Materials and Structures (2012) 45:977994

DOI 10.1617/s11527-011-9811-0

ORIGINAL ARTICLE

Resistance of self-consolidating concrete to ammonium

sulphate attack
M. T. Bassuoni M. L. Nehdi

Received: 28 June 2009 / Accepted: 4 December 2011 / Published online: 3 January 2012
 RILEM 2011

Abstract With the growing use of self-consolidating

concrete (SCC) in various infrastructure applications,
it has become necessary to establish reliable data on its
long-term durability since its mixture design, rheology
and consolidation are different than that of normal
concrete. Several applications of SCC involve its
exposure to chemical attack, particularly to sulphaterich media such as wastewater treatment facilities,
industrial and agricultural zones which also encompass ammonium cations. Thereby, this study aims at
investigating the resistance of a wide scope of SCC
mixture designs to ammonium sulphate attack. Specimens from 23 SCC mixtures were continuously
immersed in a high concentration ammonium sulphate
solution with controlled pH (6.08.0) up to 54 weeks.
The main test variables included the type of binder
(single, binary, ternary and quaternary), air-entrainment, sand-to-total aggregates mass ratio, and the
inclusion of fibre reinforcement (single and hybrid).
The SCC specimens showed variable degrees of
deterioration after 54 weeks, indicating different
modes of degradation. The study highlights the role

M. T. Bassuoni
Department of Civil Engineering, University of Manitoba,
Winnipeg, MB R3T 5V6, Canada
M. L. Nehdi (&)
Department of Civil and Environmental Engineering,
University of Western Ontario, London, ON N6A 5B9,
Canada
e-mail: mnehdi@eng.uwo.ca

of ternary and quaternary binders in improving the

resistance of SCC to ammonium sulphate attack. It is
also emphasized that multiple performance indicators
are needed to achieve a reliable assessment of cementbased materials under ammonium sulphate attack.
Keywords Self-consolidating concrete  Sulphate
attack  Ammonium sulphate  Composite binders 
Fibre reinforcement

1 Introduction
1.1 Background
Self-consolidating concrete (SCC) readily flows and
consolidates under its own weight with little or no
vibration. It is particularly suitable for precast applications, hard-to-reach areas and heavily reinforced
sections. The mixture design of SCC usually incorporates an efficient superplasticizer, relatively high
amounts of fine materials, low water-to-powder ratio,
and controlled proportions of coarse aggregates.
Viscosity-modifying admixtures (VMAs) can be used
in SCC to enhance stability of the flowable material
and inhibit segregation and bleeding. Hence, the fresh
properties and rheological characteristics of SCC are
different than that of normal concrete, but both SCC
and normal concrete may exhibit comparable mechanical properties if designed for similar strength grades
[1]. Yet, due to the difference in mixture design,

978

Materials and Structures (2012) 45:977994

placement and consolidation techniques, the durability

of SCC may be different than that of normal concrete,
and still needs thorough investigation [1, 2].
Dissolved sulphate salts existing in nature or from
industrial and agricultural sources enter into chemical
reactions with cement-based materials causing expansion, cracking and spalling and/or softening and
disintegration. Hence, the action of sulphates on
concrete has been a key durability issue, and a subject
of extensive investigation for many decades [3].
Attack of concrete by sodium sulphate (Na2SO4) and
magnesium sulphate (MgSO4) has been extensively
investigated [3, 4]. Yet, the exact mechanism of
expansion and the role of gypsum and ettringite in the
deterioration process remain subjects of controversy
[4]. Sulphate resistant cements with low C3A content
(less than 8%) and cements incorporating supplementary cementitious materials (SCMs) that consume
portlandite (CH) in pozzolanic reactions typically
have high resistance, but are not completely immune
to this type of attack [3, 4].
Despite that there is limited data on the effect of
ammonium sulphate ((NH4)2SO4) on concrete, there is
a general consensus that it has more detrimental
consequences than that of sodium and magnesium
sulphates [36]. In addition to the attack by sulphate
ions, there is a deleterious action caused by ammonium NH
4 ions. Portlandite is highly soluble in
ammonium solutions, and there is a neutralization
action of ammonium cations due to acid hydrolysis
with the cementitious matrix [6]:

NH
4 OH ! NH3 H2 O

Corrosion of concrete due to ammonium sulphate

has a combined acidsulphate attack nature and can be
generally characterized by the following reactions [4]:
CaOH2 NH4 2 SO4 ! CaSO4 2H2 O 2NH3
2
xCaSiO2 aq xNH4 2 SO4 xH2 O
! SiO2 aq xCaSO4 2H2 O 2xNH3

Consumption of CH in Eq. 2 to form gypsum leads

to a reduction in pH and an increase in porosity of the
concrete surface in contact with the ammonium
sulphate solution [7]. After or prior to the complete
depletion of CH, decalcification of calcium silicate
hydrate (CSH) in Eq. 3 occurs due to the reduction
of pH in the pore solution. This reaction starts with a

reduction of the calcium-to-silicate ratio (C/S) of the

CSH structure with its progressive loss of cohesion
and eventually its disintegration [4]. Also, formation
of gypsum and to a less extent ettringite along with
gaseous NH3 in the pore structure can lead to volume
swelling [3, 4, 7].
Literature on the role of SCMs in improving the
resistance of concrete to the attack of ammonium
sulphate is contradictory. While some authors [8, 9]
reported adverse effects, others [10, 11] indicated the
efficacy of SCMs in improving the resistance of concrete
to this type of attack. This discrepancy may be attributed
to different modes of deterioration of various binders
exposed to ammonium sulphate. Yet, this has not been
established and needs to be investigated.
1.2 External sulphate attack on SCC
Nehdi et al. [12] evaluated the sulphate resistance
(ASTM C 1012) of cost-effective SCC prepared with
binary (two-component), ternary (three-component) and
quaternary (four-component) composite cements and
water-to-cementitious materials ratio (w/cm) of 0.38.
Bars of mortar extracted from SCC mixtures were
immersed in a 5% Na2SO4 solution to observe the length
change with time for 9 months. It was observed that
quaternary SCC mixtures made with 50% OPC, 24%
Class F fly ash or slag, and 6% silica fume (SF) or rice
husk ask had the lowest expansion compared to that of
other mixtures [12]. In its state-of-the-art report on the
durability of SCC [13], RILEM Technical Committee
TC 205-DSC emphasized the physical resistance, i.e.
low penetrability, of SCC due to the incorporation of
large amounts of pozzolanic materials and fillers. This
may prolong the initiation time of sulphate attack
compared to ordinary concrete. With respect to the
chemical resistance (type of binder) of SCC, it was found
that the incorporation of large amounts of limestone filler
(LF) in SCC may make it more vulnerable to thaumasite
sulphate attack (TSA), when the conductive conditions
(continual supply of sulphate ions with high moisture
and low temperature) prevail in the field. Limited
information, though, has been given regarding the effect
of mixture design parameters of SCC on its resistance to
chemical attack such as by ammonium sulphate [13].
Several applications of SCC involve its exposure to
chemical attack, particularly to sulphate-rich media
associated with ammonium cations such as wastewater
treatment facilities and industrial and agricultural

Materials and Structures (2012) 45:977994

zones. However, there is dearth of information on the

resistance of SCC, which usually comprises a large
volume fraction of SCMs, to ammonium sulphate
attack. On the other hand, the effect of SCMs on the
resistance of concrete to such a type of chemical attack
has not yet been established. On the basis of multiple
performance indicators, this study aims at investigating the resistance to ammonium sulphate attack of a
variable range of SCC mixtures incorporating single,
binary, ternary and quaternary binders, with and
without fibre reinforcement. The effects of the coarse
aggregate volume fraction and air-entrainment were
also explored. The choice of the mixture design
variables was based on key parameters controlling the
characteristic fresh properties and rheology of SCC. In
addition to their effects on the fresh state of SCC, it is
of equal importance to investigate the influence of
these parameters on the durability of SCC, particularly
under chemical attack.

2 Experimental programme
2.1 Materials
Twenty-three SCC mixtures with a w/cm of 0.38 were
prepared using single, binary, ternary and quaternary
Table 1 Chemical and
physical properties of
cement and SCMs

979

binders. The mixtures were divided into three groups:

Group A (non-air-entrained SCC with sand-to-total
aggregates mass ratio [S/A] of 50%); Group B (airentrained SCC mixtures with S/A of 40 and 60%); and
Group C (air-entrained SCC with fibre reinforcement
and sand-to-total aggregate plus fibres mass ratio of
50%). The binders used included CSA Type 10
(ASTM Type I) ordinary portland cement (OPC),
CSA Type 50 (ASTM Type V) sulphate resistant
portland cement (SRPC), SF, Class F fly ash (FA), slag
(S), and LF. The chemical and physical properties for
these materials are listed in Table 1. Single binders
were 100% OPC (control) or 100% SRPC, binary
binders contained 92% OPC and 8% SF, and ternary
binders contained 50% OPC, 5% SF and 45% S.
Quaternary binders comprised 50% OPC, 15% LF,
20% S and 15% FA, or 50% OPC, 5% SF, 25% S and
20% FA.
The total powder content was kept constant at
470 kg/m3 to provide a relatively high volume fraction
of fine materials (paste volume of 325350 l/m3),
conforming to common SCC mixture design guidelines [14]. The fine aggregate was natural siliceous
sand with a fineness modulus of 2.80, a saturated
surface dry specific gravity of 2.65 and water absorption of 1.5%. Crushed stone with a maximum nominal
size of 19 mm, a saturated surface dry specific gravity

OPC

SRPC

Silica
fume

Fly ash

Slag

Limestone
filler

SiO2 (%)

19.8

22.3

94.0

48.9

35.0

0.3

CaO (%)

63.2

63.8

0.4

3.8

36.1

Al2O3 (%)
Fe2O3 (%)

5.0
2.4

3.5
4.0

0.1
0.1

23.3
14.9

11.2
0.5

MgO (%)

3.3

2.8

0.4

0.7

11.4

K2O (%)

1.2

0.4

0.9

1.7

0.5

SO3 (%)

3.0

2.1

1.3

0.2

3.3

Na2O (%)

0.1

0.1

0.1

0.6

0.5

TiO2 (%)

0.3

0.3

0.6

CaCO3 (%)

99.0

Loss on ignition (%)

2.5

0.9

4.7

0.3

Specific surface area (m2/kg)

410

377

19,530

280

468

3,200

Specific gravity

3.17

3.15

2.12

2.08

2.90

2.70

C3S

61

54

C2S

11

23

C3A

C4AF

14

980

Materials and Structures (2012) 45:977994

mixture, triplicates of 75 mm 9 150 mm concrete

cylinders and 75 mm 9 75 mm 9 285 mm concrete
prisms were prepared without compaction. All specimens were demolded after 24 h and moist cured at
20C and 95% RH for 56 days.
In a similar procedure to the ASTM C 1012
(standard test method for length change of hydrauliccement mortars exposed to a sulphate solution) [16],
specimens were constantly immersed in an aggressive
solution of 5% ammonium sulphate to investigate the
resistance of SCC to such a type of chemical attack.
The exposure lasted for 12 months (54 weeks). The
temperature of the solutions was around 22 2C
(room temperature). Solutions were renewed each
18 weeks, and the pH was controlled at a range of
6.08.0 by titration with diluted sulphuric acid solutions at regular time intervals (5 days), if needed.
Regular stirring was conducted to allow for a homogenous distribution of the solutions. While controlling
the pH of the sulphate solution was not specified in the
ASTM C 1012, the importance of a controlled pH
(6.08.0) of a sulphate solution correlates to field
conditions in which concrete exists in a neutral
environment with continual supply of sulphate ions
[3, 17].
Before exposure, the initial physico-mechanical
properties of the SCC specimens were determined
(Table 4). For the concrete cylinders, the initial mass
and compressive strength (ASTM C 39 standard test
method for compressive strength of cylindrical concrete specimens [18]) at 56 days were recorded to
calculate the mass change versus time of exposure and
the residual compressive (RC) strength of specimens
at the end of exposure to the ammonium sulphate
solution, respectively. For the concrete prisms, the
initial mass, length (ASTM C 1012), dynamic

of 2.68 and water absorption of 0.8% was also used.

A polycarboxylate-based high-range water-reducing
admixture (HRWRA) with a solid content of 43%
was incorporated in all mixtures. A VMA based on
a solution of modified polysaccharide was also
employed to enhance stability of the SCC mixtures.
The dosages of HRWRA and VMA were adjusted to
maintain a slump flow of 650 30 mm and L-box
(3 10 mm bars with 50 mm gaps) ratio (H2/H1) not
less than 0.70 for all mixtures. A multi-component
synthetic air-entraining admixture (AEA) based on tall
oil, was used in groups B and C mixtures to obtain a
fresh air content of 5 1%, while achieving the target
slump flow and L-box values.
In group C mixtures, micro-reinforcement of polypropylene fibrillated fibres with a specific gravity of
0.91 and graded length, complying with ASTM C
1116 Type III 4.1.3, were added at a single dosage of
0.1% by volume. In addition, macro-reinforcement of
crimped steel fibres complying with ASTM A 820
with a specific gravity of 7.85, length of 38 mm, aspect
ratio of 34, and tensile strength of 9661,242 MPa
were used at dosages of 0.4 and 0.5% by volume.
Based on trial mixtures, these fibre shapes, lengths,
and low volumes of micro- and macro-fibre reinforcement have proven adequate for achieving the characteristic flowability and passing ability of SCC with
minimal clustering of fibres.
2.2 Experimental procedures
Constituent materials were mixed in a mechanical
mixer in accordance to ASTM C 192 (standard
practice for making and curing concrete test specimens in the laboratory) [15]. Tables 2 and 3 show the
proportions of the SCC mixtures tested. For each
Table 2 Proportions of binders per cubic meter of concrete
Binder description

Binder
code

Cement
(kg)

100% SRPC

SRPC

470

100% OPC

A1, B1 or C1

92% OPC, 8% SF

A2, B2 or C2

50% OPC, 5% SF, 45% S

Silica fume
(kg)

Slag
(kg)

Fly ash
(kg)

Limestone
filler (kg)

470

430

40

A3, B3 or C3

235

25

210

50% OPC, 15% LF, 20% S, 15% FA

A4, B4 or C4

235

95

70

70

50% OPC, 5% SF, 25% S, 20% FA

A5, B5 or C5

235

25

120

90

SRPC sulphate resistant portland cement; OPC ordinary portland cement; SF silica fume; S slag; LF limestone filler; FA fly ash

Materials and Structures (2012) 45:977994

Table 3 Proportions of
groups A, B and C mixtures
per cubic meter of concrete

Binder
code

981

Mix.
ID

Steel
fibres
(kg)

Polypropylene
fibres (kg)

Fine
aggregate
(kg)

Coarse
aggregate
(kg)

Air-entraining
agent (ml/100 kg
of binder)

SRPC

SRPC

870

870

A1

A1-N-50

870

870

A2

A2-N-50

860

860

A3

A3-N-50

855

855

A4

A4-N-50

845

845

A5

A5-N-50

840

840

B1-A-40

655

1,015

45

B1-A-60

1,015

655

35

B2-A-40

640

990

70

B2-A-60

990

640

50

B3-A-40

640

985

70

B3-A-60

985

640

60

B4

B4-A-40
B4-A-60

625
970

970
625

110
95

B5

B5-A-40

625

965

120

B5-A-60

965

625

100

C1-A-S

40

830

805

35

C1-A-P

825

820

40

Group A

Group B
B1
B2
B3
SRPC 100% sulphate
resistant portland cement;
A1, B1, or C1 100% OPC;
A2, B2, or C2 92%
OPC ? 8% SF; A3, B3, or
C3 50% OPC ? 5%
SF ? 45% S; A4, B4, or C4
50% OPC ? 15%
LF ? 20% S ? 15% FA;
A5, B5, or C5 50%
OPC ? 5% SF ? 25%
S ? 20% FA; N non airentrained; A air-entrained;
50, 40, 60 S/A is 50, 40 and
60%, respectively; S, P,
H steel, polypropylene and
hybrid fibres, respectively

Group C
C1

C1-A-H

30

830

805

40

C2

C2-A-H

30

825

795

60

C3

C3-A-H

30

820

790

65

C4

C4-A-H

30

805

780

100

C5

C5-A-H

30

800

775

105

modulus of elasticity, Ed (ASTM C 215 standard test

method for fundamental transverse, longitudinal, and
torsional resonant frequencies of concrete specimens
[19], calculated as the square of the fundamental
transverse frequency as per the ASTM C 666 guidelines [20]), and flexural strength (ASTM C 78 standard
test method for flexural strength of concrete, using
simple beam with third-point loading [21]) were
recorded to calculate the change in mass, length and
Ed (REd) versus time of exposure, and the residual
flexural (RF) strength of deteriorating specimens at the
end of exposure, respectively.
To analyse the reaction phases, differential scanning calorimetry (DSC) at a heating rate of 10C/min
was conducted on powder samples passing #200 sieve
(75 lm) collected from the surface (015 mm) of
selected specimens. To examine the deterioration of

microstructure, scanning electron microscopy (SEM)

with energy dispersive X-ray (EDX) analyses were
conducted on fracture surfaces of selected specimens.

3 Results
The SCC specimens immersed in the 5% ammonium
sulphate solution showed marked features of damage
(e.g. Fig. 1) during the period of exposure. The
specimens showed various levels of swelling, cracking, softening and disintegration. This indicated the
severe aggression of this exposure, in which the SCC
specimens tested had variable performance depending
on the mixture design parameters. Apparently, the role
of physical resistance (low penetrability) [13] of this
type of high-performance concrete (high powder

982
Table 4 Hardened
properties of groups A, B
and C specimens at 56 days

Materials and Structures (2012) 45:977994

Mix.
ID

Unit
weight
(kg/m3)

Compressive
strength
(MPa)

Flexural
strength
(MPa)

Ed
(GPa)

SRPC

SRPC

2,383

43.69

5.20

43.8

A1

A1-N-50

2,371

47.12

4.99

45.40

A2

A2-N-50

2,388

54.6

5.06

44.02

A3

A3-N-50

2,365

50.18

4.40

41.38

A4

A4-N-50

2,351

41.71

3.91

41.08

A5

A5-N-50

2,335

46.78

4.20

40.04

B1-A-40

2,291

38.91

4.41

41.25

B1-A-60

2,269

35.16

4.05

37.02

B2-A-40

2,284

43.64

4.23

40.94

B2-A-60

2,267

41.43

4.18

39.20

B3-A-40

2,253

39.56

3.76

34.37

B3-A-60

2,246

36.23

3.91

36.95

B4

B4-A-40
B4-A-60

2,259
2,239

32.34
31.18

3.61
3.43

35.07
33.01

B5

B5-A-40

2,218

34.23

3.50

34.04

B5-A-60

2,213

32.39

3.32

36.14

C1-A-S

2,295

38.21

4.10

37.68

C1-A-P

2,273

32.15

4.27

34.79

Binder
code

Group A

Group B
B1
B2
B3

Group C
C1

C1-A-H

2,290

37.44

4.70

39.13

C2

C2-A-H

2,297

42.34

4.65

39.95

C3

C3-A-H

2,276

39.86

4.22

35.86

C4

C4-A-H

2,264

33.41

3.74

36.64

C5

C5-A-H

2,237

36.71

4.08

35.12

content, low water-to-powder ratio and homogenous

microstructure) diminished under the ammonium
sulphate exposure, and other factors controlled the
degradation process.
3.1 Relative dynamic modulus of elasticity (REd)
The results of the relative dynamic modulus of
elasticity (REd) versus time of exposure (54 weeks)
for groups A, B and C specimens are illustrated in
Fig. 2ac, respectively. This parameter should capture
any internal degradation in concrete specimens
exposed to chemical and/or physical damage. Specimens from all mixtures showed a downward trend in
REd, reflecting the progression of damage with time of
exposure. After 54 weeks of immersion in the 5%

(NH4)2SO4 solution, REd for specimens from groups

A, B, and C mixtures ranged from 75 to 88%, from 48
to 87%, and from 71 to 95%, respectively.
For group A specimens, the type of binder did not
seem to have a significant effect on REd results.
Specimens from mixtures with binary binders
(B specimens), ternary binders (T specimens), and
quaternary binders (Q specimens) showed comparable
behaviour over the time of exposure (Fig. 2a) with
REd values of 78, 77, and an average of 76%,
respectively. After 54 weeks of exposure, the control
SCC (A1-N-50) specimens from mixtures prepared
with 100% OPC had REd of 80%, whereas SRPC
specimens from mixtures prepared with 100% SRPC
had a relatively higher REd of 87%. All SCC
specimens in group A had REd values above the

Materials and Structures (2012) 45:977994

983

Figure 2b shows that the type of binder had a

notable effect on REd results. At S/A of 40%, the
control (B1-A-40) and the Q specimens (B4-A-40 and
B5-A-40) showed good performance (REd of at least
85% after 54 weeks). The B specimens (B2-A-40)
showed adverse performance after 32 weeks, while
the T specimens (B3-A-40) showed a high rate of
reduction in REd from the start of exposure, reaching
REd values of 48 and 57%, respectively. At S/A of
60%, only the B specimens (B2-A-60) showed an
adverse performance after 41 weeks and had REd of
55% after 54 weeks.
Group C specimens yielded relatively higher REd
values in the range of 7195% indicating lower
degradation, irrespective of the binder type (Fig. 2c).
This suggests that using fibres generally reduced the
internal degradation of specimens. For mixtures
prepared with 100% OPC, the incorporation of 0.5%
steel fibres (C1-A-S) achieved higher REd than that of
adding 0.1% polypropylene fibres (C1-A-P). The
addition of hybrid fibres (0.4% steel and 0.1%

Fig. 1 Visual characteristics of some SCC specimens after the

ammonium sulphate exposure

60% threshold recommended by the ASTM C 666

guidelines for durable concrete undergoing severe
freezing-thawing cycles.

130

(a)

120

SRPC
A1-N-50
A2-N-50
A3-N-50
A4-N-50
A5-N-50

110

REd (%)

100

B1-A-40
B2-A-40
B3-A-40
B4-A40
B5-A-40

110

90
80

B1-A-60
B2-A-60
B3-A-60
B4-A-60
B5-A-60

100
90
80

70

70

60

60

50

50

40

40
0

10

15

20

25

30

35

40

45

50

55

130

(c)

120

10

15

20

25

C1-A-S
C1-A-P
C1-A-H
C2-A-H
C3-A-H
C4-A-H
C5-A-H

110
100
90
80
70
60
50
40
0

10

15

20

25

30

35

30

35

Time (Weeks)

Time (Weeks)

REd (%)

(b)

120

RE d (%)

130

40

45

50

55

Time (Weeks)

Fig. 2 Relative dynamic modulus of elasticity versus time for groups: a A, b B, and c C specimens

40

45

50

55

984

Materials and Structures (2012) 45:977994

50) in group A. Conforming to the results of REd, the B

specimens (B2-A-40 and B2-A-60) had the highest
expansion values in group B of 0.54 and 0.38%,
respectively. The effect of quaternary binder blends on
reducing expansion was evident at both S/A levels. For
instance, the Q specimens (B5-A-60) had a reduction
in expansion of about 77% relative to the control
specimens (B1-A-60). The S/A had a mixed effect on
expansion results for group B mixtures.
Figure 3 shows that the inclusion of fibre reinforcement had a notable effect on reducing the
expansion of SCC specimens. Conforming to the
results of REd, specimens from mixture C1-A-S with
0.5% steel fibres had an expansion value (0.04%) that
is about 64% lower than that of the corresponding
specimens from mixture C1-A-P with 0.1% polypropylene fibres. Also, specimens from mixture C1-A-H
with hybrid fibre reinforcement showed an expansion
value (0.08%) between that of the corresponding
specimens from mixtures with pure steel and polypropylene fibres. For mixtures comprising hybrid
fibres, the effect of the binder type on expansion
results was not noticeable likely because it was
masked by the effect of fibres.

polypropylene) in mixture C1-A-H achieved comparable performance to that of mixture B1-A-S until
41 weeks of exposure. Subsequently, it had deterioration between that of mixtures made with pure steel
and polypropylene fibres.
3.2 Expansion
The results of expansion of the SCC specimens are
shown in Fig. 3. In group A, specimens from mixture
SRPC showed the lowest expansion of 0.07%,
whereas the control SCC (A1-N-50) specimens had
the highest expansion of 0.34%. The binary, ternary
and quaternary binder blends led to a marked reduction in expansion. For example, the T specimens
(A3-N-50) prepared with 50% OPC, 5% SF and 45% S
yielded an expansion value of 0.09%, which is about
25% of the corresponding value of the control
specimens.
In most cases, it seems that air-entrainment in group
B mixtures had a positive effect on reducing the
expansion of SCC specimens compared to that of the
corresponding non-air-entrained specimens from
group A mixtures. However, such an effect depended
on the type of binder. For instance, the control airentrained specimens (B1-A-40 and B1-A-60) at S/A of
40 and 60% had expansion values of 25% and 38%,
respectively, of the expansion value of the corresponding control non-air-entrained specimens (A1-N-

3.3 Mass change

Figure 4 shows a typical behaviour of mass change of
cylindrical SCC specimens immersed in the

0.80
Group A

Group B

Group C

0.70
0.54
0.34

0.40

0.38

0.50

0.07

0.06

0.06

C4-A-H

C5-A-H

0.06

C2-A-H

C3-A-H

0.11

0.08

0.04

0.06

0.03

0.05

B5-A-40

0.02

B4-A-60

0.13
0.06

0.13

0.08

0.12

0.07

0.10

0.09

0.20

0.10

0.30
0.16

Expansion (%)

0.60

C1-A-H

C1-A-P

C1-A-S

B5-A-60

B4-A-40

B3-A-60

B3-A-40

B2-A-40

B2-A-60

B1-A-40

B1-A-60

A4-N-50

A5-N-50

A3-N-50

A2-N-50

A1-N-50

SRPC

0.00

Fig. 3 Expansion of SCC specimens after 54 weeks of immersion in the ammonium sulphate solution

Materials and Structures (2012) 45:977994

985

ammonium sulphate solution for 54 weeks. Specimens from all SCC mixtures initially had a slight gain
in mass followed by a decreasing trend at advanced
stages of the test. The final mass for some specimens
was slightly lower than the corresponding initial value
before exposure (maximum mass loss of 1%), but no
severe mass loss was observed. The trends of mass
change primarily depended on the cementitious materials type with marginal effects of S/A and fibre
inclusion. Generally, the prismatic specimens (larger
in size) had a similar trend of mass change, but with
lower rate compared to that of the corresponding
cylindrical specimens. This behaviour is ascribed to
the relative difference in size since the cylindrical
specimens had a surface-to-volume ratio that is 10%
larger than that of the prismatic specimens. This is
supported by the analysis of variance (ANOVA) for
the mass change at the end of exposure (54 weeks).
For instance, at a significance level a = 0.05,
ANOVA of the mass change of specimens from group
B mixtures at S/A of 40% shows that the difference in
surface-to-volume ratio between cylindrical and prismatic specimens had an F value of 23.2, which is
larger than the corresponding critical F value of 4.4.
According to Montgomery [22], in this statistical
approach of random variables, exceeding the critical
value for an F-distribution density function reflects
that the tested variable affects the mean of results.
In group B mixtures, the B specimens (B2-A-40 and
B2-A-60), which showed the highest reduction in REd
with significant expansion, did not yield severe mass
loss. This may suggest that degradation consisted of
cracking and softening rather than surface scaling.

10
SRPC
A1-N-50
A2-N-50

Mass Change (%)

A3-N-50
A4-N-50
A5-N-50

3.4 Relative compressive (RC) and flexural (RF)

strength
The results of RC and RF strength after 54 weeks of
exposure to the ammonium sulphate solution are listed
in Table 5. The range of RC for groups A, B and C
specimens was 4290, 095, and 49117%, respectively. The range of RF for groups A, B and C
specimens was 7497, 095, and 3352%, respectively. In group B mixtures, the B specimens (B2-A-40
and B2-A-60) had low or zero RC and RF conforming
to the results of REd and expansion. It can be observed
that there was a general trend of strength loss due to
softening and cracking of samples.
Except for group C mixtures, the loss in compressive strength (measured on cylindrical specimens) was
generally higher than that in flexural strength
Table 5 Results of RC and RF for specimens immersed in the
ammonium sulphate solution after 54 weeks
Mix. ID

RC (%)

RF (%)

Group A
SRPC

67

86

A1-N-50

45

89

A2-N-50

49

74

A3-N-50

90

85

A4-N-50

42

97

A5-N-50

52

93

B1-A-40

44

95

B1-A-60

39

82

B2-A-40

31

B2-A-60

B3-A-40

56

79

B3-A-60

53

91

B4-A-40
B4-A-60

66
62

84
91

B5-A-40

71

90

B5-A-60

95

79

C1-A-S

49

33

C1-A-P

63

38

C1-A-H

50

33

C2-A-H

49

52

C3-A-H

53

36

C4-A-H

117

50

C5-A-H

86

42

Group B

Group C

4
2
0
0

10

15

20

25

30

35

40

45

50

55

Time (Weeks)

Fig. 4 A typical behaviour of mass change versus time for

group A specimens

986

Materials and Structures (2012) 45:977994

(measured on prismatic specimens from the same SCC

mixtures). As discussed earlier, this may be attributed
to the faster rate of deterioration of cylinders due to
their higher surface-to-volume ratio relative to that of
prisms. Testing complications on deteriorated specimens may have contributed to this difference as will be
discussed later. It can be noted that the inclusion of
fibres in group C led to the opposite trend as the
flexural strength declined relatively more in fibrereinforced specimens than compressive strength.

4 Discussion
The ammonium sulphate solution had deleterious
effects on the SCC specimens from the various
mixtures tested. The trends of mass change suggest
that the first stage of exposure to the ammonium
sulphate solution involved progressive absorption of
the solution and formation of reaction products, which
generally led to mass gain. Subsequently, dissolution
of the cementitious matrix and removal of reaction
products led to mass decrease, but at different rates
depending on the type of binder used. Fracture
surfaces from specimens subjected to flexural strength
tests were immediately sprayed with 1 N phenolphthalein solution to detect the depth of reduction in
alkalinity, as shown in Fig. 5. This depth ranged from
4 to 10 mm depending on the type of binder. However,
the zone of reduced alkalinity may or may not serve as
a protective zone, which can reduce the rate of

Reduction of alkalinity

Fig. 5 Reduction of alkalinity shown by the uncoloured zone of

the cross-section

degradation, depending on the binder components and

their dosages, as will be explained later in this text.
Gypsum crystals formed in the zone of reduced
alkalinity as shown by SEM micrographs (e.g.
Fig. 6a), particularly in SCC specimens made from
ternary and quaternary binders. The absence of
monosulphate and ettringite in this zone is likely due
to their instability at low pH. Diffusion of sulphate
ions towards the inner concrete core, where the pH is
still high (about 12.0) (zone with dark color in Fig. 5),
can lead to formation of gypsum and ettringite due to
the reaction with portlandite and monosulphate
formed during the hydration of the cementitious
matrix. For instance, Fig. 6b shows a cluster of
secondary ettringite growing in an air-void at the
boundary of the intact concrete zone. In general, the
interaction of ammonium sulphate with the hydrated
cementitious paste led to loss of cohesion, volume
swelling, cracking and spalling.
4.1 Effect of binder type
The type of binder had a pronounced effect on the
overall behaviour of group A specimens immersed in
the ammonium sulphate solution. Specimens prepared
with 100% SRPC had higher REd, and lower expansion than that of the control specimens prepared with
100% OPC, but both types of specimens were
susceptible to softening as indicted by the strength
loss results. Dissolution of CH and CSH was caused
by the acidic action of ammonium ions leading to
notable loss of mechanical properties (REd, compressive and flexural strength). Since SRPC and OPC
pastes encompass large volume fractions of CH and
CSH, both are vulnerable to acidbase reactions.
This is depicted by DSC results before exposure to
chemical attack (Fig. 7), which show that pastes from
A1-N-50 and SRPC mixtures had large calcium
hydroxide endothermic peaks (at a temperature of
about 450C) corresponding to enthalpies of 63.6 and
52.65 J/g, respectively. However, because of the
lower C3A content in SRPC (Table 1), the action of
sulphate ions is discounted since ettringite formation
is reduced resulting in lower expansion. This is shown
by DSC curves after exposure to the ammonium
sulphate solution (Fig. 8). While both SRPC and OPC
specimens had corresponding phases of gypsum (at a
temperature range of 136144C) and portlandite (at a
temperature range of 438442C), ettringite traces

Materials and Structures (2012) 45:977994

987

Fig. 6 SEM micrographs and associated EDX spectra after 54 weeks of exposure to ammonium sulphate showing a gypsum in outer
surface and b secondary ettringite growing in air voids

(at a temperature range of 8590C) occurred only

within the exposed surface of 100% OPC specimens.
Generally, the incorporation of SCMs with OPC in
ternary and quaternary blends had a significant effect
on improving the resistance of SCC specimens to the
ammonium sulphate solution compared to that of
specimens from the control and binary mixtures.
SCMs react with CH and water to produce secondary
CSH (pozzolanic reaction). Increasing the replacement level of OPC by SCMs is inversely proportional
to the amount of CH generated. This is depicted by the
DSC results in Fig. 7, where enthalpies of CH for the

ternary and quaternary binder mixtures were significantly less than that of the control and binary mixtures.
Thereby, the ternary and quaternary binder mixtures
had limited CH contents available for the acidbase
reaction compared to that of the control and binary
mixtures. In general, the replacement of OPC with
higher dosages of SCMs has a dilution effect of
reducing the portlandite and C3A contents in the
hydrated cementitious matrix, which decreased the
rate of formation of gypsum and ettringite, leading to
lower swelling relative to the control (100% OPC) and
binary (92% OPC ? 8% SF) cementitious systems.

988

Materials and Structures (2012) 45:977994

0.2

70
63.6

60

Heat Flow (W/g)

Enthalpy (J/g)

52.65

50
40

37.31

30
18.87

18.62

20

11.48

A1-N-50
A2-N-50
A3-N-50
A4-N-50
A5-N-50

0
-0.2
-0.4
-0.6

10
0
SRPC

A1-N-50 A2-N-50 A3-N-50 A4-N-50 A5-N-50

Fig. 7 DSC results for calcium hydroxide heat flow peaks of

group A mixtures at 56 days (before exposure)

0.2
A1-N-50
SRPC

Heat Flow (W/g)

0
-0.2
-0.4
-0.6
-0.8
-1
0

100

200

300

400

500

600

Temperature (C)

Fig. 8 DSC curves for 100% OPC and SRPC specimens after
54 weeks of exposure to ammonium sulphate

This is shown by the DSC curves in Fig. 9, where the

intensity of the gypsum heat flow peaks (at a
temperature range of 136144C) after exposure to
ammonium sulphate was the highest for specimens
from the control mixture (A1-N-50) corresponding to
an enthalpy of 124 J/g followed by the B (A2-N-50)
specimens (115 J/g). Comparatively, the T (A3-N-50)
and Q (A4-N-50 and A5-N-50) specimens had
enthalpy values of 78 and an average of 90 J/g,
respectively. Also, the T and Q specimens showed
minute traces of CH (at a temperature range of
438442C) and no ettringite, while the control and B
specimens showed relatively larger peaks of CH and
traces of ettringite (at a temperature range of
8590C). This implied different modes of deterioration of the single and binary binders in contrast to the
ternary and quaternary binders. It is worth noting that
the amount of reaction products detected by DSC was
also dependent on the rate of leaching to the

-0.8
0

100

200

300

400

500

600

Temperature (C)

Fig. 9 DSC curves for group A specimens after 54 weeks of

exposure to ammonium sulphate

surrounding solution; however, the effect of this

parameter was not quantified in the present study.
In group B mixtures, the B specimens (B2-A-40 and
B2-A-60) with 92% OPC and 8% SF exhibited the
highest degradation at the end of exposure compared
to that of the control, T and Q specimens. This
generally agrees with the trend observed for the B
specimens in group A which showed a higher loss of
flexural strength relative to the control, T and Q
specimens. Madej [8] and Mehta [9] also reported
negative effects of binary binders comprising 85%
OPC and 15% SF on the behaviour of concrete
exposed to 5% ammonium sulphate solutions. Lee
et al. [23] reported that binary binders incorporating
OPC and 515% SF led to a significant strength loss of
mortar samples immersed in a 5% magnesium
sulphate solution. CSH is susceptible to decomposition by ammonium and magnesium sulphate solutions [3, 4]. Hence, it may be deduced that binary
blends comprising OPC and SF, which are rich in
CSH with moderate or low CH content can be
significantly decomposed, particularly when corrosive
ions such as Mg2? and NH4?, which directly attack
CSH, exist in the solution. In the aggressive
ammonium sulphate solution, faster dissolution of
such binary matrices increased damage by allowing
further penetration of the solution resulting in marked
swelling and loss of cohesion. On the contrary, the T
and Q specimens (very rich in CSH with little CH)
had better performance than that of the B specimens
indicating different modes of deterioration of the
blended binders in the ammonium sulphate solution, in
compliance to the trends observed for group A
mixtures.

Materials and Structures (2012) 45:977994

989

Detwiler et al. [24] reported that CSH produced

by pozzolanic reactions of SCMs has a different
structure and calcium-to-silicate ratio (C/S of about
1.1) than that of conventional CSH (C/S of about 1.7
[3]) produced by the hydration of OPC. Also, CSH
produced by SCMs has a higher binding capacity for
alkalis and alumina [3, 24]. Significant decalcification
(release of Ca2?) of CSH occurs at a pH of about 9.0
[3], which may have contributed to loss of mechanical
properties in the case of exposure to a high concentration ammonium sulphate solution. Generally, progressive consumption of CH (Eq. 2) to form gypsum
leads to a reduction of pH in the pore solution and
decalcification of CSH (Eq. 3) [4]. Also, when CH
is locally depleted, local decalcification of CSH
may occur [25]. This aspect is important because
changes in the local concentration, particularly in the
reaction zone can possibly influence the dissolution
kinetics and lead to inconsistent physico-mechanical
behaviour until a steady-state is reached. However,
such an effect was not investigated in the present study
and warrants further research.
Figure 10ac exhibit fracture cross-sections for
100% OPC, B, and Q specimens, in which a thick
(whitish) layer is evident in the Q specimen and almost
absent in the B and 100% OPC specimens. It was
observed that gypsum crystals (Fig. 6a) accumulated
on the exposed surface of the control and B specimens,
which had a high OPC content (100 and 92%,
respectively), without notable deposition with depth
in the cementitious matrix. For example, at 500 lm
from the exposed surface, Fig. 11 shows a deteriorated
and porous cementitious matrix of the B specimen
shown in Fig. 10b without evident precipitation of

(b)

gypsum crystals. Similar observations were reported

by Rendell and Jauberthie [7] for pure OPC and SRPC
mortar specimens. They concluded that the effect of
ammonium sulphate is primarily through leaching of
Ca2? since no new products were found under the
exposed surface [7]. Hence, based on previous
research [7] and direct observations from the current
study, it can be stated that cementitious matrices
having high OPC content (CSH with higher C/S) are
vulnerable to continual attack by ammonium sulphate
solutions.
Comparatively, in the T and Q specimens with a
lower OPC content (50%) and combinations of SCMs
(pozzolanic CSH with lower C/S), there was a zone
under the exposed surface (zone of reduced alkalinityFig. 5) that retained gypsum in its structure, and
thus had some blocking effect for further attack by the
surrounding solution. Micro-analysis by SEM and
EDX (e.g. Fig. 12a, b) for the T and Q specimens

(c)
Passivating layer

(a)

Fig. 11 Degraded matrix (500 lm under the exposed surface)

for a B specimen

Fig. 10 Fracture cross-sections of specimens made from a single, b binary, and c quaternary binders

990

Materials and Structures (2012) 45:977994

Fig. 12 SEM micrographs for the passivating layer in Q specimens: a B5-A-60 and b C4-A-H

showed that this layer is comprised of gypsum

encapsulated by a rich media of hydrous silica with
aluminum and magnesium. This dense silico-aluminous gel incorporating gypsum formed a passivating
layer in the T and Q specimens, which relatively
sealed the intact concrete from the surrounding
ammonium sulphate solution.
Diffusion of the solution through this protective
layer might have slowed down the rate of penetration
of the ammonium sulphate solution, thus reducing the
rate of degradation of the cementitious matrix. It
appears that the effectiveness of the abovementioned
protection mechanism is a function of the dosage and
combination of SCMs. Generally, the higher (50%)
replacement level of OPC along with the combination
of SCMs in ternary and quaternary binder blends led to
better performance of the cementitious matrix in the

ammonium sulphate solution. T and Q specimens with

50% OPC replacement with combinations of SF, S,
FA, and LF achieved lower expansion and better
maintained flexural strength after 54 weeks of exposure to the ammonium sulphate solution than specimens made with a binary binder incorporating only
8% SF.
4.2 Effect of air-entrainment
Irrespective of the binder type, air-entrainment had a
positive effect on reducing the expansion of group B
specimens. It is well-documented that microscopic air
bubbles act as relief chambers for accommodating
expansive pressures induced in hydrated cement
matrices exposed to freezing-thawing cycles [26]. It
is conceivable that a network of closely spaced air

Materials and Structures (2012) 45:977994

bubbles accommodated swelling pressures, exerted by

the growth of reaction products and NH3, on the
cementitious matrix. This was supported by SEM
micrographs showing air voids hosting clusters of
gypsum or ettringite crystals (e.g. Fig. 6b), thus
providing a space for the growth of reaction products
and delaying the expansion and cracking of the
cementitious matrix. Similar observations were
reported for air-entrained mortar specimens immersed
in sodium and magnesium sulphate solutions [27].
Hence, air-entrainment can improve the resistance to
expansion induced by sulphate ions, but without a
significant effect on improving the resistance of SCC
to the acidic action of ammonium ions.
4.3 Effect of sand-to-aggregates mass ratio
The effect of S/A did not have a clear trend on the
results of group B mixtures. For instance, ANOVA, at
a significance level a = 0.05, for the expansion results
shows that the change of S/A (4060%) had an F value
of 0.3, which is significantly lower than the corresponding critical F value of 4.4, indicating that the
change of S/A from 40 to 60% was an insignificant
parameter in the case of attack by ammonium sulphate
solutions. This is attributed to the nearly constant paste
volume fraction (325350 l/m3) vulnerable to chemical attack by ammonium sulphate at both S/A levels.
4.4 Effect of fibres
Specimens from mixture (C1-A-S) with 100% OPC
and 0.5% steel fibres had lower rate of expansion, and
reduction in REd than that of the corresponding
specimens from mixture (C1-A-P) with 0.1% polypropylene fibres. The rationale behind incorporating
hybrid fibres is the combination of the advantageous
effects of polymer micro-fibres and steel macro-fibres.
While polymer micro-fibres contribute to pre-crack
behaviour by impeding the nucleation and coalescence
of micro-cracks, steel macro-fibres can bridge macrocracks, thus contributing to the control of crack
propagation [28].
The mechanical effect of 0.4% steel and 0.1%
polypropylene fibres was effective at improving the
resistance of SCC specimens to the ammonium
sulphate solution. Hybrid fibre reinforcement
restrained swelling and controlled cracking leading
to increased matrix integrity, and in turn it discounted

991

the rate of deterioration. After 54 weeks of exposure to

the ammonium sulphate solution, fibre-reinforced
SCC specimens generally had a reduction in both
compressive and flexural strengths. As discussed
earlier, the loss of strength is ascribed to the decalcification of the cementitious matrix leading to softening
and loss of cohesion. Group C specimens with hybrid
fibre reinforcement achieved comparable or better
(particularly the Q specimens) RC strength than that of
their corresponding non-fibre-reinforced specimens in
groups A and B.
Conversely, SCC specimens with hybrid fibre reinforcement had a considerable loss in flexural strength
(evaluated as the first-crack stress) after 54 weeks of
exposure to the ammonium sulphate solution compared
to that of the corresponding specimens in groups A and
B. However, unlike specimens without fibres, such
specimens retained their post-peak behaviour and did
not break in a brittle manner owing to the effect of fibres
in the core of specimens. Such drop in the first-crack
flexural strength is ascribed to the behaviour of polymer
and steel fibres in the near surface zone. Decomposition
of the cementitious matrix in this zone caused softening
and loss of bonding between fibres (polymer or steel)
and the cementitious matrix. Also, it was observed that
steel fibres near the surface were severely corroded, and
some were dislodged from the surface of specimens,
leaving notch-like locations. Since flexural tests subject
the bottom surface (fibre) of specimens to the maximum
tensile stress at the middle-third location, the existence
of severely corroded steel fibres and/or dislodged fibre
locations as well as loosely bonded fibres in this zone
triggered the first-cracking at a lower flexural stress than
that of specimens without fibres. For example, Fig. 13
shows the failure plane of a hybrid fibre-reinforced SCC
specimen, in which corroded fibres and dislodged fibre
locations initiated cracking at the bottom and the lower
side of the specimen.
4.5 Assessment of concrete exposed
to ammonium sulphate attack
In the case of attack of concrete by ammonium
sulphate, there is need to assess multiple performance
parameters such as change in mass, length, and
strength to mitigate the shortcomings of relying only
on one individual criterion. For instance, change of
mass may not reflect the occurrence of swelling.
Similarly, expansion may not indicate softening and

992

Fig. 13 Failure plane of a hybrid fibre-reinforced SCC

specimen

decomposition of the cementitious matrix. The change

of the dynamic modulus of elasticity is also subject to
local defects in the specimen.
In determining the RC and RF strengths of specimens after ammonium sulphate attack, the situation is
complicated by testing conditions. The change of the
aspect ratio and geometry of specimens after exposure
can have significant effects on the RC strength. For
instance, the effective cross-section of specimens may
become variable across the height, resulting in planes
of stress concentration, load eccentricity and non
uniform stress distribution on cylindrical specimens
during the compression test. Thus, it is somewhat
difficult to compare compressive strength values of
specimens after exposure to their initial values
obtained from intact samples.
It was stated [29] that the uniaxial compression test
tends to close cracks perpendicular to the direction of
loading and may not be sensitive to the damage of
concrete subjected to sulphate attack. In the present
study, compression tests could not reveal performance
risks associated with SCC specimens incorporating
fibre reinforcement under the ammonium sulphate
attack. Since concrete in practice frequently fails
under tensile stresses, direct tension, splitting or
flexural tests have been recommended for assessing
damage after exposure to chemical attack [10, 29].
However, in the case of attack by ammonium sulphate,
testing complications may arise in the flexure test
similar to those for the compression test. For instance,
unlike intact samples, swelled/deteriorated prismatic
specimens under flexural loading may be susceptible
to uneven stress distribution due to the non-uniform

Materials and Structures (2012) 45:977994

cross-sectional area over the middle-third span subjected to the maximum bending moment. In addition,
the residual strength is influenced by the crosssectional area of the sound concrete core of the
specimen or the effective-cross sectional area. Considering the cross-sectional area of the sound core
instead of the initial cross-sectional area, the strength
after exposure can be higher than its initial value. In
addition, in chemical immersion tests, there are two
processes with conflicting effects simultaneously
occurring in the cementitious matrix: decomposition
of the surface and continual hydration/densification of
the core (bulk cross-section). With the existence of a
passivating layer, the latter can cause strength increase
in the core, particularly for binders comprising high
dosages of SCMs that exhibit long-term strength gain.
Hence, in evaluating the deterioration of cement-based
materials in acidic media, compressive or flexural
strength loss should be interpreted with care in light of
other performance parameters.
Under attack by ammonium sulphate, the damage
to concrete is of a dissolution type accompanied by
expansion and cracking. Hence, multiple parameters
should be corroborated to draw reliable conclusions.
In the present study, the change in mass was an
insignificant indicator of damage. The changes in
length, REd, and to a less extent the relative compressive and flexural strengths, along with thermal analysis
by DSC and micro-analysis by SEM supported by
EDX gave a more comprehensive evaluation of the
performance of SCC under such a type of chemical
attack. This approach discounted the limitations of
each individual test from the overall evaluation of the
performance of SCC mixtures. Based on the present
study, the following qualitative failure limits may be
suggested as damage indicators for similar highperformance concretes under the attack of aggressive
ammonium sulphate: a maximum expansion of 0.2%,
a minimum relative dynamic modulus of elasticity
(REd) of 60%, and a minimum RF strength of 65%.
Such proposed limits can be refined in future research
to establish performance-based specifications for
concrete exposed to ammonium sulphate.

5 Concluding remarks
With the growing use of SCC in various infrastructure
applications exposed to ammonium sulphate, it has

Materials and Structures (2012) 45:977994

993

become necessary to investigate the suitability of a

wide range of SCC mixture designs to such an
exposure. Considering the mixture design parameters,
concentration and type of sulphate solution and the
testing approach implemented in this study, the
following conclusions can be drawn:

The role of physical resistance (low penetrability)

of SCC diminished in the ammonium sulphate
solution. Exposure of SCC specimens to ammonium sulphate caused detrimental effects on the
physico-mechanical properties.
Ammonium sulphate causes a combined acid
sulphate attack characterized by decomposition/
softening associated with expansion, cracking and
spalling.
The incorporation of air-entrainment in SCC
reduced/delayed expansion of specimens in most
cases.
While SRPC reduced the expansion of SCC
specimens compared to OPC, both types of
specimens were vulnerable to decomposition by
the ammonium sulphate solution. In spite of the
absence of a passivating layer, softening of these
matrices was limited due to the abundance of CH.
Binary binders with a low replacement level of
OPC (92% OPC ? 8% SF) led to inferior performance of SCC specimens in the ammonium
sulphate solution due to the lack of CH, abundance
of CSH, and the absence of a protective zone.
While ternary and quaternary binders at a high
OPC replacement level (50%) typically led to
matrices rich in CSH and poor in CH, they
performed well (low expansion, high REd and RF)
in the ammonium sulphate solution. The diffusion
of ammonium sulphate was slowed down in these
systems due to the formation of a thick passivating
layer of hydrous silica encapsulating gypsum.
Since cementitious binders respond differently to
ammonium sulphate attack, future research work is
recommended for assessing the relative diffusion
of ammonium and sulphate ions within cementitious systems made from different binders so that a
better understanding of the competition between
leaching and precipitation processes could be
gained.
While the incorporation of hybrid fibres improved
the overall integrity of SCC specimens in the
ammonium sulphate solution, the first-cracking

flexural strength of fibre-reinforced SCC declined

significantly after the ammonium sulphate exposure. This warrants caution for structural applications under sustained flexural loading and severe
ammonium sulphate exposure since the occurrence of first-cracking at low stress levels can serve
as a damage accelerator.
The present study pinpointed that multiple performance parameters along with thermal and microanalyses are needed to characterize the behaviour
of cement-based materials under ammonium sulphate attack.

References
1. Zhu W, Bartos J (2003) Permeation properties of selfcompacting concrete. Cem Concr Res 33(6):921926
2. Boel V, Audenaert K, De Schutter G, Heirman G, Vandewalle L, Desmet B, Vantomme J (2007) Transport properties of self compacting concrete with limestone filler and
fly ash. Mater Struct 40(5):507516
3. Hewlett PC (ed) (1998) Leas chemistry of cement and
concrete. Arnold, London
4. Skalny J, Marchand J, Odler I (2002) Sulfate attack on
concrete. Spon Press, London
5. Rols S, Ambroise J, Pera J, Chabannet M (2000) Calcined
paper sludge-pozzolanic admixture to improve sulfate
resistance of concrete. In: Malhotra VM (ed) Proceedings of
the fifth CANMET/ACI international conference on durability of concrete, Barcelona, Spain, ACI SP-192, vol I,
pp 353367
6. Nagele E, Hillemeier B, Hilsdorf H (1984) Attack of
ammonium salt solutions of concrete. Betonwerk FertigteilTechnik 50(11):742751
7. Rendell F, Jauberthie R (1999) The deterioration of mortar in
sulphate environments. Constr Build Mater 13(6):321327
8. Madej J (1992) Corrosion resistance of normal and silica
fume-modified mortars made from different types of
cement. In: Malhotra VM (ed) Proceedings the fourth
CANMET/ACI international conference on fly ash, silica
fume, slag, and natural pozzolans in concrete, Istanbul,
Turkey, ACI SP-132, vol II, pp 11891207
9. Mehta PK (1985) Studies on chemical resistance of low
water/cement ratio concrete. Cem Concr Res 15(6):969978
10. Miletic S, Ilic M, Ranogajec J, Marinovic-Neducin R,
Djuric M (1998) Portland ash cement degradation in
ammonium-sulfate solution. Cem Concr Res 28(5):713725
11. Mbessa M, Pera J (2001) Durability of high-strength concrete in ammonium sulfate solution. Cem Concr Res 31(8):
12271231
12. Nehdi M, Pardhan M, Koshowski S (2004) Durability of
self-consolidating concrete incorporating high-volume
replacement composite cements. Cem Concr Res 34(11):
21032112

994
13. De Schutter G, Audenaert K (eds) (2007) Durability of selfcompacting concrete, State-of-the-art report of RILEM
Technical Committee 205-DSC. RILEM Report 38, RILEM
Publications S.A.R.L., p 185
14. BIBM, CEMBUREAU, ERMCO, EFCA, EFNARC (2005)
The European guidelines for self-compacting concrete,
p 63. www.efnarc.org
15. American Society for Testing and Materials (2007) ASTM
C 192: practice for making and curing concrete test specimens in the laboratory, vol 4.02. American Society for
Testing and Materials, West Conshohocken
16. American Society for Testing and Materials (2007) ASTM C
1012: standard test method for length change of hydrauliccement mortar exposed to a sulfate solution, vol 4.01. American Society for Testing and Materials, West Conshohocken
17. Mehta PK (1992) Sulfate attack on concretea review. In:
Skalny J (ed) Materials science of concrete III. The American Ceramic Society, Westerville, pp 105130
18. American Society for Testing and Materials (2007) ASTM
C 39: standard test method for compressive strength of
cylindrical concrete specimens, vol 4.02. American Society
for Testing and Materials, West Conshohocken
19. American Society for Testing and Materials (2007) ASTM
C 215: standard test method for fundamental transverse,
longitudinal, and torsional resonant frequencies of concrete
specimens, vol 4.02, West Conshohocken, PA
20. American Society for Testing and Materials (2007) ASTM
C 666: standard test method for resistance of concrete to
rapid freezing and thawing, vol 4.02. American Society for
Testing and Materials, West Conshohocken

Materials and Structures (2012) 45:977994

21. American Society for Testing and Materials (2007) ASTM
C 78: standard test method for flexural strength of concrete
using simple beam with third-point loading. American
Society for Testing and Materials, Philadelphia
22. Montgomery D (2001) Design and analysis of experiments.
Wiley, New York
23. Lee S, Moon Y, Swamy R (2005) Sulfate attack and role of
silica fume in resisting strength loss. Cem Concr Compos
27(1):6576
24. Detwiler R, Bhatty J, Bhattacharja S (1996) Supplementary
cementing materials for use in blended cements. Portland
Cement Association, RD112.01T, p 96
25. Taylor HFW, Gollop RS (1997) Some chemical and
microstructural aspects of concrete durability. In: Scrivener
KL, Young JF (eds) Mechanisms of chemical degradation of
cement-based systems. E & FN Spon, London, pp 177184
26. Bassuoni MT, Nehdi M (2005) The case for air-entrainment
in high performance concrete. ICE Struct Build 158(5):
311319
27. Santhanam M (2001) Studies on sulfate attack: mechanisms,
test methods, and modeling. PhD Dissertation, Purdue
University, USA
28. Johnston C (2001) Fiber-reinforced cements and concretes.
Gordon and Breach Science Publications, Amsterdam
29. Ju J, Weng L, Mindess S, Boyd A (1999) Damage assessment and service life prediction of concrete subjected to
sulfate attack. In: Marchand J, Skalny J (eds) Materials
science of concrete special volume: sulfate attack mechanisms. The American Ceramic Society, Westerbrook,
pp 265282