Accepted Manuscript

Effects of activator and aging process on the compressive strengths of alkali-activated

glass inorganic binders

Tai-An Chen, Ji-Hsien Chen, Jong-Shin Huang

PII: S0958-9465(16)30785-5
DOI: 10.1016/j.cemconcomp.2016.11.011
Reference: CECO 2744

To appear in: Cement and Concrete Composites

Received Date: 14 December 2015

Revised Date: 18 October 2016
Accepted Date: 27 November 2016

Please cite this article as: T.-A. Chen, J.-H. Chen, J.-S. Huang, Effects of activator and aging process on
the compressive strengths of alkali-activated glass inorganic binders, Cement and Concrete Composites
(2016), doi: 10.1016/j.cemconcomp.2016.11.011.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Effects of activator and aging process on the compressive strengths of

alkali-activated glass inorganic binders
Tai-An Chen, Ji-Hsien Chen, and Jong-Shin Huang*

Department of Civil Engineering

PT
National Cheng Kung University

Tainan, 70101 Taiwan R.O.C.

RI
For submission to: Cement and Concrete Composites

SC
Revised May 25, 2016

U
Abstract: The compressive strengths of alkali-activated glass inorganic binders

AN
(AAGIBs), produced under a normal process (i.e., without aging) by using various

activators composed of both sodium hydroxide and sodium silicate, were measured.
M
The optimal alkali-equivalent content and silicate modulus needed in the activators to

produce the maximum compressive strengths of AAGIBs were determined

D

empirically. Next, the compressive strengths of AAGIBs, produced under an aging

TE

process by mixing with various activators composed of only sodium hydroxide, were

measured. The effects of alkali-equivalent content, aging temperature, and aging

EP

duration on the compressive strengths of AAGIBs were investigated. Moreover, the

compressive strengths of AAGIBs obtained under the aging process were compared to
C

those under the normal process. It was found that the compressive strengths of
AC

AAGIBs can be enhanced by using the aging process, even if only sodium hydroxide

is introduced in the activators.

Keywords: Waste Glass; Alkali-activated; Inorganic binder; Aging process

*Author to whom all correspondence should be addressed.

Fax: 886-6-2358542. E-mail: jshuang@mail.ncku.edu.tw

1
 ACCEPTED MANUSCRIPT

1. Introduction

Taiwan produces a total of more than 2x 105 tons of waste container glass

annually, which might have some negative ecological and environmental impacts if

PT
not handled or disposed of adequately. Only a small portion of waste container

glass can be efficiently reused, with the remainder mostly treated as solid waste.

RI
Waste container glass is utilized primarily as a supplementary cementitious material

SC
in concrete, due to its high volume fraction of silicate. When crushed waste glass is

used as fine aggregates for concrete the slump and strength of the resulting material

U
decrease, but the air content increases along with the percentage of waste glass, due

to its angular grain shape [1].

AN
Both alkali-silica and pozzolanic reactions may occur

when waste glass is used as an aggregate or cement replacement in concrete. If

M
waste glass is ground into a fine glass powder as fine aggregates [2, 3] or a partial

cement replacement [4] in concrete, then pozzolanic reactivity becomes higher and
D

alkali-silica reaction expansion reduces, leading to an increase in strength. The

TE

main factors affecting the alkali-silica and pozzolanic reactions for waste glass in

concrete include glass powder size, paste pore solution and chemical composition [5].
EP

The setting and early strength development of cement paste are unaffected, because

only a small proportion of the sodium ions is released from glass powder into the
C

paste pore solution [6]. Meanwhile, the chemical composition of waste glass meets
AC

the requirements needed for use as raw materials for cement production. The

cement produced using waste glass as a raw material shows no significant differences

compared with cement produced without waste glass [7]. Although waste glass can

be used as a supplementary material in concrete, or as a raw material for cement

production, the consumption of waste glass in Taiwan remains limited, and thus both

new and improved applications for this material are needed.

2
 ACCEPTED MANUSCRIPT

Inorganic polymers, also called geopolymers or alkali-activated binders, can be

made from a variety of raw materials containing aluminosilicate minerals, such as

metakaolin [8-13], blast furnace slag [14-19], fly ash [19-23], bottom ash [24],

construction and demolition waste [25] and calcined reservoir sludge [26-28], by

PT
mixing with an activator that is normally composed of alkali metal hydroxide and

sodium silicate. The microstructure and properties of the resulting inorganic

RI
polymers depend on the chemical nature of the raw material used and the process

SC
variables employed, such as the dosage of alkali metal hydroxide and sodium silicate

in the activator and the process temperature and time. In the alkali activation

U
process of aluminosilicates, the dissolution reaction, including the breakage of

Si − O − Si , Si − O − Al and Al − O − Al covalent bonds and the release of Si

AN
and Al ions is mainly caused by the high OH − concentration provided by alkali
M
metal hydroxide, while the subsequent polycondensation reaction is caused by the

sufficient amounts of some silicate precursors partially supplied by sodium silicate.

D

Both dissolution and polycondensation reactions are affected by the OH −

TE

concentration in the activator and the process temperature and time. Consequently,

different aluminosilicate raw materials, dosages of activators and processing methods

EP

result in various microstructures that have different physical and mechanical

properties. In particular, the compressive strengths of inorganic binders produced

C

by mixing with activators of sodium hydroxide and sodium silicate are much better
AC

than those made from ordinary Portland cement [15]. Alkali-activated binders can

thus be used as a replacement for Portland cement in concrete when the aims are to

reduce energy use and carbon emissions. However, the production of sodium

silicate has a greater environmental impact than ordinary Portland cement [29], and

its price is high in Taiwan, thus limiting the application of alkali-activated binders in

concrete. Therefore, the reduction in the use of sodium silicate in activators to

3
 ACCEPTED MANUSCRIPT

produce inorganic binders for concrete from solid waste is an issue of considerable of

importance, and one that needs to be explored in more detail. Presumably, waste

container glass with the primary chemical constituents of SiO2 , Na2O , Al 2O3 and

CaO might be alkali-activated by mixing with activators of both sodium hydroxide

PT
and sodium silicate, or only sodium hydroxide, and could be potentially utilized as a

RI
raw material in the production of inorganic binders.

SC
When glass powders are mixed with sodium hydroxide solution in a container,

some Si − O − Si and Si − O − Al covalent bonds on their surfaces are destroyed

U
under a high OH − concentration attack, and Si and Al ions are released into the
AN
solution to form species of Si − OH and Al − OH groups. During the dissolution

reaction, an ionic species of Si(OH ) 4 forms on the surface of glass powders and
M
−
then further reacts with sodium hydroxide to become SiO4 H 3 [30]. The ionic

species of Si − OH and Al − OH groups in the solution could later condense to

D

form Si − O − Si and Si − O − Al bonds. The dissolution and polycondensation

TE

chemical reactions are influenced by the OH − concentration in the solution and its

temperature. The feasibility of using soda-lime silicate glass powders, composed of

EP

roughly 68.2% SiO2 , as a raw material in the production of inorganic polymers was

studied by Ke and Gong [31], using an activator composed of only calcium hydroxide.
C

The compressive strengths of the alkali-activated glass inorganic binders (AAGIBs)

AC

reached 55 MPa when they were cured under a high steam pressure of 1.2 MPa for 10

hours. However, the compressive strength decreased to less than 15 MPa when the

steam pressure was less than 0.6 MPa, and in some cases the AAGIB specimens had

no compressive strength. The compressive strength of soda-lime silicate glass,

alkali-activated by an activator with an SiO2 / Na2O weight ratio of 1.636 and then

4
 ACCEPTED MANUSCRIPT

cured at a temperature of 800 C for 8 hours, was 92.19 MPa, as measured by Jin

[32]. In addition, Jing [33] reported that the compressive strengths of

aluminosilicate glass powders, alkali-activated by only sodium hydroxide and cured

at a temperature higher than 120 0 C for 8 hours, was around 60 MPa. When the

PT
powders were cured at a temperature of less than 100 0 C , the alkali-activation

reaction was severely retarded, and thus no strength development was observed.

RI
Based on experimental results reported in the literature [31-33], the strength

SC
development of AAGIBs is very limited and their compressive strengths much lower

if glass powders are alkali-activated at room temperature by mixing with an activator

U
without sodium silicate.

AN
The dissolution and polycondensation reactions of waste glass powders are

affected by the OH − concentration and the process temperature and time when they
M
are alkali-activated by mixing with an activator of sodium hydroxide and sodium
D

silicate. In the present study, the compressive strengths of AAGIBs alkali-activated

TE

by mixing with various activators under a normal process with stirring at room

temperature for 30 minutes are measured and compared. The optimal activator for
EP

producing maximum compressive strength is thus determined. The dissolution and

polycondensation reactions are expected to become faster at a higher temperature,

C

and more complete with a longer duration of stirring, even when only sodium
AC

hydroxide is used in the activator. Next, AAGIB specimens are produced under an

aging process by mixing glass powders with various activators without sodium

silicate at different aging temperatures and for various durations. The effects of

activator, aging temperature, and aging duration on the compressive strengths of

AAGIBs are thus evaluated. The compressive strengths of AAGIBs produced under

an aging process are compared with those of AAGIBs obtained under a normal
5
 ACCEPTED MANUSCRIPT

process to evaluate the effects of aging.

2. Materials and experimental methods

PT
2.1 Waste glass powders

RI
Waste glass containers were first washed and then crushed using a vertical

pulverizer to pass through a 300 µm sieve.

SC
The resulting glass particles were

further ground down to become glass powders, with a specific surface area of

U
4303 cm2 / g , using a disk mill (KAWASAKI, T-100). The size distribution of the

AN
glass powders, as characterized using an INSITEC laser diffraction particle size

analyzer, is shown in Fig. 1. The geometrical size of the glass powders ranges
M
mostly from 0.5 to 200 µm , with a mean value of 20.12 µm and a standard deviation

of 25.5 µm . The primary chemical constituents of the glass powders with a specific
D

gravity of 2.55 and a water absorption capacity of 0.2% are 71-74% SiO2 ,
TE

12-14% Na2O , 9-12% CaO and 1-3% Al 2O3 .

EP

2.2 Activator
C

The glass powders were then alkali-activated by mixing with an activator to

AC

evaluate the feasibility of using waste glass as a raw material in the production of

inorganic binders. The activator used here was a mixture of water, sodium

hydroxide, and sodium silicate, alternatively named water glass, with a chemical

composition of 29.1% SiO2 and 9.5% Na2O . Since the chemical reaction,

microstructure, and properties of an inorganic binder can be affected dramatically by

6
 ACCEPTED MANUSCRIPT

the amounts of Na2O , SiO2 and water in an activator, three parameters, namely

alkali-equivalent content, silicate modulus, and water/binder ratio are typically

employed. The alkali-equivalent content, denoted by AE % , is defined as the

weight fraction of Na2O to glass powders, and the silicate modulus, denoted by

PT
M s , is the molar ratio of SiO2 to Na2O in an activator. The water/binder ratio,

denoted by W / B , is the weight ratio of water to the sum of glass powders, and

RI
SiO2 and Na2O in an activator. In this study, W / B was fixed at 0.3 for all

SC
AAGIB specimens.

U
2.3 Normal process

AN
The activator plays an important role in determining the microstructure and
M
properties of AAGIB specimens. To evaluate the effects of sodium hydroxide and

sodium silicate on the compressive strengths of AAGIBs, activators with various

D

alkali-equivalent content levels of AE % =1%, 2%, 3%, 4%, 5%, 6%, and 8% and
TE

various silicate moduli of M s =0, 0.5, 1.0, 1.5, and 2.0 were used in the production of

AAGIB specimens, with a water/binder ratio of W / B =0.3. For each activator, the
EP

required amounts of sodium hydroxide, sodium silicate, and water were weighed,

mixed, and then placed in a container at room temperature for one day. Next, glass
C

powders were added to the container and stirred vigorously for 30 minutes. After
AC

complete mixing, the AAGIB slurry was poured into 3cmx3cmx3cm steel molds and

further compacted on a vibrating table (CONTROLS, frequency of 60 Hz) to get rid

of any air bubbles. The steel molds were covered with plastic foil to prevent the

evaporation of moisture, and then cured in an oven at a temperature of 700 C for 24

hours. One day later, the cubic specimens were demolded and cured at room

temperature for an additional 3 days. The 4-day AAGIB specimens were loaded
7
 ACCEPTED MANUSCRIPT

uniaxially in a universal testing machine (SHIMADZU, UHC-100A) under a constant

compressive stress rate of 300KPa/s. The resulting compressive strengths of the

4-day AAGIB specimens were recorded and compared to investigate the effects of

activators with different AE % and M s values.

PT
2.4 Aging process

RI
SC
It is expected that the dissolution rate of Si ions, subsequent polycondensation

reaction, and resulting microstructure and compressive strength of AAGIB specimens

U
depend on the AE % in the activator without any sodium silicate (i.e., M s =0), as

AN
well as the aging temperature and duration. To evaluate the effects of sodium

hydroxide and the aging process on the compressive strengths of AAGIBs, activators
M
with various alkali-equivalent contents of AE % = 1%, 2%, 3%, 4%, 5%, 6%, and 8%

and various aging temperatures of 290 C , 400 C , 500 C , 600 C , and 700 C were
D

employed in the production of AAGIB specimens with W / B =0.3. Intuitively, the

TE

dissolution and polycondensation reactions of glass powders in an activator become

more severe and complete as the aging temperature is increased and the aging
EP

duration is extended. The interval between any two consecutive aging durations

was thus reduced as the aging temperature was raised. Here, the aging duration
C

intervals were 120, 60, 30, 15, and 5 min for aging temperatures of 29, 40, 50, 60,
AC

and, 700 C respectively, regardless of the AE % in the activator.

The aging process employed in the production of alkali-activated glass inorganic

binder specimens was as follows. Glass powders were mixed with the activator, and

then stirred vigorously at a high temperature for a long duration. The chemical

reactions of dissolution and polycondensation were faster at a higher temperature, and

8
 ACCEPTED MANUSCRIPT

were more complete with a longer duration of stirring, even when only sodium

hydroxide was used in the activator. The amounts of sodium hydroxide and water

required for each mix of activator were first mixed in a container and then cooled to

room temperature. Next, a bath tank was heated to the prescribed aging temperature

PT
by using a heater equipped with a temperature controller. Glass powders and the

activator were then poured into a separate reaction tank with an inner diameter of

RI
82 mm and a height of 110 mm , and stirred vigorously by using a DC mixer with a

SC
blade of 72 mm in diameter at a speed of 500 rpm for 30 seconds at room

temperature. The reaction tank was then placed in the center of the bath tank at the

U
prescribed aging temperature, and stirred again using the DC mixer at a constant

AN
speed of 500 rpm for various aging durations.

the aging process is illustrated in Fig. 2.

The set-up of the apparatus used for

After the prescribed aging duration, the

M
AAGIB slurry was cooled to room temperature and poured into 3cmx3cmx3cm steel

molds. The subsequent procedures of specimen preparation, curing, and mechanical

D

testing for the AAGIB specimens obtained under an aging process were exactly the
TE

same as those used for the specimens obtained under a normal process.
EP

3. Results and discussion

C

3.1 Optimal AE % and M s for normal process

AC

The variations of the measured 4-day compressive strengths of the AAGIB

specimens, produced under a normal process by mixing glass powders with various

activators with different AE % and M s values, and stirring for only 30 min at

room temperature, are presented in Figs. 3(a) and (b), respectively. In Fig. 3(a), the

9
 ACCEPTED MANUSCRIPT

compressive strengths of AAGIB specimens alkali-activated by activators with a

given M s (e.g., M s =0, 0.5, 1.0, 1.5, and 2.0) initially increase dramatically and

then decrease steadily after reaching a maximum as AE % is increased from 1% to

8%. When AE % is very low, for example AE % = 1%, the dissolution of

PT
aluminosilicate minerals and the release of Si ions are limited. The solubility and

RI
concentration of inorganic polymer precursors are thus too low to give a dense

microstructure, leading to a lower compressive strength. The microstructures of the

SC
AAGIB specimens become denser and the compressive strengths become higher

when AE % is increased slightly. However, the compressive strengths of the

U
AAGIBs decrease significantly when AE % exceeds an optimal value (i.e., that at

AN
which maximum compressive strength is obtained). From Fig. 3(a), the optimal

AE % depends on the prescribed M s in the activator. When M s =0, (i.e., only

M
sodium hydroxide is in the activator), the compressive strength of the AAGIB
D

specimens increases to a value of 70.4 MPa when AE % approaches 5%, then drops
TE

significantly when AE % is further increased to 8%. When M s is raised to 0.5

and 1.0, the maximum compressive strengths are 101.0 and 120.4 MP, respectively,
EP

with an optimal AE % of 4%. The compressive strengths of AAGIBs with

M s =1.5 and 2.0 increase to 138.3 and 126.3MPa, respectively, with an optimal
C

AE % of 3% in the activators. Hence, the variations in the compressive strengths of

AC

AAGIBs with AE % are similar for various M s values, but the optimal AE %

decreases from 5% to 3% when M s is increased from 0 to 2.

Figure 3(b) shows that the compressive strengths of AAGIB specimens

10
 ACCEPTED MANUSCRIPT

alkali-activated by the activators of sodium hydroxide and sodium silicate ( M s = 0.5,

1.0, 1.5, and 2.0) are much higher than those for specimens activated with only

sodium hydroxide ( M s = 0), regardless of the AE % in the activators. When glass

PT
powders are mixed with the activator with only sodium hydroxide and then stirred at

room temperature for 30 min, the dissolution of aluminosilicate minerals and release

RI
of Si ions are very limited, and the resulting concentration of inorganic polymer

precursors is too dilute to form a dense microstructure, leading to lower compressive

SC
strength. Therefore, the addition of sodium silicate to activators is necessary to

enhance the compressive strengths of AAGIB specimens alkali-activated under a

U
normal process. From Fig. 3(b) it is also noted that the compressive strengths of the
AN
AAGIB specimens initially increase and then gradually decrease after reaching a

peak value as M s is increased from 0 to 2.0. For instance, the compressive

M

strengths of AAGIB specimens with AE % = 4% increase significantly as M s is

D

increased from 0 to 0.5, and then reach a maximum compressive strength of 128.2
TE

MPa as M s is further increased from 0.5 to 1.5, but decrease slightly when M s is

further increased to 2.0. This occurs because the ionic silicate species are more
EP

polymerized with the addition of a high molar fraction of sodium silicate, and the

reactions between themselves and with other ionic species are slower [34].
C

The

polycondensation reaction is triggered in a short time before the dissolution of

AC

aluminosilicate minerals and the release of Si ions are complete. As a result, the

compressive strengths of AAGIB specimens decrease slightly when M s is increased

to 2.0. The optimal M s value depends on the prescribed AE % in a mix. When

AE % is high, the dissolution of aluminosilicate minerals and the release of Si ions

11
 ACCEPTED MANUSCRIPT

are faster, and the amount of sodium silicate required to enhance the subsequent

polycondensation reaction is thus lower. In contrast, the amount of sodium silicate

required in the activators is greater, because the dissolution of aluminosilicate

minerals and the release of Si ions are much slower when AE % is low. From

PT
Fig. 3(b), it is found that the best values in the activator to produce the AAGIB

specimens with a maximum compressive strength of 138.3 MPa are M s = 1.5 and

RI
AE % = 3%.

SC
3.2 Effects of aging temperature and duration in the aging process

U
AN
The variations in the 4-day compressive strengths of the AAGIB specimens,

produced by mixing glass powders with activators and with various aging durations
M
and different temperatures, are shown in Figs. 4-10 for AE % =1%, 2%, 3%, 4%,

5%, 6%, and 8%, respectively. Here, the aging duration intervals are 120, 60, 30,
D

15, and 5 min, respectively, for aging temperatures of 29, 40, 50, 60, and 700 C .
TE

The experimental results in these figures indicate that the compressive strengths of

the AAGIB specimens are affected dramatically by the AE % in the activators, as

EP

well as the temperature and duration of the aging process. Moreover, it is found

that the compressive strengths of the AAGIB specimens alkali-activated by the

C

aging process at the tested temperatures initially remain stationary, later increase
AC

steeply, and then decrease significantly after reaching a peak value as the aging

duration is increased. The trends of the variation of compressive strengths with

aging duration are consistent for all aging temperatures and AE % values, as

illustrated in Figs. 4-10.

12
 ACCEPTED MANUSCRIPT

Some of the measured 4-day compressive strengths of the AAGIB specimens are

listed in Tables 1 and 2 for AE % =1%-3% and AE % =4%-8%, respectively. The

compressive strengths of the AAGIB specimens alkali-activated under the aging

process remain initially unchanged as compared to those of specimens obtained

PT
under the normal process, and then start to increase along with the aging duration

after a critical time of stirring, defined as the threshold aging duration and denoted

RI
by <> in Tables 1 and 2. In other words, the 4-day compressive strengths of the

SC
AAGIB specimens obtained under the aging process with stirring for the threshold

aging duration at a specific temperature start to become slightly larger than those

U
obtained under the normal process with 30 stirring min at room temperature. The

AN
threshold aging durations for all aging temperatures and AE % values can thus be

directly determined from Figs. 4-10, and the results are given in Tables 1 and 2. As
M
expected, the threshold aging durations decrease with increasing AE % and aging

temperature. For example, the threshold aging duration decreased from 30 to 15

D

min when AE % was increased from 2% to 3% for an aging temperature of 600 C ,

TE

and it decreased from 60 to 15 min when the aging temperature was increased from

40 to 500 C with AE % =3%. When the aging temperature was 700 C , threshold
EP

aging durations of less than 5 min were found in some cases, but were not

determined empirically because the interval between any two consecutive aging
C

durations was set to be 5 min in the initial phase of the study. These threshold
AC

aging durations are thus regarded as 5 min in Tables 1 and 2.

When the stirring time exceeds the threshold aging duration, the compressive

strengths of the AAGIB specimens initially increase along with the aging duration,

and then start to drop after another critical time of stirring, defined as the optimal

aging duration and denoted by {} in Tables 1 and 2. The increase in compressive

13
 ACCEPTED MANUSCRIPT

strength after the threshold aging duration but before the optimal one is due to the

enhancement of the dissolution rate of Si ions and subsequent polycondensation

reaction with vigorous stirring at a high temperature for a long duration. In contrast,

the reduction in compressive strength after the optimal aging duration of stirring is

PT
due to the subsequent polycondensation reaction being interfered with or even

prevented by further vigorous stirring. In Table 1, the compressive strengths of the

RI
AAGIB specimens reached 135.98 MPa when AE % = 3%, the aging temperature

SC
was 600 C , and the optimal aging duration was 75 min, but remained almost

unchanged when the aging duration was increased to 90 min. It is also found that

U
the optimal aging durations for a given aging temperature decreased with increasing

AE % .
AN
However, the corresponding maximum compressive strengths increased

initially and then decreased as AE % was increased from 1% to 8%. The optimal
M
aging duration can be greatly reduced for the AAGIB specimens with a given AE %

value if the aging temperature is increased slightly. The compressive strength of the
D

AAGIB specimens with AE % =3% was 129.35MPa when they were alkali-activated
TE

at 400 C by stirring vigorously for 420 min. The compressive strength reached

147.07MPa when the aging temperature was raised to 500 C but the optimal aging
EP

duration decreased to 180 min. Meanwhile, it was found that the maximum

compressive strength of the AAGIB specimens with AE % =2% under the aging
C

temperature of 290 C after 480 min stirring was 74.67 MPa, much lower than that
AC

under other aging temperatures of 40, 50, 60 or 700 C . Therefore, the interval

between any two consecutive aging durations under the aging temperature of 290 C

was shortened from 120 to 60 min, and the maximum compressive strength of 95.97

MPa under the aging temperature of 290 C after 540 min stirring was obtained, as

shown in Table 1.

14
 ACCEPTED MANUSCRIPT

After a long period of stirring, the resulting AAGIB slurry becomes highly

viscous and thus more difficult to pour into steel molds, leading to a less dense

microstructure and a lower compressive strength. The longest time of stirring,

before which the AAGIB specimens can be produced and their compressive strengths

PT
are measured, is defined as the ultimate aging duration of the aging process, and

denoted by [] in Tables 1 and 2. For example, the compressive strength was 45.18

RI
MPa when AE % = 2%, the aging temperature was 600 C , and the threshold aging

SC
duration was 30 min. Although the compressive strength could be increased to

102.47 MPa by stirring vigorously for an optimal aging duration of 105 min, it

U
decreased significantly to 57.26 MPa with stirring for 190 min (the ultimate aging

duration).
AN
From Tables 1 and 2, the ultimate aging duration decreases with

increasing aging temperature and AE % in the activators. The reasons for these
M
effects are similar to those for the threshold and optimal aging durations.
D

3.3 Effects of AE% in aging process

TE

From Tables 1 and 2, it is noted that the compressive strengths of the AAGIB
EP

specimens, alkali-activated with a given AE % , but stirred vigorously at different

aging temperatures for the corresponding optimal aging durations, are almost
C

identical. Here, the average of the compressive strengths of the AAGIB specimens,
AC

under a given aging process but at different aging temperatures for the corresponding

optimal aging durations, is defined as the ultimate compressive strength under the

aging process with optimal aging duration. When AE % = 3%, for example, the

compressive strengths of the AAGIB specimens in Table 1 are 147.20 MPa under an

aging process at room temperature for 480 min, 129.35 MPa at 400 C for 420 min,

147.07 MPa at 500 C for 180 min, 135.98 MPa at 600 C for 75 min, and 124.51
15
 ACCEPTED MANUSCRIPT

MPa at 700 C for 40 min, thus giving the ultimate compressive strength under the

aging process with optimal aging duration of 136.82 MPa, as reported in Table 1.

The basic compressive strength is defined as the compressive strength of the AAGIB

specimen obtained under the normal process with stirring for 30 min at room

PT
temperature. Based on the experimental results, the basic compressive strengths

under the normal process of the AAGIB specimens with AE % =1%, 2%, 3%, 4%,

RI
5%, 6% and 8%, are 11.10, 26.24, 34.47, 51.26, 70.39, 56.25 and 47.26 MPa,

SC
respectively. From Tables 1 and 2, it can be observed that the difference between

the basic and ultimate compressive strengths is significant, and depends on the

U
amount of AE % in the activators. When AE % is within the range of 1-3%, the

AN
compressive strengths of the AAGIB specimens can be enhanced 3.5- to 5.9-fold by

using an aging process with stirring at an aging temperature for the optimal aging
M
duration. When AE % = 4-8%, the basic compressive strengths of the AAGIB

specimens are more than 47 MPa under the normal process with stirring at room
D

temperature. Although the ultimate compressive strengths are still enhanced by

TE

using an aging process with stirring at higher aging temperatures for longer aging

durations, they are only 1.4- to 2.3-fold higher than the basic compressive strengths
EP

under the normal process.

C

The variations of the basic and ultimate compressive strengths of the AAGIB
AC

specimens with AE % are plotted in Fig. 11. Both the basic and ultimate

compressive strengths initially increase significantly, and then drop gradually after

reaching a peak value as AE % is increased from 1% to 8%. When AE % is low,

for instance AE % = 1% in Fig. 11, the concentration of inorganic polymer precursors,

after a long aging duration with stirring at a specific aging temperature, is still too

dilute to form a dense microstructure, resulting in a lower compressive strength.

16
 ACCEPTED MANUSCRIPT

When AE % is much higher than the optimal value, for instance AE % = 5-8%, an

approximately proportional reduction in the compressive strengths of the AAGIB

specimens with respect to AE % is observed in Fig. 11. This reduction can be

attributed to the formation of sodium carbonate resulting from the reaction of extra

PT
sodium ions with carbon dioxide [35]. The FTIR results for the glass powder and

the AAGIB specimens with AE % =1% and 8% are shown in Fig. 12. The peak at

RI
the wavenumber of 1560 cm−1 for the AAGIB specimen with AE % = 8% in Fig. 12

SC
is responsible for the existence of sodium carbonate [36]; no peaks were observed for

either the glass powder or the AAGIB specimen with AE % =1%. Hence, the

U
optimal AE % for obtaining the maximum compressive strengths of AAGIB

AN
specimens obtained under an aging process is around 3%.
M
3.4 Effects of aging process
D

The maximum compressive strengths with respect to various AE % of the

TE

AAGIB specimens, alkali-activated by the activators with sodium hydroxide and

sodium silicate under the normal process, can be determined from Fig. 3(a), as plotted
EP

in Fig. 11, and further compared to those of specimens obtained under the aging

process. In Fig. 11, it is noted that the compressive strengths of the AAGIB
C

specimens obtained under the aging process are much larger than those of specimens
AC

obtained under the normal process when AE % is within the range of 1-2%. When

AE % = 3%, the ultimate compressive strength of the AAGIB specimens obtained

under the aging process (136.82 MPa) is roughly identical to that (138.27 MPa) of the

specimens obtained under the normal process with the optimal silicate modulus

( M s =1.5). In Fig. 11, the compressive strengths of the AAGIB specimens obtained

17
 ACCEPTED MANUSCRIPT

under the aging process become slightly lower than those of specimens obtained

under the normal process if AE % ≥ 4% . Based on the experimental results

presented in Fig. 11, the compressive strengths of the AAGIB specimens can be

increased by employing an adequate aging process, even when glass powders are

PT
mixed with an activator composed of only sodium hydroxide. Meanwhile, the pH

values with respect to the AE % in the activators used and the AAGIB slurry

RI
obtained under normal and aging processes were measured, and the results are shown

SC
in Fig. 13. It is seen that the reduction in pH values between the activator and the

resulting AAGIB slurry under the aging process is much larger than that under the

U
normal process. In other words, the consumption of OH − ions during the

AN
dissolution reaction for AAGIB specimens obtained under the aging process is more

severe than that under the normal process. As expected, the effect of the aging
M
process on the increase in compressive strength of AAGIB specimens is more

efficient than that of the normal process.

D
TE

4. Conclusions
EP

The compressive strengths of the AAGIB specimens, as alkali-activated by

activators with a given M s under a normal process, initially rise dramatically and
C

then decrease steadily after reaching a maximum value as AE % is increased. The

AC

optimal AE % for obtaining the maximum compressive strength depends on the

prescribed M s . The best values in the activator to produce the AAGIB specimens

under a normal process are M s = 1.5 and AE % = 3%, giving the maximum

compressive strength of 138.3 MPa. Furthermore, the compressive strengths of the

AAGIBs under an aging process are affected significantly by the AE % in the

18
 ACCEPTED MANUSCRIPT

activators, aging temperature, and aging duration. The compressive strengths of

AAGIBs initially remain unchanged, then start to increase along with the aging

duration after the threshold aging duration is reached, and finally drop slightly when

the stirring time exceeds the optimal aging duration. It is also found that the

PT
threshold aging duration decreases with increasing aging temperature, the optimal

aging duration decreases with increasing AE % , and the resulting maximum

RI
compressive strengths depend on AE % . The AE % for obtaining the maximum

SC
ultimate compressive strengths of 136.82 MPa for the AAGIB specimens under an

aging process with optimal aging duration is around 3%. Moreover, the results

U
show that an increase in the compressive strengths of the AAGIB specimens can be

AN
achieved by employing an adequate aging process, even when glass powders are

mixed with an activator that contains only sodium hydroxide.

M
ACKNOWLEDGEMENTS
D
TE

The financial support of the Ministry of Science and Technology, Taiwan,

R.O.C., under grant NSC 102-2211-E-006-153-MY3 is gratefully acknowledged.

C EP
AC

19
 ACCEPTED MANUSCRIPT

References

[1] S.B. Park, B.C. Lee, J.H. Kim, Studies on mechanical properties of concrete
containing waste glass aggregate, Cem. Concr. Res. 34 (12) (2004) 2181-2189.
[2] A. Shayan, A. Xu, Value-added utilization of waste glass in concrete, Cem.
Concr. Res. 34 (1) (2004) 81-89.

PT
[3] C. Shi, Y. Wu, C. Riefler, H. Wang, Characteristics and pozzolanic reactivity of
glass powders, Cem. Concr. Res. 35 (5) (2005) 987-993.
[4] Y. Shao, T. Lefort, S. Moras, D. Rodriguez, Studies on concrete containing

RI
ground waste glass, Cem. Concr. Res. 30 (1) (2000) 91-100.
[5] L.M. Federico, S.E. Chidiac, Waste glass as a supplementary cementitious
material in concrete- Critical review of treatment methods, Cem. Concr. Comp.

SC
31 (8) (2009) 606-610.
[6] N. Schwarz, N. Neithalath, Influence of a fine glass powder on cement
hydration: comparison to fly ash and modeling the degree of hydration, Cem.
Concr. Res. 38 (4) (2008) 429-436.

U
[7] G. Chen, H. Lee, K.L. Young, P.L. Yue, A. Wong, T. Tao, K.K. Choi, Glass
recycling in cement production—an innovative approach, Waste Manage. 22 (7)
(2002) 747753. AN
[8] J. Davidovits, Geopolymers: Inorganic polymeric new materials, J. Therm. Anal.
37 (8) (1991) 1633-1656.
[9] A. Palomo, F.P. Glasser, Chemically-bonded cementitious materials based on
M
metakaolin, Br. Ceram. Trans. J. 91 (4) (1992) 107-112.
[10] A. Palomo, M.T. Blanco-Varela, M.L. Granizo, F. Puertas, T. Vazquez, M.W.
Grutzeck, Chemical stability of cementitious materials based on metakaolin,
D

Cem. Concr. Res. 29 (7) (1999) 997-1004.

[11] M.L. Granizo, M.T. Blanco-Varela, A. Palomo, Influence of the starting kaolin
on alkali-activated materials based on metakaolin. Study of the reaction
TE

parameters by isothermal conduction calorimetry, J. Mater. Sci. 35 (24) (2000)

6309-6315.
[12] P. Duxson, J.L. Provis, G.C. Lukey, F. Separovic , J.S.J. van Deventer, 29Si NMR
EP

Study of Structural Ordering in Aluminosilicate Geopolymer Gels, Langmuir. 21

(7) (2005) 3028-3036.
[13] P. Duxson, S.W. Mallicoat, G.C. Lukey, W.M. Kriven, J.S.J. van Deventer. The
effect of alkali and Si/Al ratio on the development of mechanical properties of
C

metakaolin-based geopolymers. Colloids Surf., A: Physicochemical and

Engineering Aspects 292 (1) (2007) 8-20.
AC

[14] T.W. Cheng, J.P. Chiu, Fire-resistant geopolymer produced by granulated blast
furnace slag, Miner. Eng. 16 (3) (2003) 205-210.
[15] Davidovits J, Sawyer JL. Early high-strength mineral polymer. U. S. Patent
4509985 A, 1985.
[16] Davidovits J. Method for obtaining a geopolymeric binder allowing to stabilize,
solidify and consolidate toxic or waste materials. U. S. Patent 5349118 A, 1994.
[17] S. Hu, H. Wang, G. Zhang, Q. Ding, Bonding and abrasion resistance of
geopolymeric repair material made with steel slag, Cem. Concr. Comp. 30 (3)
(2008) 239-244.
[18] C.K. Yip, G.C. Lukey, J.S.J. van Deventer, The coexistence of geopolymeric gel
and calcium silicate hydrate at the early stage of alkaline activation, Cem.
20
 ACCEPTED MANUSCRIPT

Concr. Res. 35 (9) (2005) 1688-1697.

[19] D. Ravikumar, S. Peethamparan, N. Neithalath, Structure and strength of NaOH
activated concretes containing fly ash or GGBFS as the sole binder, Cem. Concr.
Comp. 32 (6) (2010) 399-410.
[20] A. Palomo, M.W. Grutzeck, M.T. Blanco, Alkali-activated fly ashes: A cement
for the future, Cem. Concr. Res. 29 (8) (1999) 1323–1329.
[21] J.G.S. van Jaarsveld, J.S.J. van Deventer, G.C. Lukey, The characterisation of

PT
source materials in fly ash-based geopolymers, Mater. Lett. 57 (7) (2003)
1272-1280.
[22] A. Buchwald, M.Schulz, Alkali-activated binders by use of industrial
by-products, Cem. Concr. Res. 35 (5) (2005) 968-973.

RI
[23] P. Chindaprasirt, T. Chareerat, V. Sirivivatnanon, Workability and strength of
coarse high calcium fly ash geopolymer, Cem. Concr. Comp. 29 (3) (2007)
224-229.

SC
[24] P. Chindaprasirt, C. Jaturapitakkul, W. Chalee, U. Rattanasak, Comparative
study on the characteristics of fly ash and bottom ash geopolymers, Waste
Manage. 29 (2) (2009) 539-543.
[25] C. Lampris, R. Lupo, C.R. Cheeseman, Geopolymerisation of silt generated

U
from construction and demolition waste washing plants. Waste Manage. 29 (1)
(2009) 368-373.
AN
[26] J.H. Chen, J.S. Huang, Y.W. Chang, A preliminary study of reservoir sludge as a
raw material of inorganic polymers, Constr Build Mater 23 (10) (2009)
3264-3269.
[27 J.H. Chen, J.S. Huang, Y.W. Chang, Use of reservoir sludge as a partial
M
replacement of metakaolin in the production of geopolymers, Cem. Concr.
Comp. 33 (5) (2011) 602-610.
[28] K.H. Yang, C.W. Lo, J.S. Huang, Production and properties of foamed reservoir
D

sludge inorganic polymers, Cem. Concr. Comp. 38 (2013) 50-56.

[29] G. Habert, J.B. d’Espinose de Lacaillerie, N. Roussel, An environmental
TE

evaluation of geopolymer based concrete production: reviewing current

research trends, J. Clean. Prod. 19 (11) (2011) 1229-1238.
[30] R. Dron, F. Brivot, Thermodynamic and kinetic approach to the alkali-silica
reaction. Part 1: concept, Cem. Concr. Res. 22 (5) (1992) 941-948.
EP

[31] C.J. Ke, P. Gong, Study on autoclaved performance of waste glass, J. Yangtze
Univ Nat Sci 3 (2) (2006) 95-99.
[32] Jin W. Inorganic binders employing waste glass. U. S. Patent 6296699 B1,
2001.
C

[33] Z.P. Jing, The alkali activated glass, Housing Materials & Applications 31 (5)
(2003) 19-20.
AC

[34] A. Fernandez-Jimenez, M. Monzo, M. Vicent, A. Barba, A. Palomo, Alkaline

activation of metakaolin–fly ash mixtures: Obtain of Zeoceramics and
Zeocements, Microporous Mesoporous Mater. 108 (1-3) (2008) 41-49.
[35] J.G.S. van Jaarsveld, J.S.J. van Deventer, Effect of the Alkali Metal Activator on
the Properties of Fly Ash-Based Geopolymers, Ind. Eng. Chem. Res. 38 (10)
(1999) 3932-3941.
[36] J.C. Swanepoel, C.A. Strydom, Utilisation of fly ash in a geopolymer material,
Appl. Geochem. 17(8) (2002) 1143-1148.

21
Table 1. Measured 4-day compressiveACCEPTED
strengths ofMANUSCRIPT
AAGIB specimens alkali-activated under the aging
process at various temperatures and for various durations with AE%=1-3%.
AE%=1% AE%=2% AE%=3%
Aging Aging Compressive Aging Compressive Aging Compressive
Temperature Duration Strength Duration Strength Duration Strength
( ℃) (min) (MPa) (min) (MPa) (min) (MPa)
15 10.95 15 25.10 15 31.93
<360> 30.46 <240> 62.25 <120> 43.43
960 71.00 360 51.61 360 127.91
29

PT
1080 66.60 {540} 95.97 {480} 147.20
{1200} 71.90 600 56.80 600 119.60
1440 59.73 [693] 34.75 [691] 102.22

RI
15 11.16 15 24.94 30 27.93
<60> 22.67 <60> 29.90 <60> 45.23

SC
{420} 61.48 360 86.49 300 125.18
40
480 30.91 {420} 103.04 360 112.88
540 30.58 480 85.57 {420} 129.35

U
[720] 15.32 [554] 35.96 [516] 59.41
30 10.25 15 24.30 <15> 41.28
<60>
240
25.58
46.88
AN<60>
150
38.89
87.17
30
120
55.01
113.67
50
270 54.44 180 87.77 150 120.35
M
{300} 73.51 {210} 94.59 {180} 147.07
[375] 37.79 [300] 24.98 [263] 54.95
D

15 10.13 15 33.24 <15> 48.80

<45> 33.99 <30> 45.18 30 60.15
TE

{165} 64.56 60 89.10 45 131.77

60
225 63.35 75 91.22 {75} 135.98
240 62.31 {105} 102.47 90 134.84
EP

[270] 39.89 [190] 57.26 [125] 68.85

<5> 15.84 <5> 38.89 <5> 56.42
10 18.37 10 42.73 10 72.74
C

40 13.85 40 72.84 35 117.20

70
AC

70 31.87 45 72.18 {40} 124.51

{105} 58.22 {55} 82.29 45 111.68
[120] 31.20 [85] 60.30 [55] 69.44
Ultimate Compressive
Strength under Aging
65.93 95.67 136.82
Process with Optimal
Aging Duration (MPa)

22
 ACCEPTED MANUSCRIPT
Table 2. Measured 4-day compressive strengths of AAGIB specimens alkali-activated under the aging
process at various temperatures and for various durations with AE%=4-8%.

AE%=4% AE%=5% AE%=6% AE%=8%

Aging Aging Compressive Aging Compressive Aging Compressive Aging Compressive
Temperature Duration Strength Duration Strength Duration Strength Duration Strength
(℃) (min) (MPa) (min) (MPa) (min) (MPa) (min) (MPa)
15 50.07 15 66.81 15 53.28 15 47.00
<120> 106.95 <120> 74.18 <120> 66.10 <120> 49.11

PT
360 104.11 240 91.77 240 89.18 240 53.47
29
{480} 111.01 {360} 92.65 {360} 101.20 {360} 65.01
600 108.60 480 89.09 480 66.74 480 55.05

RI
[650] 65.30 [628] 71.99 [577] 21.60 [485] 39.28
<15> 62.29 <15> 81.12 <15> 72.15 <15> 56.14

SC
60 60.33 60 80.14 60 59.99 60 41.28
180 110.89 {300} 102.30 {180} 79.96 {120} 59.41
40
360 112.68 360 95.99 240 76.21 180 59.25

U
{420} 118.00 420 96.62 360 69.39 240 52.71
[505]
<15>
84.21
61.64
[500]
<15>
AN83.10
74.37
[443]
<15>
65.07
66.59
[360]
<15>
44.98
47.86
90 73.80 30 93.81 30 67.12 30 53.21
M
150 109.23 90 71.46 60 64.54 60 59.68
50
{180} 117.66 {180} 100.81 90 66.76 90 48.30
240 114.70 240 97.35 {120} 72.96 {120} 64.92
D

[252] 65.98 [252] 53.41 [180] 26.12 [150] 56.77

TE

<15> 86.78 <15> 81.21 <15> 71.14 <15> 53.09

30 102.49 60 75.49 30 87.10 30 59.92
45 97.35 {75} 99.69 45 75.11 45 56.45
60
EP

{75} 119.99 90 85.59 {60} 92.72 {60} 62.83

90 106.89 105 84.86 75 75.82 75 62.17
[120] 55.81 [120] 80.55 [105] 35.93 [90] 52.83
C

<5> 100.50 <5> 96.51 <5> 84.69 <5> 47.47

25 96.76 10 96.01 10 72.03 15 55.91
AC

{35} 117.01 15 91.52 20 90.66 {20} 76.90

70
40 112.82 20 96.42 {25} 99.66 25 74.30
45 105.57 {35} 97.81 30 77.94 30 64.76
[55] 62.44 [50] 73.53 [50] 48.75 [40] 42.73
Ultimate Compressive
Strength under Aging
116.73 98.65 89.30 65.81
Process with Optimal
Aging Duration (MPa)
23
 ACCEPTED MANUSCRIPT

3.5 100
Volume Percentage (%)
Cumulative Volume Percentage (%)
3.0

Cumulative Volume Percentage (%)

80

2.5

Volume Percentage (%) 2.0

60

PT
1.5
40

1.0

RI
20
0.5

SC
0.0 0
0.1 1 10 100 1000

Particle Size (µm)

U
Fig. 1. Particle size distribution of glass powders obtained from a laser diffraction
particle size analyzer.
AN
M
D
TE
C EP
AC

Fig. 2. Set-up of apparatus used for the aging process employed in the production
of the alkali-activated glass inorganic binder specimens.

24
 ACCEPTED MANUSCRIPT

160
Ms=0
Ms=0.5
140 Ms=1.0
Ms=1.5
120 Ms=2.0

Compressive Strength (MPa) 100

PT
80

60

RI
40

20

SC
0
0 2 4 6 8

AE%

U
(a)
AN
160
AE%=1%
AE%=2%
M
140 AE%=3%
AE%=4%
AE%=5%
120
Compression Strength (MPa)

AE%=6%
D

AE%=8%
100
TE

80

60

40
EP

20

0
C

0
0.0 0.5 1.0 1.5 2.0

Ms
AC

(b)
Fig. 3. Variations of compressive strength of AAGIBs, produced under a normal
process by mixing glass powders with various activators with a W / B
ratio of 0.3 and various AE % and M s values and stirring for 30 min at
room temperature, with (a) AE % and (b) M s , respectively.

25
 ACCEPTED MANUSCRIPT

80

70

60

Compressive Strength (MPa)

PT
50

40

RI
o
30 29 C
o
40 C
20 o
50 C

SC
o
10 60 C
o
70 C
0
0 200 400 600 800 1000 1200 1400 1600

U
Aging Duration (min)

Fig. 4. Variations of 4-day compressive strength of AAGIBs with aging duration at

various aging temperatures with AE % =1%.
AN
M
120
o
29 C
o
100 40 C
D

o
50 C
Compressive Strength (MPa)

o
80 60 C
TE

o
70 C

60
EP

40

20
C

0
0 200 400 600 800
AC

Aging Duration (min)

Fig. 5. Variations of 4-day compressive strength of AAGIBs with aging duration at

various aging temperatures with AE % =2%.

26
 ACCEPTED MANUSCRIPT

160

140

PT
Compressive Strength (MPa) 120

100

80

RI
o
60 29 C
o
40 C

SC
40 o
50 C
o
20 60 C
o
70 C
0

U
0 200 400 600 800

Aging Duration (min)

AN
Fig. 6. Variations of 4-day compressive strength of AAGIBs with aging duration at
various aging temperatures with AE % =3%.
M

140
D

120
TE
Compressive Strength (MPa)

100

80
EP

60
o
29 C
o
40 40 C
C

o
50 C
20 o
60 C
AC

o
70 C
0
0 200 400 600 800

Aging Duration (min)

Fig. 7. Variations of 4-day compressive strength of AAGIBs with aging duration at

various aging temperatures with AE % =4%.

27
 ACCEPTED MANUSCRIPT

120

100

PT
Compressive Strength (MPa)
80

60

RI
o
40 29 C
o
40 C

SC
o
50 C
20
o
60 C
o
70 C
0

U
0 200 400 600 800

Aging Duration (min)

AN
Fig. 8. Variations of 4-day compressive strength of AAGIBs with aging duration at
various aging temperatures with AE % =5%.
M
120
o
29 C
D

o
100 40 C
o
50 C
Compressive Strength (MPa)

TE

o
60 C
80
o
70 C

60
EP

40
C

20
AC

0
0 200 400 600 800

Aging Dutation (min)

Fig. 9. Variations of 4-day compressive strength of AAGIBs with aging duration at

various aging temperatures with AE % =6%.

28
 ACCEPTED MANUSCRIPT

100
o
29 C
o
40 C
80 o
50 C

PT
Compressive Strength (MPa)
o
60 C
o
60 70 C

RI
40

SC
20

U
0 100 200 300 400 500 600

Aging Duration (min)

AN
Fig. 10. Variations of 4-day compressive strength of AAGIBs with aging duration at
various aging temperatures with AE % =8%.
M
D

160
TE

140

120
Compressive Strength (MPa)

100
EP

80

60
C

40
AC

Ultimate compressive strength under

20 aging process with optimal aging
Basic compressive strength under normal process
Maximum compressive strength under normal process
0
0 2 4 6 8 10

AE%

Fig. 11. Variations of basic, maximum and ultimate compressive strengths of

AAGIB specimens with AE % under normal and aging processes.

29
 ACCEPTED MANUSCRIPT

100
1025 AE%=1%
AE%=8%
Glass powder
80

PT
Absorptance(%)

60
778 3429
1660

RI
40

1560

SC
20

U
1000 2000 3000 4000

Wavenumbers(cm-1 )

Fig. 12.
AN
FTIR spectra of the glass powder and the AAGIB specimens with
AE % =1% and 8%.
M
D

15.0
TE

14.5

14.0
EP

13.5
pH

13.0
C

12.5
AC

Ultimate compressive strength, aging process

12.0 Activator, aging process
Maximum compressive strength, normal process
Activator, normal process
11.5
0 2 4 6 8

AE%

Fig. 13. Variation of the pH values in the activators used and the AAGIB slurry
obtained under normal and aging processes.

30