INFLUENCE OF AMORPHOUS COLLOIDAL SILICA ON THE

PROPERTIES OF SELF-COMPACTING CONCRETES

Prof. Mario Collepardi Engineering Faculty, Politecnico of Milan, Milan, Italy

Ing. J.Jacob Ogoumah Olagot Enco, Engineering Concrete, Spresiano (TV), Italy
Ing. Ulf Skarp Eka Chemicals, Bohus, Sweeden
Ing. Roberto Troli Enco, Engineering Concrete, Spresiano (TV), Italy
ABSTRACT. Self-Compacting Concretes (SCCs) were manufactured by using the
following ingredients particularly suitable for structures which need low heat of hydration
such as mass concrete structures: CEM III/A 32.5R (with 60% of GGBS) at a dosage of 300
kg/m3, ground limestone, PFA, ground PFA, aggregate with maximum size of 20 mm, ultra-
fine amorphous colloidal silica (UFACS), and acrylic polymer superplasticizer.
The following properties were determined on the fresh concrete mixtures: slump flow,
bleeding capacity, and bulk density.
The following properties were measured on the hardened concrete specimens: compressive
strength at 1, 3, 7, 28, and 60 days, water permeability at 28 days on wet cured concrete
specimens, chloride diffusion and carbonation rate, shrinkage, and thermal heating.
All these SCCs resulted to be suitable for low-heat cement hydration mixtures devoted to
mass concrete structures with very congested reinforcement.

Keywords: Self-Compacting Concrete, Fly Ash, Superplasticizer, Heat of Hydration

Mario Collepardi ACI member, is Professor of Materials Science and Technology at the
Politecnico of Milan, Italy. He is author or co-author of numerous papers on concrete
technology and cement chemistry. He is also the recipient of several awards for his
contributions to the knowledge of superplasticizers and their use in concrete.

J.J. Ogoumah Olagot is a civil engineer working as researcher in the field of concrete at
Enco, Spresiano (TV), Italy.

Ulf Skarp is a civil engineer working in Eka chemicals, Bohus, Sweeden as technical
manager. He is the author of some papers in the area of concrete technology.

Roberto Troli is a research civil engineer and director of the Enco laboratory. He is author
of several papers in the field of concrete technology and in particular of chemical and
mineral admixtures.
Mo

1
 INTRODUCTION

Mass concrete structures are particularly vulnerable to the risk of cracking due to the thermal
gradient (δΤt) between the nucleus of the structure and its periphery. This is related to the
total heat of cement hydration developed in the concrete (Ht) at the time t, and the difference
in dissipating this heat from the nucleus and the outer part (Fig. 1).

Figure 1 Schematic trend of the temperature of the nucleus and the periphery of a mass
concrete structure in real or adiabatic conditions and complete dissipation of
heat changes in temperature in real and extreme conditions.

From a practical point of view (by taking account the thermal expansion, the elastic strain as
well as the creep of the concrete) a thermal gradient not higher than 20°C is needed to avoid
the risk of thermal cracking [1]:

δT3 ≅ δTmax ≤ 20°C (1)

where δTmax is the maximum difference in temperature between the nucleus and the
periphery of the structure (Fig. 1), which, in general occurs at a time of 3 days (δT3).
It is very difficult to calculate δTmax since it depends on many parameters such as the thermo-
insulating properties of the forms, the size and the shape of the structure, the environmental
temperature and wind conditions after placing, etc. However, we can easily calculate the
maximum gradient δTmax in extreme conditions such as:
- no thermal dissipation (adiabatic heating = ∆T) for the nucleus;
- complete thermal dissipation of the periphery (∆T = 0).
In these extreme and conservative conditions

δTmax = δT3 = ∆T3 (2)

where ∆T3 is the adiabatic heating of the nucleus after 3 days from the placing.

2
In the case of adiabatic heating for the nucleus, the maximum increase in the temperature of
the structure at 3 days can be calculated as:

∆T3 = H3/(m·sh) = c·h3/(m·sh) (3)

where h3 is the unitary heat of cement hydration at 3 days (referred to 1 kg of cement), c is

the cement content in 1m3 of concrete, m is the specific weight of 1m3 of concrete (generally
2400 kg/m3), and sh is the specific heat of concrete (generally 1.1 kJ ·kg-1·C-1). Therefore,
∆T3 can be also written as:

∆T3 ≅ c·h3/(2400·1.1) ≅ c·h3/2600 (4)

Therefore, if there is a limit of 20°C in the adiabatic heating of the nucleus at 3 days (∆T3 ≤
20°C) in real (not-extreme) conditions, the maximum thermal gradient between nucleus and
periphery (δT3) is even lower than 20°C:

δT3 << ∆T3 ≤ 20°C (5)

Then, in order to avoid any risk of thermal cracking, by taking into account the equations (4)
and (5), a cement content (c) and a cement type, with a given unitary heat of hydration (h3),
must be adopted so that:

c ⋅ h3
≤ 20°C (6)
2600

c·h3 ≤ 52MJ/m3 (7)

In order to reduce the thermal cracking risk, a cement content as low as 200-250 kg/m3 is
usually adopted for very massive concrete structures such as dams. This means that
aggregates with maximum size as large as 100-150 mm are usually used to reduce the
amount of mixing water (w) and then the cement content at a given w/c. For the same reason,
concretes at a stiff consistency are usually adopted to reduce w and then c at a given w/c.
For both the large size of aggregates and the stiff consistency of the fresh mixture, this
concrete is manufactured in special mixers and placed under very effective compaction
methods.

PURPOSE OF THE WORK

The purpose of the present work was to study self-compacting concrete (SCC), then with a
maximum size of coarse aggregate relatively small (20 mm), which should be thermally
crack-free since it would be devoted to mass concrete structures such as slab foundations
with very congested metallic reinforcement (Figure 2).

3
Figure 2 Slab foundation with very congested metallic reinforcement where SCC with
low heat of hydration is needed

The production of a low-heat and self-compacting concrete seems to be a very difficult task.
Low heat and self-compaction of concrete mixes are apparently antithetic properties for the
need of low and high amount of cement, respectively. However, due to the availability of
European blended cements, very effective superplasticizers, and new anti-bleeding agents
this objective could be achieved.

EXPERIMENTAL

In the following sections the materials used and the methods adopted will be presented.

Materials

The following materials were used. A blast high-furnace slag cement according to the EN
197/1 (CEM III/A 32.5R) with 60% of slag and 40% of Portland cement was used for its
relatively low unitary heat of hydration (Table 1).

Table 1 Unitary heat of cement hydration at different times for CEM III/A-32.5R

Heat (kJ/kg) Time (days)

70 1
170 3
210 7
260 28

Three different mineral additions as fine powder material to replace part of the cement and
reduce the heat of cement hydration were used. Table 2 shows some typical characteristics of

4
these mineral additions: ground limestone, fly-ash, and ground fly-ash in form of a liquid
slurry with 48% of water manufactured by Solena, (Italy). Figure 3 shows the ESEM
(Environmental Scanning Electron Microscopy) micrograph of the ground fly ash with the
loss of spherical particles larger than 10 µm.

Figure 3 Environmental Scanning Electron Microscopy of ground fly ash (by courtesy
of Davide Salvioni, Mapei, Milan, Italy).

Table 2 Chemical composition and physical properties of mineral additions

Ground Ground Fly-Ash

Characteristics Fly-Ash
Limestone slurry
CaCO3 (%) 98.1 - -
SiO3 (%) 1.1 59.9 31.1
Al2O3 (%) - 22.9 11.9
Fe2O3 (%) - 4.7 2.4
Na2O (%) - 0.6 0.3
K2O (%) - 2.2 1.1
SO3 (%) - 0.4 0.2
H2O (%) - - 48.1
Specific Surface
0.31 0.35 2.1
area (m2/g)
Mean size (µm) 30 25 6

Natural aggregates in the form of 0-4 mm sand (fineness modulus of 2.80) and 4-20 mm
gravel were used.
A superplasticizer based on acrylic polymer (AP) specially studied for low-slump loss ready-
mixed concrete (Dynamon 5R1 by Mapei, Milan, Italy) was used.
An inorganic Viscosity Modifying Agent in the form of an ultra-fine amorphous colloidal
silica (UFACS) dispersed in water was used (Cembinder 88, by Eka Chemicals, Sweeden) to
reduce the bleeding water in SCC. Figure 4 shows the typical morphology of spherical
particles (5-50 nm) of the UFACS used in this work.

5
 Figure 4 Micrograph of spherical particles of UFACS

Concrete mixtures. Table 3 shows the composition of 9 concrete mixtures all with about
300 kg/m3 of CEM III/A 32.5R corresponding to about 120 kg/m3 of Portland cement and
180 kg/m3 of ground granulated blast-furnace slag (GGBFS). Three concrete mixtures for
each mineral addition were manufactured with different dosage of UFACS (0-1-2% by
weight of cement + mineral addition) in order to reduce the bleeding capacity of the SCC.
The code indicating the concrete mixture contains the symbol of the mineral addition (GL for
ground limestone, FA for fly ash, GFA for ground fly-ash) and the percentage (0-1-2) of the
UFACS admixture (Table 3). A proper dosage of the AP-based superplasticizer (in the range
of 0.7-1.6% by cement + mineral addition) was adopted in order to produce SCCs all with a
slump flow of at least 750 mm at a given w/c of 0.58.

Table 3 Composition of concrete mixtures at w/c of 0.58

Mineral* AP UFACS
Cement addition Sand Gravel Water
Mix (% by (% by
(kg/m3) type/content (kg/m) (kg/m2) (kg/m3)
powder) powder)
(kg/m3)
GL/0 306 GL / 157 964 824 178 0.70 0
GL/1 306 GL / 157 964 824 178 0.96 1
GL/2 306 GL / 157 964 824 178 1.17 2
FA/0 307 FA / 128 965 824 178 0.96 0
FA/1 300 FA / 125 944 806 174 1.14 1
FA/2 304 FA / 127 964 822 176 1.31 2
**
GFA/0 312 GFA / 130 981 838 181*** 1.12 0
GFA/1 312 GFA / 129** 981 838 181*** 1.20 1
GFA/2 312 GFA / 130** 981 838 181*** 1.56 2
* GL (Ground Limestone; FA (Fly Ash); GFA (Ground Fly Ash)
** as dry content
*** including the water of the slurry of ground fly ash

6
 METHODS

The fresh mixtures were characterized by the measurements of slump flow just at the end of
the mixing and after 30 minutes. Moreover, the bleeding capacity was determined by
collecting the total amount of bleeding water (every 20 minutes) on the top of a concrete into
a container (5.2 liters) up to the setting time.
Concrete mixtures after mixing were placed without vibration in an insulating polystyrene
container of 500mmx500mmx500mm and, by the use of a thermocouple in the nucleus of the
concrete, the change in the temperature (∆T) was recorded for about 1 week in a quasi-
adiabatic condition. These tests were carried out to confirm whether or not these SCCs are
suitable for mass concrete structures according to equation (7) where, for these concretes, c =
300 kg/m3 and h3 = 170 kJ/kg:

c·h3 = 300·170 = 51,000 < 52,000 kJ/m3

The concrete mixtures were placed without any compaction in cubic forms for determining
the specific gravity and the compressive strength from 1 day up to 90 days by curing at 20°C
and 95% RH.
Permeability of water was determined by a penetration test of water under pressure (1 to 7
bar) through concrete prismatic specimens (200mmx200mmx120mm) placed without
vibration and cured at 20°C under water [2].
Shrinkage tests were carried out on prismatic specimens (100mmx100mmx500mm) placed
without vibration and exposed to dry air (RH = 50%) at 20±3°C after demoulding at 1 day.
Penetration of CO2 was determined by the phenolphthalein colorimetric test on cube
specimens exposed to dry air (RH = 60-70%) at 20°C after demoulding.
Penetration of chloride was determined by the fluorescein + AgNO3 colorimetric test [3] on
cube specimens immersed at 20°C in a 3.5% NaCl aqueous solution after curing the
demoulded specimens in water for 7 days.

RESULTS

Workability and Bleeding. Table 4 shows the slump flow and the bleeding capacity results.
The addition of UFACS makes the mix more cohesive, reduces the bleeding water and the
segregation tendency. However , in order to keep the same slump flow, a higher amount of
superplasticizer is needed by increasing the UFACS dosage (Table 3).

7
 Table 4 Properties of concrete mixtures in the fresh state

SLUMP FLOW Bleeding

Specific Aspect**
capacity (%
Mix gravity After mixing After 30 min
by vol. of (Visual Rating)
(kg/m3)
Mm sec* mm sec* concrete)
GL/0 2429 790 20 720 25 0.12 Slight segregation
GL/1 2429 780 20 650 30 0.06 Fair
GL/2 2429 790 20 670 27 0.06 Good
FA/0 2402 790 30 750 35 0.11 Slight segregation
FA/1 2349 790 29 660 40 0.09 Fair
FA/2 2393 800 30 740 36 0.07 Good
GFA/0 2442 790 20 700 25 0.06 Fair
GFA/1 2442 780 20 700 26 0.04 Good
GFA/2 2442 790 20 730 25 0.03 Excellent
*time needed to research the final slump flow
**fair=cohesive, good=very cohesive, excellent=extremely cohesive

The workability loss, in terms of decrease in the slump flow during the first 30 minutes after
mixing, is negligible independently of the mineral addition and UFACS dosage (Table 4)
The viscosity of the fresh mixture in terms of time needed to reach the final slump flow is
higher for the concrete with fly ash than with the other mineral additions.
The use of ground fly ash makes the cohesiveness of the fresh mixture much better than the
mixtures with the other mineral additions. However, the ground fly ash needs a little higher
dosage of superplasticizer for the concrete due to the higher fineness of this mineral addition
(Table 3).
Thermal Change. The thermal change in the nucleus of the SCC with GFA/3 as a function
of time is shown in Figure 5. The temperature increase (∆T3) at 3 days is lower than 20°C.
Similar results were obtained for the other SCCs. These results are in agreement with the
equation (7) and confirm that these special SCCs are also suitable for mass concrete
structures.

Figure 5 Temperature increase in the nucleus of SCC with GFA/3 in a quasi-adiabatic

condition.

8
Compressive strength. Figure 6 shows the results of the compressive strength. Concretes
with ground fly ash (GFA) and, to a lower extent, with fly ash (FA) give better results than
those with ground limestone (GL). This behaviour is due to the pozzolanic activity of the fly
ash which increases when ground because of the higher reactivity related with the higher
fineness (Table 2).
The addition of UFACS causes a slight reduction in the strength development of concretes
with GL, but does not affect the compressive strength of mixtures with FA or GFA.
The combined use of GFA (130 kg/m3), AP-based superplasticizer (1.5%) and UFACS (2%)
gives the best result in terms of both compressive strength (55MPa at 28 days, Figure 6) and
rheological properties in the fresh state: lowest bleeding capacity, highest cohesiveness,
highest fluidity and lowest loss of slump flow (Table 4).

Figure 6 Compressive strength as a function of time of SCC with different numeral addition
and FACS dosage.

Shrinkage Test. Figure 7 shows the length change of the specimens kept in a dry
environment (RH = 50%). These results indicate that the drying shrinkage of these SCCs is
of the same order of magnitude of ordinary concretes [4].

Figure 7 Drying shrinkage of concrete mixture as a function of time for the SCCs with
different mineral additions and UFACS dosage.

Durability Test. Figures 8 and 9 show the penetration of CO2 and Cl-, respectively, in the
concrete mixtures. The penetration of both CO2 and Cl- is slightly lower in FA-SCCs and
GFA-SCCs mixtures with respect to the SCCs mixtures with GL. This is in agreement with

9
the available data [5] indicating that the fly ash addition reduces the CO2 penetration for the
lower porosity and there is a certain tendency for Cl- ions to be adsorbed on the surface of
the pozzolanic material or that of hydration products [6].
There is no change in the penetration rate of both CO2 and Cl- caused by the presence of
UFACS.

Figure 8 Penetration of CO2 as a function of time ( t ) in SCCs with different mineral

addition and UFACS dosage.

Figure 9 Penetration of Cl- ions as a function of time ( t ) in SCCs with different

mineral addition and UFACS dosage.

CONCLUSIONS

The results of the present paper indicate that the combined use of CEM III/A 32.5R (300
kg/m3), AP-based superplasticizer (0.8-1.5%), powder mineral addition (130-150 kg/m3),
ultre-fine amorphous colloidal silica (1-2%), and aggregate with a maximum size of 20 mm
allow to manufacture self-compacting concretes characterized by low heat development
particularly suitable for mass concrete structures.
The use of the AP-based superplastizer is useful for self-compacting concretes with
negligible workability loss.
The presence of UFACS reduces bleeding water and segregation, and improves the
cohesiveness of the fresh mixes.

10
The performance of SCCs in terms of compressive strength appear to be better when ground
fly ash is used as powder mineral addition: 55 MPa at 28 days, versus 40 MPa with fly ash
and 35 MPa with ground limestone.
The drying shrinkage of SCCs is of the same order of magnitude as for ordinary concretes.
The durability performance in terms of carbon dioxide or chloride penetration is slightly
better in the presence of ground fly ash with respect to the other mineral addition. Ultra-fine
amorphous colloidal silica does not affect the durability performance.

REFERENCES

1. European Standard, EN 206-1, “Concrete Performance, Production, Placing and

Compliance Criteria”, 2000.
2. ISO DIS 7031 “Permeability of Concrete. Penetration of Water under Pression.”
3. M. Collepardi, “Quick method o determine free and bound chlorides in concrete”,
RILEM Workshop on Chloride Penetration into Concrete, Saint Remy-les-Chevreuse,
pp. 10-16, 1995.
4. A.M. Neville, Properties of Concrete, 4th edition, Longman Group Limited, Harlow
Essex, England, 1995.
5. A. Borsoi, S. Collepardi, L. Coppola, R. Troli and M. Collepardi, “Effect of
Superplasticizer Type on the Performance of High-Volume Fly Ash Concrete”,
Proceedings of the Sixth CANMET/ACI International Conference on Superplasticizer
and Other Chemical Admixtures in Concrete, pp. 17-27, 2000.
6. M. Collepardi, A. Marcialis and R. Turriziani, “Penetration of chloride ions in cement
pastes and in concretes”, Journal of American Ceramic Society, 55, pp. 534-535, 1972.

11