Content uploaded by Daewook Park
Author content
All content in this area was uploaded by Daewook Park on Feb 26, 2016
Content may be subject to copyright.
Mitigating effect of chloride ions on sulfate attack
of cement mortars with or without silica fume
Seung-Tae Lee, Dae-Wook Park, and Ki-Yong Ann
Abstract: This paper presents a detailed experimental study on the sulfate attack of mortar specimens with or without
silica fume exposed to sulfate and sulfate–chloride solutions (with the same concentration of SO42– ions) up to 510 d. The
overall aim of the study is to investigate the beneficial effect of chloride ions on sulfate attack. In addition, the role of
silica fume and water–binder ratio (w/b) in resisting sulfate attack is also reported. To qualitatively assess the performance
of mortar specimens exposed to test solutions, visual examination and compressive strength and expansion tests were car-
ried out. X-ray diffraction (XRD) and mercury intrusion porosimetry (MIP) techniques were also used to evaluate the
products formed by hydration and chemical reaction and the change of porosity for paste samples. Results indicated that
the presence of chloride ions in sulfate environments mitigated the deterioration of ordinary Portland cement mortar speci-
mens, especially with a higher w/b, due to sulfate attack. It seems that the mitigating effect of chloride ions on sulfate at-
tack is attributable to the increased solubility of sulfate products in the chloride-bearing sulfate solution, and the chemical
binding of the ions to form Friedel’s salt.
Key words: chloride ions, sulfate attack, silica fume, Friedel’s salt.
Re
´sume
´:Cet article pre
´sente une e
´tude expe
´rimentale de
´taille
´e de l’attaque des sulfates sur des e
´chantillons de mortier,
contenant ou non de la fume
´e de silice, expose
´sa
`une solution de sulfates et de chlorure de sulfates (avec une concentra-
tion identique d’ions SO42–) pendant une pe
´riode pouvant atteindre 510 jours. L’objectif global de cette e
´tude est d’exa-
miner l’effet be
´ne
´fique des ions chlorures sur l’attaque des sulfates. De plus, le ro
ˆle de la fume
´e de silice et du rapport
eau-liant a
`contrer les attaques des sulfates a e
´galement e
´te
´aborde
´. Afin d’e
´valuer quantitativement le comportement des
e
´chantillons de mortier expose
´s aux solutions d’essai, un examen visuel, des essais de re
´sistance en compression et des es-
sais d’expansion ont e
´te
´re
´alise
´s. La diffraction aux rayons X et la porosime
´trie au mercure ont e
´galement e
´te
´utilise
´es
pour e
´valuer les produits forme
´s par l’hydratation et la re
´action chimique, de me
ˆme que le changement de porosite
´des
e
´chantillons de pa
ˆte. Les re
´sultats indiquent que la pre
´sence d’ions chlorure dans des milieux sulfate
´sare
´duit, en raison
de l’attaque des sulfates, la de
´te
´rioration des e
´chantillons de mortier faits de ciment Portland ordinaire, particulie
`rement
avec un rapport eau-liant plus e
´leve
´. Il semble que l’effet d’atte
´nuation des ions chlorures sur l’attaque des sulfates peut
e
ˆtre attribuable a
`la solubilite
´accrue des produits de sulfates dans la solution de sulfate contenant des chlorures et a
`la liai-
son chimique des ions pour former du sel de Friedel.
Mots-cle
´s:ions chlorures, attaques aux sulfates, fume
´e de silice, sel de Friedel.
[Traduit par la Re
´daction]
1. Introduction
The most aggressive environmental agents that affect the
durability of concrete structures are chlorides and sulfates.
There have been many studies related to steel corrosion due
to ingress of chlorides and concrete deterioration due to sul-
fate attack (Poupard et al. 2006; Ampadu and Torii 2002;
Brown and Hooton 2002; Hartshorn et al. 2002). It has
been widely accepted that both deterioration phenomena are
the most severe problems influencing the service life of con-
crete structures.
The sulfate attack mechanism is still a subject of contro-
versy although many research studies have been carried out
to unravel its causes. It has been widely accepted that ce-
ment compositions (i.e., C3A content and silicate ratio) and
mineral admixtures play a critical role in resisting sulfate at-
tack (Al-Amoudi 1998; Lee et al. 2005; Moon et al. 2002).
In particular, C3A content of cement is found to be an im-
portant parameter determining sulfate resistance of concrete,
because it can control the quantity of calcium hydroxide in
the cement paste. Furthermore, the use of mineral admix-
tures is recommended in sulfate environments, because of
their pore refining ability, C3A dilution, and removal of cal-
cium hydroxide by pozzolanic reaction.
Chlorides break down the initial passive layer that is
Received 16 September 2007. Revision accepted 12 May 2008.
Published on the NRC Research Press Web site at cjce.nrc.ca on
5 November 2008.
S.-T. Lee.1Department of Civil Engineering, Kunsan National
University, 68 Miryong-dong, Kunsan, Jeonbuk 573-701, South
Korea.
D.-W. Park. Department of Civil and Environmental
Engineering, Kunsan National University, San 68 Miryong-dong,
Kunsan, Cheollabuk-do 573-701, South Korea 573-791, South
Korea.
K.-Y. Ann. School of Civil and Environmental Engineering,
Yonsei University, 134 Sinchon-dong, Seodaemun-gu, Seoul
120-749, South Korea.
Written discussion of this article is welcomed and will be
received by the Editor until 31 March 2009.
1Corresponding author (e-mail: stlee@kunsan.ac.kr).
1210
Can. J. Civ. Eng. 35: 1210–1220 (2008) doi:10.1139/L08-065 #2008 NRC Canada
formed on the steel bar surface that is further protected due
to the highly alkaline environment of the cement matrix. Re-
garding sulfate attack, the chemical process has often been
discussed in terms of the interaction between the hydrates
in cement pastes and the dissolved compounds from attack-
ing sources (Taylor 1997). Obviously, the attack process
leads to expansion, softening, and microcracks in the cement
paste or interfacial zone. Therefore, the co-presence of both
chloride and sulfate ions as the external source in ground-
water or marine environment causes rebar corrosion and sul-
fate deterioration in reinforced concrete structures. However,
with respect to sulfate attack, only a few studies of the dam-
age to and corresponding mechanism of the cement matrix
have been done in the conditions of co-presence of Cl–and
SO42– ions.
Al-Amoudi et al. (1994) reported that the increased solu-
bility of calcium aluminate hydrate leads to the nonexpan-
sive crystals formed by sulfate attack in the case of the
presence of chloride ions. They added that the transforma-
tion of aluminate hydrate phases into calcium chloro-alumi-
nates (e.g., Friedel’s salt) reduces the amount of ettringite
formation. Harrison (1990) also insisted that the reaction of
monosulfate to form erringite should be retarded in a chlor-
ide–sulfate solution, because chloride ions substitute for sul-
fate ions in the structures of monosulfate, thereby modifying
the morphology of ettringite.
Actually, the effect of the conjointed presence of chloride
and sulfate ions on the deterioration has been highly de-
bated. Moon et al. (2002) carried out a long-term exposure
test to evaluate the resistance of Portland cement with dif-
ferent C3A contents and silicate ratios to chloride solution.
They stated in their publication that Friedel’s salt may not
damage the microstructure of the cement matrix.
Similarly, Lea (1970) confirmed that the expansion of
concrete due to SO42– attack in seawater is retarded due to
the concomitant presence of chloride and sulfate ions. He at-
tributed this retardation to the increased solubility of sulfate
phases in chloride solution. Ogawa and Roy (1982) have
also revealed in their work that expansive phases formed by
sulfate attack are apt to be reduced in the presence of chlor-
ides. Another possible reason for less sulfate deterioration
may contribute to the increased solubility of gypsum forma-
tion in a chloride environment. It has been reported that ten-
sile stresses developed during gypsum formation may lead
to significant expansion (Tian and Cohen 2000). However,
in a chloride-bearing sulfate solution, the quantity of gyp-
sum formation causing expansion and cracking inthe micro-
structure of hardened cement can be significantly reduced or
eliminated. Al-Amoudi et al. (1994) also reported a similar
study, indicating the beneficial effect of chloride ions on
the sulfate attack.
The diffusion coefficients for chloride and sulfate ions are
different owing to different ionic mobilities and ion valen-
ces. When compared with sulfate ions, the chloride ions
penetrate more rapidly into the cement matrix. Subse-
quently, the higher diffusivity allows the chloride ions to re-
act with C3A in the cement paste to form Friedel’s salt, and
eventually lead to the reduction in the amount of sulfate
products, such as ettringite, gypsum, and thaumasite.
This study was carried out to investigate the different be-
haviours of the deterioration in mortar and paste specimens,
with or without silica fume, exposed to both a sulfate solu-
tion and a chloride–bearing sulfate solution with the same
concentration of SO42– ions for 510 d at ambient tempera-
ture. Ultimately, the aim of the study is to assess the benefi-
cial effect of chloride ions on sulfate attack of plain and
silica-fume-blended cement matrices.
2. Experimental programme
2.1. Materials used
In this study, ordinary Portland cement (OPC) conforming
to ASTM C150 (ASTM 2007) was used throughout, which
had been produced by a local cement company in South
Korea. The C3S, C2S, C3A, and C4AF contents of the OPC
by Bogue calculation were 54.9%, 16.6%, 10.3%, and
9.1%, respectively. Silica fume (SF) with a specific surface
area of 20 470 m2/kg was used at replacement levels of 5%,
10%, and 15% by mass of cement. The chemical composi-
tion of OPC and SF used in this investigation is shown in
Table 1.
River sand, with a maximum size of 5 mm, was used as
the fine aggregate. The fine aggregate, which is immune to
most chemical agents and has little organic material, was
employed for manufacturing mortar specimens. The specific
gravity, absorption, and fineness modulus of the fine aggre-
gate were 2.60, 0.80%, and 2.80, respectively. To obtain a
suitable workability, a polycarbonic acid-based superplasti-
cizing chemical admixture (SP) was used in all mortar mix-
tures. The SP was added to the mixing water at a level of
0.5% by mass of total binder; it was well stirred until com-
pletely dissolved before it was added to the mortar mixtures.
2.2. Preparation of mortar and paste samples
The mass ratio of fine aggregate to binder materials (OPC
and SF) was kept invariant at 2.0 in all mortar mixtures. The
mixing of the mortar was carried out using the appropriate
amount of binder materials, fine aggregate, and water with
the added SP. The sequence of mixing was as follows: mix-
ing of all materials for 30 s, resting for 1.5 min., followed
by further mixing for 1 min. The details of mortar mixtures
used in this study are given in Table 2. For example, the
symbol ‘‘SF10–45’’ indicates a mortar mixture that has 10%
of OPC by mass replaced with SF and a water–binder ratio
(w/b) of 0.45. Flow and air content tests of fresh mortar
conforming to ASTM C230 (ASTM 2003) and C185
(ASTM 2002) were carried out and the detailed data are pre-
sented in Table 3.
All mortar specimens were demoulded after 24 h of cast-
ing, and were then cured in water for an additional 6 d.
Thereafter, some of the specimens were moved to test solu-
tions and were continuously immersed in the solutions.
Water and test solution temperatures were kept at 20 ± 3 8C
during the test period.
Paste samples with 0.45 of w/b were made using 100 g of
total binder materials (OPC + SF) and 45 g of deionized
water with no SP. After hand-mixing, the paste samples
were cast and rotated in their moulds for 8 h to avoid segre-
gation and demoulded after 1 d. All paste samples were cast
into plastic cylinders 13 mm in diameter. After demoulding,
the paste samples were cured in water for 6 d and then
coated with epoxy on all surfaces except for the upper side.
Lee et al. 1211
#2008 NRC Canada
2.3. Test solutions
The mortar and paste specimens were immersed in water
and test solutions maintained at ambient temperature. Ac-
cording to ASTM C1012 (ASTM 2004), the sulfate solution
concentration was determined at 33 800 ppm as SO42– ions
by using regent-grade sodium sulfate. To produce the sul-
fate–chloride solution, the sodium chloride solution with a
concentration of 3.5% was added to the sulfate solution.
The concentration of SO42– ions in the sulfate–chloride solu-
tion was also fixed at 33 800 ppm. The test solutions were
replaced once per month. During the test period, the solu-
tions and specimens were kept in plastic containers with
space between specimens.
2.4. Test techniques
The mortar specimens tested in this study are divided into
cubes (50 mm 50 mm 50 mm) for compressive
strength measurement and prisms (25 mm 25 mm
285 mm) for expansion measurement.
The compressive strength of mortar specimens was meas-
ured after 0, 28, 91, 180, 270, 360, and 510 d of exposure.
Companion mortar specimens cured in water for the same
duration were also tested. The strength deterioration of mor-
tar mixtures was evaluated by compressive strength loss, as
expressed in the following equation:
Compressive strength loss ð%Þ¼100½ðf0
cw f0
csÞ=f 0
cw
where f’
cw is the average compressive strength of mortar
specimens cured in water and f’
cs is the average compres-
sive strength of mortar specimens exposed to test solutions.
Three mortar specimens from water and test solutions were
tested at each test period and then the values were averaged.
Based on ASTM C1012 (ASTM 2004), expansion meas-
urements of mortar specimens exposed to test solutions
were carried out and the values of three specimens were
averaged.
Paste samples immersed in water and test solutions for
510 d were used for XRD analysis. The upper-side portions
exposed to the test solutions were sliced 5–10 mm in thick-
ness. Thereafter, the selected paste portions were ground by
hand using acetone. The XRD was conducted as using the
RINT D/max 2500 (Rigaku, Japan) X-ray diffractometer.
For the XRD tests, CuKaradiation with a wavelength of
l.5405 A
˚(1 A
˚= 0.1 nm) at a voltage of 30 kV and current
of 30 mA was used. The analysis was performed between 58
and 4082qat a scanning speed of 28/min.
Mercury intrusion porosimetry (MIP) with a contact angle
of 1308and maximum pressure of about 413 MPa was used
to investigate the porosity of paste samples exposed to test
solutions.
3. Results
3.1. Visual examination
A visual examination was carried out to check the visible
surface damages of spalling, expansion, softening, delamina-
tion, and cracking in the cube mortar specimens exposed to
test solutions. Figures 1 and 2 show typical examples of
damage of OPC and SF mortar specimens immersed in a
sulfate or sulfate–chloride solution for 510 d. Regardless of
test solutions and w/b of mortar mixtures, the first sign of
sulfate attack was the deterioration of the corners of the
mortar samples followed by damage at the edges and the
faces.
Figure 1, related to the surface damages of OPC mortar
specimens, confirms that the degree of deterioration due to
sulfate attack is greatly dependent on the attacking sources
(sulfate solution or sulfate–chloride solution) supplying sul-
fate ions to the mortar specimens. The OPC mortar speci-
mens exposed to a sulfate solution showed progressive
deterioration. The negative effect of higher w/b on the dete-
rioration of OPC mortar mixtures in a sulfate solution was
observed more remarkably. For example, the OPC-55 mortar
specimen exposed to the solution suffered a serious deterio-
ration showing a great amount of material loss (about 40%),
whereas there was about 1.5% of material loss in addition to
cracks along the edges in OPC-35 mortar specimens. Com-
paratively, in the case of the OPC-55 mortar specimen ex-
posed to a sulfate–chloride solution for 510 d, there was
only some spalling and cracks at the corners and edges of
the specimen. Moreover, the visual examination of OPC-35
and OPC-45 mortar specimens in the sulfate–chloride solu-
tion showed little visible deterioration with no detectable
material loss. The beneficial effects of the sulfate–chloride
solution on visual surface damage of OPC mortar mixtures
were thoroughly investigated. However, a different trend
was observed for SF mortar specimens exposed to test solu-
tions, as shown in Fig. 2. In other words, regardless of re-
placement levels of SF, surface damages of the SF mortar
specimens were negligible, even in a sulfate solution. This
is because SF leads to low permeability in the hardened ce-
ment matrix and eventually has a good resistance against
sulfate attack (Lee et al. 2005; Lawrence 1992).
Table 1. Chemical composition and physical properties
of ordinary Portland cement (OPC) and silica fume
(SF).
Properties OPC SF
Chemical composition (%)
Silicon dioxide 20.2 91.2
Aluminium oxide 5.8 1.3
Ferric oxide 3.0 0.8
Calcium oxide 63.3 0.7
Magnesium oxide 3.4 0.3
Sulfur trioxide 2.1 —
Loss on ignition 2.1 2.3
Mineralogical components (%)
C3S 54.9 —
C2S 16.6 —
C3A 10.3 —
C4AF 9.1 —
Setting time (min)
Initial set 250 —
Final set 400 —
Specific gravity 3.13 2.20
Specific surface area (m2/kg) 312 20 470*
*Nitrogen adsorption.
1212 Can. J. Civ. Eng. Vol. 35, 2008
#2008 NRC Canada
3.2. Compressive strength of mortar
The compressive strength values of OPC and SF-blended
cement mortars with 0.35, 0.45, and 0.55 of w/b exposed to
water, sulfate, and sulfate–chloride solutions are summarized
in Table 4. It should be noted that the duration means the
exposure period to test solutions after a 7 d pre-curing in
water. These compressive strength results are discussed be-
low in terms of compressive strength loss defined earlier, to
highlight the damage caused to mortar specimens by expo-
sure to a sulfate solution or sulfate–chloride solution. Based
on the results of the compressive strength tests, compressive
strength loss of OPC and SF mortar specimens exposed to
test solutions are shown in Figs. 3 and 4, respectively.
Figure 3 confirms that the w/b is a critical factor in the
compressive strength loss of OPC mortar specimens, partic-
ularly when exposed to a sulfate solution. It was clearly evi-
dent that OPC mortar specimens with a higher w/b recorded
greater values in compressive strength loss at a longer dura-
tion. On the other hand, the data indicated that OPC mortar
specimens stored in a sulfate–chloride solution exhibited a
better resistance, showing relatively lower values of com-
pressive strength loss after 510 d of exposure. However,
after that duration, the compressive strength loss of OPC
mortar specimens with a 0.55 w/b exposed to a sulfate solu-
tion was about 91.7%. The compressive strength losses of
OPC mortar specimens in a sulfate–chloride solution were
significantly lower than those of the similar mortar speci-
mens in a sulfate solution, and the difference of the values
between both solutions was significant as the w/b increases.
Figure 4 shows the compressive strength loss data for SF-
blended mortar specimens (5% to 15% of replacement level)
stored in sulfate and sulfate–chloride solutions. The duration
was also up to 510 d. These results emphasize that the in-
corporation of SF led to lower compressive strength loss in
a sulfate solution, compared with OPC mortar mixtures. Ad-
ditionally, it can be seen that the compressive strength loss
of SF10–55 mortar specimens in a sulfate solution was
somewhat significant. This suggests that the w/b is the crit-
ical parameter influencing the resistance to sulfate attack.
For the mortars with the same w/b (0.45), however, the data
on compressive strength loss in the solution didn’t show sig-
nificant differences. The compressive strength loss of SF
mortar specimens with a 0.45 w/b stored in a sulfate–
chloride solution was not appreciable, even up to 510 d.
The overall indication from the data shown in Figs. 3 and
4 is that the strength deterioration of OPC mortar mixtures
due to a sulfate solution was much more significant com-
pared with that due to a sulfate–chloride solution. In con-
trast, the strength loss of SF mortar mixtures was not
significant in both test solutions, regardless of replacement
levels of SF.
3.3. Expansion of mortar
Expansion data of mortar specimens with or without SF
are presented in Figs. 5 and 6, respectively. Figure 5 shows
the expansion of OPC mortar specimens with different w/b
exposed to a sulfate or sulfate–chloride solution. These re-
sults also emphasize the key role of w/b of OPC mortar
specimens in resisting sulfate attack. The OPC-45 and OPC-
55 mortar specimens exposed to a sulfate solution disinte-
grated after 210 and 180 d, respectively, whereas the expan-
sion of OPC-35 mortar specimens was relatively stable up to
510 d. A different trend was observed in expansion results
for OPC mortar specimens exposed to a sulfate–chloride sol-
ution. Although OPC-55 mortar specimens were continu-
ously exposed to a sulfate–chloride solution, the specimens
expanded about 0.44% after 360 d of exposure and then
eventually disintegrated. There was a small expansion for
OPC-35 and OPC 45 mortar specimens in a sulfate–chloride
solution.
Expansion due to aggressive ions of mortar specimens
with different replacement levels of SF is shown in Fig. 6.
All mortar specimens with SF exhibited expansion values
below 0.07% after 510 d of exposure, irrespective of test
solutions. Unlike the expansion results for OPC mortar
specimens shown in Fig. 5, the general expansion trend of
SF mortar specimens exposed to a sulfate solution were sim-
ilar to those exposed to a sulfate–chloride solution. Much
more importantly, the beneficial effect of the presence of
chloride ions on expansion by sulfate attack was observed
to be insignificant for the mortar mixtures incorporating SF,
as already mentioned in the results of visual examination
and compressive strength.
Table 2. Details of mortar mixtures. OPC, ordinary Portland cement; SF, silica fume.
Mix
number Symbols
Cementitious
materials w/b (%)
Replacement levels
of SF (%)
1 OPC-35 OPC 35 —
2 OPC-45 OPC 45 —
3 OPC-55 OPC 55 —
4 SF5–45 95% OPC + 5% SF 45 5
5 SF10–35 90% OPC + 10% SF 35 10
6 SF10–45 90% OPC + 10% SF 45 10
7 SF10–55 90% OPC + 10% SF 55 10
8 SF15–45 85% OPC + 15% SF 45 15
Table 3. Properties of fresh mortars.
Symbol Flow (mm) Air content (%)
OPC-35 175 5.5
OPC-45 202 6.6
OPC-55 213 7.0
SF5–45 190 6.1
SF10–35 161 5.3
SF10–45 186 5.9
SF10–55 191 6.4
SF15–45 166 5.0
Lee et al. 1213
#2008 NRC Canada
3.4. X-ray diffraction analysis
Figure 7 shows the results of XRD analysis performed on
the powders obtained from OPC pastes with 0.45 w/b. The
paste samples tested using XRD analysis were continuously
stored in water or test solutions for 510 d. The XRD result
of the paste sample cured in water shows the typical diffrac-
togram of a normal hydrated portland cement, indicating the
intensive peaks for portlandite as well as the small amount
of ettringite, gypsum, and calcite (Fig. 7a). The diffracto-
gram for the sample drawn from the deteriorated part of the
OPC paste exposed to a sulfate solution shows the presence
of thaumasite and a small amount of portlandite, along with
relatively strong intensity peaks for gypsum and ettringite
produced by sulfate attack, as shown in Fig. 7b. Although
thaumasite formation is frequently regarded as a low-
temperature phenomenon, it can be also be detected in
paste specimens exposed at ambient conditions (Brown
and Hooton 2002; Hartshorn et al. 2002).
The formation of gypsum and ettringite when exposed to
a sulfate solution seems to have been associated with con-
Fig. 1. Deterioration of ordinary Portland cement mortar specimens after 510 d of exposure. w/b, water–binder ratio.
Fig. 2. Deterioration of silica fume (SF) mortar specimens after 510 d of exposure. w/b, water–binder ratio.
1214 Can. J. Civ. Eng. Vol. 35, 2008
#2008 NRC Canada
siderable expansion and compressive strength loss in the
OPC mortar system, as shown in Figs. 3 and 5.
The diffractogram of the sample of a similar paste made
with OPC but immersed in a sulfate–chloride solution
(Fig. 7c) exhibits the absence of thaumasite and gypsum
even though immersion was for 510 d. The XRD trace indi-
cates a weak intensity peak, at 11.282qfor calcium chloro-
aluminate hydrate, known as Friedel’s salt. Furthermore the
results of XRD analysis clearly showed that OPC pastes im-
mersed in a sulfate–chloride solution were less deteriorated,
indicating a large amount of portlandite and a small amount
of ettringite, compared with those in a sulfate solution.
The XRD pattern, as shown in Fig. 8a, for the powders of
the SF10–45 paste sample shows relatively weak intensity
peaks, especially for portlandite, compared with that of
100% OPC pastes in water. There is no doubt that these
weak intensity peaks resulted from the pozzolanic reaction
of SF and the reduction in plain cement content. Apart from
the results obtained from XRD analysis of OPC paste sam-
ples, the diffractogram pattern for the powdered samples of
the SF10–45 paste sample stored in a sulfate solution was
very similar to that for the paste sample stored in a sulfate–
chloride solution. The paste samples in sulfate and sulfate–
chloride solutions contained ettringite, gypsum, portlandite,
and calcite, as shown in the diffractograms of Fig. 8band
8c, but there was no evidence for the presence of thaumasite
or Friedel’s salt in both solutions.
3.5. Porosity
The pore structure of a cement system affects its mechan-
Table 4. Compressive strength of mortar specimens exposed to test solutions.
Compressive strength (MPa)
Duration (d) Symbol Water Sulfate solution Sulfate–chloride solution
Prior to exposure OPC-35 57.4 57.4 (0) 57.4 (0)
OPC-45 39.7 39.7 (0) 39.7 (0)
OPC-55 33.2 33.2 (0) 33.2 (0)
SF5–45 38.0 38.0 (0) 38.0 (0)
SF10–35 51.9 51.9 (0) 51.9 (0)
SF10–45 37.2 37.2 (0) 37.2 (0)
SF10–55 29.1 29.1 (0) 29.1 (0)
SF15–45 45.1 45.1 (0) 45.1 (0)
180 OPC-35 78.4 69.8 (10.9) 76.8 (2.0)
OPC-45 57.9 49.8 (13.9) 57.1 (1.2)
OPC-55 51.2 33.9 (33.9) 46.1 (10.0)
SF5–45 68.9 66.3 (3.7) 68.8 (0.1)
SF10–35 90.8 90.6 (0.2) 91.4 (–0.7)
SF10–45 70.6 70.2 (0.6) 69.4 (1.7)
SF10–55 60.9 54.1 (11.2) 55.8 (8.4)
SF15–45 69.8 68.9 (1.3) 70.2 (–0.6)
270 OPC-35 78.6 67.9 (13.6) 77.4 (1.4)
OPC-45 55.6 41.0 (26.2) 51.3 (7.7)
OPC-55 51.5 15.7 (69.5) 40.1 (22.2)
SF5–45 73.3 61.2 (16.4) 68.3 (6.8)
SF10–35 93.4 86.2 (7.7) 92.3 (1.1)
SF10–45 73.1 60.0 (17.9) 69.9 (4.3)
SF10–55 63.5 45.3 (28.6) 53.1 (16.4)
SF15–45 81.8 70.2 (14.2) 81.5 (0.4)
360 OPC-35 78.7 64.5 (18.0) 73.7 (6.4)
OPC-45 58.4 32.1 (45.0) 52.6 (9.9)
OPC-55 52.0 9.4 (81.9) 37.6 (27.7)
SF5–45 73.6 62.5 (15.1) 68.1 (7.5)
SF10–35 92.9 85.0 (8.5) 90.2 (2.9)
SF10–45 74.5 57.4 (22.9) 70.8 (5.0)
SF10–55 63.3 44.3 (30.0) 52.1 (17.7)
SF15–45 81.6 68.6 (15.9) 80.2 (1.7)
510 OPC-35 79.6 49.7 (37.6) 65.5 (17.7)
OPC-45 59.2 21.4 (63.8) 46.9 (20.7)
OPC-55 52.7 4.4 (91.7) 31.3 (40.5)
SF5–45 75.8 61.6 (18.7) 71.3 (5.9)
SF10–35 95.4 85.7 (10.2) 88.9 (6.8)
SF10–45 74.7 58.2 (22.1) 67.9 (9.2)
SF10–55 64.3 40.9 (36.3) 51.5 (19.8)
SF15–45 83.8 69.1 (17.5) 81.2 (3.0)
Note: Numbers in parentheses represent the compressive strength loss (%).
Lee et al. 1215
#2008 NRC Canada
ical properties (Czerni 1980). The pore structure of a hard-
ened cement system changes with the hydration process. Ad-
ditionally, cement-based materials exposed to aggressive
media are vulnerable and changeable in pore volume. Thus,
the MIP technique was used to measure the porosity of the
paste samples exposed to test solutions. The MIP measure-
ment was performed on paste samples after 0, 270, and
510 d of exposure to the test solutions.
Figure 9 shows the curves for pore size distribution of
OPC–45 and SF10–45 paste samples prior to exposure to
test solutions. There is no marked difference of capillary po-
rosity between both paste samples. However, coarser pores
in OPC paste samples were observed.
Figure 10 presents the results of cumulative pore volume
of OPC–45 paste samples exposed to a sulfate or sulfate–
chloride solution. Results indicate that the cumulative pore
volume of OPC–45 paste samples increased with duration
in a sulfate–solution due to sulfate attack. However, there
was a reverse trend for OPC–45 paste samples in a sulfate–
chloride solution, showing the decrease of cumulative pore
volume with duration. After 510 d of exposure, the ultimate
cumulative pore volumes of OPC–45 paste samples were
about 0.487 and 0.169 cm3/g for sulfate and sulfate–chloride
solutions, respectively. In the conjointed environment of sul-
fate and chloride ions, it appeared that the increase of pore
volume by sulfate attack was less than the decrease of pore
volume by cement hydration.
The variation of cumulative pore volume with duration of
SF10–45 paste samples is presented in Fig. 11. As expected,
the difference of cumulative pore volume between the sam-
ples in both test solutions was not significant. With respect
to duration, there was only a small variation in pore volume.
4. Discussion
Sulfate attack on cement-based materials is characterized
by the chemical reaction of sulfate ions with the cement hy-
drates and the corresponding products formed during the re-
action. Many studies to minimize the attack and to prolong
the service life of concrete structures exposed to sulfate en-
vironments have been carried out (Akoz et al. 1995; Khatri
et al. 1997; Moon et al. 2001).
In this study, the different behaviour of samples exposed
to sulfate and sulfate–chloride solutions was examined. Ex-
perimental results suggest that the mechanisms of deteriora-
Fig. 3. Compressive strength loss of ordinary portland cement
(OPC) mortar specimens.
Fig. 4. Compressive strength loss of silica fume (SF) mortar speci-
mens.
Fig. 5. Expansion of ordinary portland cement (OPC) mortar speci-
mens.
Fig. 6. Expansion of silica fume (SF) mortar specimens.
1216 Can. J. Civ. Eng. Vol. 35, 2008
#2008 NRC Canada
tion in these environments are significantly different from
each other.
From visual examination, it is clear that OPC mortar
specimens exposed to a sulfate–chloride solution behave
very differently from those stored in a sulfate solution. The
degree of deterioration of OPC mortar specimens in the
chloride-containing sulfate solution was much less compared
with that of the specimens in a sulfate solution with respect
to visible surface damage. Additionally, it was also observed
that w/b of OPC mortar mixtures can be an important factor
determining resistance to sulfate attack. This was strongly
supported by the compressive strength loss (Fig. 3) and ex-
pansion (Fig. 5) of OPC mortar specimens. Furthermore, the
compressive strength data presented in Table 4 and Fig. 4
emphasize the more pronounced positive effect of SF
against sulfate attack. In fact, the superior resistance to sul-
fate attack of the SF system has been widely reported (Al-
Amoudi 1998; Lee et al. 2005; Turker et al. 1997). The im-
provement of sulfate resistance of the cement matrix incor-
porating SF can probably be attributed to the combined
effects of reduced permeability and the decrease in calcium
hydroxide in the hardened cement paste.
Mineralogical analyses based on XRD suggest that the
primary cause for deterioration of OPC paste samples due
to sulfate attack may be attributed to the formation of gyp-
sum and ettringite as well as thaumasite. Additionally, there
was considerable reduction in peak intensities for portlandite
in the XRD pattern (Fig. 7b) compared with the pattern cor-
responding to OPC paste sample in water (Fig. 7a). For the
XRD pattern of OPC samples exposed to a sulfate–chloride
environment, however, the peaks for gypsum and thaumasite
were not detected or negligible. In addition, the reduced et-
Fig. 7. X-ray difraction analysis of OPC-45 paste samples (510 d): (a) water; (b) sulfate solution; (c) sulfate–chloride solution.
Fig. 8. X-ray diffraction analysis of SF10–45 paste samples (510 d): (a) water; (b) sulfate solution; (c) sulfate–chloride solution.
Lee et al. 1217
#2008 NRC Canada
tringite formation in the sulfate–chloride solution is prob-
ably attributable to the increased solubility of sulfate ions
due to the presence of chloride ions and being combined as
calcium chloro-aluminate (Harrison 1990). Consequently,
when chloride is present in a sulfate solution, the ettringite
can crystallize from the pore solution, which has a charac-
teristic of little expansion, because of its higher solubility.
This was confirmed by the expansion results shown in
Fig. 5, which was similar to the data published by Al-
Amoudi and Maslehuddin (1993).
It has been generally accepted that SF reacts with cal-
cium hydroxide in the presence of moisture to form secon-
dary C–S–H. In addition, the replacement of cement by SF
leads to a dilution of the C3A phase due to the reduction in
the content of cement and to the decrease of aluminate-
bearing phases. The effects of SF under a sulfate environ-
ment have been reported by many researchers (Mehta
1985; Rasheeduzzafar et al. 1990; Hooton 1993; Cohen
and Bentur 1988). The formation of gypsum and ettringite
in concrete incorporating sufficient SF, if exposed to a so-
dium–sulfate environment, would be migrated. Moreover,
as the use of SF produces a lower pH in the pore solution
due to the consumption of alkalies and calcium hydroxide;
ettringite formed by sulfate attack becomes less expansive
(Al-Dulaijan et al. 2003).
The beneficial role of SF in a sulfate solution was exper-
imentally confirmed by Wee et al. (2000). They also empha-
sized the role of the secondary C–S–H formation due to
pozzolanic reaction of SF in resisting sulfate attack. Mangat
and Khatib (1995) demonstrated that neither porosity nor
pore structure are the primary reasons for the enhancement
of sulfate resistance of SF, and added that the improved re-
sistance is attributable to the reduction in calcium hydroxide
content in SF.
As previously mentioned, comparison of the visual exami-
nation, compressive strength loss, and expansion data for
mortar specimens exposed to test solutions with the same
concentration of SO42– ions indicates that the presence of
chlorides in the sulfate solution mitigates the deterioration
due to sulfate attack. his beneficial effect of chloride ions is
observed to be significant for the OPC mortar specimens
with no replacement of SF. However, the beneficial effect is
almost insignificant for SF mortar specimens. To elucidate
the beneficial effect of chloride ions on compressive strength
loss of mortar specimens by sulfate attack, a mitigation fac-
Fig. 9. Pore size distribution of paste samples prior to exposure to
test solutions.
Fig. 10. Cumulative pore volume of OPC-45 paste samples.
Fig. 11. Cumulative pore volume of SF10–45 paste samples.
Fig. 12. Mitigation factor of mortar mixtures.
1218 Can. J. Civ. Eng. Vol. 35, 2008
#2008 NRC Canada
tor (MF), which is the difference between the compressive
strength loss value due to sulfate solution exposure and this
value due to sulfate–chloride solution exposure, was calcu-
lated using following equation and presented in Fig. 12.
Mitigation factor ð%Þ¼LSS ð%ÞLSC ð%Þ
where LSS is the compressive strength loss of mortar speci-
mens immersed in a sulfate solution and LSC is the compres-
sive strength loss of mortar specimens immersed in a
sulfate–chloride solution, respectively.
Figure 12 shows the MF of all mixtures used in this ex-
periment after 270 and 510 d of exposure. Certainly, OPC
mortar mixtures with a higher w/b clearly show the mitigat-
ing effect of sulfate attack because of the presence of chlor-
ide.
5. Conclusions
In this paper, results are reported that demonstrate that the
presence of chloride ions during sulfate attack play a posi-
tive role on the performance of OPC mortar specimens, es-
pecially with a higher w/b.
This may be due to following reasons: (i) the increased
solubility of ettringite and gypsum in the chloride-bearing
solution can reduce the amount of sulfate products formed
in cement paste, (ii) the ettringite formed in the sulfate–
chloride solution is less expansive, and (iii) the higher diffu-
sivity of chloride ions compared with sulfate ions permits
the chloride ions to react with C3A in cement paste to form
Friedel’s salt, which will reduce the conventional sulfate at-
tack.
Another important observation is that the use of SF, even
at a low replacement level, has a beneficial effect in terms
of controlling the surface damage, compressive strength
loss, and expansion of plain mortar specimens due to its
strong pozzolanic reaction and the consequently smaller
amount of Ca(OH)2. Because of the better resistance of SF
to sulfate attack, the mitigation effect due to the chloride-
bearing solution was not significant, even after 510 d of ex-
posure.
References
ASTM. 2002. Standard test method for air content of hydraulic ce-
ment mortar. ASTM standard C185-02. American Society for
Testing and Materials, West Conshohocken, Pa.
ASTM. 2003. Standard specification for flow table for use in tests
of hydraulic cement. ASTM standard C230-03. American So-
ciety for Testing and Materials, West Conshohocken, Pa.
ASTM. 2004. Standard test method for length change of hydraulic-
cement mortars exposed to a sulfate solution. ASTM standard
C1012-04. American Society for Testing and Materials, West
Conshohocken, Pa.
ASTM. 2007. Standard specification for Portland cement. ASTM
standard C150-07. American Society for Testing and Materials,
West Conshohocken, Pa.
Akoz, F., Turker, F., Koral, S., and Yuzer, N. 1995. Effects of so-
dium sulfate concentration on the sulfate resistance of mortars
with and without silica fume. Cement and Concrete Research,
25(6): 1360–1368. doi:10.1016/0008-8846(95)00128-Y.
Al-Amoudi, O.S.B. 1998. Sulfate attack and reinforcement corro-
sion in plain and blended cements exposed to sulfate environ-
ments. Building and Environment, 33(1): 53–61. doi:10.1016/
S0360-1323(97)00022-X.
Al-Amoudi, O.S.B., and Maslehuddin, M. 1993. The effect of
chloride and sulfate ions on reinforcement corrosion. Cement
and Concrete Research, 23(1): 139–146. doi:10.1016/0008-
8846(93)90144-X.
Al-Amoudi, O.S.B., Rasheeduzzafar, Maslehuddin, M., and Abdul-
jauwad, S.N. 1994. Influence of chloride ions on sulphate dete-
rioration in plain and blended cements. Magazine of Concrete
Research, 46(167): 113–123.
Al-Dulaijan, S.U., Maslehuddin, M., Al-Zahrani, M.M., Sharif,
A.M., Shameem, M., and Ibrahim, M. 2003. Sulfate resistance
of plain and blended cements exposed to varing concentrations
of sodium sulfate. Cement and Concrete Composites, 25(4–5):
429–437. doi:10.1016/S0958-9465(02)00083-5.
Ampadu, K.O., and Torii, K. 2002. Chloride ingress and steel cor-
rosion in cement mortars incorporating low-quality fly ashes.
Cement and Concrete Research, 32(6): 893–901. doi:10.1016/
S0008-8846(02)00721-4.
Brown, P., and Hooton, R.D. 2002. Ettringite and thaumasite for-
mation in laboratory concretes prepared using sulfate-resisting
cements. Cement and Concrete Composites, 24(3–4): 361–370.
doi:10.1016/S0958-9465(01)00088-9.
Cohen, M.D., and Bentur, A. 1988. Durability of Portland cement-
silica fume pastes in magnesium sulfate and sodium sulfate solu-
tions. ACI Materials Journal, 85(3): 148–157.
Czerni, W. 1980. Cement chemistry and physics for civil engineers.
2nd English ed. Foreign Publication Inc., New York.
Harrison, W.H. 1990. Effect of chloride in mix ingredients on sul-
phate resistance of concrete. Magazine of Concrete Research,
42(152): 113–126.
Hartshorn, S.A., Sharp, J.H., and Swamy, R.N. 2002. The thauma-
site form of sulfate attack in Portland-limestone cement mortars
stored in magnesium sulfate solution. Cement and Concrete
Composites, 24(3–4): 351–359. doi:10.1016/S0958-9465(01)
00087-7.
Hooton, R.D. 1993. Influence of silica fume replacement of cement
on physical properties and resistance to sulfate attack, freezing
and thawing, and alkali-silica reactivity. ACI Materials Journal,
90(2): 143–151.
Khatri, R.P., Sirivivatnanon, V., and Yang, J.L. 1997. Role of per-
meability in sulphate attack. Cement and concrete research,
27(8): 1179–1189.
Lawrence, C.D. 1992. The influence of binder types on sulfate at-
tack. Cement and Concrete Research, 22(6): 1047–1058. doi:10.
1016/0008-8846(92)90035-T.
Lea, F.M. 1970. The chemistry of cement and concrete. 3rd ed. Ed-
ward Arnold, London.
Lee, S.T., Moon, H.Y., and Swamy, R.N. 2005. Sulfate attack and
role of silica fume in resisting strength loss. Cement and Con-
crete Composites, 27(1): 65–76. doi:10.1016/j.cemconcomp.
2003.11.003.
Mangat, P.S., and Khatib, J.M. 1995. Influence of fly ash, silica
fume, and slag on sulfate resistance of concrete. ACI Materials
Journal, 92(5): 542–552.
Mehta, P.K. 1985. Studies on the chemical resistance of low water/
cement ratio concrete. Cement and Concrete Research, 15(6):
969–978. doi:10.1016/0008-8846(85)90087-0.
Moon, H.Y., Lee, S.T., and Kim, H.S. 2001. The selection of effec-
tive Korean cement for sulfate environments. In Proceeding of
the 3rd International Conference on Concrete Under Severe Con-
ditions, CONSEC’01, Vancouver, B.C., 18–20 June 2001. The
University of British Columbia, Vancouver, B.C. pp. 349–356.
Moon, H.Y., Lee, S.T., Kim, H.S., and Kim, S.S. 2002. Evaluation on
Lee et al. 1219
#2008 NRC Canada
the resistance of Portland cement mortars due to chemical attack.
Journal of the Korean Society of Civil Engineers, 22: 181–191.
Ogawa, K., and Roy, D.M. 1982. C4A3S hydration, ettringite for-
mation, and its expansion mechanism. III. Effect of CaO,
NaOH and NaCl. Cement and Concrete Research, 12(2): 247–
256. doi:10.1016/0008-8846(82)90011-4.
Poupard, O., L’Hostis, V., Catinaud, S., and Petre-Lazar, I. 2006.
Corrosion damage diagosis of a resinforced concrete beam after
40 years natural exposure in marine environment. Cement and
Concrete Research, 36(3): 504–520. doi:10.1016/j.cemconres.
2005.11.004.
Rasheeduzzafar, Dakhil, F.H., Al-Gahtani, A.S., Al-Saadoun, S.S.,
and Bader, M.A. 1990. Influence of cement composition on the
corrosion of reinforcement and sulfate resistance of concrete.
ACI Materials Journal, 87(2): 114–122.
Taylor, H.F.W. 1997. Cement chemistry. 2nd ed. Thomas Telford,
London.
Tian, B., and Cohen, M.D. 2000. Does gypsum formation during
sulfate attack on concrete lead to expansion? Cement and Con-
crete Research, 30(1): 117–123. doi:10.1016/S0008-8846(99)
00211-2.
Turker, F., Akoz, F., Koral, S., and Yuzer, N. 1997. Effects of
magnesium sulfate concentration on the sulfate resistance of
mortars with silica fume. Cement and Concrete Research, 27(2):
205–214. doi:10.1016/S0008-8846(97)00009-4.
Wee, T.H., Suryavanshi, A.K., Wong, S.F., and Anisur Rahman,
A.K.M. 2000. Sulfate resistance of concrete containing mineral
admixtures. ACI Materials Journal, 97(5): 536–549.
1220 Can. J. Civ. Eng. Vol. 35, 2008
#2008 NRC Canada
