Cement and Concrete Composites 72 (2016) 168e179

Contents lists available at ScienceDirect

Cement and Concrete Composites

journal homepage: www.elsevier.com/locate/cemconcomp

Inuence of the slag content on the chloride and sulfuric acid

resistances of alkali-activated y ash/slag paste
N.K. Lee a, H.K. Lee b, *
a
Structural Engineering Research Division, Korea Institute of Civil Engineering and Building Technology, 283 Goyangdae-ro, Ilsanseo-gu, Goyang 102-23,
South Korea
b
Department of Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology, Guseong-dong, Yuseong-gu, Daejeon 305-701,
South Korea

a r t i c l e i n f o a b s t r a c t

Article history: This study aims to investigate the inuence of the slag content on the chloride and sulfuric acid re-
Received 29 July 2015 sistances of alkali-activated y ash/slag (AFS) paste. A series of tests were conducted to examine the
Received in revised form effects of reaction products and their contents on the chloride and sulfuric acid resistance capabilities. It
2 March 2016
was shown from the tests that the deterioration of the AFS binder due to a sulfuric acid attack was caused
Accepted 7 June 2016
Available online 9 June 2016
by 1) the corrosion by means of SO2
4 penetration through the surface of the AFS binder, which is
associated with permeable voids and a rate of water absorption and 2) the corrosion of the reaction
products resulting from the different degrees of resistance to sulfuric attack between the C-(A)-S-H and
Keywords:
Alkali-activation
the N-A-S-H. Variation of the slag content led to differences in the reaction product content of the AFS
Chloride resistance binder, clearly affecting the chloride-binding capacity and the resistance to chloride penetration.
Sulfuric acid resistance 2016 Elsevier Ltd. All rights reserved.
Reaction products
Fly ash
Slag

1. Introduction (AAS) was reported to have better durability compared to ordinary

Alkali-activated binders have been investigated by many re-
searchers over the past decade. In particular, alkali-activated, two-
source binders, which consist of y ash/slag or metakaolin/slag,
have been studied recently in an effort to improve the performance
of alkali-activated, one-source binder. There are two main types of
reaction products, N-A-S-H and C-A-S-H, in alkali-activated, two-
source binders [1,34]. The main product of alkali-activated binders
with a y ash content of 50 wt% and a slag content of 50 wt% was a
N-C-A-S-H gel with a chemical composition intermediate between
C-A-S-H and N-A-S-H [1]. Since the nature of the gel in alkali-
activated, two-source binders is highly dependent on the y ash/
slag ratio, the mechanical properties and the durability of two-
source binders should be investigated in view of the mixture ra-
tio of the raw materials.
Acid resistance is one of the important performance parameters
for structural materials when they are applied in aggressive envi-
ronments such as urban sewer structures [2]. Alkali-activated slag
* Corresponding author.
E-mail address: leeh@kaist.ac.kr (H.K. Lee).
Portland cement (OPC), with one study nding that AAS concrete
showed a strength reduction of about 33% as compared to the 47%
reduction of OPC in an acid environment (pH 4) after a time of one
year [2]. The high resistance of AAS to an acid attack was attributed
to the low Ca content (40% CaO) of the slag compared to Portland
cement (~65% CaO) and to the presence of glassy slag, which is
practically insoluble in an acid solution [2,3]. Likewise, alkali-
activated y ash binder also showed better performance in some
aggressive environments compared to OPC concretes [4,5], result-
ing from a more stable cross-linked aluminosilicate gel which
formed in the binder [2]. Allahverdi et al. [6] suggested the corro-
sion mechanism of geopolymer paste at a relatively high concen-
tration of sulfuric acid (pH 1); the rst step is an ion exchange
reaction between the charge compensating cations of the frame-
work, and the second step is a reaction between the exchanged
calcium and the sulfate anion, resulting in the formation of gypsum
crystals inside the corroding layer. In alkali-activated, two-source
binders, it was found that the binders underwent surface corrosion
due to an acid attack [7]. The presence of calcium in the binder and
a high alkali concentration resulted in high resistance to an acid
attack due to the decreased mass transport rate through the
tortuous pore structures of the binders [7].

http://dx.doi.org/10.1016/j.cemconcomp.2016.06.004
0958-9465/ 2016 Elsevier Ltd. All rights reserved.
 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179
Alkali-activated y ash binders may be a favorable alternative in
the manufacturing of acid resistant concrete since it consists mostly
of aluminosilicate rather than calcium silicate hydrate [8]. On the
other hand, the resistance of an alkali-activated, two-source binder
to an acid attack may be more complicated as the reaction products
in the binders are C-A-S-H and N-A-S-H and possibly the hybrid C-
N-A-S-H [1]. Very few studies of the acid resistance of alkali-
activated, two source binders have been conducted thus far, and
the relationship between the reaction products and the acid
resistance capability still needs to be elucidated.
Chloride resistance is another important factor with regard to
the durability of concrete. Although chlorides are not signicantly
harmful to concrete, chlorides may cause corrosion of an embedded
reinforcement through a depassivation process, leading to serious
damage to the concrete [9,10]. Ismail et al. [10] found in their study
of the chloride penetration of alkali-activated slag/y ash binder
(y ash contents at 0%, 25%, 50%, and 75% of the total binder) that an
addition of y ash in slag binder caused an increase in the porosity
and chloride permeability due to the formation of more porous
sodium aluminosilicate (NeAeSeH) type gels. Their study showed
better chloride resistance in alkali-activated slag dominated by
CeAeSeH gels compared to alkali-activated y ash [10]. The
addition of metakaolin to the slag binder also led to a reduction in
the chloride permeability compared to an alkali-activated slag
binder, showing that the addition of metakaolin as a secondary
aluminosilicate precursor contributes to the renement of the pore
network and accordingly a reduction in the absorptivity [11].
Studies of chloride resistance in alkali-activated, two-source
binders are very limited, and the relationship between the reaction
products and the chloride resistance capability needs to be eluci-
dated. This study is thus to investigate the inuence of the slag
content on the chloride and acid resistances of the AFS paste. A
series of tests related to the chloride and sulfuric acid resistance
capabilities were conducted and the effects of reaction products
and their contents on the chloride and acid resistance capabilities
were discussed. Chloride penetration depth and sulfuric acid-
resistant reaction product were quantitatively evaluated through
a steady-state chloride penetration test and NMR analysis,
respectively.
2. Materials and mix proportion
In this study, Class F y ash and blast furnace slag (BFS) were
used as binder materials. For comparison, ordinary Portland
cement (OPC) samples were manufactured. The Class F y ash
(classied according to ASTM C 618) and blast furnace slag were
obtained from a power plant and from a steel plant both located in
South Korea, respectively. Their chemical compositions are listed in
Table 1.
A sodium silicate liquid solution (SiO2/Na2O 1.0) was prepared
as an alkali-activator. It was made by mixing a NaOH solution with a
Table 1
Chemical composition of the binder materials.
Oxide (wt. %) Fly ash (FA) Blast furnace slag (BFS)
CaO 4.41 42.47
SiO2 67.26 35.17
Al2O3 14.76 13.93
Fe2O3 4.07 0.58
SO3 e 2.03
MgO 1.29 4.12
Na2O 2.04 0.15
K2O 1.39 0.46
LOI 3.57 0.18
169
concentration of 4 mol/L and water glass (Korean industrial stan-
dards (KS) Grade-3; SiO2 (29%), Na2O (10%), H2O (61%), specic
gravity 1.38 g/mL) in accordance with the procedure developed by
Lee and Lee [33]. The NaOH solution with a concentration of 4 mol/L
was prepared by adding solid NaOH with a purity level of 98% to
distilled water.
The alkali-activated paste prepared was cast into 50 mm cubic
molds. All of the samples were cured at a temperature of 20  C and
at a relative humidity of 50% in a room with a constant temperature
and humidity. After one day, all of the samples were removed from
their molds and stored in the conditioning room until the day of
testing.
Sample names (S0, S10, S30, and S50) are denoted with specic
codes. The label S represent blast furnace slag, and the numbers,
0, 10, 30, and 50 indicate the percentage of slag amount in the
AFS binder by weight, respectively. For comparison, an OPC sample
was manufactured using ordinary Portland cement with a water to
cement ratio of 0.4.
3. Experimental details
3.1. Sulfuric acid test
The sulfuric acid test was conducted as follows. After 28 days of
curing, the samples were immersed in a 10% sulfuric acid solution
(H2SO4) for 28 days and for 56 days to investigate the resistance of
the AFS binder to an acid attack. The solution was refreshed once a
week.
X-ray diffraction (XRD), Fourier transform infrared (FT-IR)
spectroscopy, and 29Si MAS NMR spectrometry were utilized to
investigate the microstructures of the AFS samples immersed in the
sulfuric acid solution. Mass changes, water absorption rates, vol-
ume of permeable voids, and compressive strength were also
measured to investigate the physical changes of the AFS samples.
The powdered samples were taken from the middle of the AFS
samples by mechanical grinding. The XRD data were recorded on a
Rigaku D/MAX-2500 machine using Cu Ka radiation at a scanning
rate of 2 /min from 2 to 160 in the 2 mode. The powdered
samples were also analyzed by Fourier transform infrared (FT-IR)
spectroscopy (Model FT-IR 4100, JASCO, Japan).
A29Si MAS NMR spectrometer (Varian NMR system, WB500)
was used to compare the solid-state NMR spectra of the AFS sam-
ples immersed and not immersed in sulfuric acid solution. The 29Si
resonance frequency was 99.31 MHz and the spinning rate was
3 kHz. The spectra were acquired using a pulse length of 2.0 ms, and
a relaxation delay of 20s was chosen. Tetramethylsilane (TMS) was
used as reference for the 29Si NMR spectra. The reaction product
contents (N-A-S-H and C-A-S-H) of the sample immersed in the
sulfuric acid solution were evaluated through a NMR analysis. The
method used to determine N-A-S-H and C-A-S-H can be found in
Lee and Lee [12]. This was helpful to understand the resistance of
the reaction products against the acid attack.
The mass change was measured to investigate the degradation
of the samples resulting from the formation of a new phase after
sulfuric attack. The compressive strength was measured to
compare the difference between the samples immersed and not
immersed in sulfuric acid solution. Each value was measured three
times, and the average value of these three measurements was used
as the nal value.
3.2. Chloride penetration test
A steady-state chloride penetration test was conducted in
accordance with a previously published method [13], as follows.
After 28 days of curing, the side and top surfaces of 50-mm cubic
170 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179

samples were sealed with a polymer sealant to prevent unwanted

chloride penetration from the other surface except for the exposure
surface. Each sample was immersed in a 10 wt% sodium chloride
solution after 28 days of curing. The level of the solution was set to
be 10 mm from the bottom surface of the samples. After 28 and 91
days of immersion, the samples were cut using a diamond saw
perpendicular to the bottom surface and sprayed with a 0.1 N silver
nitrate solution to measure the chloride penetration depth. When
the white precipitation of silver chloride on the surface was clearly
visible, the penetration depths at seven spots were measured and
the average measured value was calculated.
After the chloride immersion test, X-ray diffraction (XRD) and
29
Si MAS NMR spectrometry analyses were conducted to investi-
gate the microstructures of the AFS samples. Water absorption
rates, volume of permeable voids, the acid- and water-soluble
chloride contents, and compressive strength were also measured
to investigate the physical changes of the AFS samples.
The acid- and water-soluble chloride measurements were ac-
cording to the ASTM C 1152 and 1218 standards, respectively. The
samples immersed in the sodium chloride solution for 28 days were
cut to a thickness of 10 mm and to a height of 30 mm from the
bottom surface along the vertical direction of the bottom surface
using a diamond saw. The sliced samples were crushed into a ne
powder with particles smaller than 100 mm by means of mechanical
grinding. The concentrations of the acid- and water-soluble chlo-
ride extracted from the powder were measured using residual
chlorine meter (DKK-TOA Corporation).
The water absorption rates and the volume of permeable voids
were assessed according to the ASTM C1585 and ASTM C 642
standards, respectively.

4. Sulfuric acid resistance of AFS paste

4.1. Reaction products identied by X-ray diffraction

Fig. 1 shows the XRD patterns of the AFS samples immersed in

the H2SO4 (sulfuric acid) solution for 28 days. The main peaks are
related to gypsum (PDF#00-033-0311), quartz (PDF#00-046-1045)
and mullite (PDF#00-015-0776). The formation of gypsum can be
possibly due to the reaction of calcium hydroxide or C-S-H with
sulfuric acid solution [35]. However, calcium hydroxide was not
detected in the AFS sample by XRD analysis, whereas C-(A)-S-H
phase was found in that sample [12]. Thus, the formation of gypsum
is most likely due to the decalcication of the C-(A)-S-H phase in
the sample as caused by the H2SO4 attack. The peak intensity of the
gypsum increased as the slag content increased. It was reported in Fig. 1. XRD patterns of AFS samples after acid treatment during 28 days.
the literature that the amount of the C-(A)-S-H phase increased as
the slag content increased in AFS binders [1]. Thus, the intensity
change of the gypsum peak supports the contention that the for- 4.2. Reaction products identied by FT-IR
mation of gypsum is due to the decalcication of the C-(A)-S-H
phase in the AFS paste. In addition, diffuse hump at 29e30 of the Fig. 6 shows the FT-IR spectra of AFS samples immersed and not
AFS paste, which is related to the presence of C-(A)-S-H phase or immersed in 10% sulfuric acid solutions. The absorption peaks in
calcite, as shown in previous research [14] were not observed due Fig. 2(a) clearly prove the presence of gypsum in the AFS samples
to the decalcication of the C-(A)-S-H phase. immersed in H2SO4 solution. The gypsum crystal showed O-H
The formation of gypsum and ettringite is generally the main stretching bands ranging from 3550 to 3400 cm
1 and bending
cause of the deterioration of OPC mortar subjected to sulfuric acid bands at 1684 and 1621 cm
1 [17-20]. In addition, the bands which
exposure. However, ettringite was not present in the AFS binders, appeared at 600, 665 and 1120e1133 cm
1 are associated with the
as shown in the XRD patterns, due to the low Ca/Si ratio of these stretching modes of SO4 [21]. These results are consistent with the
binders [15] and the low amount of Al available for the formation of results of the XRD patterns, in which the presence of gypsum was
ettringite [16]. Since the calcium hydroxide was not also detected detected.
here, the reaction between the C-(A)-S-H gel and the H2SO4 is the Fig. 2(b) presents the spectra between 800 and 1500 cm
1 of the
main cause of the formation of gypsum. AFS samples immersed in H2SO4 solution. It was noted that the S0
 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179 171

NS30
S50

Transmission (%)
Transmission (%)

S30
S30
S10
NS0

S0

1684 980 1090

600 700 1120-1133 1621 3243 3407 3532 1120 1133
0 1000 2000 3000 4000 5000 800 900 1000 1100 1200 1300 1400 1500
Wavenumbers (cm-1) Wavenumbers (cm-1)
(c) NS30 and NS0 samples not immersed in H2SO4 solution and S30 sample immersed in
(a) Overall spectra of AFS samples immersed in H2SO4 solution
H2SO4 solution

S50 Fly ash

Transmission (%)
Transmission (%)

S30

S10
S0
S0

1120 1133
800 900 1000 1100 1200 1300 1400 1500 800 900 1000 1100 1200 1300 1400 1500
Wavenumbers (cm-1) Wavenumbers (cm-1)

(b) Spectra between 800 and 1500 cm-1 of AFS samples immersed in H2SO4 solution (d) S0 sample immersed in H2SO4 and raw fly ash

Fig. 2. FT-IR spectra of AFS samples immersed and not immersed in H2SO4.

sample showed the strongest band at 1090 cm
1, unlike the other 4.3. Measured mass change of AFS paste
samples. The bands related to gypsum were not present in the S0
sample since the sample contains a very low amount of calcium. Fig. 3 shows the rate of the mass change of the AFS and OPC
However, the S0 sample also showed a considerable reduction in its samples immersed in the sulfuric acid solution. Except for the OPC
residual strength due to the degradation of the binder despite the and S50 samples, the weights of the S0, S10 and S30 samples were
fact that the formation of gypsum stemming from the decalcica- nearly constant. As shown in Fig. 4, delamination of the OPC sample
tion of C-(A)-S-H was not possible in the S0 sample. This indicates occurred due to the degradation of the sample, reducing its mass,
that the N-A-S-H gel present in the S0 sample showed weak while the mass of the AFS sample increased slightly or remained
resistance to an acid attack due to its porous characteristics, as mostly unchanged. While the mass of the AFS sample was nearly
noted in the aforementioned previous study [1], or that the S0 constant over time, a serious surface crack was observed in the S30
sample was very porous due to very low level of alkali-reaction at and S50 samples, as shown in Fig. 4. The formation of gypsum in the
ambient temperature. Moreover, the raw y ash and the S0 sample AFS sample, as detected in the XRD patterns, caused an expansion
showed very similar band absorption characteristics, as shown in and dimensional instability in the AFS sample [23]. The cracks may
Fig. 2(d), indicating that microstructural changes related to the have occurred due to the dimensional instability of the binder.
asymmetric stretching of the Si-O-Si (Al) bonds did not arise in the
S0 sample. It was reported that a higher degree of the substitution
of tetrahedral Al for Si in a silicate network results in a shift of the
main Si-O-Si (Al) asymmetric stretching band toward lower
wavenumbers [22]. However, the band at 1100 cm
1 in the S0 40
sample did not shift to a lower peak after alkali-reaction, as the 35
Al2O3 content in raw y ash was relatively low (14.76%). 30
Mass change ratio (%)

Fig. 2(c) compares the spectra of the S30 samples immersed and
25 C
not immersed in sulfuric acid solution. There was an absorption OPC
20 S0
band at 980 cm
1 for the S30 sample not immersed in sulfuric acid
15 S10
solution, while there was an absorption band between 1120 and
S30
1133 cm
1 for the S30 sample immersed in sulfuric acid solution. 10
S50
This difference is evidence of the formation of gypsum caused by 5
the decalcication of the C-(A)-S-H gel. It was also found from the 0
spectra that aluminosilicate gel existed in the binder even after the -5
immersion in the H2SO4 solution, which supports the contention
-10
that C-(A)-S-H gel is vulnerable to an acid attack while alumino- 0 7 14 21 28 35 42 49 56 63
silicate gel is less susceptible to such an attack. Time (days)

Fig. 3. Mass change ratio of AFS and OPC samples immersed in sulfuric acid: the plus
and minus values mean the mass loss and gain, respectively.
172 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179

Fig. 4. AFS and OPC samples after immersion in sulfuric acid solution for 56 days.

4.4. Measured compressive strength of AFS paste
Fig. 5 shows the compressive strength of the AFS and OPC
samples immersed in the H2SO4 solution during 28 and 56 days.
Before the immersion in the H2SO4 solution, the S50 sample with a
slag content of 50% showed the highest compressive strength. The
compressive strength was decreased by reducing the slag content
in the mixture. As the sample age stretched from 28 days to 56 days,
the compressive strengths of all samples increased. Specically, the
S30 sample showed the highest rate of increase in the compressive
strength from 28 days to 56 days, even higher than that of the S50
sample. After the immersion in the H2SO4 solution, the compressive
strength of the AFS sample was greatly reduced. The S30 sample
with a slag content of 30% showed the highest rate of decrease in
the compressive strength, while the S50 sample showed greater
residual strength than the OPC sample. The AFS paste showed
improved resistance to sulfuric acid attack than the OPC paste. This
was attributed to the formation of calcium sulfate by the reaction of
sulfate ion with calcium being inhibited due to low content of
calcium hydroxide in the AFS systems, and the denser micro-
structure of the AFS paste compared with that of the OPC paste. In
addition, the decalcication of N-A-S-H gel by the H2SO4 ion attack
did not occur in the AFS paste. This gel contributed to the formation
of more stable structure than C-S-H gel even after the sulfuric acid
 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179 173

45
40 28 days

Permeable void (%)

35 91 days

56 30
25
56
20
15
10
5
0
S0 S10 S30 S50 OPC
C
Samples
Fig. 5. Compressive strength of AFS and OPC samples not immersed and immersed in Fig. 7. Permeable void (%) of AFS samples at the age of 28 and 56 days.
H2SO4 solution during 28 days and 56 days.

was possibly N-C-(A)-S-H gel with a certain amount of calcium

attack.
incorporated in the structure, which is consistent with the previous
Fig. 6 shows the residual compressive strength ratio which is
result [1]. It seems that the resistance of this gel against an acid
dened as the ratio of compressive strength after the immersion in
attack is lower than that of the C-(A)-S-H gel which is a main
the H2SO4 solution to the compressive strength before the im-
product in OPC or alkali-activated slag.
mersion in the H2SO4 solution. The residual strength ratio was the
Although the S10 and S0 samples not immersed in the H2SO4
highest for the S10 sample, followed by the S50, OPC, S0, and S30
solution had much lower compressive strength than the OPC
samples. The S50 and S30 samples showed higher and similar re-
sample, the residual strength ratios for the S10 and S0 samples
sidual compressive strengths compared to the OPC sample,
immersed in the H2SO4 solution were higher than that of OPC, as
respectively. In the literature, it was reported that C-(A)-S-H gel is
shown in Figs. 5 and 6. This indicates that a substantial amount of
vulnerable in acid environments, while aluminosilicate gel is less
N-A-S-H gel, which formed in the S10 and S0 samples and which
susceptible due to the stability of its cross-linked aluminosilicate
was less susceptible to an acid attack than the C-(A)-S-H gel,
structure [2,24]. By increasing the slag content in the mixture, the
resulted in higher resistance of the binder to an acid attack, despite
AFS sample is expected to be more vulnerable to acid environ-
the fact that the contribution of the N-A-S-H gel to the compressive
ments, since the C-(A)-S-H gel content increased and the alumi-
strength was lower than that of C-(A)-S-H.
nosilicate gel content decreased as found in Lee and Lee [12].
The S10 and S50 samples showed higher residual strengths than
However, as the slag content increased, the pore volume became
the OPC and S30 samples, as shown in Fig. 6. However, the causes of
low and the matrix became dense, as shown Fig. 7. As a conse-
this may differ between the S10 sample and the S50 sample. The
quence, the resistance of the binder to an acid attack was improved,
fewer permeable voids and the denser matrix are the main causes
resulting in a low reduction in the compressive strength. This is
in the S50 sample, while the inherent high resistance of the N-A-S-
consistent with results which showed that alkali-activated slag is
H gel to an acid attack is the main cause in the S10 sample, despite
less porous than alkali-activated y ash [25]. Although the S30
the fact that the S10 sample showed a high number of permeable
sample had slightly fewer permeable voids (23.76%) at the age of 28
voids and a high rate of water absorption, as shown in Figs. 7 and
days compared to the OPC sample (28.4%), the residual strength
10, respectively. On the other hand, the S30 sample showed a
ratio of the S30 sample (14.39%) was much lower than that of the
higher volume of permeable voids and a higher rate of water ab-
OPC (21.89%). Thus, it can be said that the gels formed in the S30
sorption compared to the S50 sample. Moreover, the amount of N-
sample was less resistant to an acid attack. The gels appeared to be
A-S-H gel was less and that of N-C-A-S-H gel (which has lower
different from the N-A-S-H gel [12] which was found to be resistant
resistance) was more than those in the S10 sample, thus resulting in
to acid attack in a previous study [24]. The gel in the S30 sample
a lower residual strength than that of the OPC sample. Accordingly,
the resistance capability of the S30 sample to an acid attack was
lower than that of the S10 and S50 samples.
0.7
Rate of residual compressive strength

Immersion during 28 days

0.6 Immersion during 91 days 4.5. Inuence of reaction products on sulfuric acid resistance
0.5
The presence of C-(A)-S-H and N-C-A-S-H with certain amounts
0.4
of Ca in the AFS binders was reported in a previous study conducted
0.3

0.2 Table 2
Mix proportions of alkali-activated y ash/slag paste.
0.1 Sample AAa/(Fly ash slag)b SiO2/Na2O ratio of AA Slag/(y ash slag)b

0 S0 0.4 1.0 0.0

S0 S10 S30 S50 C
OPC S10 0.4 1.0 0.1
S30 0.4 1.0 0.3
Samples
S50 0.4 1.0 0.5
a
Fig. 6. Ratio of residual compressive strength of AFS and OPC samples immersed in AA indicates an alkali activator.
b
H2SO4 solution. All values are given as mass ratios.
174 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179

Table 3
Relative area (%) obtained from deconvolution results of29Si NMR spectra.

Acid treatment Sample Reaction products derived from Reaction products derived from y ash
slag

Signal (ppm)
77, etc.a Q1 Q2(1Al) Q2 Q4(4Al) Q4(3Al) Q4(2Al) Q4(1Al) Q4(0Al)

79
81
86
89
94
98
104
109, etc.b

Beforec AS10 0.40 0.35 0.53 2.55 4.81 3.77 3.5 9.86 55.6
After 0 0 0 3 0 4 0 30 63
Beforec AS30 1.15 4.27 12.89 18.55 2.66 7.21 3.41 6.53 34.24
After 0 0 0 0 0 17 0 44 39
Beforec AS50 14.58 10.65 22.98 17.64 0.96 0.64 1.48 8.10 12.76
After 0 0 5 9 0 5 0 24 57
a
The peaks at
77,
75,
72,
70,
68,
65 ppm are attributed to reaction of slag in AFS sample.
b
These are the peaks above
109 ppm. Most of the peaks appearing above
108 ppm were assigned to different crystalline phases of silica (Q4(0Al) signals) (Engelhardt and
Michel, 1987).
c
Reproduced from Lee and Lee (2015) for comparison.

by the authors [12]. In the previous study, the amount of these

phases and the degree of reaction were quantitatively evaluated
through 29Si NMR spectroscopic investigation [12]. Based on the
results, the relationship between these phases and the resistance of
the AFS binder to sulfuric acid attack were investigated in the
present study. The amounts of the N-C-A-S-H and N-A-S-H gels of
AFS samples, which were not immersed in H2SO4 solution, with
different amounts of slag can be found in Lee and Lee [12].
The 29Si NMR spectra were analyzed and listed in Table 3. The
detailed procedure for peak analysis was explained in the previous
work [12]. With a slag content of 10%, the amount of C-(A)-S-H with
Q1 and Q2 sites was little, while the amount of N-A-S-H with Q4
sites, which is known to be relatively more resistant to an acid
attack, was high, resulting in a high residual strength. With a slag
content of 30%, the peaks assigned to the Q1 and Q2 sites of C-(A)-S-
H nearly disappeared after the immersion in the H2SO4 solution, as
shown in Fig. 8(b), while the peaks assigned to the Q4 (4Al) and Q4
(1Al) sites of N-A-S-H were still present. The S30 sample was more
easily affected by the sulfuric attack due to its higher void volume
compared to the S50 sample, resulting in the decalcication of the
C-(A)-S-H gel and a signicant reduction in the compressive
strength of the binder.
In Fig. 8, there are clear differences between the peaks for the
samples immersed and not immersed in H2SO4 solution. In
particular, it was noted that a new broad band appeared at

95 ppm to
110 ppm. In Table 2, after the immersion in the H2SO4
solution, the increase in the intensity assigned to Q4 (1Al) may be
due to the combination of alumina and silica gels which remained
after the decalcication of the C-(A)-S-H gel. With a slag content of
50%, although little N-C-A-S-H gel was formed, the S50 sample was
less affected by the sulfuric attack due to its dense matrix and low
pore volume, leading to a high residual strength, as shown in Fig 5.
Consequently, there are two causes of the deterioration of the
AFS binder due to a sulfuric acid attack. The rst is corrosion by
means of SO2
4 penetration through surface of the AFS binder,
which is associated with permeable voids and a rate of water ab-
sorption. An attack of sulfuric acid ion can take place easily in a
more porous structure. As the slag amount increases, the C-(A)-S-H
gel increases, resulting in a lower permeable void fraction and a
denser structure. Accordingly, it makes an inltration of sulfuric
acid ion into the AFS paste difcult, which delays the disintegration
of N-A-S-H and C-(A)-S-H gels.
The second is corrosion of the reaction products present in the
AFS binder, resulting from the difference in the degree of resistance
to the sulfuric acid attack between the C-(A)-S-H gel and the N-A-S-
H gel. When sulfuric acid ion penetrates into the AFS paste, it reacts
with C-(A)-S-H gel, resulting in the decalcication of the gels and
the formation of calcium sulfate, as expressed in the following
Fig. 8. 29Si NMR spectra of AFS samples immersed (S10, S30, S50) and not immersed
equation [35]: (NS10, NS30, NS50) in H2SO4 solution.
 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179
CaO$SiO2$2H2O H2SO4 / CaSO4 Si(OH)4 H2O
Due to the decalcication, the C-(A)-S-H gel was disintegrated,
and the compressive strength of AFS paste was decreased. In case of
N-A-S-H gel, when sulfuric acid ion penetrates into the AFS paste,
the aluminum in N-A-S-H gel framework was exchanged for H or
H3O from a sulfuric acid solution [36]. After the dealumination,
the gel framework was transformed into an imperfect silica gel
structure [36]. The dealumination phenomenon by the acid attack
is supported by the results in Table 3; after the sulfuric acid attack,
Q4(4Al) site almost disappeared while Q4 sites with a lower level of
Al substitution increased in all of the samples. The aluminum
leached from the N-A-S-H gel reacted with amorphous silicate gels
Fig. 9. XRD patterns of AFS samples exposed to 10% NaCl solution during 91 days.
175
derived from the decalcication of C-S-H gel, resulting in the for-
mation of Q4(1Al) site. However, unlike the C-S-H gel, the severe
erosion of AFS samples was not observed, as shown in Fig. 4. Hence,
it appears that the N-A-S-H gels were only partially dealuminated,
not completely disintegrated, showing a clear distinction from the
complete decalcication of C-(A)-S-H gel. For this reason, S10
sample showed higher residual strength than S30 and S50 samples.
It should be noted that among the AFS samples, the residual
strength of S30 sample was lowest and was similar to that of OPC.
This indicates that S30 sample was most severely affected by the
aforementioned two causes of corrosion. In particular, it appears
that N-C-A-S-H gel as a main reaction product of S30 sample was
weaker against sulfuric acid attack than N-A-S-H gel, as this gel
underwent both decalcication and dealumination phenomenon.
However, given that the S30 sample showed the highest
compressive strength among all of the samples, as shown in Fig. 5,
the resistance of the S30 binder to a sulfuric attack can be improved
by means of an addition of a mineral admixture such as silica fume,
which leads to densication of matrix and an increase in the
amount of N-A-S-H gel.
5. Chloride resistance of AFS paste
5.1. Reaction products identied by X-ray diffraction
Fig. 9 shows XRD patterns of the AFS samples immersed in 10%
NaCl solutions for 91 days. There were no differences between the
AFS samples immersed and not immersed in these solutions,
indicating that the C-(A)-S-H gel physically adsorbs chloride ions
Fig. 10. Rate of water absorption of AFS samples.
176 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179
without a change of the mineralogical structure [10]. However, it
was noted that the peaks at 27.3, 31.7, 45.5, and 56.5 in the 2q
mode associated with halite (PDF # 01-072-1668) existed in the S0
sample, while these peaks were not detected in the S10, S30, and
S50 samples. Ismail et al. [10] stated that the presence of halite in
the S50 sample with a slag content of 50% is due to the precipitation
of pore-solution salt after the drying of powdered samples; the
presence of a sufcient amount of this pore solution stems from the
more highly porous structure in these binders, where N-A-S-H gels
and C-(A)-S-H gels exist, in comparison with the samples with a
slag content of 75% or 100%. However, in the present study, peaks
related to halite were not detected when a slag content of 50% was
used, which is not consistent with that earlier study [10]. These
results signify that most of the sodium in the binder participated in
the formation of the C-N-A-S-H or N-A-S-H gels.
5.2. Measured water absorption and permeable voids of AFS paste
Fig. 10 presents the water absorption rates of the AFS samples
after 28 and 56 days. The OPC sample showed a water absorption
rate very similar to that of the S30 sample after the age of 28 days.
The rate of water absorption changed over time; the AFS binder
showed a steep increase in the water absorption rate after the age
of 28 days compared to the OPC sample. Lee et al. [14] showed that
the AFS sample had a higher pore volume in pores greater than
200 nm than the OPC sample, while the OPC samples had a higher
pore volume in pores between 20 nm and 200 nm than the AFS
sample. In pores less than 30 nm, the AFS sample was much higher
than the OPC sample. There appeared to be a correlation between
the pore size distribution and the rate of water absorption. In
Fig. 10, there are three stages of the rate of water absorption. First,
water was mainly absorbed through macropores which were
greater than 200 nm in size, followed by mesopores between 30 nm
and 200 nm. Finally, water absorption occurred through nanopores
less than 30 nm in size. In fact, the occurrence of this phenomenon
was more complex, and water absorption through the pores can
occur even at different stages. In the beginning, the rate of water
absorption in the AFS sample was high due to water absorption by
pores greater than 200 nm. After that, the rate of increase was
drastically reduced. In contrast, the increase in the rate of water
absorption in the OPC sample was relatively low in the beginning,
as there were fewer macropores larger than 200 nm compared to
the AFS sample. However, the rate of water absorption of the OPC
sample continually increased after water absorption started.
The tendency shown at the age of 56 days is quite different from
that at the age of 28 days. In Fig. 10(b), the S0 sample shows the
lowest rate of water absorption. This is in contrast to the results of
the S0 sample, which had the highest volume of permeable voids,
as shown in Fig. 7. This may be a result of the hydrophobicity of the
raw y ash used here [26,27]. The hydrophobicity of the S0 sample
was stronger at 56 days rather than at 28 days, resulting in a low
rate of water absorption at the age of 56 days. This will be inves-
tigated in detail later. As the slag content increased, the volume of
permeable voids decreased. It is likely that an increase in the
amount of C-(A)-S-H leads to a denser matrix of the binder and thus
a reduction in the volume of permeable voids.
5.3. Measured chloride penetration depth in a steady-state test
Fig. 11 shows the cross-sections of AFS samples reacted with the
silver nitrate solution after the immersion for 91 days. The S0 and
S10 samples did not show a clear penetration depth, while the S30,
S50, and OPC samples showed a clear penetration depth due to the
presence of AgCl caused by the reaction with the AgNO3 solution. It
is certain from the results listed in Table 4 that a substantial amount
of chloride penetrated into the S10 and S0 samples. In addition,
much of the aluminosilicate gels present in the S0 and S10 samples
likely reacted with the AgNO3 solution, even in regions where the
chloride did not penetrate. As a consequence, no boundary could be
observed in the cross-section of the binder (see Table 4).
As the slag content increased, the penetration depth decreased
signicantly from 16.59 mm to 7.74 mm, as listed in Table 4. The
penetration depth of the S30 sample was very similar to that of the
OPC sample, while the S50 sample showed a relatively low pene-
tration depth compared to the OPC sample, possibly due to the fact
that the S50 sample showed fewer permeable voids and a low rate
of water absorption, as mentioned above. The dense matrix of the
AFS samples which derived from an increase in the slag content
hindered chloride penetration, resulting in a low depth of pene-
tration, as noted also by Ismail et al. [10]. Meanwhile, the decrease
in the slag content resulted in an increase in the N-A-S-H gels and a
decrease in the C-(A)-S-H gels. As a result, the less dense matrix
caused a greater chloride penetration depth in the AFS binder.
Therefore, the addition of slag to the AFS binder resulted in higher
resistance to chloride penetration.
5.4. Measured free and bound chloride contents in AFS paste
The chloride contents results (mg/g) of the AFS sample
immersed in the 10% sodium chloride solution for 91 days are listed
in Table 4. The chloride content of the S0 sample was not measured
due to hydrophobicity of the y ash. Total chloride content means
the acid-soluble chloride, while free chloride refers to water-
soluble chloride. Bound chloride is thus dened as the free chlo-
ride content subtracted from the total chloride content.
The S10 sample among the AFS samples showed the highest
total chloride content (0.1511 mg/g and 0.3961 mg/g) at the ages of
28 and 91 days, respectively, as shown in Table 4. As the immersion
time increased from 28 days to 91 days, the total and free chloride
contents also increased. The total chloride content of the OPC
sample was similar to that of the S30 sample. For the samples
immersed for 91 days, the OPC sample showed the highest total
chloride content, followed by the S10, S30, and then the S50 sam-
ples. It should be noted that both the total and free chloride con-
tents at the age of 91 days decreased as the amount of added slag
increased. The OPC sample showed a higher free chloride content
than the AFS samples. As shown in Fig. 12, most of the chloride was
bound in the AFS sample. The bound chloride content increased as
the amount of slag decreased. It was noted that the chloride
adsorption capacity increased as the aluminosilicate content in the
AFS binder increased. These chlorides can physically or chemically
be bound to the gels, and they do not signicantly affect the
corrosion of the reinforcement in concrete [28]. The bound chloride
content was similar in both the S30 and OPC samples immersed for
28 days, while this value was the highest in the OPC samples
immersed for 91 days. It is noted that OPC sample showed a higher
bound chloride content than AFS samples, even though AFS sample
contained a high slag content (i.e., 50%). This result is not consistent
with the tendency in OPC/slag blended systems where the chloride
binding capacity is increased by the slag. The reason is not clearly.
However, it is well known that in OPC/slag blended systems, the
formation of Friedels salt (calcium chloroaluminate hydrate) is
increased by the addition of slag to OPC [37]. Accordingly, the
chloride-binding capacity is improved by the increase in Friedels
salt. On the other hand, Friedels salt was not detected in XRD
pattern after the sulfuric acid attack in AFS systems, as shown in
Fig. 9, even though AFS paste contained a slag content of 50% of
total binder by weight. This may be the reason that AFS paste
showed a lower bound chloride content at 91 days than OPC paste.
 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179 177

Fig. 11. Cross section of AFS samples reacted with silver nitrate solution after exposure to 10% NaCl solution during 91 days.

Table 4
Chloride penetration depth (mm) and chloride contents (mg/g) of AFS samples exposed to NaCl solution during 28 days and 91 days.

Slag (y ash slag) (wt.%) Penetration depth Total chloride (acid Free chloride (water soluble) Bound chloride
(mm) soluble)

28days 91days 28days 91days 28days 91days 28days 91days

0 a a b b b
10 a a 0.1511 0.3961 0.0061 0.1449
0.0223 0.3738
30 8.867 16.594 0.0275 0.1465 0.0063 0.0211
0.0165 0.1299
50 3.867 7.736 0.0082 0.0632 0.0048 0.0035
0.0146 0.0486
OPC 8.361 14.428 0.0294 0.6666 0.0076 0.0218
0.0302 0.6364
a
Samples with slag contents of 0% and 10% were not measured because the penetration depth was not clear.
b
Sample with slag content of 0% was not measured due to the hydrophobic property of the powdered sample.
178 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179

100 16 30
90 91 days
Bound Cl / Total Cl (wt.%)

14
25

Water absorption rate (%)

80

Penetration depth (mm)

12 Penetration depth
70 R2=0.9513 20
10
60 28 days
8 15
50
40 6
10
30 4
20 Water absorption rate 5
2 R2=0.9955
10
0 0
0 15 20 25 30 35 40 45
S10 S30 S50 C
OPC Permearble void (%)
Samples
Fig. 14. Water absorption rate (%) and penetration depth (mm) over the permeable
Fig. 12. Bound chloride over total chloride in AFS samples at the age of 28 and 91 days void (%) in AFS samples.

5.5. Inuence of reaction products on chloride resistance
Fig. 13 shows the relationship between the ratio of bound-
chloride to total-chloride (%) or the penetration depth (mm) and
the aluminosilicate amount (%) in the AFS binder. It is clear that the
chloride-binding capacity and the penetration depth increased as
the aluminosilicate gel content in the AFS binder increased. There is
a correlation between the amount of aluminosilicate and the bound
chloride content. This is in good agreement with earlier results
[29,30] which found that the chloride-binding capacity increased
with an increase in the y ash content and with a decrease in the
slag content. The addition of y ash in the AFS binder results in
greater physical absorption of chloride into the aluminosilicate gel
compared to the C-(A)-S-H gel, which is attributed to the higher
surface area of the aluminosilicate gel in the AFS binder [31].
Hydrotalcite is known to be the main cause of chloride binding in
alkali-activated slag binders [32]; however, this phase was not
detected in the XRD patterns, as shown in Fig. 9.
Fig. 14 shows the relationship between the water absorption
rate (%) or penetration depth (mm) and the volume of permeable
voids (%) in the AFS samples. The results showed that both the
water absorption rate and the penetration depth increased as the
volume of permeable voids increased in the AFS binder. Thus, the
difference in the reaction product contents in the AFS binder clearly
affects the chloride-binding capacity and the resistance to chloride
penetration. As the slag content increases, the matrix becomes
denser due to an increase in the amount of C-(A)-S-H gel, resulting
in greater resistance to chloride penetration. On the other hand, the
100
90
Bound Cl / Total Cl (wt.%)
chloride-binding capacity increases as the slag content decreases,
as shown in Table 4. The N-A-S-H gel content increases as the slag
content decreases, resulting in the higher rate of chloride adsorp-
tion by the gels.
There are pros and cons when adding slag in view of the dura-
bility of the AFS binder. Thus, identifying the causes of this phe-
nomenon would be meaningful. As the slag content decreased in
the AFS paste, the N-A-S-H gel amount increased while the C-(A)-S-
H gel amount decreased [12]. Accordingly, the chloride binding
capacity (bound chloride content/total chloride content, %)
increased since the chloride binding capacity of N-A-S-H gel was
higher than that of C-(A)-S-H gel, as shown in Fig. 13. Meanwhile,
the structure of AFS paste became more porous due to the forma-
tion of C-(A)-S-H gel in lesser extent, as shown in Fig. 7, and the
penetration of chloride ions into the AFS paste was thus easier.
Therefore, as the slag content decreased, the chloride binding ca-
pacity increased and the resistance to chloride penetration
decreased. On the other hand, as the slag content increased in the
AFS paste, the C-(A)-S-H gel amount increased while the N-A-S-H
gel amount decreased. The AFS paste became denser due to the
increase in the C-(A)-S-H gel, which made the penetration of
chloride ions more difcult. However, the chloride binding capacity
decreased due to a decrease in the N-A-S-H gel which had a higher
binding capacity than C-(A)-S-H gel. As a result, improved dura-
bility could not be achieved through the simple control of the
mixture ratio of the raw materials in the AFS binder. However, an
addition of a mineral admixture such as silica fume may improve
both performances (the chloride-binding capacity and the resis-
tance to chloride penetration). Such a study will be conducted at a
later date.
30
6. Concluding remarks
25

Penetration depth (mm)

80
This paper has presented the results of an experimental study
70 conducted to examine the chloride and sulfuric acid resistance
20
60 Bound Cl
R2=0.9955
capabilities of AFS paste. The AFS paste samples were immersed
50 15 into a 10% sodium chloride solution and a 10% sulfuric acid solution
40 until the day of testing. Chloride penetration depth and sulfuric
10 acid-resistant reaction product were quantitatively evaluated
30
Penetration depth through a steady-state chloride penetration test and NMR analysis,
20
R2=0.986 5 respectively. The following conclusions can be drawn from the re-
10
sults presented here.
0 0
0 5 10 15 20 25 30 35 40
(1) From the XRD and FT-IR analyses, it was found that gypsum
N-A-S-H gel amount (%)
was formed due to the decalcication of C-(A)-S-H gel in AFS
Fig. 13. Bound chloride-to-total chloride ratio (%) and penetration depth (mm) over samples immersed in an H2SO4 solution, while halite was
the N-A-S-H gel amount (%) in AFS samples. formed only in the S0 sample immersed in the NaCl solution.
 N.K. Lee, H.K. Lee / Cement and Concrete Composites 72 (2016) 168e179
(2) As the slag content increased in the y ash/slag mixture, the
volume of permeable voids decreased. It is likely that an
increase in the amount of C-(A)-S-H leads to a denser matrix
of the binder and therefore a reduction in the volume of
permeable voids.
(3) A substantial amount of N-A-S-H gel, which formed in the
S10 and S0 samples and was less susceptible to a sulfuric acid
attack than C-(A)-S-H gel, resulted in higher resistance of the
binder despite the fact that the contribution of the N-A-S-H
gel to the compressive strength was lower than that of the C-
(A)-S-H gel.
(4) There are two causes of the deterioration of the AFS binder
due to a sulfuric acid attack. The rst one is corrosion by
means of SO2
4 penetration through the surface of the AFS
binder, which is associated with permeable voids and a water
absorption rate. The second one is corrosion of the reaction
products present in the AFS binder resulting from the dif-
ference in the resistance to a sulfuric attack between the C-
(A)-S-H gel and the N-A-S-H gel.
(5) As the slag content increased in the mixture, the chloride
penetration depth was greatly decreased. The dense matrix
of the AFS samples derived from an increase in the slag
content hindered chloride penetration, thus resulting in a
low penetration depth. Meanwhile, a decrease in the slag
content resulted in an increase in the N-A-S-H gels and a
decrease in the C-(A)-S-H gels. As a result, the less dense
matrix caused a greater chloride penetration depth in the
AFS binder.
(6) Variation of the slag content in the mixture changed the
reaction product content of the AFS binder, which clearly
affected the chloride-binding capacity and the resistance to
chloride penetration. It was found that the chloride-binding
capacity increased with the amount of N-A-S-H gels and that
the resistance to chloride penetration increased with the
amount of C-(A)-S-H gels. Accordingly, with an increase in
the slag content in the AFS binder, the chloride-binding ca-
pacity decreased and the resistance to chloride penetration
increased.
Acknowledgements
This research was supported by the National Research Founda-
tion of Korea (NRF) grant funded by the Korean government
(Ministry of Science, ICT & Future Planning)
(2015R1A2A1A10055694).
References
[1] I. Ismail, S.A. Bernal, J.L. Provis, R. San Nicolas, S. Hamdan, J.S. van Deventer,
Modication of phase evolution in alkali-activated blast furnace slag by the
incorporation of y ash, Cem. Concr. Compos. 45 (2014) 125e135.
[2] T. Bakharev, Resistance of geopolymer materials to acid attack, Cem. Concr.
Res. 35 (4) (2005) 658e670.
[3] T. Bakharev, J.G. Sanjayan, Y.B. Cheng, Resistance of alkali-activated slag
concrete to acid attack, Cem. Concr. Res. 33 (10) (2003) 1607e1611.
[4] X.J. Song, M. Marosszeky, M. Brungs, R. Munn, Durability of y ash based
geopolymer concrete against sulphuric acid attack, in: Proceedings of the
International Conference on Durability of Building Materials and Components,
Lyon, France, 2005, pp. 369e375.
[5] S. Thokchom, P. Ghosh, S. Ghosh, Acid resistance of y ash based geopolymer
mortars, Int. J. Recent Trends Eng. 1 (6) (2009a) 36e40.
[6] A. Allahverdi, F. Skvara, Sulfuric acid attack on hardened paste of geopolymer
cements-Part 1. Mechanism of corrosion at relatively high concentrations,
Ceram. Silik. 49 (4) (2005) 225.
[7] R.R. Lloyd, J.L. Provis, J.S. van Deventer, Acid resistance of inorganic polymer
binders. 1. Corrosion rate, Mater. Struct. 45 (1-2) (2012) 1e14.
[8] S. Thokchom, P. Ghosh, S. Ghosh, Effect of Na2O content on durability of
179
geopolymer mortars in sulphuric acid, Int. J. Chem. Biomol. Eng. 2 (1) (2009b)
20e25.
[9] A. Neville, Chloride attack of reinforced concrete: an overview, Mater. Struct.
28 (2) (1995) 63e70.
[10] I. Ismail, S.A. Bernal, J.L. Provis, R. San Nicolas, D.G. Brice, A.R. Kilcullen, J.S. van
Deventer, Inuence of y ash on the water and chloride permeability of alkali-
activated slag mortars and concretes, Constr. Build. Mater. 48 (2013a)
1187e1201.
[11] S.A. Bernal, R. Meja de Gutie
rrez, J.L. Provis, Engineering and durability
properties of concretes based on alkali-activated granulated blast furnace
slag/metakaolin blends, Constr. Build. Mater. 33 (2012) 99e108.
[12] N.K. Lee, H.K. Lee, Reactivity and reaction products of alkali-activated, y ash/
slag paste, Constr. Build. Mater. 81 (2015) 303e312.
[13] K.S. Chia, M.H. Zhang, Water permeability and chloride penetrability of high-
strength lightweight aggregate concrete, Cem. Concr. Res. 32 (4) (2002)
639e645.
[14] N.K. Lee, J.G. Jang, H.K. Lee, Shrinkage characteristics of alkali-activated y
ash/slag paste and mortar at early ages, Constr. Concr. Compos. 53 (2014)
239e248.
[15] I. Ismail, S.A. Bernal, J.L. Provis, S. Hamdan, J.S. van Deventer, Microstructural
changes in alkali activated y ash/slag geopolymers with sulfate exposure,
Mater. Struct. 46 (3) (2013b) 361e373.
[16] T. Bakharev, J.G. Sanjayan, Y.B. Cheng, Sulfate attack on alkali-activated slag
concrete, Cem. Concr. Res. 32 (2) (2002) 211e216.
[17] V.N. Nosov, N.G. Frolova, V.F. Kamyshov, IR spectra of calcium sulfate semi-
hydrates, J. Appl. Spectrosc. 24 (4) (1976) 509e511.
[18] S. Borhan, S. Hesaraki, S. Ahmadzadeh-Asl, Evaluation of colloidal silica sus-
pension as efcient additive for improving physicochemical and in vitro
biological properties of calcium sulfate-based nanocomposite bone cement,
J. Mater. Sci. Mater. Med. 21 (12) (2010) 3171e3181.
[19] C. Yue, S. Li, K. Ding, N. Zhong, Thermodynamics and kinetics of reactions
between C1-C3 hydrocarbons and calcium sulfate in deep carbonate reser-
voirs, Geochem. J. Jpn. 40 (1) (2006) 87.
[20] B.C. Smith, Infrared Spectral Interpretation: a Systematic Approach, CRC Press,
Florida, 1998, pp. 105e108. LCC no. 98-37190.
[21] J.C. Doadrio, D. Arcos, M.V. Cabanas, M. Vallet-Reg, Calcium sulphate-based
cements containing cephalexin, Biomaterials 25 (13) (2004) 2629e2635.
[22] C.A. Rees, J.L. Provis, G.C. Lukey, J.S. van Deventer, Attenuated total reectance
fourier transform infrared analysis of y ash geopolymer gel aging, Langmuir
23 (15) (2007) 8170e8179.
ser, J. Stark, Inuence of sulfate solution concentration on
[23] F. Bellmann, B. Mo
the formation of gypsum in sulfate resistance test specimen, Cem. Concr. Res.
36 (2) (2006) 358e363.
[24] A.M. Izzat, A.M.M. Al Bakri, H. Kamarudin, L.M. Moga, G.C.M. Ruzaidi,
M.T.M. Faheem, A.V. Sandu, Microstructural analysis of geopolymer and or-
dinary Portland cement mortar exposed to sulfuric acid, Mater. Plast. 50 (3)
(2013) 171e174.
[25] J.L. Provis, R.J. Myers, C.E. White, V. Rose, J.S. van Deventer, X-ray micro-
tomography shows pore structure and tortuosity in alkali-activated binders,
Cem. Concr. Res. 42 (6) (2012) 855e864.
[26] T. Sen, U. Mishra, Usage of industrial waste products in village road con-
struction, Int. J. Environ. Sci. Dev. 1 (2) (2010) 122e126.
[27] T. Sakthivel, D.L. Reid, I. Goldstein, L. Hench, S. Seal, Hydrophobic high surface
area zeolites derived from y ash for oil spill remediation, Environ. Sci.
Technol. 47 (11) (2013) 5843e5850.
[28] C. Arya, N.R. Buenfeld, J.B. Newman, Factors inuencing chloride-binding in
concrete, Cem. Concr. Res. 20 (2) (1990) 291e300.
[29] M.H. Shehata, M.D. Thomas, Alkali release characteristics of blended cements,
Cem. Concr. Res. 36 (6) (2006) 1166e1175.
[30] W. Chen, H.J.H. Brouwers, Alkali binding in hydrated Portland cement paste,
Cem. Concr. Res. 40 (5) (2010) 716e722.
[31] I. Ismail, S.A. Bernal, J.L. Provis, S. Hamdan, J.S. van Deventer, Drying-induced
changes in the structure of alkali-activated pastes, J. Mater. Sci. 48 (9) (2013c)
3566e3577.
[32] W. Chen, H.J.H. Brouwers, A method for predicting the alkali concentrations in
pore solution of hydrated slag cement paste, J. Mater. Sci. 46 (10) (2011)
3622e3631.
[33] N.K. Lee, H.K. Lee, Setting and mechanical properties of alkali-activated y
ash/slag concrete manufactured at room temperature, Constr. Build. Mater. 47
(2013) 1201e1209.
[34] F. Puertas, A. Fern
andez-Jime
nez, Mineralogical and microstructural charac-
terisation of alkali-activated y ash/slag pastes, Cem. Concr. Compos. 25 (3)
(2003) 287e292.
[35] J. Monteny, E. Vincke, A. Beeldens, N. De Belie, L. Taerwe, D. Van Gemert,
W. Verstraete, Chemical, microbiological, and in situ test methods for biogenic
sulfuric acid corrosion of concrete, Cem. Concr. Res. 30 (4) (2000) 623e634.
[36] A. Allahverdi, F. Skvara, Nitric acid attack on hardened paste of geopolymeric
cements. Part 2, Ceramics 45 (4) (2001) 143e149.
[37] R. Luo, Y. Cai, C. Wang, X. Huang, Study of chloride binding and diffusion in
GGBS concrete, Cem. Concr. Res. 33 (1) (2003) 1e7.