the wet density of cellular
concrete mixes. In practice the approximate wet density range is considered
to be from 320 to 1920 kg/m a (20 to 120 lb/ft3).
838
Types of Cellular Concretes

The various kinds of cellular concretes may be classified and briefly
described as follows.
1. Neat Cement Cellular Concrete is made using portland cement, water,
and preformed foam. It contains no solid aggregates. Neat cement cellular
concrete usually is limited to the low density range (wet density up to about
720 kg/m 3 (45 lb/ft3).
2. Sanded Cellular Concrete is a cellular concrete that contains fine aggregate (sand) in addition to cement, water, and preformed foam. These concretes are usually limited to a wet density range from about 800 to 1920
kg/m 3 (50 to 120 lb/ft3).
3. Lightweight Aggregate Cellular Concrete is similar to sanded cellular
concrete but with lightweight aggregate replacing part of the sand. The
lightweight aggregates used should be of structural grade in order to increase
the strength-density ratio.
4. Modified Cellular Concretes--Any of the above types of cellular concrete mixtures can be modified beneficially by the use of additives. The
following additives have been used successfully in cellular concrete.
(a) Cement Dispersing Agent--This additive acts to disperse the cement
particles in the batch, and to increase compressive strength especially for wet
densities of 1440 kg/m 3 (90 lb/ft 3) and above. The strength increase varies
from about 10 percent at a wet density of 1440 kg/m 3 (90 lb/ft 3) to about 40
percent at a wet density of 1760 kg/m 3 (110 lb/ft3).
(b) Chopped Fiber Glass--Alkali resistant fiber glass is added to low density cellular concrete to increase the tensile strength and thus help to control
shrinkage cracking. The amount used is usually a compromise based on effectiveness, desired or required workability of the concrete, and cost. A suggested lower limit for fiber glass, based on effectiveness, is a volume of glass
fiber equal to about 0.5 percent of the absolute volume of cement. The upper
limit depends upon the projected use of the concrete, its required workability, and the cost. The range of fiber length is from about 20 to 40 mm (0.75 to
1.5 in.). The lower limit length is based on bond with the cement paste. The
upper length limit is related to the increasing difficulty of dispersing fibers in
the mix as their length is increased. Balanced against the dispersion problem, any claimed advantage is questionable.
(c) Expansive Additive--There are advantages to be gained by using an
expansive additive or a shrinkage compensating cement in making reinforced
cellular concrete members. The tendency of the member to expand is
restrained by the reinforcing steel, causing a modest level of prestress which
partially compensates for the effects of drying shrinkage. However, the centroid of the area of reinforcing steel must be at or close enough to the centroid of the section to avoid unwanted curvature of the member due to eccentric prestress.
839
Physical Properties
The density of cellular concrete may be varied over a wide range, from 320
to 1920 kg/m 3 (20 to 120 lb/ft3). The density is an additional variable
encountered in working with cellular concrete. This additional variable is
accompanied by advantages and disadvantages. The lower densities and the
accompanying lower values of thermal conductivity, than stone aggregate
concrete, are among the important advantages of cellular concrete. The principal disadvantage is at least an expected one--the strength properties are
reduced by adding air cells (preformed foam) to reduce the density. However,
there is a compensation for this disadvantage. It can be considered as the
price that must be paid for the ability to reduce both density and thermal
conductivity by adding stable air cells.
The principal factors affecting the compressive strength of cellular concrete are: (a) density, (b) cement content, (c) aggregate, type and amount,
(d) total water content, (e) additives, and (/3 curing conditions. Sufficient test
results have not been collected to document the compressive strength and
other physical properties across the full density range, including also all of
the mixes modified by ,carious additives. However, information is at hand
9(see appended figures) for neat cement cellular concretes up to wet density of
720 kg/m 3 (45 lb/ft 3) with no additives and sanded cellular concretes from
800 to 1920 k g / m 3 (50 to 120 lb/ft3). The sanded mixes up to 1280 kg/m 3 (80
lb/ft 3) had no additives while those above 1360 kg/m 3 (85 lb/ft 3) contained a
cement dispersing agent. All of the mixes were made with Type I cement.
Those at or above wet density 800 k g / m 3 (50 lb/ft 3) had a cement content of
390 k g / m 3 (658 lb/yd3). The cement factor for neat cement mixes is
calculated after selection of the w/c ratio, as will be discussed under "Design
of Mixes." Figure 1 shows a curve of compressive strength versus wet density
(unit weight) for the cellular concrete mixes described above. Similar curves
can be drawn for different cement factors and for various additives.
The modified cellular concrete mixes shown in Table 1 demonstrate that
special admixtures can be quite beneficial in increasing the compressive
strength.
Tensile Strength
The tensile strength of cellular concrete bears about the same relation to
its compressive strength as does stone aggregate concrete, which is 10 to 15
percent. Therefore, the cellular concretes of lower density have very low tensile strength. Thus the use of an additive such as chopped fiber glass to inCopyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:47:39 EDT 2014
840
641
320
961
1282
SANDED
1602
1922kg/m3
MIXES
SACKS~Y-D 3.
i
27.6
DISPERSING
I-- 3
o
z
- - ~
20.7
7~
13.8s
Ill
ae
I2
Ill
1
-
E
eL
6.9
20 30 40 50 60 70 80 90 100 110 20
UNIT
FIG.
WEIGHT
(WET),
Iblft 3
1--Compressive strength versus unit weight.
crease tensile strength is common practice in the application of low density

cellular concretes. Of course the fiber glass should be alkali-resistant.
Shear Strength
Beam specimens of dimensions 14.3 by 17.5 by 127 cm (5 % by 6 7/a by 50
in.) were used for shear tests. They were reinforced with two No. 4 bars
placed at an effective depth of 12.7 cm (5 in.). The beams were supported at
a span of 91.4 cm (36 in.) and were loaded at midspan. Test results are
shown in Fig. 2 where the ultimate shear stress is plotted against the wet density of the concrete. Shown also is the nominal permissible ultimate shear
stress for lightweight concrete according to American Concrete Institute
(ACI) Building Code Requirements for Reinforced Concrete (ACI 318).
Comparison indicates that the ACI Code is applicable to cellular concrete.
Bond Strength
T h e data for bond strength are, unfortunately, quite old and are the
results of pull-out tests of No. 4 and No. 5 bars embedded in the full length of
~These values are for a sanded mix.
MN/m 2
psi
296
5
0
86
volume as required for density
3.8
550
0
0
0
0
3
0
51
694
429
45
721
412
255
lb/yd 3
Wet density
Cement, Type I
Water (total)
Aggregate
sand
ceramic beads
Additives
fiber glass
25 m m (1 in.) long
Liquid latex
Expansive additives
Preformed foam
(with no additives)
lb/ft 3
kg/m 3
Item
Mix A
0
105
0
0
262
1041
418
247
kg/m 3
65
lb/ft 3
TABLE 1--Typical modified cellular concrete mixes quantities are for 1 m 3 (1 yd3).
0
176
o
0
442
805
416
lb/yd 3
Mix B
1630
460 a
3.2 ~
psi
11.2
MN/m 2
co
Fn
.-n
0
0
z
c
i-:IJ
t'n
r'i--
O
Z
O
(9
f-/'11
842
200
190
180 - 170
o.
1281
1442
1602
e1922 kg/m3
~.z
A: - Average Ultimate From Beam Tests

/P" _
/
of Cellular Concrete.
B: Permissible Shear Stress For
~
1.24
Sand Lightweight Concrete
f
(ACl-318 71),
/~
o ~/~
160
'
150
I-
1762
Avg.-15% ~ . / ,
140
o e
me /
0.97
0
Z
lm
130
120
/?~-
110
n'
< 100
W
Z
90
1/)
///
g 15%
0.69
8O
I
70~
6o!
~
5O
70
80
UNIT
~c a
90
WEIGHT
100
(WET),
FIG. 2--Shear strength versus
110
Ib/ft 3
120
unit weight.
152 by 305-mm (6 by 12-in.) cylinders of cellular concrete of various densities. The dotted line in Fig. 3 is intended as a reasonable approximation of
the lower boundary of the test results. It would seem that a new test series for
development bond is needed.
Modulus of Elasticity
The modulus of elasticity of cellular concrete is a function of its density
and compressive strength. It is reasonable to assume that cellular concrete
should have a somewhat lower modulus of elasticity than concrete of the
same density but made with a more rigid aggregate whose density is reduced
by enclosed voids (manufactured lightweight aggregate). In this statement
the air cells are treated as aggregate.
In a laboratory study of modulus of elasticity of cellular concretes whose
density (wet) varied from 1281 to 1842 kg/m 3 (80 to 117 lb/ft 3) the measured
LEGATSKI
ON CELLULAR
843
CONCRETE
kg/m 3
1282
1442
1602
1762
700
e - No.
a-No.
+-No.
a-No.
I"
4
5
4
5
Ba'r,
Bar,
Bar,
Bar,
No Dispersing Agent. +
No Dispersing Agent.
with Dispersing Agent.+
with Dispersing A g e n t . ;
-p ~
600
i
Z
W
E
I- 500
4.83
,~
~"
4.14
m
o4
E
Z
#4+\
a
Z
0
W 400
I# 4 o & #4+
O
I 300
.,I
,,J
L
~'
c~
1922
,oo ~
,//~-.~
9
-/
_ _
db
4*
2.76
(Not more than 425)
2.07
70
7'5
CELLULAR
80
85
CONCRETE
90
95
UNIT
100
105
WEIGHT
110
115
120
(WET),
Iblft
F I G . 3--Pull-out bond strength versus unit weight9
modulus was plotted against the measured density (wet) as shown in Fig. 4.
Most of the specimens tested were from mixes containing a cementdispersing agent. Points indicated on the graph by unshaded circles represent mixes without a dispersing agent.
The equation selected to best interpret the trend of the test results is
Ec = W 1.s 37.04 fx/~'--~2, (W L5 28.6 x/j~" psi)
(2)
where
W
=
(wet density - - 80) kg/m 3, (wet density -- 5) lb/ft 3, and
f~' = 28-day compressive strength, N / m 2 (psi).
The dotted lines on each side of the Ec line are intended as a guide by which
to judge the scatter of data.
The study was not extended to the densities below 1281 kg/m 3 (80 lb/ft 3)
because, at that time, it was considered likely that the very low density concretes would be used principally for insulation and that there would be little
844
2.5
1281
.
D,
?
o
v-
1442
.
kg/m31602
1922
1762
*---Mixes containing a cement - dispersing agent.

ec= W '5 2 8 . 6 v ~
I
w = Unit Weight (wet)-5 pcf"
f'c = Comp. Strength
.[ / /
/7~ . /
>.
i-
Ec+ 10%
/"_"
; / / ~~,/
_\
1.6
17.24
13.79
"9"
10.34
/g-o
Ec
I/I
.i
ul
--///I
--. / y ~,,-'%~x__Ec 10%
/ 2 S~"
"
Ul
Q
O
Z
,.I
i
6.89
3.45
0.5
0
76
80
86
UNIT
90
96
WEIGHT
100
105
(WET),
110
116
Iblft
120
FIG. 4--Modulus of elasticity versus unit weight.

interest in their stiffness. That prediction is being proven false by a growing
interest in low density cellular concrete to resist calculated loads.
T h e r m a l Conductivity, k
The thermal conductivity of a material is the time rate of transfer of heat

by conduction, through unit thickness, across unit area for unit difference of
temperature. The units of k are watts/metre-Kelvin or W / m k (BTU 9 in/h 9',
ft 2 9 ~
Information from several sources is included in Fig. 5. The dashed line,
National Bureau of Standards curve, was developed from thermal conductivity tests on concrete made using no-fines gravel and lightweight manufactured aggregate. The test results plotted represent determinations of thermal
conductivity of cellular concretes for a limited range of densities from the
sources noted on Fig. 5. This information, assembled from four sources,
indicates that a single curve can be used to show the relation of thermal
conductivity to oven-dry density for both cellular and other concretes whose
density can be varied over a wide range.
10'
Notional Bureau of Standards curve

for No-fines gravel and lightweight
aggregate concrete
~
1.442
" Univ.of Michigan tests

>-P
>
8, 1.154
/,I
Rensseioer tests
//
o Pittsburgh Testing Lob.

0.865
" ~ /i
To=
845
i/
11
//
tests
*
~
0.577 ~II
ua I 9I- en
I"
//
II
2
0.288
Ib/ft 3 25
40
60
80
100
120
k g / m 3 400
644
961
1281
1602
1922
140
2242
DENSITY (OVEN DRY)

FIG.
S--Thermal conductivity versus density.
Design of Mixes
Following are two examples of mix designs. Example 1 is for a neat cement
cellular concrete mix, and Example 2 is for a sanded cellular concrete mix. It
is necessary to have at hand the foam gun calibration. This consists of the
following factors, values for which are taken from a calibration test.
foam volume/air volume ---- 1.05
(3)
unit weight of foam = 44.05 k g / m 3 (2.75 lb/ft 3)
(4)
foam rate = 0.454 m3/min (16.04 ft3/min)
(5)
Also required is an approximate or trial value of w/c by weight. Table 2 may

be used as a guide.
Example No. 1 (SI Units)

Design a neat cement cellular concrete mix to have a wet density of 641
k g / m 3. Use Type I cement and w/c = 0.57.
846
TABLE 2--Trial guide of w/c by weight.

Wet Density of Concrete
kg/m 3
641
721
961
1281
1602
1922
lb/ft 3
Trial Value of w/c

for Type I Cement
40
45
60
80
100
120
0.58
0.57
0.53
0.49
0.48
0.59
C a l c u l a t e w e i g h t o f c e m e n t , c, r e q u i r e d f o r 1 m 3
c + (w/c)c
= 641
from which c = 641/1.57 = 408.3 kg/m 3 of concrete.
Calculate the weights and absolute volumes of materials for 1 m a of

concrete
Absolute
m3
Weight, k g
Cement
Water
(total)
=0.57
408.3
408.3
Volume,
= 0.1295
3.15 1000
_ 232.7
1
-- 0.2326
641.0
1 1000
A b s o l u t e v o l u m e -----0 . 3 6 2 1
1 m 3 ~--- 1 . 0 0 0 0
Air volume required = 0.6379 m 3
Foam volume = 0.6379 1.05" = 0.6698 m 3

W e i g h t o f w a t e r in f o a m = 0 . 6 6 9 8 4 4 . 0 5 * = 2 9 . 5 k g / m 3 o f c o n c r e t e
Net water added at mixer = 232.7 -- 29.5 = 203.2 kg = 0.2032 m 3
F o a m t i m e r e q u i r e d = 0 . 6 6 9 8 / 0 . 4 5 4 * = 1.48 m i n = 1 r a i n 2 9 s
T h e t r i a l m i x , f o r 1 m 3 o f c o n c r e t e , is
Cement
Net water
F o a m v o l u m e -----0 . 6 6 9 8 m 3
Foam time
408.3 kg
203.2 kg = 0.2032 m 3
1 min 29 s
*See Eqs 3 to 5.
847
Example No. I (U. S. Customary Units)

Design a n e a t c e m e n t cellular concrete mix with a wet density of 40 l b / f t 3.
Use T y p e I cement a n d w / c = 0.57.
Calculate weight o f c e m e n t , e, r e q u i r e d for 1 yd 3
C+(w/c)
c =40
27=
1080
where
c = 1080/1.57 = 688 l b / y d 3 of concrete.
Calculate the weights a n d absolute volumes of m a t e r i a l s for 1 yd 3 of concrete.
Absolute Volume, i t 3
Weight, lb
Cement
W a t e r (total) = 0.57 688
688
3.15 62.4
392
1
3.500
X
6.282
1080
62.4
A b s o l u t e volume = 9.782
1 yd 3 = 27.000
-
Air volume r e q u i r e d = 17.218 ft 3

F o a m volume = 17.218 1.05" = 18.08 ft 3
W e i g h t o f water in f o a m = 18.08 2.75* = 49.7 l b / y d 3 o f concrete
Net water a d d e d at m i x e r = 392 - 49.7 = 342.3 lb = 41 gal
F o a m t i m e r e q u i r e d = 18.08/16.04" = 1.127 min = 1 m i n 08 s
T h e trial mix, for 1 yd 3 of concrete is
Cement
688 lb
Net water
342.3 lb = 41 gal
F o a m volume = 18.08 ft 3
Foam time
1 m i n 08 s
Example No. 2 (SI Units)

Design a cellular concrete mix to have a wet density of 1602 k g / m 3. Use
T y p e I cement, 418.26 k g / m 3 ; e s t i m a t e w / c = 0.46; a s s u m e the sand contains 2 weight p e r c e n t moisture.
Calculate the weights a n d absolute volumes of the m a t e r i a l s for 1 m 3 of
concrete
848
Absolute
Volume, m 3
Weight, kg
Cement
418.26
W a t e r (total) = 0.46 X 418.26 =
192.40 X
C a l c i u m ligno-sulfanate
1.11
611.77
W e i g h t of 1 m 3
S a n d (dry)
3.15 1000
1
1
1000
= 0.1328
- - 0.1924
neglect
= 1602.0
990.23 X
1
- - 0.3737
2.65 X 1000
A b s o l u t e volume = 0.6989
1 m 3 = 1.0000
Air volume r e q u i r e d = 0.3011 m 3

F o a m V o l u m e = 0.3011 1.05" = 0.3162 m 3
W e i g h t of water in f o a m ---- 0.3162 X 44.05* -= 13.93 k g / m 3 o f concrete
Correction for s a n d moisture = 0.02 X 990.23 = 19.80
Net water a d d e d as mixer = 192.40 - 13.93 - 19.80 --- 158.67 kg
-- 0.1587 m 3
S a n d correction; A d d to s a n d an a m o u n t equal to weight o f s a n d moisture:
990.23 + 19.80 = 1010.03 kg
F o a m t i m e r e q u i r e d ---- 0 . 3 1 6 2 / 0 . 4 5 4 " = 0.6965 min = 0 rain 42 s
T h e trial mix, for 1 m3 of concrete, is
Cement
418.26
Net water a d d e d
158.67
1.11
Calcium ligno-sulfanate
S a n d (with 2 p e r c e n t water) 1010.03
F o a m t i m e = 0 min 42 s
W e i g h t of f o a m water
13.93
kg
kg
kg
kg
S u m o f weights = 1602.00 k g / m 3
Example No. 2 ( U.S. Customary Units)

Design a cellular concrete mix to have a wet density of 100 l b / f t 3. Use
Type I cement, 7.5 sacks p e r yd 3. Use 1A lb of c a l c i u m lignosulfanate per
sack o f cement. E s t i m a t e w/c = 0.46; sand c o n t a i n s 2 percent moisture.
Calculate the weights a n d a b s o l u t e volumes o f the m a t e r i a l s for 1 yd 3 of
concrete.
Absolute
Volume, f l 3
Weight, lb
Cement 7.5 X 94
Water 705 X 0.46
1
3.15 X 62.4
1
324.3 X 62.4
1.9
1031.2
705 X
849
--
3.587
5.197
Weight of 1 yd 3 = , 2 7 X 100 2700.0

1
Sand (dry)
1668.8 X 2.65 X 62.4-- 10.092
Absolute volume = 18.876
1 yd 3 : 27.000
Air volume required = 8.124 ft 3
Foam volume ---- 8.124 X 1.05" = 8.53 ft 3
Weight of water in foam = 8.53 X 2.75* = 23.5 lb/yd 3 of concrete
Correction for moisture in sand: 0.02 X 1668.8 ---- 33.4 lb
Net water added at mixer = 324.3 -- 23.5 -- 33.4 = 267.4 lb = 32.1 gal
Foam time required = 8.53/16.04 = 0.5318 min = 0 min 32 s
Add to sand an amount equal to weight of moisture in the sand = 168.8 +
33.4 = 1702 = total weight of sand
The trial mix, for 1 yd 3 of concrete, is
Cement
705 lb
Water
267.4 lb = 32.1 gal
1.9 lb
Sand (moist)
1702.0 lb
2676.3 lb
Water in foam
23.5
2699.8 lb versus 2700 lb, check
Foam time 0 rain 32 s
Notes on Mix Design

1. If Type III cement is used, increase w/c by 0.03 to 0.05.
2. If fiber glass is added, increase w/c about 0.02 or more, depending on
amount of fiber. Also account for weight and volume of fiber in calculations.
3. Mixes designed according to the above procedures will often have a wet
density about 48 k g / m 3 (3 lb/ft 3) below the design density. This is largely
due to the presence of accidentally entrapped air in the mix. After the concrete has been pumped into place most of the accidental air will have been
removed. If the concrete is placed by means other than pumping a minor correction can be made in the design to account for about 2 percent of the
volume.
850
Applications
The number and kinds of applications for cellular concrete seem to be
limited only be one's imagination. The present applications are limited
almost entirely to nonstructural uses because of the general lack of published
information on properties of cellular concrete. It is used widely as a sound insulating fill in floors and as a heat insulating fill in roofs. These applications
are near the opposite ends of the density range, the floor fills being about
1682 kg/m3/(105 lb/ft 3) and the roof fills being about 561 k g / m 3 (35 lb/ft3).
Cellular concrete at a low density has been used recently as a road fill on
bridge approaches over a very poor subgrade which could not support an
earth fill. It can also be used as floats on water if care is taken to use a foam
chemical which produces a discrete bubble system.
A system has been designed and tested, which permits casting low density
cellular concrete in large floating fabric envelopes of controlled thickness,
with no support other than the water. The envelopes are filled while floating
on the water. The unit envelopes then are fastened together after curing to
form a floating cover for reservoirs to minimize algae growth and evaporation loss. For this kind of system, in particular, it is necessary to use a foam
chemical which produces a foam made up of a discrete bubble system. In
other words, the low density cellular concrete must float indefinitely.
Cellular concrete of various unit weights for different parts of the structure
has been used successfully to build a three-story apartment house in
Honolulu. It is expected that a demonstration house will be built soon in
Michigan.
Cellular concrete has many uses in construction of all types.
Standards
ASTM Testing Foaming Agents for Use in Producing Cellular Concrete
Using Preformed Foam (C 796) now has also been designated an American
National Standard by ANSI. This standard, ASTM/ANSI Method C 796, is
significant because it provides a method of testing foaming chemicals (foam
concentrates) by measuring common physical properties of a standard test
batch of cellular concrete made using the foam chemical to be tested. The
test batch contains only cement, water, and foam. Thus, the factors that can
influence physical properties are reduced to a minimum.
The ASTM Specification for Foaming Agents Used in Making Preformed
Foam for Cellular Concrete (C 869) was published in 1977. This specification
assigns limits to the physical properties of the test batch used in ASTM
Method C 796 making it possible to specify the quality of the foaming
chemical. This is the basic reason for developing both ASTM Method C 796
and the ASTM Specification C 869.
Three guides also have been developed from 1967 to 1975 by the ACI
851
which are applicable to cellular concrete. They are: (1) Guide for Cast-inPlace Low Density Concrete (ACI 64-44), (2) Guide for Low Density Precast
Concrete Floor, Roof, and Wall Units (ACI 65-38), and (3) Guide for
Cellular Concrete Above 50 lb/ft 3, and for Aggregate Concretes Above 50
lb/ft 3, with Compressive Strengths Less than 2500 psi (ACI 72-7).
R. J. Schutz
Chapter 49--Organic Materials for

Bonding, Patching, and Sealing of
Concrete
Introduction
Concrete is one of the most durable materials of construction. However, because concrete is inherently porous and alkaline, it can be attacked both physically and chemically by certain injurious solutions including acids, salts, and water. Penetration of concrete by water or injurious solutions is especially severe where the concrete is exposed to alternate cycles of wetting and drying or freezing and thawing, or both.
Since such attack can only occur in a wet environment, coatings quite
often are applied to concrete to prevent ingress of the solution. There is
no frost damage to concrete if it is dry [1],2 and all chemical attack requires the presence of water [2].
Absorption and subsequent chemical or physical attack also can be
reduced by polymer impregnation [3] or by inclusion of the polymer in
the plastic-concrete mixture in the form of a latex or unreacted premixed
epoxy resin and curing agent [4].
Organic materials which have been used for bonding, patching, and
sealing of concrete include epoxy resins, silicones, bitumens, linseed oil,
oil-based paints, acrylics, urethanes, polyvinyls, and so-called rubberbased coatings. For these uses the epoxy resin systems have proven to be
the most versatile.
Bonding and Patching Materials

Materials in common use for bonding and patching of concrete fall
into three general groups: (1) bonding admixtures, (2) adhesives, and
(3) resinous mortars.
1Construction materials consultantl RR 1, Augusta, N.J. 07822.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
lntcrnational
www.astm.org
SCHUTZ ON ORGANIC MATERIALS FOR COATING CONCRETE
853
Admbctures
Admixtures used to improve the adhesion of plastic mortar or concrete
to hardened concrete are classified by the American Concrete Institute
(ACI) as bonding admixtures. These admixtures are latices of either
acrylic, butadiene styrene, or polyvinyl acetate. These admixtures are generally used in one of three methods: (1) applied directly as an adhesive; (2)
mixed with portland cement in sand and applied as a bonding grout for
conventional mortar or concrete; or (3) used in a similar manner with concrete or mortar containing the latex as an admixture.
The inclusion of the latex as an admixture in the mortar or concrete
will improve both freeze/thaw and chemical resistance. However, the
working time of such mortars and concretes is extremely short--the latex
will tend to coalesce in 10 to 20 min depending on the type, temperature,
and other factors. Handling beyond that time generally results in profuse cracking on drying. In order to be effective as an admixture, these
latices must be used at a concentration of at least 12 percent polymer
solids based on the cement content.
These latices may be water-reemulsifiable after hardening of the concrete, and care should be taken in choosing the proper latex in areas where
the installation will be subject to excessive moisture or weathering. The
water resistance of the butadiene styrenes and acrylic latices is generally
good while the polyvinyl acetate latices are generally inferior.
Adhesives, Patching, and Overlaying Materials

Epoxy resins are the most widely used organic systems for bonding,
patching, and overlaying concrete. Suitable systems are available for use
as adhesives for bonding new plastic mortar or concrete to existing hardened
concrete and also as binders for the manufacture of resinous mortar and
concrete patches or overlays. For these purposes, liquid low-molecularweight epoxy resins of the bisphenol A type, reaction products of bisphenol
A epichlorohydrin, are used universally. These resins are mixed with a
curing agent immediately prior to use. These curing agents are generally
of the chemical groups such as amines, polyamines, or polyamides.
The epoxy resins cure by cross-linking; therefore, curing shrinkage is
much less here than in those resin systems such as the polyesters which
cure by polymerization. The cured-resin systems will have tensile strengths
and compressive strengths between 10 and 20 times that of good concrete.
They will be resistant to alkalies, mild acids, solvents, and oils.
A specification is available for epoxy resin systems for use with concrete; namely, ASTM Specification for Epoxy Resin Base Bonding Systems
for Concrete (C 881).
This specification classifies epoxy resin systems by type, class, grade,
and color based on the following uses: (1) bonding hardened concrete and
854
other materials to concrete, (2) bonding freshly placed concrete to hardened

concrete, and (3) bonding skid-resistant materials to hardened concrete,
and as a binder in epoxy mortar or epoxy concrete.
Each type is divided into three grades based upon its flow characteristics;
that is, Grade I has low viscosity; Grade II has moderate viscosity; Grade
III is for nonsagging applications. This specification defines an epoxy
resin as a resin that contains or did contain epoxy groups that are principally responsible for its reaction. It also requires that the epoxy resins shall
be furnished in two components for combining immediately prior to use.
Component A shall contain a bisphenol A, epichlorohydrin epoxy resin
with or without a reactive diluent. Component B shall contain one or more
curing agents which, on mixing with Component A, will cause the system
to harden. A suitable inert filler may be incorporated uniformly in one
or both components. The specification stipulates limits for viscosity, consistency, gel time, creep, compressive strength, bond strength, and percent of volatile matter.
ASTM specifications are available for testing the suitability of epoxy
resin systems for use with concrete. ASTM Bond Strength of Epoxy Resin
Systems Used with Concrete (C 882) covers the test method developed by
Kriegh and Nordby for determining the bond strength of epoxy resin systems to concrete. In this test method, a standard concrete cylinder is cast
or cut at a 60-deg angle, and plastic or hardened concrete is bonded to the
original hardened concrete using an epoxy-resin adhesive. After 14 days
cure, the composite cylinder is tested in compression. A strength of 10
MPa (1500 psi) based on the elliptical area (glue line) is considered satisfactory.
ASTM Effective Shrinkage of Epoxy Resin Systems Used with Concrete
(C 883) determines whether or not an epoxy resin system will have excess
volume changes on curing. In this test method, a J/8 in. (3 mm) thick
film of the epoxy resin under test is cast on a glass plate and cycled 10 times
between 124.9 + 3.6~ (51.6 2~
and 73.4 _ 1.8~ (23 I~
If the shrinkage is such that the glass plate breaks, the resin system is
considered unsuitable for use with concrete.
ASTM Thermal Compatibility Between Concrete and Epoxy Resin Overlay (C 884) is a test method in which a 1/2 in. thick layer of epoxy resin
sand mortar is cast on a standard concrete block. After 14 days' curing,
the overlay and block are cycled 5 times between 77 5 ~ (25 + 2.8~
and --6 + 3 ~ (--21.1 ___ 1.7~
Any crack or delamination of the system is considered a failure.
Epoxy-resin systems when used as adhesives have a relatively high elastic modulus, since a structural bond is required in most adhesive applications.
In the thin glue lines usually used, the differences in linear thermal coefficient of expansion between the epoxy resin and the concrete is inconsequential because the total tensile strength of the thin glue line (1 mm or
855
less) is less than the tensile strength of the concrete. In thin glue lines,
creep of the epoxy resin is not a problem because the effective modulus
of elasticity of a material in a thin glue line will be extremely high [6].
For applications as an adhesive, the epoxy resin should contain no solvent since solvent may be entrapped during cure and cause both a rubbery
cure and later shrinkage of the adhesive. The epoxy-resin systems may be
applied by brush, roller, or spray to properly prepared concrete surfaces.
Where reinforcing steel is exposed, the steel should be cleaned of rust,
oil, and foreign material and coated with the epoxy-resin adhesive to prevent further corrosion.
Epoxy-resin systems are used quite commonly as the sole binder for
mortar and concrete patches and overlays. Since such patches and overlays are generally thick, a resin with a low modulus of elasticity must be
used. If high-modulus binders are used, the differences in thermal coefficient of expansion between the thick resin mortar and the base concrete will result in failure of the overlay or patch by shearing of the concrete adjacent to the resinous mortar or concrete. When applied in layers
40 mm (1189 in.) or greater in thickness, it is desirable to include coarse
aggregate in the system, both for reasons of economy and thermal compatibility. Figure 1 illustrates the effect of aggregate loading on linear coefficient of expansion of several epoxy-resin systems [7].
Some epoxy-resin systems are sensitive to water before cure. Therefore,
they can only be used under dry conditions; and these dry conditions must
be maintained at least until initial cure of the system. Epoxy-resin systems
are also available which will not only cure under wet conditions, but will
u._
-80
42
-70~
36
30
E
-50 E
24
'~
~
18
~[_
-40 x
Sand-filledepoxy
E
Lq>
-30 ~:=
._o
-20 o~
12
I.-
-10 "~
Concrete
!o
1
e--
10
Aggregate-binderratio
FIG. 1--The effect of changes in the sand aggregate-binder ratio on the thermal coefficient of an epoxy system.
856
displace water from the surface. Patching and other repairs using these
systems can be carried out even under water.
Where large cavities are to be repaired by use of an epoxy-resin concrete, the most economical method is to prepack aggregates into the void
and then inject the epoxy resin as a grout from the bottom of the void.
This approach displaces air in the case of water-sensitive systems or it
displaces water in the event the repair is made under water with a waterinsensitive system.
Proportioning epoxy-resin mortar or concrete follows the same guidelines as proportioning portland cement concrete mixtures. However, since
the resin systems are sticky and viscous, the sand/aggregate ratio is usually
the reverse of that which would be used for portland cement concrete.
The total voids in the system should be less than 12 percent by volume.
Voids in excess of 12 percent will result in a permeable mass with resultant
poor chemical and freeze-thaw resistance. When using forms for epoxy
resin mortar or concrete, the forms should be coated with polyethylene
rather than oil. Oil may be absorbed into the resin impairing the cure
of the system.
ACI Committee 403 has issued a guide covering the use of epoxy resin
in concrete construction [7]. This report should be studied by those planning to use these materials. It stresses the necessity for preparing the concrete surface to be treated before application of the resin system. Two
surface conditions must be met if an application is to be successful: (1) the
surface must be strong and sound, and (2) the surface must be clean; that
is, free of oil, grease, or other contaminants, such as residues of curing
compounds, waxes, or polishes that may have been applied to the surface
of the concrete.
Coatings and Impregnations

Bituminous materials include both asphaltic- and coal-tar-based materials.
There is considerable difference between asphalts and coal tars. Asphalts
are resistant to many mildly corrosive materials, but are attacked readily
by solvents, oils, and gasoline. Further, unless they are specially formulated, asphaltic coatings may be permeable to water. Asphaltic coatings
generally are used to protect foundations, basements, and similar structures from water.
Coal-tar coatings possess a high degree of water resistance and should
be considered wherever continuous immersion is encountered. Coal-tar
coatings have a tendency to crack and craze when exposed to ultraviolet
light or high temperature. The most common use for coal-tar coatings
is for protection of underground pipelines. When formulated with epoxyresin systems, excellent long-term performance has been recorded.
SCHUTZ ON ORGANIC MATERIALS FOR COATING CONCRETE 857
Bituminous coatings may be used with reinforcing fabric and tape to

increase their resistance to impact on backfilling and handling. Such reinforcing also increases their thickness and boosts both impermeability
and service life. Total coating thickness for either system will vary between
0.1 and 3.2 mm (5 and 125 mils).
Oleo-Resinous Materials
Penetrating Sealers
Numerous tests carried out by state highway departments and private
laboratories indicate that one of the most efficient penetrating sealers is
linseed oil [8]. This may be either the solution or emulsion type.
The solution type is a mixture of raw or polymerized (boiled) linseed
oil and solvent usually in a 50-percent concentration. The solution types
are more economical and more widely used.
The emulsion type contains polymerized linseed oil, kerosene, detergent,
and trisodium phosphate in water.
The protective effect is obtained from penetration of the linseed oil into
the concrete to a depth of approximately 3 mm (l/a in.). On evaporation
of the solvent or water, the linseed oil oxidizes and an effective seal is developed. The use of linseed oil impregnations is especially desirable where
the concrete is suspected of having a low air content and, therefore, poor
durability.
Oleo-Resinous Paints
Oleo-resinous paints, while attractive in appearance, are only capable
of satisfactory performance when applied to dry, well-aged concrete (3 to
12 months). They have limited resistance to moisture. Moisture behind
an oleo-resinous paint will cause premature failure due to blistering and
flaking [9].
Latex Coatings
Latex coatings are relatively insensitive to dampness and alkalinity and
have become the most widely used coatings for interior and exterior concrete surfaces. Two types of latices in common use are acrylic and polyvinylacetate latices. While both are suitable for interior concrete coatings, the
acrylic gives the best results on exterior exposure due to their superior
ultraviolet resistance. Since these latices are water-based, surface dampness
is not a problem. When heavily filled or pigmented, they form permeable
coatings, and presence of vapor pressure does not result in failure.
858
Synthetic-Resin Coatings
Synthetic-resin coatings other than latex-based coatings which are successfully used for concrete are generally applied as solvent solutions and
include: coumarone-indene, styrene-butadiene, chlorinated rubber, chlorosulfonated polyethylene, polyurethane, neoprene, and epoxy resins.
The most commonly used materials in this group are the so-called chlorinated-rubber-based coatings, the polyurethane, and the epoxy-resin-based
coatings. For interior use for protection where aggressive solutions are
encountered, the epoxy-resin-based coatings have proven very successful.
These may be either polyamine, polyamide, amine adduct, or acrylic amine
systems [9]. They may be applied as solvent solutions, water-based, or
100-percent solids systems.
Because these epoxy-resin systems will tend to chalk, they are not suitable as exterior architectural coatings. The new acrylic-amine epoxy-resin
systems exhibit good resistance to ultraviolet radiation and may be used
in both interior and exterior applications. However, they are not quite as
chemically resistant as the conventional epoxy-resin systems.
The polyurethane-based coatings exhibit extremely good scuff and abrasion resistance and form the basis for many coatings for concrete floors.
They often are used in conjunction with an epoxy-resin based prime coat.
Chlorinated-rubber-based clear coatings are used as both curing compounds and so-called floor hardeners. Applied after the concrete has
hardened, usually 24 h after placing the concrete, these chlorinated rubber coatings retain moisture efficiently and aid in curing the concrete.
Although these systems do not actually harden the concrete surface, they
do protect the concrete from abrasion and subsequent dusting.
Silicones
Silicone resins in solvent solution or in aqueous solution as alkali metal
methyl siliconates impart water repellency to concrete and masonry surfaces.
Both types are flooded onto the surface in dilute solutions containing
3 to 6-percent silicone solids. The alkali metal methyl siliconates are more
effective on calcareous surfaces than the solvent-borne silicone resins.
Due to the water repellency imparted by the silicones, dirt, soot, and
similar contaminants tend to wash off architectural concrete. These materials have not proven effective in protecting nondurable concrete from
freezing and thawing damage. The service life of these impregnations is
rather short, 2 to 5 years being average.
859
References
[1] Woods, Hubert, "Durability of Concrete Construction," ACI Monograph 4, American
Concrete Institute, 1968.
[2] Kleinlogel, A., Influences on Concrete, Frederick Ungar Publishing Company, 1950.
[3] "Concrete-Polymer Materials," Third Topical Report, Bureau of Reclamation, Brookhaven National Laboratory, Jan. 1971.
[4] Nawy, E. G., Sauer, J. A., and Sun, P. F., "Polymer Cement Concrete," American
Concrete Institute, Nov. 1975.
[5] Kriegh, J. D. and Nordby, G. M., "Methods of Evaluation of Epoxy Compounds Used
for Bonding Concrete," Epoxies with Concrete, SP-21, American Concrete Institute.
[6] Schultz, R. J., "Epoxy Adhesives in Prestressed and Precast Concrete Bridge Construction," Journal. American Concrete Institute, March 1976, Title No. 73-14.
[7] "Use of Epoxies with Concrete," Manual of Concrete Practice. American Concrete Institute, No. 503-1, Part 3.
[8] "Linseed Oil in Road Up," Chemical Week. 21 Nov. 1964, p. 47.
[9] Roberts, A. D., "Organic Coatings: Properties, Selection, and Use," Building Sc&nce
Series 7, U.S. Department of Commerce, Feb. 1968.
R. C. Valore, Jr. ~
Chapter 50
Concrete
Pumpability Aids for
Introduction
An essential requirement of concrete for any use is that it be transported
from mixer to forms without separation of water, cementing materials, or
aggregates, with sufficient retention of consistency to permit economical
consolidation. To achieve that end, it is an objective in proportioning the
ingredients of concrete to obtain a blend of solid materials, from the finest
cement grain to the largest particle of coarse aggregate, which, when consolidated, will have an interstitial void structure of minimum volume, and
relatively low permeability and surface area. The latter requirements would
appear to be contradictory since permeability of a bed of granular solids is
inversely related to its surface area. Permeability of the aggregate void
system should be low enough to prevent permeation by water or cement
paste under pressure. Surface area should be low enough to permit the use
of economical amounts of cement and suitable water/cement (w/c) ratios
to provide a volume of paste sufficient to saturate the aggregate voids and
form workable cohesive concrete.
In pumped concrete these general requirements become critically important; the price for failure to achieve them is failure to achieve pumpability. Discussion of the details of the influence of specific size gradings
and particle shape of aggregates, upon the interstitial void structure and
the ways in which they affect pumpability of concrete are beyond the scope
of this paper. For more specific information, reference is made to a study by
Kempster and its elaboration in Guide to Concrete P u m p i n g of the British
Building Research Establishment (BRE) [1,2]. 2 These references describe
direct measurements in the laboratory of aggregate void volume and water
permeability.
Freshly mixed concrete is a suspension of solids in water of which only
l Principal, Valore Research Associates, Ridgewood,N.J. 07450.
Universidade
do Gois pursuant
Agreement.
authorized.
Copyright9
1978tobyLicense
ASTM
International
www.astm.org
VALORE ON PUMPABILITY AIDS FOR CONCRETE
861
the latter is inherently pumpable. Pumpable concrete must be mobile or

workable enough to be moved through changes in shape and direction of
lines without a buildup of excessive pressure. Also it must be sufficiently
stable or cohesive to resist separation or segregation of its constituents
under the pressures of pumping.3
Forced bleeding of water or cement paste from concrete under pressure
is an example of segregation which, in excess, can make pumping impossible. This is illustrated in Fig. 1. The workable and cohesive pumpable
concrete possesses a sufficient volume of cement paste to act as a lubricant
between concrete and walls of the line to reduce frictional resistance.
Pumping can be impaired, however, when contents of cement and other
subsieve fines are excessive.
Kempster and the British Guide defined optimum relationships between
total aggregate interstitial void volume and cement contents for pumpable
concrete [1,2]. Using the expedient of considering the density of dry portland cement to be 1440 k g / m 3 (90 lb/ft3), it is shown that, in laboratory
trials, when the volume of cement was equal to or greater than the aggregate interstitial void volume, the concrete mixtures generally were pumpable.
Exceptions were mixtures in which the volumes of total fines, consisting
of cement plus minus 150-#m (No. 100 sieve) aggregate, were excessively
high.
The British Guide illustrates effects on pumpability, not only of aggregate
void volumes and cement contents, but also of size of voids as influenced
by particle gradation and shape.
Admixtures
Pumpability aids for concrete are admixtures whose sole or primary
function is the improvement of pumpability; they will not normally be
Particle
Uigration of
High Fr~cfional Aesistance

Courtesy American Concrete Institute
FIG. 1--Dewatering of concrete under pressure in pipeline. From Browne and B a m f orth [3].
3"Mobile" means "movable" while one definition for the word "stable" is "difficult to
move." These terms are not opposites in the present context; pumpable,concrete is an internally stable mixture that is pumpable, that is, "movable."
862
used in concrete that is not pumped or in concrete that can be readily

pumped.
The primary purpose of using admixtures to enhance pumpability of
concrete is to overcome difficulties, unavoidable at reasonable cost, of
obtaining satisfactory aggregate interstitial void volume and individual
void size when using available materials. As in the use of any ingredient
in concrete, the objective is economic.
Pumpability admixtures can make some, but not all, unpumpable concretes pumpable, or can improve pumpability of marginally pumpable
concretes, as shown by the BRE experiments.
The principle involved in virtually all of the admixtures marketed for
improving the pumpability of concrete is that of thickening or increasing
the viscosity of water. Since the objective is to inhibit forced bleeding of
water from cement paste or permeation of paste through interstitial
aggregate voids, permeability of the void structure is effectively reduced
by increasing the viscosity of the permeating fluid. Therefore the objective
is to prevent dewatering of the paste under pressure of pumping. Dewatering has two cumulatively adverse effects: it decreases mobility of the concrete and depletes the lubricating fluid.
Therefore the function of the pumpability admixture is that of imparting
water retentivity to the cement paste under forces tending to separate
water. In masonry mortars, thickening agents provide water retentivity and
retention of plasticity against the forces of suction of porous brick or
concrete block. In polymer-latex and other paints, they serve as suspending
aids for inorganic pigments and fillers, prevent running on vertical surfaces,
and minimize rapid absorption of vehicle into substrate, while providing
undiminished brushability and integrity of the applied coating. The
requirements in pumped concrete are analogous; the desired inhibition of
segregation must not be obtained at a cost of decreased mobility. Under
optimum conditions, these agents impart a degree of thixotropy to suspensions of mineral solids in water: stability of the suspension at rest and
compliability when worked upon.
The Standards Association of Australia has classified thickening admixtures for concrete and mortar [4]. The following list is based in part on
that draft.
A. Water-soluble synthetic and natural organic polymers which increase
the viscosity of water:
1. Cellulose derivatives (methyl, ethyl, hydroxyethyl, and other
cellulose gums
2. Polyethylene oxides
3. Acrylic polymers
4. Polyacrylamides
5.
6.
7.
8.
863
Carboxyvinyl polymers
Natural water-soluble gums
Starches
Polyvinyl alcohol
B. Organic flocculents:
1. Carboxyl-containing styrene copolymers
2. Other synthetic polyelectrolytes
3. Natural water-soluble gums
C. Emulsions of various organic materials: paraffin, coal tar, asphalt.
and acrylic and other polymers.
D. High-surface area inorganic materials:
1.
2.
3.
4.
Bentonites and organic-modified bentonites

Pyrogenic silicas
Milled asbestos and silica-coated milled asbestos
Asbestos short fibers and other fibrous materials
E. Finely divided inorganic materials which supplement cement in

cement paste.
1. Fly ash; various raw or calcined pozzolanic materials
2. Hydrated lime
3. Natural or precipitated calcium carbonates; various rock dusts
This list does not include all of the possibilities, some of which are listed
in McCutcheon's Functional Materials [5]. Classification may be misleading since the performance of a given admixture can change drastically
when dosage rates, cement composition, mixing temperature and time, and
other factors change. An example is provided by the polyethylene oxides.
When used in small amounts of 0.01 to 0.05 percent of cement weight,
they improve pumpability. Smaller amounts make water slipperier; larger
amounts produce thickening which may or may not disappear upon
prolonged mixing. Other examples are provided by synthetic polyelectrolytes
which act as flocculents or thickeners, depending upon dosage levels. It
would appear to be highly undesirable to induce flocculation and increase
bleeding in pumped concrete. Nevertheless, it is held that these admixtures
are effective in pumped concrete because they lower bleeding capacity or
total bleeding, despite causing increases in initial rates of bleeding [4].
Other problems with the listing given above occur with the natural gums
(algins, tragacanth, arabic). They can function as thickeners or flocculents,
depending upon dosage levels and other factors. These agents and some of
the synthetic materials also can have dispersing or water-reducing effects.
Gum arabic is a powerful water reducer for gypsum hemihydrate plasters
but in portland cement pastes it can produce a virtual glue-like stickiness.
864
Factors to consider in the use of emulsions (paraffins, polymers) is

whether they function in the desired way in cement paste by remaining
stable or by breaking of the emulsion. Both types of paraffin emulsion are
considered to be useful in Australian concrete technology.
Despite their inclusion in the BRE Guide, the listing given above does
not include air-entraining agents or surface-active agents widely used in
concrete such as hydroxylated carboxylic acid derivatives, lignosulfonates
and their derivatives, formaldehyde-condensed naphthalene sulfonates and
melamine polymers, and other set-retarding or water-reducing admixtures.
The omission here is deliberate because a substantial proportion of concrete
that is to be pumped in North America will be specified as air-entrained
concrete and will probably also contain a water-reducing or set-retarding
admixture. Therefore, such admixtures may be considered to be normal
constituents of concrete. Evaluation of and experience with these admixtures
are well established and specifications and testing methods for them are
well known. These types of agents are not considered specifically, in this
paper, as pumping aids; they will be present in many cases, in combination
with other agents introduced for the specific purpose of improving pumpability. In such cases, evaluation of effects of such combinations on pumpability and other properties of concrete will be required in order to determine
whether or not adverse interactions occur between admixtures.
Side Effects
A side effect of a concrete pumpability admixture is any effect that it
has on fresh concrete other than that on pumpability, and any effect that
changes the characteristics of the hardened concrete. Since the main effect
of a water thickener is to increase viscosity, substantial thickening can
increase water requirements with the usual consequence of reduced strength.
By using a suitable dispersant in combination with a thickening agent, no
increase in water may be required. At certain dosage levels, some thickeners
act as dispersants of solid particles.
Many of the thickening agents cause entrainment of air; to control air
content, a defoamer (for example, tributyl phosphate) may need to be
used, especially for higher concentrations of pumpability admixture in
mortars and concretes.
Many of the synthetic and natural organic thickening agents retard the
setting of portland cement pastes. For dosage of methyl or hydroxyethyl
cellulose of 0.1 percent or more, by weight of portland cement, retardation
may be substantial. In any case, the particular concrete system in which a
pumpability admixture is incorporated must be evaluated in terms of side
effects upon the fresh and hardened concrete in addition to assessing the
effectiveness of the admixture in performing its intended function.
VALORE ON PUMPABILITY AIDS FOR CONCRETE 865
Test Methods
The BRE Guide contains the statement: "No methods o f . . . reliably
predicting pumpability have yet been devised . . . . " Inquiries by the present
author of specialists in pumping concrete have elicited general agreement
with that statement. This is to say that pumpability of a given concrete
cannot be evaluated without pumping trials.
The situation is not much different with evaluation of some of the other
properties of concrete and mortar. The water retention test method for
masonry mortars in ASTM Specification for Masonry Cement (C 91)
provides information that will not necessarily predict working properties of
a mortar with a highly absorptive concrete block in dry, windy, and hot
weather. Similarly, ASTM Test for Resistance of Concrete to Rapid Freezing and Thawing (C 666) provides results that show the relative ability of
specific concrete specimens to survive a specified regimen in a particular
type of machine. Correlation of such results with durability in service is
not perfect. In general, the test methods are satisfactory if results provide
a reasonably accurate indication of performance.
In pumping concrete, the pumping action is itself the crucial test, and it
is not necessary to wait for years for confirmation of laboratory test results
as may be required to learn the actual durability of concrete in service.
In order to evaluate admixtures for pumping concrete in the laboratory,
small-scale test methods are needed. If pumpability of concrete containing
such admixtures cannot be tested directly in the laboratory, a number of
other properties of mortars or concretes which correlate in some degree
with pumpability can be determined to provide indirect indications of
effectiveness of the admixtures.
Bleeding Tests
Several test methods provide data on bleeding of cement pastes, mortars,
and concretes.
ASTM Test for Bleeding of Cement Pastes and Mortars (C 243) based
on a carbon tetrachloride displacement method developed by Valore et al
[6] provides data on initial bleeding rate and bleeding capacity and was
used successfully, when modified, in determining the bleeding characteristics of concrete. The apparatus is shown in Fig. 2.
Ritchie [7] adapted the carbon tetrachloride displacement method to
testing of concrete mixtures with apparatus depicted in Fig. 3. In this and
the Valore et al studies, it was concluded that only by obtaining bleeding
data for concrete and for the corresponding cement paste can a complete
picture of the potential dewatering of concrete be seen, including effects of
aggregate interference in causing internal bleeding.
866
1,1,1
1 0
1 2
4 5
INCHES
FIG. 2--Carbon tetrachloride displacement bleeding test apparatus. From Valore et al [6].
Water seal --~

Burette tube-'P
Collected bleeding water
Carbon letrachloride
Pycnometer jar
$ompll of r
paste or concrete
I~'ii:1
Courtesy American Society of Civil Engineers

FIG. 3--Carbon tetrachloride displacement bleeding test apparatus modified for tests of
concrete. From Ritchie [7].
867
Test methods which cover determination of water loss under pressure

have been applied by Schupack [8] to portland cement grouts for posttensioned tendons in concrete structures, and by the American Petroleum
Institute in a fluid-loss test of oil-well cements and cement additives (API
Recommended Practice for Testing Oil-Well Cements and Cement Additives; Fluid Loss Test (API RP 10B)). These methods are based on the use
of a filter press; filter presses of 0.12, 1, and 3 litre capacity are shown in
Fig. 4. A vertical cylindrical container with a filter, support, and fluid
outlet at the bottom is filled with grout. An air-tight fitting containing an
inlet is attached to the top of the cylinder and a gas or liquid is introduced
through the inlet. The pressure then can be increased gradually or in
increments and the volume of water forced from the grout can be collected
and measured.
Browne and Bamforth [3] described a new pressure bleeding test and
apparatus for the purpose of establishing pumpability of concrete. A
drawing reproduced from the paper is shown in Fig. 5. The concrete sample
is subjected to a pressure of 3.5 MPa (500 psi) by the piston, actuated by a
hydraulic jack. The volume of bleeding water collected at 140 s (V140),
minus that collected at 10 s (V~0), is said to correlate directly with pumpability for concrete of a given slump. This was interpreted as showing that
concretes that dewater quickly (in the first 10 s) under pressure tend to be
unpumpable. The authors developed a calibration curve of (V~40 -- Vl0)
versus slump which forms a line of demarcation between zones of "pumpability" and "unpumpability." Results of this work tend to negate the
principles involved in using flocculating admixtures to improve pumpability
by increasing initial bleeding rates and decreasing bleeding capacities.
hi'
,.
III
.11
lit
,li
Ill
tll
"-"iF-o"
Courtesy Fisher Scientific C o m p a n y

FIG. 4--Filter-presses of O. 12, 1, and 3-litre capacities9
868
TESTS
AND
PROPERTIES
OF OTHER
Iflllf
~Ukt
(~(
MATERIALS
Double Acting
IJHJIH ydroulic
....
Colibrofed
Cyl
I',lf Ij
......
Rubber "0" Ri no s
125 ram.Diameter
Cyhader
'
~--- Top Cop
~,iJ~a
,,ce,L J,
i', '1 r
Bose
inder
i i ,,,,J
I ..~--- Piston
l~r"2")~:~.,,,d
,"-'"~'.~':b-"'r '~
Concrete
Sample
-~"--"----7"--~ -"
/
'\
,;, Measuring
~-'~ Cylinder
5OMes,Gouz.
II
Gauze Rei'oinin~ --~ ' ~

Plate
FIG. 5--Pressure bleeding test of concrete devised by Browne and Bamforth [3].
Other Tests
A water-retentivity test for masonry cement mortars (ASTM Specification C 91) has been mentioned. It combines a type of forced bleeding
under a small partial vacuum with a measure of workability retention,
using a flow table. The principles involved might well be applied to development of tests of pumpability admixtures, combining measures of bleeding
rates and capacities under pressure with some measure of the decrease in
workability accompanying dewatering of test samples. Tests could be
conducted on mortars or concretes especially proportioned to have lower
than optimum pumpability, in order to show effects of admixtures.
Gray [9] developed a laboratory procedure for comparing pumpability
of concrete mixtures which involved pumping concrete samples upward
through a 1.2 m (4 ft) vertical section of pipe with a curved offset of 305
mm (12 in.) and measuring the pressure required to move the concrete.
The testing arrangement is shown in Fig. 6.
Tanaka [10] and Ozaki and Emura [11] presented results of pumping
tests on lightweight aggregate mortars and concretes, respectively. Test
mortars were made with natural sands and fine lightweight aggregates, and
included mixtures containing methyl cellulose, paraffin emulsion, and fly
869
I11
To Voct
Purn
FIG. 6--Schematic drawing of pumping apparatus for concrete, due to Gray [9].
ash. Evaluation of pumpability was based on the rate of pumping in litres

per minute with equipment not otherwise described. The admixtures used
were found to reduce segregation; rates of pumping were reduced when
methyl cellulose or paraffin emulsion were used and increased when fly
ash was used in mortar mixtures at dosage levels required to prevent
segregation. Lightweight mortars were more difficult to pump than sand
mortars. In pumping tests of lightweight concretes pressures were measured
at several stations in a U-shaped pumping circuit about 45 m (150 ft) in
length. Slump and flow were measured before and after pumping. The
testing set-up provided a good indication of pumpability in the field but,
because of the space required, it is not well-suited for use as a laboratory
method for evaluating admixtures.
Kempster [12] described a similar but smaller concrete pumping circuit,
about 23 m (75 ft) in length, with three pressure measuring stations to
870
determine the effects of aggregate interstitial void size and total volume
and to evaluate the effectiveness of admixtures on unpumpable or marginally
pumpable concretes. Results of Kempster's tests are summarized in Refs 1
and 2. The pumping circuit is shown in Fig. 7. The voidmeter developed
in Kempster's studies is shown in Fig. 8.
Kempster [13] also studied effects of sand gradation and admixtures
on the pumpability of mortars.
A study in progress by Lane [14] involves a small pumping circuit suitable
for laboratory installation. Grouts, mortars, and concretes are being tested
for pumping characteristics with this equipment.
Many papers have been published on various aspects of workability of
concrete. Slump and remolding (Vebe) tests have survived and the flow
test, widely used in testing of mortars, has reappeared in recent years in
testing concretes containing "superplasticizers" or high-dosage water
reducers [15]. Herschel and Pisapia [16] developed harshness, segregation,
shear resistance, and stickiness test methods that showed recognition of
important aspects of workability, but the methods were not very discriminatory.
9auge
los ,t
Strain gouge "C'
~{ ~.Flexible
rubber pipe
Im
Reduction
pipes
$fraln
g,,~u,ge
Distance A- B:1,~m
Distance I1 _ C : 10 m
pump
CourtesyBritish BuildingResearchEstablishment
FIG. 7mPurnpingcircuit used by Kempster [12].
871
Reservoir
Position| ~
Inilio! wole~_Level
Meosuring
Tube~
Col ibroted
Scote~
~Top
ir-tight
. ,qDJZ Lid
~[~
Aggregate
[I 9
.. ,.:~.~.~ '-.-.--.s o . , , .
Reservoir
Position
2~,
II
Courtesy American Concrete Institute

FIG. 8--Voidmeter developed by Kempster [12]. From paper by Browne and Bamforth [3].
Conclusion
Admixtures for pumping concrete and potential methods for evaluating
them have been described.
It has been recognized that precise prediction of pumpability characteristics of concrete in the field may not be attainable with presently known
bench-scale laboratory methods.
Admixtures for pumping of concrete might be evaluated indirectly by
means of determination of their effects on initial bleeding rates and bleeding capacities of cement pastes, mortars, or concretes under pressure,
using methods now available.
Measurement of workability of concrete appears to be a necessary
adjunct to bleeding or dewatering tests.
Since admixtures for improving pumpability need not be used in concrete
that can be pumped readily, their main area of use in marginally pumpable
or "somewhat" nonpumpable mixtures requires that standard test mortars
or concretes be formulated with inherently mediocre or poorer levels of
pumpability. The principles affecting pumpability, as elucidated by
Browne and Bamforth, Kempster, and others, seem to make it a simple
matter to proportion pumpable concretes, or for that matter, concretes
or mortars lacking in pumpability.
In this discussion it has been assumed that proportioning of concrete
mixtures for pumping has been done properly with the materials at hand.
The question then becomes one of asking, when everything else possible
has been done and the concrete still cannot be pumped, can an admixture
872
help? To answer that question, a test method or combination of test

methods is required that will provide some indication, even if indirect, of
pumpability in the field.
Potential side effects of concrete pumping admixtures of air entrainment
and retardation of set have been mentioned. Tests for these characteristics
are now incorporated in specifications for admixtures for concrete and
need not be repeated in a specification for pumping admixtures.
References
[1] Kempster, E., "Measuring Void Content: New Apparatus for Aggregates, Sands, and
Fillers," Contract Journal, Vol. 228, 27 March 1969, p. 409.
[2] Crowe, F. W., Guide to Concrete Pumping, Building Research Establishment, Department of the Environment, London, Her Majesty's Stationery Office 1972.
[3] Browne, R. D. and Bamforth, P. B., "Tests to Establish Concrete Pumpability," Journal,
American Concrete Institute, Proceedings, Vol. 74, No. 5, May 1977, pp. 193-203.
[4] Committee BD/33-Concrete Admixtures, "Information on Thickening Admixtures
for Use in Concrete and Mortar," Standards Association of Australia, Draft for Comment, Miscellaneous Publication, DR 73146, October 1973, unpublished.
[5] "Functional Materials," McCutcheon Division, MC Publishing Co., Ridgewood, New
Jersey, 1975.
[6] Valore, R. C., Jr., Bowling, J. E., and Blaine, R. L., "The Direct and Continuous
Measurement of Bleeding in Portland Cement-Water Mixtures," Proceedings, American
Society for Testing and Materials, Vol. 49, 1949, pp. 891-908.
[7] Ritchie, A. G. B., "Stability of Fresh Concrete Mixes," Proceedings, American
Society of Mechanical Engineers, Journal of Construction Division, Vol. 92, No. CO 1,
Jan. 1966, pp. 17-35.
[8] Schupack, M., "Development of a Water-Retentive Grouting Aid to Control the Bleed
in Cement Grout Used for Post-Tensioning," FIP Seventh Congress, New York, Feb. 1974.
[9] Gray, J. E., "Laboratory Procedure for Comparing Pumpability of Concrete Mixtures,"
[10] Tanaka, I., "On the Artificial Lightweight Aggregate Mortar-Mesalite Mortar," Journal
of Research, Onoda Cement Company, Vol. 21, Dec. 1969, p. 93 (in Japanese; English
translation arranged by subcommittee).
[11] Ozaki, Y. and Emura, K., "Lightweight Aggregate Concrete Placing with Pump
Method," Journal of Research, Onoda Cement Company, Vol. 20, No. 73, May 1968,
p. 62 (in Japanese; English translation arranged by subcommittee).
[12] Kempster, E., "Pumpable Concrete," Building Research Station, Current Papers, CP
29/69, Aug. 1969, pp. 1-14.
[13] Kempster, E.,"PumpabilityofMortars,"ContractJournal, Vol. 217, May4,1967, p. 28.
[14] Lane, R. O., Unpublished Report on Study of Pumpability of Grouts, Mortars, and
Concretes, Tennessee Valley Authority, Knoxville, Tenn., 1974, 1977.
[15] "Superplasticising Admixtures in Concrete," Cement and Concrete Association,
Wexham Springs, Slough, England, 1976.
[16] Herschel, W. H. and Pisapia, E. A., "Factors of Workability of Portland Cement
Concrete," Journal, American Concrete Institute, Proceedings Vol. 32, May-June 1936,
p. 641.

ASTM (1978) - Significance of Tests and Properties of Concrete and Concrete-Making Materials

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ASTM (1978) - Significance of Tests and Properties of Concrete and Concrete-Making Materials

Uploaded by

Copyright:

Available Formats

Copyright by ASTM Int'l (all rights reserved); Sun Apr 6 21:48:22 EDT 2014

AMERICAN SOCIETY FOR TESTING AND MATERIALS

Copyright 9 by American Society for Testing and Materials 1978

Printed in Baltimore, Md.

Chapter 2--Techniques, Procedures, and Practiees of Sampling of

Chapter 4--Quality in Concrete Testing--J. R. DISE

TESTS AND PROPERTIES OF CONCRETE

Chapter 24--Resistance to Chemical Attack--L. H. TUTHILL

Chapter 33--Petrographic Examination--R. c. MIELENZ

Chapter 37--Porosity--w. L. DOLCH

Chapter 41--Chemical Reactions of Carbonate Aggregates in Cement Paste--H. N. WALKER

43--Mixing and Curing Water for Concrete--w. J. MCCOY

Chapter 48--Cellular Concrete--L M. LEGATSKI

When an ASTM Committee undertakes the preparation of a statement on

SIGNIFICANCE OF THE PROPERTIES OF CONCRETE

Chapter 2nTechniques, Procedures,

Millions upon millions of dollars change hands daily in various segments

ABDUN-NUR ON SAMPLING OF CONCRETE

actually taking the sample. A major reorientation of ideas on the subject

The relationship of a sampling plan to the actual taking of a sample

Significance of Various Facets of Sampling

ABDUN-NUR ON SAMPLING OF CONCRETE

Adams (Chapter 8) shows the importance of sampling and of the making

Sampling and Inspection

ABDUN-NUR ON SAMPLING OF CONCRETE

on concrete properties due to the use of bags contaminated with sugar or

Sampling and Quality Control

cesses or operations. It cannot be imposed by the purchaser or owner from

Sampfing Concrete and Its Constituents

ABDUN-NUR ON SAMPLING OF CONCRETE

locations such as quarries, pits, railroad cars, bins, and stockpiles. It is

Admixtures, Curing, Patching, Bonding, and Sealing Materials

gineer to determine in each case. But if proper original or continuous

ABDUN-NUR ON SAMPLING OF CONCRETE

Needed Improvements and Their Benefits

also for the sampling of hardened concrete. In most cases, modifications or

ABDUN-NUR ON SAMPLING OF CONCRETE

ABDUN-NUR ON SAMPLING OF CONCRETE

SAMPLING AND TEST METHODS

Unit Weight of Aggregate

Compressive Strength of Concrete Using Portions

ABDUN-NUR ON SAMPLING OF CONCRETE

Effectiveness of Mineral Admixtures in Preventing

Concrete Joint Sealers, Cold-Application Type

ABDUN-NUR ON SAMPLING OF CONCRETE

Jet-Fuel-Resistant Concrete Joint Sealer, Hot-Poured

Petrographic Examinations of Aggregates for Concrete

It is not the purpose of this chapter to provide a short text on general

The first consideration in any process of evaluation of characteristics

ARNI ON STATISTICAL CONSIDERATIONS

tively small n u m b e r of tests. After preliminary analysis, the properties of

Conversionfactors-1 in. = 25.4 mm, and

ARNI ON STATISTICAL CONSIDERATIONS

Mineral Admixtures--Pozzolans and other mineral admixtures may be

ARNI ON STATISTICAL CONSIDERATIONS

of Strength Test Results of Concrete (ACI 214-77). First published as a

Evaluating Standard Cylinders

Coefficient of Variation versus Standard Deviation

of measurements, and that which one should be used in a given situation

Accelerated and In-Place Tests for Evaluating Concrete

ARNI ON STATISTICAL CONSIDERATIONS

thing about remedying the situation if serious difficulties are found,

Testing of concrete specimens for compressive strength occupies a