Construction and Building Materials 117 (2016) 88–98

Contents lists available at ScienceDirect

Construction and Building Materials

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

Study of deterioration of concrete exposed to different types of sulfate

solutions under drying-wetting cycles
Lei Jiang a,⇑, Ditao Niu b
a
School of Civil Engineering and Architecture, Anyang Normal University, Anyang 455000, China
b
College of Civil Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, China

h i g h l i g h t s

Durability of concrete under different types of sulfate solutions is investigated.

SEM, XRD and TG-DSC are used to study microstructure of concrete after corrosion.

Large amount of corrosion products can be formed before obvious damage is occurred.

The damage process of concrete exposed to different sulfate solutions is analyzed.

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

Article history: Properties of concrete subjected to drying-wetting cycles in different types of sulfate solutions were
Received 12 October 2015 investigated in this paper. The corrosion solution includes three types, namely, 10% sodium sulfate solu-
Received in revised form 6 April 2016 tion, 10% magnesium sulfate solution and the composite solution of 10% sodium sulfate and 3.5% sodium
Accepted 24 April 2016
chloride (by mass). Through the experiment, visual change, relative dynamic modulus of elastically,
Available online 9 May 2016
weight loss, compressive strength loss and the damage layer thickness of concrete were measured. To
identify the products formed by sulfate attack, analytical techniques, including X-ray diffraction,
Keywords:
scanning electron microscopy and thermal analysis were performed on the selected samples. Test results
Concrete
Sulfate attack
show that the deterioration degree of concrete in magnesium sulfate solution is more severe than that in
Chloride the other sulfate solutions. The existence of chloride ions in the composite solution reduces the deterio-
Drying-wetting cycles ration rate of concrete, and the damage degree of concrete could be inhibited effectively. Test results also
Thermal analysis show that the quantity of corrosion products in magnesium sulfate solution is higher than that in the
other sulfate solutions. While, the quantity of corrosion products is the lowest when exposed to the
composite solution. Furthermore, the quantity of gypsum in concrete is less than that of ettringite in test,
and some of gypsum can be observed only after a certain corrosion extent. The quantity of corrosion prod-
ucts does not correspond well with the observations of the physical properties, and extensive amount of cor-
rosion products can be detected before the deterioration degree of concrete increased significantly.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
In the sulfate environment, sulfate ions, magnesium ions and
chloride ions are the common corrosive ions exist widely in seawa-
ter, groundwater, salt lake and sulfate soil. Sulfate attack on
cement materials is one of the most severe problems affecting
the durability and service life of concrete structures in the natural
climate. Sulfate attack can manifest in the form of expansion and
cracking of concrete. Sometimes the expansion of concrete may
cause serious structural problems. Sulfate attack can also take
⇑ Corresponding author.
E-mail address: djjianglei@163.com (L. Jiang).
the form of a progressive decrease in the strength and weight loss
due to the cohesiveness loss of the cement hydration products
[1,2]. In sulfate environment, the elements of concrete construc-
tion situated in water table fluctuation, tidal zone and splash zone
will suffer from more complex attack because of the drying-
wetting cycles, which can accelerate the deterioration of concrete.
Sulfate attack on concrete structures caused the attention of the
scholars earlier. After a long period of explorative research, there
already have a certain understanding to the mechanism, influence
factors, evaluation methods and evaluation indexes of concrete
under sulfate attack exposed to sodium sulfate solution. Many
works for the damage process of concrete exposed to sulfate attack
[3–6] and the coupling function of sulfate attack and

http://dx.doi.org/10.1016/j.conbuildmat.2016.04.094
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98
drying-wetting cycles [7–9] such as expansion, weight loss, rela-
tive dynamic modulus of elastically (RDME) loss and the changes
of mechanical properties of concrete have been reported.
However, little studies of the damage process are done on
cement materials attacked by magnesium sulfate solution or the
composition salt solution of chloride and sulfate ions. And it is in
the little studies that most of them are conducted on cement paste
or mortar. Salah et al. [10] investigated the performance of plain
and blended cements exposed to magnesium sulfate solutions with
varying sulfate concentrations. Siad et al. [11] investigated the
effects of mineral admixture type on the behavior of self-
compacting concrete exposed to magnesium sulfate solutions over
4 years of exposure. Diab et al. [12] suggested the guidelines in
compressive strength assessment of concrete modified with silica
fume as results of magnesium sulfate attack. Geng et al. [13] inves-
tigated the stability of bound chlorides in chloride-contaminated
cement pastes with and without fly ash and ground granulated
blast-furnace slag when subjected to sodium and magnesium sul-
fate attack. Zhang et al. [14] studied the expansion and subsequent
damage of concrete immersed in sulfate solutions of different sul-
fate concentrations, and mixed solutions of different sulfate and
chloride concentrations. Xu et al. [15] reported the results of a
study conducted to evaluate the releases of bound chlorides from
chloride-admixed plain and blended cement pastes exposed to sul-
fate attacks. However, the role of chloride ions in the presence of
sulfate solution on the deterioration of concrete is not well known.
In addition, the damage mechanism of concrete exposed to drying-
wetting cycles under magnesium sulfate or composite solution of
sulfate and chloride environment need further studied.
In the study of microscopic-test, little literatures have been con-
ducted concerning the quantity of corrosion products in concrete
under sulfate attack, especially quantitatively compared in differ-
ent corrosion layer of concrete exposed to magnesium sulfate solu-
tion or composite solution of sulfate and chloride. Santhanam et al.
[16] discussed the thermal analysis results for mortars immersed
in solutions of seawater and groundwater. Mathias et al. [17]
investigated the amount of ettringite and gypsum using quantita-
tive Rietveld analysis after X-ray diffraction (XRD) measurements
for motar exposed to a combined sodium sulfate and sodium chlo-
ride solution. Chen et al. [18] investigated the microstructure of
interior concrete by using XRD and thermal analysis exposed to
combined attack of chloride and sulfate under drying-wetting
cycles. Jiang et al. [19] discussed the XRD and thermal analysis
results of concrete under sulfate attack exposed to freeze-thaw
cycles. Chen et al. [20] studied the quantity of ettringite in concrete
under the dry-wet cycling condition and sodium sulfate
environment.
Furthermore, most researches have focus on raising the temper-
ature of sulfate solution in the drying state to accelerate the dam-
age process of concrete under sulfate attack. For example, test of
specimens drying under 100 °C for 24 h was used by Sahmaran
et al. [21], drying under 60 °C was used by Niu et al. [22], drying
under 70 °C was used by Yuan et al. [23], drying under 80 °C was
used in literatures [24,25], and drying under 105 °C for 24 h was
used by Yang et al. [26]. But, the high temperature has a great
influence on the quantity of corrosion products in concrete, prob-
ably altering the mechanism of sulfate attack. In the test, the dry-
ing state was natural drying in the air, which was more relevant to
a real environment.
This study simulated concrete exposed to sodium sulfate, mag-
nesium sulfate and composite solution of sodium sulfate and
sodium chloride under cyclic environmental condition. Basic
experimental research on the performance of concrete in different
sulfate solutions was conducted based on the macroscopic and
microscopic test. Visual change, weight loss, RDME loss, compres-
sive strength loss and the damage thickness of concrete were
89
conducted. Furthermore, the corrosion products of the concrete
were distinguished and quantitatively compared by the scanning
electron microscopy (SEM), XRD and the thermal analysis. The
deterioration process of concrete exposed to the different types
of sulfate solutions under drying-wetting cycles was also analyzed.
2. Experimental details
2.1. Materials and mix proportions
A Chinese standard Ordinary Portland Cement (OPC) of PO
42.5R produced by the Cement Factory of Tongchuan was adopted.
Grade II fly ash from the Weihe Power Station, river sand with a
fineness modulus of 2.69 and coarse aggregate of crushed basalt
stone with a diameter of 5–16 mm were used in the test. A
naphthalene-type superplasticizer was used, and the dosage was
adjusted to keep the slump of fresh mixed concrete in the range
of 50–120 mm. The tap water was adopted as mix water. The
chemical composition of cement and fly ash is shown in Table 1.
In this experiment, the water-binder ratio (W/B) was 0.45, and
the concrete with 20 wt% replacements of cement with fly ash was
used. According to the related investigations [27–29], about 20 wt
% of cement is replaced by fly ash, the resistance property of con-
crete under sulfate attack could be improved effectively. The mix-
ture proportion and corresponding compressive strength of the
concrete are presented in Table 2.
2.2. Specimens preparation and curing conditions
Concrete specimens were prepared in a 0.3 m3 forced action
mixer, and all the test specimens were fabricated from a single
batch of concrete. The components of the concrete mixture were
batched by weight, the cement was premixed with fly ash, sand
and coarse aggregate before adding the water and the admixtures
for 1 min. Then, the entire amount of mixing water with the dis-
solved superplasticizer were added and mixed for 3 min. Finally,
concrete mixture was mixed for another 2 min. The concrete spec-
imens were cast in steel moulds and compacted on a vibration
table. All specimens were demolded after 24 h of casting and were
cured in a condition of 20 ± 3 °C and 95% relative humidity.
2.3. Experiment methods
In this paper, the concrete specimens exposed to sulfate
solution under drying-wetting cycles up to 360 days (Fig. 1). The
drying state was natural drying in the air, which was more relevant
to in-service conditions. One drying-wetting cycle lasted 15 days.
First, the specimens were continuously immersed in the test
solution for 7 days, and then were moved into the air to dry
naturally for another 8 days. In this study, concentrations (by
mass) of 10% sodium sulfate solution, 10% magnesium sulfate solu-
tion and the composite solution of 10% sodium sulfate and 3.5%
sodium chloride were used. For each case, 3 specimens of
100 mm  100 mm  400 mm prisms were used for the study of
the RDME loss, weight loss and damage layer thickness of concrete.
For each case, 30 specimens of 100 mm  100 mm  100 mm
cubes were used for the study of the compressive strength loss
and the analysis of attack products of concrete. All experiments
were performed on three specimen replicates. The average values
were used for the discussion of the test results. According to the
test procedure, the deterioration of the specimens was investigated
by determining the RDME loss, weight loss and compressive
strength loss. The specimen was considered to be a failure if the
RDME dropped to 60%, the weight loss exceeded 5% or the
compressive strength dropped to 75%.
90 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98

Table 1
Chemical composition of OPC and fly ash.

Constituent (wt%) SiO2 Al2O3 CaO MgO SO3 Fe2O3 Na2O K2O Loss
OPC 21.66 5.13 64.37 1.06 2.03 5.25 0.25 1.35 4.12
Fly ash 49.02 31.56 4.88 0.83 1.2 6.97 0.43 1.36 3.65

Table 2
Mix proportion and compressive strength of concrete.

Water-binder Cement/(kg m3) Fly ash/(kg m3) Water/(kg m3) Aggregate/(kgm3) Sand/(kg m3) Superplas-ticizer/% Compressive
ratio strength/MPa
28d 90d
0.45 285 70 160 1300 585 0.5 50.3 61.6

The measurement of compressive strength property was con-

ducted according to the GB/T50081-2002 [31] method. The follow-
ing Eq. (4) was used to calculate the compressive strength:
f cc ¼ F=A ð4Þ
where fcc is the compressive strength (MPa), F is the maximum load
(N) and A is the area of the cube loading face (mm2).
The damage thickness of concrete is detected by using a high
accuracy nonmetal ultrasonic analyzer according to the CECS
21:2000 [32] and literature [33]. Select the prism-side of
100 mm  400 mm prism as test surface, as shown in Fig. 2. The
transmitting transducer A is placed on the surface, and then a
receiving transducer B moves along the concrete surface to record
the corresponding time of sound propagation. The test distance l
between A and B is 50 mm, 75 mm, 100 mm, 150 mm, 200 mm
and 250 mm, respectively.
Fig. 1. Concrete specimens exposed to sulfate solutions under drying-wetting At first, the receiving transducer B close to the transmitter will
cycles. only sense the top layer (the damage layer of concrete), but as the
distance increases, the influence from the lower layer (the sound
concrete) is felt. When the distance between two transducers
According to the GB/T50082-2009 [30] method, the following
reaches a certain range l0, the ultrasonic propagation time from A
Eq. (1) was used to calculate the weight loss:
to B through the damage layer and the sound concrete is equal to
the time propagated only in the damage layer as shown in Fig. 2.
DW n ¼ ½ðW 0  W n Þ=W 0   100 ð1Þ There are:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
where DWn is the weight loss of specimens at every 30 exposure l0 2 H2f þ x2 l0  2x
days (%), W0 is the average weight of concrete specimens before sul- ¼ þ ð5Þ
Vf Vf Va
fate attack (kg) and Wn is the average weight of concrete specimens
at every 30 exposure days in the sulfate solutions (kg). sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
The dynamic modulus of elasticity (Ed), which was determined l0 Va  Vf
Hf ¼ ð6Þ
by Eq. (2), was measured by using a high-accuracy nonmetal ultra- 2 Va þ Vf
sonic analyzer as the initial value. The RDME, which was deter-
mined by Eq. (3), is the ratio of Ed value to the initial Ed value where Hf is the damage layer thickness of concrete (mm), x is the
after every 30 exposure days. horizontal projection of the ultrasonic propagation path through
the damage layer (mm), Vf is the ultrasonic velocity in the damage
layer (km/s); Va is the ultrasonic velocity in the sound concrete
ð1 þ tÞð1  2tÞqV 2
Ed ¼ ð2Þ
1t
l3
Edn V 2n l0
RDME ¼ ¼ ð3Þ l2
Ed0 V 20 l1
Where Ed is dynamic modulus of elasticity (GPa), V is ultrasonic A B1 B2 B0 B3 B4 B5
speed (m/s), q is density of specimen (kg/m3) and t is Poisson’s
ratio, Ed0 is the dynamic modulus of elasticity of concrete speci- Vf hf
mens before sulfate attack (GPa), Edn is the dynamic modulus of x
elasticity of concrete specimens at every 30 exposure days (GPa),
Va
V0 is the ultrasonic speed of concrete specimens before sulfate
attack (m/s), Vn is the ultrasonic speed of concrete specimens at
every 30 exposure days (m/s). Fig. 2. Setting of the transducers.
 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98
(km/s), l0 is the distance from A to B when the ultrasonic hits the
interface between damage layer and sound concrete (mm).
The experimental results are plotted in a time vs. distance plot,
which will be a straight line similar to Fig. 3. It is shown that l0 is
the intersection of two straight lines. The regression equations of l-
t in damage layer and in sound concrete by using the linear regres-
sion method as follows:
l f ¼ Af þ V f t f
l a ¼ Aa þ V a t a
where lf is the distance before l0 in Fig. 3, such as l1 and l2 (mm); la is
the distance after l0, such as l3, l4 and l5 (mm); tf is the time corre-
sponding to lf before l0, such as t1 and t2 (ls); ta is the time corre-
sponding to la after l0, such as t3, t4 and t5 (ls); Af and Aa are the
intercept of the two lines, respectively; Vf and Va are the slop of
the two lines, respectively.
Therefore, l0 can be written as:
A1 V a  A2 V f
l0 ¼
Va  Vf
and then the damage thickness Hf can be calculated by Eq. (6).
The deteriorated surfaces of selected samples were examined
by SEM and equipped with an energy dispersive spectroscopy
(EDS) detector. In addition, XRD (Cu-Ka) was performed on
samples to identify any compounds formed during the exposure
to sulfate solutions. Finally, thermal analysis including thermo-
gravimetric analysis (TGA) and differential scanning calorimetry
(DSC) was conducted on powder samples collected from the sur-
face layer (0–2 mm depth from the exposed surface) and second
layer (2–4 mm inside the surface layer of concrete) of selected
specimens.
3. Results and discussion
3.1. Visual inspection
It is clearly observed from Fig. 4 that the degree of surface dete-
rioration is obviously different in the three types of solutions. The
corrosion products are formed as veins sub-parallel to the surface
of specimens when exposed to sodium sulfate solution after
210 days, cracks have formed along the corners and edges
(Fig. 4a). The surface of concrete in magnesium sulfate solution
becomes uneven and part of coarse aggregates exposed in some
severely scaled specimens (Fig. 4b). Compared with concrete in
sodium and magnesium sulfate solution, concrete specimens exhi-
bit slightly attack with only a surface layer of mortar scaled when
exposed to the composite solution (Fig. 4c). Visual inspection
l5
l4
l/mm
l3
l0
l2
l1
t1 t2 t0 t3 t4 t5
t/μs
Fig. 3. Relationship between travel time and test distance.
91
revealed that the damage of concrete specimens exposed to mag-
nesium sulfate solution is more severe than that exposed to the
other sulfate solutions. The results corroborate the findings of
other researchers [34,35]. However, these studies have been con-
ducted on cement pastes or mortars.
3.2. RDME loss
ð7Þ
As can be seen from the Fig. 5, after 360 days, the RDME loss of
concrete in sodium sulfate, magnesium sulfate and composite
ð8Þ solution are 18.8%, 25.9% and 13.2%, respectively. The RDME loss
of concrete in the magnesium sulfate solution is greater than that
in the other solutions. The behavior of RDME loss of concrete in the
different sulfate solutions is similar. There exhibits three distinct
periods: (I) the decreased period, from the initial immersion to
90 days, (II) the steady period, from 90 to 180 days and (III) the
accelerating decreased period, from 180 to 360 days. The
emergence of the first phase may be due to the damage of
drying-wetting cycles and salt crystals crystallization. The reason
that period II occurs is that sulfate chemical reaction leads to the
ð9Þ
formation of ettringite and gypsum, increasing the compactness
of concrete to a certain extent. Therefore, the RDME loss is reduced
for a certain period of time. However, sulfate attack results in the
formation of more microcracks in concrete during later corrosion
time, which accelerates the RDME loss in period III. The RDME loss
of concrete in sodium sulfate solution in first stage is the greatest,
but the RDME exhibits a significant drop in magnesium sulfate
solution after 180 days. It is clear from the figure that the deterio-
ration by the magnesium sulfate solution covers the most aggres-
sive corrosion on concrete. Fig. 5 also shows that the RDME loss of
concrete is the least when exposed to composite solution. This
indicates that the presence of chloride ions delays the deterioration
of concrete caused by sodium sulfate attack.
The deterioration because of magnesium sulfate covers more
aggressive corrosion on concrete. Firstly, the Ca ion in the calcium
silicate hydrate (C-S-H) can be replaced by the Mg ion, leading to
the formation of non-cementitious magnesium silicate hydrate
(M-S-H) and the loss of the cementitious structure [36]. Secondly,
the formation of ettringite and gypsum due to the sulfate ion cause
expansion and cracks and thus significantly lead to the deteriora-
tion of concrete. Because of the simultaneous significant decompo-
sition of the C-S-H that accompanies the formation of ettringite
and gypsum, the corrosive action of magnesium sulfate is greater.
The existence of chloride ions reduces the diffusion rate of sul-
fate irons, and the deterioration rate of concrete could be inhibited
effectively. On the one hand, each diffusion rate of anion in the
composite solution of sulfate and chloride is less than that in single
diffusion for concrete exposed to drying-wetting cycles according
to charges-equilibrium and mass-equilibrium principle. On the
other hand, chloride ion can react with hydration product of con-
crete to form the Friedel’s salt, which could delay the corrosion
product formed in concrete, finally reduces the damage degrada-
tion of concrete under sulfate attack. The results are in a good
agreement with what were suggested by Liang et al. [37], which
showed that the presence of chloride in composite solution
decreases the quantity of ettringite in corrosion process and
reduces the rate of crack propagation. And the chlorine salt con-
centration is higher, the delayed effect is more obvious.
3.3. Weight loss
As can be seen from the Fig. 6, the weight loss of concrete spec-
imens exposed to sulfate solution under drying-wetting cycles
exhibits two distinct stages: steady stage and increased stage. In
stage I, the weight loss has little decrease when exposed to sodium
sulfate solution. In magnesium sulfate, the weight loss keeps
92 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98
Fig. 4. Visual inspection of concrete exposed to sulfate solutions under drying-wetting cycles. (a) in sodium sulfate solution; (b) in magnesium sulfate solution; (c) in
composite solution.
Fig. 5. RDME in concrete exposed to sulfate solutions under drying-wetting cycles.
nearly at a constant before 150 days and then increases at a steady
rate. As sulfate attack occurs continuously, the weight loss of con-
crete exposed to the magnesium sulfate solution is greater in later
corrosion time. The increase rate of weight loss was seen to be the
highest in magnesium sulfate solution. However, the weight loss of
concrete is least when exposed to the composite solution. The
weight loss in composite solution is less than half of that in the
other solutions after 360 days of drying-wetting cycles.
3.4. Compressive strength loss
The results from the compressive strength loss are presented in
Fig. 7. The behavior of concrete specimens in the three types of sul-
fate solutions is similar. There is an initial increase in the compres-
sive strength, indicating a deposition of corrosion products in the
pore spaces. This followed by a steady drop after 60 days. As can
be seen from the Fig. 7, the compressive strength of concrete in
sodium sulfate, magnesium sulfate and composite solution are
Fig. 6. Weight loss in concrete exposed to sulfate solutions under drying-wetting
cycles.
decreased by 12.7%, 14.8% and 8.4% after 360 days, respectively.
The damage rate of concrete in the sodium and magnesium sulfate
solution has been greatly accelerated after 180 days. Obviously, the
rate of the drop was seen to be higher when concrete exposed to
the magnesium sulfate solution. However, there is no obvious
decrease of compressive strength loss during the initial 240 days
when concrete exposed to the composite solution. The compres-
sive strength loss of concrete in composite solution is found to
be less than that in the other sullfate solutions. The results of com-
pressive strength loss agree with those of RDME loss shown in
Fig. 5.
3.5. The damage layer thickness of concrete
It is known that the presence of horizontal layers that are
formed when concrete exposed to an aggressive environment.
 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98
Fig. 7. Relatively compressive strength in concrete exposed to sulfate solutions
under drying-wetting cycles.
Fig. 8. Ultrasonic velocity in the damage layer of concrete exposed to sulfate
solutions under drying-wetting cycles.
Along with the development of ultrasonic testing technique, non-
destructive testing methods were widely used in the durability of
concrete research. The ultrasonic penetration testing can be used
to locate the interface between solid and corroded concrete. A
one-side non-destructive testing method was used to detect the
surface damage of concrete according to the propagation speed dif-
ference of wave velocity in different materials. By measuring the
damage layer thickness of concrete (defined as Hf) and the ultra-
sonic velocity in the damage layer (defined as Vf), the deterioration
degree of concrete could be estimated effectively.
The damage of concrete caused by sulfate attack is a process
from outside to inside. The damage layer of concrete is not obvious
at the beginning of the corrosion time. As the deterioration degree
increases, the damage layer of concrete could be calculated accu-
rately after 210 days in the experiment. Figs. 8 and 9 show the
results of damage layer of concrete exposed to different sulfate
solutions. Cracks initiated by sulfate attack reduce the Vf of con-
crete. It is shown in Fig. 8 that the ultrasonic velocity Vf in the dam-
age layer decreases with corrosion time. Obviously, the Vf was seen
to be the lowest in magnesium sulfate solution. Also, The Vf of con-
crete in composite solution exhibits the slowest reduction.
As seen from Fig 9, the Hf increases with an increase of the cor-
rosion time, which exhibits the opposite trend compared with Vf.
The Hf in magnesium sulfate solution is thicker than that in the
other sulfate solutions, indicating that magnesium sulfate covers
the most aggressive corrosion. The experimental results of Hf in
composite solution showed that the chloride ion has significant
93
Fig. 9. Damage layer thickness of concrete exposed to sodium sulfate solutions
under drying-wetting cycles.
Fig. 10. Relationship between RDME and damage layer thickness of concrete.
inhibiting effect on the deterioration degree of concrete. The Hf
in composite solution increased slowly with corrosion time, and
the damage is not obvious compared with concrete in the other
sulfate solutions. The results are in a good agreement with what
were obtained by Zhang et al. [38,39], which showed that the com-
pactness reduces and the deterioration degree of concrete
increases when the damage layer of concrete is thicker and the
ultrasonic speed is lower under sulfate attack.
Fig. 10 reveals the relationship between RDME and Hf of con-
crete. It is shown that the temporal variation of Hf are in good
agreement with the RDME after 210d, there is a significant relativ-
ity between them. Obviously, the damage layer thickness of con-
crete increases with reduction of RDME, indicating that the
damage degree of concrete increases. Therefore, the deterioration
degree of concrete could be estimated effectively by measuring
the damage layer.
3.6. Microstructural investigations by SEM
Microstructure of concrete specimens at different exposure
times under the combined actions of sulfate attack and drying-
wetting cycles are shown in Fig. 11. A small quantity of corrosion
products needle-like ettringite can be seen in concrete pores after
90 days in Fig. 11a. As the corrosion time growth, the amount of
products continuously increases after 180 days, which is shown
in Fig. 11b. After 270 days of drying-wetting cycles, the pores are
almost completely filled up by needle-like ettringite (Fig. 11c).
94 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98

Moreover, a large amount of columnar gypsum crystals can be seen

in the pores (Fig. 11d). The corrosion products of ettringite and
gypsum can be also found in the magnesium sulfate solution and
the composite solution during the drying-wetting cycles. When
the tensile stress generated by the corrosion products exceeds
the tensile strength of concrete, resulting in occurrence of cracking.
When concrete exposed to sodium sulfate and composite solu-
tion, there were some small pieces peeled off and a layer of white
substance was found on the concrete surface. By using SEM, a large
amount of petal shaped crystals are presented in Fig. 12a. To char-
acterize the chemical compositions of these crystals, careful obser-
vation of the EDS spectra showed that the elements O, Na and S in
Fig. 12b. This indicated that the white substance on the surface of
concrete is thenardite. In the alternate action of drying-wetting
cycles, concrete specimens are attacked by corrosion products such
as ettringite and gypsum during the wetting cycle. While, the con-
centration of the hole solution in concrete increases rapidly as a
result of the evaporation action during the drying cycle. The crys-
tallizing pressure due to the rapidly growing thenardite crystals
lead to the damage in concrete when the supersaturated state is
attained. However, this phenomenon has not appeared when con-
crete exposed to magnesium sulfate solution.
A large amount of dense popcorn-like crystals could be also
observed in the pores, as shown in Fig. 13a. The EDS spectra of
these crystals indicated that they are calcium carbonate, which
detected the elements C, O and Ca (Fig. 13b). This means that part
of the hydrated cement paste of concrete has been carbonated dur-
ing the long time test. The results are in a good agreement with
what were obtained in literature [40], which showed that the cal-
cium carbonate could also be discovered in cement paste speci- Fig. 12. SEM micrograph and EDS spectrum of the petal shaped crystals in concrete.
(a) SEM; (b) EDS.
mens when exposed to immersion tests in sulfate environment.

3.7. XRD analysis determining concrete exposed to sodium and magnesium sulfate solution, the
peaks selected for the qualitative analysis of these phases are
Fig. 14 presents the XRD patterns for concrete exposed to the ettringite, gypsum, calcium hydroxide, quartz and calcium carbon-
three types of sulfate solutions under drying-wetting cycles. When ate in these diagrams. The ettringite and gypsum peaks are strong,

Fig. 11. Microstructure of concrete specimens at different exposure times in the sodium sulfate solution. (a) ettringite at 90 days; (b) ettringite at 180 days; (c) ettringite at
270 days; (d) gypsum at 270 days.
 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98
Fig. 13. SEM micrograph and EDS spectrum of the popcorn-like crystals in concrete.
(a) SEM; (b) EDS.
indicating that they are main corrosion products. The results cor-
roborate the findings that were suggested by Jin et al. [28], which
showed that the corrosion products of concrete are ettringite and
gypsum when exposed to drying-immersion cycles in the 5%
sodium sulfate solution and a composite solution of 3.5% sodium
chloride and 5% sodium sulfate. Fig. 14 also shows that the concen-
trations of gypsum and ettringite after 270 days are higher than
those after 180 days. Also, the concentration of calcium hydroxide
gradually decreases along with corrosion time, indicating that the
deterioration rate of concrete increases. It is also observed that the
concentrations of gypsum and ettringite in concrete exposed to
magnesium sulfate solution are higher than those in sodium sul-
fate solution. In addition, the concentration of calcium hydroxide
of concrete in magnesium sulfate solution is the lowest.
When concrete exposed to the composite solution, there still
have small fractions of Friedel’s salt seem to be present in these
diagrams. Moreover, the concentrations of gypsum and ettringite
in concrete are the lowest, indicating the damage of concrete in
the composite solution is not obvious. By considering the XRD
analysis, it seems that the magnesium sulfate solution cover the
most damage degradation on concrete. The results agree with the
macro-properties of concrete, such as RDME loss and compressive
strength loss.
3.8. The thermal analysis
The instrument used for TG-DSC was the SDT-Q600 thermal
analysis system. This experiment was done to determine the
amounts of phases forming at different stages of sulfate attack.
The phases of interest are the corrosion products of ettringite
and gypsum. Fig. 15 shows the typical peak temperature from
95
Fig. 14. XRD patterns of concrete subjected to drying-wetting cycles in sulfate
solutions. (a) 180d; (b) 270d (E:ettringite; G:gypsum; CH: calcium hydroxide; Q:
quartz; C: calcium carbonate; F: Friedel’s salt).
DSC curves, which displaying three typical endothermic peaks that
lie in the ranges 90–110 °C, 130–140 °C and 430–450 °C. Previous
studies indicated that these three temperature stages correspond
to dehydration and decomposition of ettringite, dehydration and
decomposition of gypsum, and decomposition of calcium hydrox-
ide, respectively [41,42].
According to the temperature corresponding to heat absorption
peak in the DSC curves, the chemical product is identified, with the
weight loss shown simultaneously in the TG curves. The quantity
of ettringite formed in concrete is shown in Fig. 16. For concrete
in sodium and magnesium sulfate solution, the quantity of ettrin-
gite formed in the first layer is higher than that in the second layer
in early test period. However, the experiment results exhibit the
opposite trend after 180 days. Test results also show that the quan-
tity of ettringite increases firstly and then decreases in each layer
with corrosion time. The quantity of ettringite increases gradually
in the first 270 days in sodium sulfate solution, then reduces in
later test period. This phenomenon also occurs when concrete
exposed to magnesium sulfate solution, but the quantity of ettrin-
gite is reduced after 180d. This may be due to the surface scaling
and ettringite decomposition in concrete. Ettringite is not stable
when the PH falls below 11.5–12 at high sulfate concentrations
environment [43]. Calcium hydroxide most probably has reacted
with sulfate ions, which reduces the alkalinity of the pore solution.
Therefore, the ettringite could decompose to form gypsum. It is
also shown that the quantity of corrosion products in magnesium
sulfate solution is higher than the quantity in sodium sulfate
solution.
96 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98
Fig. 15. Typical peak temperature (from DSC curves). (a) The first layer of concrete;
(b) the second layer of concrete; S, M and SC denote the sodium sulfate solution,
magnesium sulfate solution and composite solution of sodium sulfate and sodium
chloride, respectively; T1, T2, T3, T4 denote the drying-wetting days of 90, 180, 270,
300, respectively.
Fig. 16. The quantity of ettringite formed in concrete; M, S and SC denote the
magnesium sulfate solution, sodium sulfate solution and composite solution of
sodium sulfate and sodium chloride, respectively.
For concrete in composite solution, the quantity of ettringite
formed is approximately proportional to the corrosion time. The
quantity of ettringite decreases from the surface to the inside, as
well as increases gradually in each layer with the increase of corro-
sion time. However, the quantity of ettringite is significantly less
than that of ettringite in sodium and magnesium sulfate solution.
The quantity of gypsum formed in concrete is shown in Fig. 17.
The results revealed that the quantity of gypsum increases gradu-
ally with corrosion time in the three types of sulfate solutions. It is
also shown that the quantity of gypsum in concrete is significantly
less than that of ettringite in the experiment. Test results show
that the quantity of gypsum in sodium sulfate solution is less than
that in magnesium sulfate solution, and some amount of gypsum
has been checked in the second layer only after 90 days. In the
composite solution, the quantity of gypsum is the lowest at the
same corrosion time, especially is no gypsum detected during the
first 180 days. Moreover, the quantity of gypsum in concrete
decreases from the surface to the inside.
The corrosion products of concrete were seen to increase with
an increase of the concentration of sulfate solution. Sulfate ions
concentration in the exposure solution is a factor not only increas-
ing sulfate attack intensity, but also altering the mechanism of sul-
fate attack. When concrete exposed in the early corrosion time or
in the second layer, sulfate attack is controlled by ettringite forma-
tion because of the lower concentration of sulfate ions. While, sul-
fate attack is controlled by ettringite and gypsum formation as
corrosion time increases or in the first layer of concrete, due to
the higher concentration of sulfate ions.
According to Fig. 18, the quantity of calcium hydroxide
decreases gradually as a result of the reaction forming corrosion
products with corrosion time increases in the three types of sulfate
solutions. Moreover, the quantity of calcium hydroxide in concrete
increases from the surface to the inside, which proves the damage
caused by sulfate attack is proceed from the exterior to the interior.
In sodium sulfate solution, the calcium hydroxide in the first layer
is completely consumed after 360 days. While in magnesium sul-
fate solution, the calcium hydroxide has been exhausted only after
270 days, which shows the deterioration extent is more serious
than that in the sodium sulfate solution. For concrete in composite
solution, the quantity of calcium hydroxide is higher than that in
the other sulfate solutions, indicating that the deterioration rate
of concrete could be inhibited effectively.
According to the above study, the quantity of corrosion prod-
ucts does not correspond well with the observations of the physical
properties, and extensive amount of corrosion products can be
detected before obvious damage occurs. It could be explained by
the formation of corrosion products in concrete under sulfate
attack. In the first corrosion time, little expansion occurs while
the corrosion products fill the voids, so the damage of concrete is
not obvious. As corrosion time increases, the continuing formation
of ettringite and gypsum cause expansion and cracks in concrete.
Fig. 17. The quantity of gypsum formed in concrete.
 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98
Fig. 18. The quantity of calcium hydroxide in concrete.
The deterioration degree of concrete increased significantly after
180 days.
4. Conclusions
When concrete exposed to sodium sulfate, magnesium sulfate
and composite solution under drying-wetting cycles, the weight
loss can be classified to two stages: the steady stage and increased
stage. The RDME of concrete exhibits three distinct periods: the
decreased period, steady period, and accelerating decreased period.
The compressive strength loss of concrete exhibits the following
three periods: increased period, decreased period and accelerating
decreased period. The degradation rate of concrete has been
greatly accelerated after 180 days of drying-wetting cycles. The
ultrasonic velocity in the damage layer decreases and the damage
layer thickness of concrete increases with corrosion time. When
the damage layer thickness of concrete is thicker as well as the
ultrasonic speed is lower, indicating that the deterioration degree
of concrete increases.
The different type of sulfate solution has a great influence on
the quantity of corrosion products. In early test period, the quan-
tity of ettringite in the first layer is higher than that in the second
layer when concrete exposed to sodium and magnesium sulfate
solution, but then the quantity exhibits the opposite trend. The
quantity of ettringite increases firstly and then decreases in each
layer with corrosion time. While, the quantity of ettringite in the
composite solution decreases from the surface to the inside, as well
as increases gradually in each layer with corrosion time increases.
Test results also show that the quantity of gypsum in concrete is
less than the quantity of ettringite. In some concrete specimens,
only a small quantity of gypsum is checked after a certain corro-
sion extent, especially exposed to the composite solution. More-
over, the quantity of gypsum in concrete decreases from the
surface to the inside, and increases at a steady rate with corrosion
time. The quantity of corrosion products in magnesium sulfate
solution is higher than the quantity in the other sulfate solutions.
In the composite solution, the quantity of corrosion products is
the lowest at the same corrosion time.
The deterioration caused by magnesium sulfate solution covers
the most aggressive corrosion on concrete due to the simultaneous
dissolution of cement hydrates that accompanies the formation of
ettringite and gypsum. In the composite solution, chloride ions
have a mitigating effect on sodium sulfate attack. The deterioration
degree of concrete is delayed by the presence of chloride ions.
The quantity of corrosion products does not correspond well
with the observations of the physical properties. There have exten-
97
sive amount of corrosion products can be formed before obvious
damage is occurred. Therefore, the deterioration degree of concrete
under sulfate attack could not be simply judged by the quantity of
corrosion products in concrete.
Acknowledgements
This project was supported by the Program for Changjiang
Scholars and Innovative Research Team in University (IRT13089)
and supported by the National Natural Science Foundation of China
(No. 51308445).
References
[1] M. Santhanam, M.D. Cohen, J. Olek, Mechanism of sulfate attack: a fresh look
Part 2: proposed mechanisms, Cem. Concr. Res. 33 (2003) 341–346.
[2] W. Mullauer, R.E. Beddoe, D. Heinz, Sulfate attack expansion mechanisms,
Cem. Concr. Res. 52 (2013) 208–215.
[3] E. Roziere, A. Loukili, R. El Hachem, et al., Durability of concrete exposed to
leaching and external sulphate attacks, Cem. Concr. Res. 39 (2009) 1188–1198.
[4] N.M. Al-Akhras, Durability of metakaolin concrete to sulfate attack, Cem.
Concr. Res. 36 (2006) 1727–1734.
[5] J.B. Wang, D.T. Niu, Y.L. Zhang, Mechanical properties, permeability and
durability of accelerated shotcrete, Constr. Build. Mater. 95 (2015) 312–328.
[6] W.Y. Ouyang, J.K. Chen, M.Q. Jiang, Evolution of surface hardness of concrete
under sulfate attack, Constr. Build. Mater. 53 (2014) 419–424.
[7] M.T. Bassuoni, M.L. Nehdi, Durability of self-consolidating concrete to sulfate
attack under combined cyclic environments and flexural loading, Cem. Concr.
Res. 39 (3) (2009) 206–226.
[8] M.T. Bassuoni, M.L. Nehdi, Durability of self-consolidating concrete to different
exposure regimes of sodium sulfate attack, Mater. Struct. 42 (8) (2009) 1039–
1057.
[9] J.M. Gao, Z.X. Yu, L.G. Song, Durability of concrete exposed to sulfate attack
under flexural loading and drying-wetting cycles, Constr. Build. Mater. 39
(2013) 33–38.
[10] Salah U. Al-Dulaijan, Sulfate resistance of plain and blended cements exposed
to magnesium sulfate solutions, Constr. Build. Mater. 21 (2007) 1792–1802.
[11] H. Siada, M. Lachemi, S.K. Bernard, M. Sahmaran, A. Hossain, Assessment of the
long-term performance of SCC incorporating different mineral admixtures in a
magnesium sulphate environment, Constr. Build. Mater. 80 (2015) 141–154.
[12] A.M. Diab, A.E.M. Awad, H.E. Elyamany, A.E.M. Abd Elmoaty, Guidelines in
compressive strength assessment of concrete modified with silica fume due to
magnesium sulfate attack, Constr. Build. Mater. 32 (2012) 311–318.
[13] J. Geng, D. Easterbrook, L.Y. Li, L.W. Mo, The stability of bound chlorides in
cement paste with sulfate attack, Cem. Concr. Res. 68 (2015) 211–222.
[14] M.H. Zhang, J.K. Chen, Y.F. Lv, D.J. Wang, J. Ye, Study on the expansion of
concrete under attack of sulfate and sulfate-chloride ions, Constr. Build. Mater
39 (2013) 26–32.
[15] J.X. Xu, C.K. Zhang, L.H. Jiang, L. Tang, G.F. Gao, Y.P. Xu, Releases of bound
chlorides from chloride-admixed plain and blended cement pastes subjected
to sulfate attacks, Constr. Build. Mater. 45 (2013) 53–59.
[16] M. Santhanam, M. Cohen, J. Olek, Differentiating seawater and groundwater
sulfate attack in Portland cement mortars, Cem. Concr. Res. 36 (12) (2006)
2132–2137.
[17] M. Maes, N.D. Belie, Resistance of concrete and mortar against combined
attack of chloride and sodium sulphate, Cem. Concr. Compos. 53 (2014) 59–72.
[18] Y.J. Chen, J.M. Gao, L.P. Tang, X.H. Li, Resistance of concrete against combined
attack of chloride and sulfate under drying–wetting cycles, Constr. Build.
Mater. 106 (2016) 650–658.
[19] L. Jiang, D.T. Niu, L.D. Yuan, Q.N. Fei, Durability of concrete under sulfate attack
exposed to freeze-thaw cycles, Cold Reg. Sci. Techol. 112 (2015) 112–117.
[20] J.H. Chen, S.B. Zhao, J.T. Yao, Thermal analysis of ettringite in concrete attacked
by sodium sulfate in dry-wet cyclic environment, J. Basic Sci. Eng. 18 (6) (2010)
950–958.
[21] M. Sahmaran, T.K. Erdem, I.O. Yaman, Sulfate resistance of plain and blended
cements exposed to wetting-drying and heating-cooling environments,
Constr. Build. Mater. 21 (2007) 1771–1778.
[22] D.T. Niu, Y.D. Wang, R. Ma, J.B. Wang, S.H. Xu, Experiment study on the failure
mechanism of dry-mix shotcrete under the combined actions of sulfate attack
and drying–wetting cycles, Constr. Build. Mater. 81 (2015) 74–80.
[23] J. Yuan, Y. Liu, Z.C. Tian, B.K. Zhang, Investigating the failure process of
concrete under the coupled actions between sulfate attack and drying–wetting
cycles by using X-ray CT, Constr. Build. Mater. 108 (2016) 129–138.
[24] F. Zhang, S.C. Li, J.S. Ye, S.K. Li, Evolution model of concrete failure surface
under coupling effect of seawater freeze-thaw and erosion, J. Southeast Univ.
Nat. Sci. Ed. 27 (2) (2011) 206–209.
[25] Y.N. Liang, Y.S. Yuan, Mechanism of concrete destruction under sodium sulfate
and magnesium sulfate solution, J. Chin. Cer. Soc. 35 (4) (2007) 504–508.
[26] Q.B. Yang, B.R. Zhu, Deterioration of concrete due to action of salt
crystallization, J. Build Mater. 10 (4) (2007) 392–396.
[27] S.U. Al-Dulaijan, M. Maslehuddin, M.M. Al-Zahrani, A.M. Sharif, M. Shameem,
M. Ibrahim, Sulfate resistance of plain and blended cements exposed to
98 L. Jiang, D. Niu / Construction and Building Materials 117 (2016) 88–98
varying concentrations of sodium sulfate, Cem. Concr. Compos. 25 (2003) 429–
437.
[28] Z.Q. Jin, W. Sun, Y.S. Zhang, J.Y. Jiang, J.Z. Lai, Interaction between sulfate and
chloride solution attack of concretes with and without fly ash, Cem. Concr. Res.
37 (8) (2007) 1223–1232.
[29] X.B. Chen, M.X. Tang, K.L. Ma, Underground concrete structure exposure to
sulfate and chloride invading environment, J. Central South Univ. Sci. Technol.
Ed. 43 (7) (2012) 2803–2812.
[30] GB/T 50082-2009, Standard for Test Method of Long-Term Performance and
Durability of Ordinary Concrete, Chinese Standard Institution Press, 2009.
[31] GB/T 50081-2002, Standard for Test Method of Mechanical Properties on
Ordinary Concrete, Chinese Standard Institution Press, 2002.
[32] CECS 21:2000, Technical Specification for Inspection of Concrete Defects by
Ultrasonic Method, Chinese Standard Institution Press, 2000.
[33] P.K. Mehta, P.J. Monteiro, Concrete: Microstructure, Properties, and Materials,
Mc Graw Hill press, New York, 2006.
[34] S.T. Lee, R.D. Hooton, H.S. Jung, D.H. Park, C.S. Choi, Effect of limestone filler on
the deterioration of mortars and pastes exposed to sulfate solutions at
ambient temperature, Cem. Concr. Res. 38 (1) (2008) 68–76.
[35] J.K. Chen, M.Q. Jiang, J. Zhu, Damage evolution in cement mortar due to erosion
of sulphate, Corros. Sci. 50 (9) (2008) 2478–2483.
[36] J. Skalny, J. Marchand, I. Odler, Sulfate Attack on Concrete, Spon Press, UK,
2002.
[37] Y.N. Liang, J.Y. Huang, X.J. Lin, The effect of chloride on concrete under sulfate
attack, J. Fuzhou Univ.-Nat. Sci. Ed. 39 (6) (2011) 1–5.
[38] F.J. Zhang, Y.S. Yuan, J.M. Du, Ultrasonic detection in concrete structures of
damage from sulfate attack, J. Chin. Univ. Min. Technol. 40 (3) (2011) 373–378.
[39] F. Zhang, J.J. Cai, S.C. Li, Ultrasonic detection of freeze-thaw damage thickness
of concrete, J. Shenzhen Univ. Sci. Eng. 29 (3) (2012) 207–210.
[40] Z.Q. Liu, D.H. Deng, Geert De Schutter, Z.W. Yu, Micro-analysis of ‘‘salt
weathering” on cement paste, Cem. Concr. Compos. 33 (2) (2011) 178–191.
[41] P. Kresten, G. Berggren, The thermal decomposition of thaumasite,
Thermochim. Acta 20 (6) (1975) 1–6.
[42] R.D. Gao, Q.B. Li, S.B. Zhao, X.M. Yang, Deterioration mechanisms of sulfate
attack on concrete under alternate action, J. Wuhan Univ. Technol. Mater. Sci.
25 (2) (2010) 355–359.
[43] M. Santhanam, Studies on Sulfate Attack: Mechanisms, Test Methods, and
Modeling, Purdue University, West Lafayette, 2001.