Cement & Concrete Composites 25 (2003) 363–369

[Link]/locate/cemconcomp

Sulfate resistance of high-performance concrete

M.J. Shannag *, Hussein A. Shaia
Department of Civil Engineering, Jordan University of Science & Technology, P.O. Box 3030, Irbid 22110, Jordan
Received 5 November 2001; accepted 16 April 2002

Abstract
High-performance concrete mixes containing various proportions of natural pozzolan and silica fume (up to 15% by weight of
cement) were prepared and stored in sodium and magnesium sulfate solutions, in Dead Sea and Red Sea waters. The progressive
deterioration and the relative sulfate resistance of these mixes were evaluated through visual observations, ultrasonic pulse velocity
measurements, and relative strength determinations. The investigation indicated that the concrete mix containing 15% natural
pozzolan, and 15% silica fume showed the best protection in sulfates solutions and sea waters. It retained more than 65% of its
strength after one year of storage in sulfates solutions and sea waters. The superior resistance of that mix against sulfate attack is
attributed to the pore refinement process and further densification of the transition zone occurring due to the conversion of lime
forming from the hydration of cement into additional binding material through lime-pozzolan reaction. This investigation rec-
ommends the use of silica fume in combination with natural pozzolan for better performance in severe sulfate environments.
Ó 2002 Elsevier Science Ltd. All rights reserved.

Keywords: High-performance concrete; Natural pozzolan; Silica fume; Portland cement mortar

1. Introduction efficient use of portland cement, which conserves natural

resources and energy. From a structural design stand-
High-performance concrete is defined as concrete that point, high-performance concrete allows more slender
meets special performance and uniformity requirements structural elements with greater rigidity (higher modulus
that cannot always be achieved routinely by using con- of elasticity) for smaller deflection, less creep, and a high
ventional materials and normal mixing, placing, and MPa/$ ratio. Unfortunately, very few material building
curing practices. Thus, a high-performance concrete is a codes have provisions for high-performance concrete
concrete in which certain characteristics are developed at present [1].
for a particular application and environment. For ex- One of the main causes of deterioration in concrete
ample, concrete that provides substantially improved structures is the corrosion of concrete due to its exposure
durability under severe service conditions, extraordinary to harmful chemicals that may be found in nature such
properties at earlier ages, or substantially enhanced as in some ground waters, industrial effluents and sea
mechanical properties are potential HPCs. These con- waters. The most aggressive chemicals that affect the long
cretes may contain materials such as fly ash, silica fume, term durability of concrete structures are the chlorides
ground granulated slags, natural pozzolan, fibers, and sulfates. The chloride dissolved in waters increase the
chemical admixtures, and other materials, individually rate of leaching of portlandite and thus increases the
or in various combinations [1–4]. porosity of concrete, and leads to loss of stiffness and
Engineers are making increasing use of high-perfor- strength. Calcium, sodium, magnesium, and ammonium
mance concrete because of its availability, greater per- sulfates are––in increasing order of hazard––harmful to
formance, including the highest possible durability, and concrete as they react with hydrated cement paste leading
to expansion, cracking, spalling and loss of strength [5].
The rate at which the hardened cement paste is de-
*
Corresponding author. Tel.: +962-27095111x22076; fax: +962-
teriorated due to the exposure to harmful chemicals
27095018. depends mainly on the concentration of the chemicals in
E-mail address: jshannag@[Link] (M.J. Shannag). water, the time of exposure and the chemical resistance
0958-9465/03/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 9 5 8 - 9 4 6 5 ( 0 2 ) 0 0 0 4 9 - 5
364 M.J. Shannag, H.A. Shaia / Cement & Concrete Composites 25 (2003) 363–369

of concrete. Extensive investigations have been carried Table 1

out on the use of fly ash and silica fume in concrete Characteristics of cements, natural pozzolan, and silica fume used in
this study
during the past two decades and have consequently led
to their widespread application in the construction in- Oxide (%) Natural Silica Cement Pozzolanic
pozzolan fume (Type I ) cement
dustry [6–11]. Many national standards also exist which
SiO2 40.8 93 20.9 22.3
determine the degree of attack, and this primarily on the
Al2 O3 12.8 0.4 5.6 6.0
basis of the concentration of the aggressive substances. Fe2 O3 10.5 1.2 3.1 4.0
However, the chemical resistance of high-performance CaO 11.8 0.2 62.7 55.8
concrete using silica fume, fly ash, natural pozzolan, and MgO 9.1 1.2 2.2 4.8
superplasticizers are issues which have not yet received Na2 O 2.3 0.1 0.2 0.4
K2 O 1.1 1.1 0.8 0.8
sufficient attention from the research community [10].
SO3 0.1 0.3 2.9 2.7
Hence, this investigation was undertaken to study and
define a better concrete mix, containing local natural Loss on 9.30 0.75 1.30 2.48
ignition (%)
pozzolan and silica fume which can produce the maxi-
Specific gravity 2.68 2.10 3.15 3.12
mum protection of the concrete in sulfate environment. Specific surface 600 20,000a 300 434
The present work examines the progressive deterioration (m2 /kg)
of concrete mixtures containing various combinations a
The specific surface of silica fume is measured by the nitrogen
of natural pozzolan and silica fume due to exposure adsorption technique.
to sulfate solutions and sea waters.

and silica fume. The laboratory program conducted in

2. Experimental investigation
this investigation focused on five basic mixes in which
silica fume and natural pozzolan contents were varied
An experimental program was designed to produce a
between 0 and 15% (by weight of cement). For com-
high-performance concrete mix for the best protection in
parison purposes, one of these mixes (Mix 2) was de-
sulfate environment by adding several combinations of
signed using commercially produced portland pozzolana
a local natural pozzolan and silica fume. The materials
cement which contains approximately 15% natural po-
used and the experimental procedures are described in
zzolan as an interground with cement clinker. The de-
the following sections.
tails of these mixes are given in Table 2. All the concrete
mixes were prepared with superplasticizer at a dosage of
2.1. Materials
approximately 2 l per 100 Kg of the total cementitious
materials. The concretes made were of similar work-
Locally available ordinary portland cement (ASTM
ability and water to cementitious materials ratio. They
Type I) and portland pozzolana cement complying with
consisted of approximately 400 Kg/m3 of portland ce-
Jordanian specification JSS 219 were used. The natural
ment (Type I) with the addition of 0, 20, 40, and 60
pozzolan was a blend of certain volcanic tuffs from Tal
Kg/m3 of silica fume and natural pozzolan (0%, 5%,
Rimah region of north eastern Jordan and the same
10%, 15% by weight of cement).
material used in the production of portland pozzolana
cement. The Silica fume was in powder form with an
average of 93% silicon dioxide. The chemical composi- 2.3. Mixing and casting
tion and some physical characteristics of these materials
are given in Table 1. The superplasticizer used was a The concrete mixes were prepared using a tilting
naphthalene formaldehyde sulfonated superplasticizer drum mixer of 0.15 cubic meter capacity. The interior of
with 41% solids content and a specific gravity of 1.21. the drum was initially washed with water to prevent
The superplasticizer was incorporated in all mixes and absorption. The coarse and medium aggregate fractions
the content was adjusted slightly for some mixes to were mixed first, followed by the cement, part of the
maintain the same workability. The coarse and medium- sand, and the water containing the required amount of
size aggregates were crushed limestone with maximum superplasticizer. The final mixing stage involved the
size of 10 and 5 mm, respectively. Natural silica sand addition of the natural pozzolan and silica fume, and the
with a fineness modulus of 2.66 was used for making remaining sand. One-forth of the superplasticizer was
mortar and concrete mixes. always retained to be added during the last 3 min of the
mixing period. The concrete prepared was poured in
2.2. Mix proportions steel molds which were oiled and placed on the vibration
table, and vibrated at low speed. After each mold was
The objective was to produce high-performance and properly filled the vibration speed was increased to en-
workable concretes containing local natural pozzolan sure good compaction.
 M.J. Shannag, H.A. Shaia / Cement & Concrete Composites 25 (2003) 363–369 365

Table 2
Concrete mix quantities (kg/m3 )
Mix designation W=ðC þ SF þ NPÞ CA (10 mm) CA (5 mm) S C NP SF SP (L) W
1 0.35 995.0 115.0 740.0 400.0 – – 10 140
2a 0.35 992.0 112.0 740.0 400.0 – – 10 140
3 0.35 955.0 106.1 707.4 400.0 20 20 10 154
4 0.35 913.1 101.5 676.6 400.0 40 40 10 168
5 0.35 871.6 96.8 645.6 400.0 60 60 10 182
CA: coarse aggregate, S: sand, C: cement, NP: natural pozzolan, SF: silica fume, SP: superplasticizer, W: water.
a
Mix 2 only contains commercially produced portland pozzolana cement; Mixes 1,3,4,5 contain ordinary portland cement.

Table 3
Compressive strength of concrete mixes
Mix designation 28-day strength Compressive strengths at 360 days (MPa)
Na2 SO4 MgSO4 Red Sea Dead Sea Tap water
1 54.8 37.2 (0.56) 28.4 (0.43) 44.5 (0.67) 25.5 (0.38) 66.5 (100)
2 67.1 65 (0.85) 35.4 (0.46) 72.8 (0.95) 42 (0.55) 76.5 (100)
3 71.9 60 (0.70) 41.5 (0.48) 75 (0.88) 48.5 (0.57) 85.7 (100)
4 79.5 76 (0.77) 49.8 (0.51) 73 (0.74) 58.5 (0.60) 98.2 (100)
5 71.8 71 (0.93) 49.8 (0.65) 57.5 (0.75) 50.2 (0.66) 76.5 (100)

2.4. Curing and testing 3. Results and discussion

After casting, the specimens were covered with wet The progressive deterioration of the specimens sub-
burlap and stored in the laboratory at 23 °C and 65% merged in sulfate solutions, sea and tap waters was
relative humidity for 24 h and then demoulded and monitored periodically by observing the visual changes
placed under water. Each specimen was labeled as to the in the cubes such as cracking, abrasions and spallings.
date of casting, mix used and serial number. The spec- Relative strength determination with respect to strengths
imens were then taken out of water a day before testing of control specimens stored in tap water were conducted
and dried in air. Some of the specimens were used for at final stages of the test. Since the strength tests meant
compressive strength determination, and the rest were the termination of the experiment, they were delayed as
kept in sulfate solutions and sea waters for sulfate re- much as possible in this investigation. The compressive
sistance evaluation. strength of the cubes was determined after one year of
submersion in sulfates solutions and sea waters. The
results are presented in Table 3. For comparison, rela-
2.4.1. Compressive strength determination tive strengths as fractions of control specimen values
The compressive strength of the concrete mixes was stored in tap water were also calculated and indicated
measured on 10-cm cubes that were cast and cured in inside parantheses beside the strength figures. Ultrasonic
steel molds, and tested using Avery Denison compres- pulse velocity (UPV) measurements were also taken at
sion machine of 2000 kN capacity at a rate of 150 kN/ the beginning of the test and periodically afterwords for
min following ASTM C 39 standards [12]. The com- all mixes.
pressive strengths were determined after 28 days, and
one year of water curing. The strength results listed in
Table 3 are the average of three specimens.
4. Exposure to sulfates

2.4.2. Resistance to sulfate attack Sulfates react with various phases of hydrated cement
To evaluate the resistance of the proposed mixes to paste leading to expansion, cracking, and spalling. The
sulfates and aggressive chemicals, 10-cm cubic speci- sulfate attack is generally attributed to the formation of
mens were prepared following the guide lines of ASTM expansive ettringite and gypsum, which are known to
C 39. After 28-day moist curing, the specimens were precipitate by through-solution mechanism [5]. Investi-
submerged in 20% magnesium sulfate solution, 20% gations by Mehta, however, suggested that the sulfate
sodium sulfate solution, sea waters and Tap water. attack is partially due to the cement paste losing its
Sulfates concentrations were kept relatively high to ac- stiffness when exposed to sulfate rich environment [5].
celerate deterioration and also to compensate for small As seen from Table 3 and Fig. 1, after one year of
variations in concentration during the test. storage in sulfate solutions (Na2 SO4 , and MgSO4 ), the
366 M.J. Shannag, H.A. Shaia / Cement & Concrete Composites 25 (2003) 363–369

The mixes containing various combinations of natu-

ral pozzolan and silica fume (Mix 3, 4, and 5), exhibited
good resistance to sulfates. Even though, the relative
strength of these mixes were reduced significantly as
shown in Table 3, still they remained within the strength
range of high-performance concretes reported in the
literature [13]. The periodic observation indicate that all
the cubes of the pozzolanic mixes performed well in
sulfate solutions. None of the cubes have shown signif-
icant deterioration or spalling even at the corners as
shown in Figs. 2 and 3. The mix containing a commer-
cially produced portland pozzolana cement (Mix 2)
Fig. 1. Relative strength of concrete mixes stored in sulfate soutions
showed better sulfate resistance compared to ordinary
and sea waters with respect to concrete mix stored in tap water. portland cement (Mix 1). Among all the mixes investi-
gated, Mix 5, that contains 15% natural pozzolan and
15% silica fume showed the best performance in sulfate
mix containing ordinary portland cement only (Mix 1) environment.
showed a sharp reduction in compressive strength The results presented in Table 3. clearly show that
compared to those stored in tap water. The periodic magnesium sulfates has a more damaging effect than
visual observation indicate that signs of deterioration sodium sulfates; this is consistent with the data available
due to sulfate attack occurred near the corners of the in the literature [14], because MgSO4 leads to the de-
cubes submerged in the previous solutions as shown in composition of the hydrated calcium silicates as well as
Figs. 2 and 3. The corners were the first to spall off of Ca(OH)2 and of hydrated C3 A; hydrated magnesium
because near the corners the intrusion will be from the silicate is eventually formed and it has no binding
two adjacent faces of the cube. properties [15].
The superior resistance of the mixtures containing
pozzolanic products against sulfate attack is attributed
to the pore refinement process occurring due to the
conversion of lime forming from the hydration of
cement into additional binding material through lime-
pozzolan reaction. In addition to the pozzolanic reac-
tion, the filler action due to the finer particle size of silica
fume (0.1–0.2 lm) further densifies the pore structure to
enhance the resistance to sulfate attack [16]. Further-
more these pozzolanic materials play an important role
in improving the aggregate-paste bond through the
densification of the transition zone and formation of
more calcium silicate hydrates [9].

Fig. 2. Relative behavior of concrete mixes stored in Na2 SO4 solution.

5. Exposure to sea waters

The Dead Sea water was obtained from the north-

eastern shores of the Dead Sea whereas the Red Sea
water was obtained from the shores near Aqaba. The
waters were taken from areas away from possible pol-
lution with organic or foreign matter. The average
compositions of the sea waters are shown in Table 4,
[15]. Among the sea waters Dead Sea water has unusual
composition as seen from Table 4, with a total salinity
around 27%, about half of it being magnesium chloride
followed by sodium and calcium chlorides. The Red Sea
water has a composition similar to that of Mediterra-
nean with slightly more total salinity reaching to about
Fig. 3. Relative behavior of concrete mixes stored in MgSO4 solution. 4.1%, [15].
 M.J. Shannag, H.A. Shaia / Cement & Concrete Composites 25 (2003) 363–369 367

Table 4
Composition of sea waters
Principal ions (g/l)
Na K Mg Ca Cl Br SO4 HCO3 pH
Dead Sea 33.5 6.3 34.5 13 180.0 4.1 0.9 0.25 6.3
Red Sea 12.2 0.44 1.88 0.51 22.7 0.07 3.16 0.15 8.2

Considering the relative strength results listed in Ta-

ble 3 and Fig. 1, it is observed that after one year of
storage in Dead Sea water, the mix containing ordinary
portland cement only (Mix 1) lost more than 60% of its
compressive strength compared to those stored in tap
water. The visual observations over the specimens stored
in Dead Sea water revealed severe signs of deterioration
of Mix 1 as shown in Figs. 4 and 5. The corners were
completely destroyed, and the cubes lost more than 50%
of their weight.
The mixes containing various combinations of
natural pozzolan and silica fume (Mix 2, 3, 4, and 5),
exhibited good performance and retained somehow Fig. 6. Relative behavior of concrete mixes stored in Red Sea water.
appreciable proportions of their strength when sub-
merged in Dead Sea water for one year. The visual ob-
servations over the specimens did not reveal any Sea water, followed by Mix 4, 3 and 2. The relatively
cracking as shown in Fig. 4. Among all the mixes in- superior behavior of the mixes containing natural po-
vestigated, Mix 5, showed the best performance in Dead zzolan and silica fume could be explained in a way
similar to that discussed previously in case of sulfate
solutions.
All the mixes behaved in a similar manner inside the
Red Sea and did not show any significant deterioration
during the test period of one year as shown in Fig. 6.
Mix 2 (commercially produced portland pozzolana ce-
ment that contains 15% natural pozzolan interground
with the cement during manufacturing) showed the best
performance in Red Sea water. It lost about 5% of its
strength at the end of the test period. Some other re-
searchers [15] have reported a weekened sulfate attack
from the Red Sea water due to increased solubility of
ettringite and gypsum.

Fig. 4. Relative behavior of concrete mixes stored in Dead Sea water.

6. Assessment of concrete deterioration using ultrasonic
pulse velocity tests

UPV tests are commonly used in attempting to define

the extent and magnitude of deterioration resulting from
fire, mechanical or chemical attack. In this investigation
UPV measurements were taken monthly for all mixes
stored in sulfate solutions, tap and sea waters. Unfor-
tunately some of these measurements did not correlate
well with the condition of the test. Some reduction in
UPV values was noticed only after the specimens were
considerably deteriorated. Therefore, this method may
not be reliable for comparative studies. However most
of the UPV measurements taken in this investigation
Fig. 5. Deterioration of Mix 1 in Dead Sea water. were rather consistent with the strength results obtained
368 M.J. Shannag, H.A. Shaia / Cement & Concrete Composites 25 (2003) 363–369

previously and agreed with the values given in the lit-

erature for high strength concrete [14]. After one year of
storage in sulfate solutions, and sea waters, all the mixes
showed a significant decrease (about 30%) in UPV val-
ues compared to those stored in tap water as seen from
Figs. 7–11. Within the range of UPV values obtained in

Fig. 10. UPV for mixes stored in Red Sea water.

Fig. 7. UPV for mixes stored in Na2 SO4 soution.

Fig. 11. UPV for concrete mixes stored in Tap water.

this investigation (3.75–5.25 km/s) it is difficult to draw

comparisons from the figures shown.

Fig. 8. UPV for mixes stored in MgSO4 solution. 7. Conclusions

Based on the results of this study it can be concluded

that:

1. After one year immersion in sulfates solutions and

sea waters, the concrete mix containing a combina-
tion of 15% silica fume, and 15% natural pozzolan
(by weight of cement) showed a maximum protection
against sulfate attack compared to those investigated
in the present study.
2. High-performance concretes that contain various
combinations of silica fume and natural pozzolan,
can provide good balance between strength and dura-
bility, and can be recommended for use in concrete
industry.
3. The present research showed that the high strength
Fig. 9. UPV for mixes stored in Dead Sea water. concrete mix made using ordinary portland cement
 M.J. Shannag, H.A. Shaia / Cement & Concrete Composites 25 (2003) 363–369 369

only did not perform satisfactorily in sulfate solutions [5] Wee TH, Suryavanshi AK, Wong SF, Rahman AK. Sulfate
and sea waters. resistance of concrete containing mineral admixtures. ACI Mater
J 2000;97:536–49.
4. The sulfate resistance of the concretes stored in Dead [6] ACI committee 226, silica fume in concrete. ACI Mater J 1987;84:
Sea water deserves further investigation, due to the 158–66.
unique composition of its water. [7] ACI committee 232, natural pozzolans in concrete. ACI Mater J
1994;91:410–26.
[8] Malhotra VM, Carette GG, Sivasundaram V. Role of silica fume
Acknowledgements in concrete: a review. In: Malhotra VM, editor. Proceedings
International Conference on Advances in Concrete Technology.
Ottawa, Canada: CANMET; 1992.
The author acknowledge the support received from [9] Goldman A, Bentur A. Bond effects in high strength silica fume
the Deanship of Scientific Research at Jordan University concretes. ACI Mater J 1989;86:440–7.
of Science and Technology. [10] Tamimi AK. High-performance concrete mix for an optimum
protection in acidic conditions. Mater Struct 1997;30:188–91.
[11] Shannag M, Brincker R, Hansen W. Interfacial fiber-matrix
References properties of high strength mortar (150 MPa) from fiber pullout.
ACI Mater J 1996;93:480–6.
[1] Aitcin P-Claude, Baalbaki M. In: Canadian experience in [12] STM, Annual book of ASTM standards. vol. 4.02, concrete and
producing and testing HPC, ACI international workshop on high mineral aggregates. Am Soc Testing Mater, 1991.
performance concrete, Bangkok, Thailand. 1994. p. 295–308. [13] Haque MN, Kayali O. Properties of high strength concrete using a
[2] Naaman AE. SIFCON: tailored properties for structural perfor- fine fly ash. Cement and Concrete Res 1998;28:1445–52.
mance, high performance fiber reinforced cement composites. In: [14] Neville AM. Properties of concrete. Second ed. New York: John
Proceedings of the International RILEM/ACI Workshop. 1992. Wiley and Sons Inc; 1997.
p. 18–38. [15] Yeginobali A, Al-Hadrami N. Sulfate resistance of Jordanian
[3] Shah S, Ahmad A. High performance concretes and applications. cements. In: Symposium on low cost housing for developing
England: Edward Arnold; 1994. countries, Bangkok, Thailand. 1987. p. 141–9.
[4] Nawy EG. Fundamentals of high strength high performance [16] Popovics S. Portland cement-fly ash, silica fume systems in
concrete. England: Longman Group Limited; 1996. concrete. Adv Cement Based Mater J 1993;1:83–91.