Construction and Building Materials 133 (2017) 171–181

Contents lists available at ScienceDirect

Construction and Building Materials

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

Durability study on engineered cementitious composites (ECC) under

sulfate and chloride environment
Hezhi Liu a,b,c, Qian Zhang d, Victor Li e, Huaizhi Su a,b,c,⇑, Chongshi Gu a,b,c,⇑
a
State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, China
b
College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing 210098, China
c
National Engineering Research Center of Water Resources Efficient Utilization and Engineering Safety, Hohai University, Nanjing 210098, China
d
Department of Civil and Environmental Engineering, University of Louisiana at Lafayette, United States
e
Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI, United States

h i g h l i g h t s

ECC remains durable after 200 days of exposure to sulfate and sulfate-chloride environment.

Long-term exposure leads to compressive and tensile strength increase of ECC.

Long-term exposure leads to slight reduction of tensile strain capacity of ECC.

Composite level behavior of ECC is a result of changed matrix fracture toughness and fiber/matrix interfacial bond.

Tensile ductility reduction of ECC is attributed to reduced crack number and width.

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

Article history: The lack of durability of concrete hydraulic structures, especially under sulfate and chloride environment,
Received 17 August 2016 has become a worldwide problem. The present paper investigated the feasibility of applying ductile engi-
Received in revised form 16 October 2016 neered cementitious composites (ECC) as an alternative to conventional concrete in hydraulic structures
Accepted 16 December 2016
to improve their durability performance. Specifically, the durability of ECC under sulfate and combined
Available online 24 December 2016
sulfate-chloride conditions were studied. Compressive and tensile behavior of ECC after long-term expo-
sure to Na2SO4 and Na2SO4 + NaCl solutions was experimentally characterized. Subsequently, microme-
Keywords:
chanical study was adopted to investigate the microscopic mechanisms underlying the composite level
Engineered cementitious composites (ECC)
Durability
behavior of ECC under those aggressive environments. The research findings demonstrated that ECC
Sulfate remains durable and maintains its high mechanical performance even after 200 days of exposure to con-
Chloride centrated sulfate and combined sulfate-chloride environments. In addition, the micromechanics-based
Micromechanics study quantitatively characterized the influence of sulfate and sulfate-chloride solutions on the microme-
Hydraulic structures chanical parameters including matrix fracture toughness and fiber/matrix interfacial bond, providing
insights into the underlying mechanisms of the composite level behavior of ECC under those
environments.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction structures even before their intended service life is reached [2].
For example, the Gutianxi II hydropower station in China faced
The durability performance of concrete hydraulic structures in severe challenges from concrete aging, curtain grouting failure
field conditions are often called into question. Most hydraulic and other problems associated with deterioration only after oper-
structures are designed for a service life of 50 to 100 years depend- ation for 30 years, and about four million dollars were spent on
ing on their specific application [1]. However, poor durability per- maintaining and repairing the structure [3]. Similarly, unsatisfac-
formance of concrete material often causes deterioration of tory performance of aging hydraulic structures also troubles the
United States. According to the 2013 report card published by
⇑ Corresponding authors at: State Key Laboratory of Hydrology-Water Resources American Society of Civil Engineers (ASCE) [4], the average age of
and Hydraulic Engineering, Hohai University, Nanjing 210098, China. dams in the United States is 52 years. The overall number of
E-mail addresses: [email protected] (H. Su), [email protected] (C. Gu). high-hazard dams reached nearly 14,000 in 2012 and the number

http://dx.doi.org/10.1016/j.conbuildmat.2016.12.074
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
172 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181

continues to grow. The required investment for repairing these sulfate ions and cement hydration products form expansive
aging and high-hazard dams is estimated to be up to 21 billion dol- compound, causing cracking of concrete. Cracks in concrete allows
lars. A worldwide engineering challenge to enhance the durability the sulfate ions to penetrate into the interior of the structure, fur-
of hydraulic structures and to lower the maintenance cost has ther accelerating the deterioration. As a result, hydraulic structures
emerged. serving in sulfate rich environment like Northwest China could fail
In field conditions, the most common cause of insufficient dura- in less than 5 years, which leads to huge economic loss and social
bility of hydraulic structures is the development of cracks within impacts [26,27].
concrete [5,6]. For concrete dams, various causes including Most previous researches on concrete (and ECC) durability only
mechanical loading, chemical attack, thermal contraction, shrink- considered a single deterioration mechanism (such as sulfate
age, and foundation settlement can result in concrete cracking attack) which is inconsistent with service condition. In reality, con-
[7]. It is generally agreed that cracking in concrete accelerates crete structures may be exposed to combined environmental load-
deterioration by providing easy access for aggressive ions (e.g. sul- ings (such as combined sulfate and chloride attack). This is
fate and chloride ions) to get into the interior of the structure [8]. particularly true for Northwest China where more than 1000 salt
Based on previous study [8,9], the influence of cracks on the corro- lakes are located and also in most marine environments. Under
sion process is negligible when the cracks are relatively narrow; those environments, the durability of hydraulic structures is
however, the corrosion rate increases significantly when cracks greatly challenged due to the combined sulfate and chloride attack.
exceed 100 lm in width. Hence, the crack width must be carefully Several studies have already pointed out that there exists a clear
controlled to achieve desirable durability performance of concrete interaction between the two phenomena [28,29]. Therefore, to
hydraulic structures. fully evaluate the potential of applying ECC in hydraulic structures,
Crack width control, however, is considered a significant chal- it is important to consider the coupled effect of sulfate and chloride
lenge for conventional concrete. Great efforts have been made to attack on the durability of ECC. However, no research on the per-
control crack initiation and propagation within concrete, such as formance of ECC under this aggressive environment has been con-
careful selection and proportioning of material, temperature con- ducted so far.
trol during construction and using steel reinforcement [10–12]. Unlike conventional concrete, ECC is designed based on
However, due to the inherent brittleness of concrete, under micromechanics. Micromechanics establishes the linkage between
mechanical and environmental loadings, cracking is almost inevi- composite level properties (e.g. tensile strength, tensile ductility,
table [13]. In addition, it is generally recognized that crack width and crack widths) and properties at microscale (e.g. matrix tough-
in concrete is difficult to control reliably. Therefore, a material with ness and fiber/matrix interfacial bond). Utilizing micromechanics
intrinsic and consistent crack width control capability is highly and by investigating the micromechanical properties of ECC, the
desirable. deterioration mechanisms of ECC composite under sulfate and
The feasibility of using Engineered Cementitious Composites chloride attack can be revealed.
(ECC) to control crack width and enhance structural durability The objective of the present research is to fully evaluate the
has been demonstrated in a number of studies [14–17]. ECC is a durability performance of ECC subjected to sulfate-chloride envi-
family of high performance fiber-reinforced cementitious compos- ronment, and to understand the underlying mechanisms of its
ites featuring high tensile ductility and tight crack width. Under deterioration under such conditions. For this purpose, the mechan-
tensile load, ECC exhibits a pseudo strain-hardening behavior by ical performance of ECC under sulfate and sulfate-chloride attack
developing multiple micro-cracks. The tensile strain capacity of were experimentally investigated. Specifically, compressive and
ECC reaches 3–5% (with less than 2% fiber content by volume), tensile properties of ECC after long-term exposure (up to 200 days)
which is 300–500 times that of normal concrete [16–18]. More were used to evaluate the durability of ECC under such conditions.
importantly, the micro-cracks are typically below 60 lm wide. In addition, micromechanical studies including matrix toughness
According to previous studies, the tight crack width of ECC is ben- tests and single-fiber pull-out tests were conducted to fully under-
eficial for preventing initiation of deterioration and slowing down stand the influence of sulfate and sulfate-chloride attack on ECC
the deterioration rate, leading to extended service life and material. The research findings here are expected to provide useful
enhanced serviceability of hydraulic structures. data and guidelines for future design and application of ECC in sur-
The durability of ECC under combined mechanical loading and face repair of hydraulic structures to improve their shielding
various environmental loadings has been extensively investigated performance.
in the past decades. Sahmaran et al. [19] and Li and Li [13]
reported that ECC remains durable and exhibits robust high ten- 2. Experiment procedures
sile ductility under chloride environment. Özbay et al. [20] inves-
tigated the durability of ECC (containing high volumes of slag) 2.1. Materials and mix proportions
subjected to sulfate attack and freeze-thaw cycles. The experi-
mental results indicate that the ductility of ECC specimens A regular ECC mix as found in the literature [30] was used in
decreases after 300 freeze-thaw cycles both in water and Na2SO4 this study. The constituent materials for this ECC mixture are Type
solution. In contrast, Lepech and Li [21] found no deterioration of I Portland cement, fine silica sand, class F fly ash, water, high range
ECC (containing fly ash) when exposed to freeze-thaw cycles. In water reducing agent (HRWRA), and poly-vinyl-alcohol (PVA)
addition, the durability of ECC under hot and humid environment fibers. The detailed mix proportion is given in Table 1. The PVA
[22], natural environment [23] and highly alkaline environment fiber used here has an average diameter of 39 lm and length of
[24] were also experimentally studied and demonstrated to be 12 mm. The tensile strength and maximum elongation of the fiber
much higher than that of conventional concrete. These prior are 1600 MPa and 6.0%, respectively. In addition, the surface of the
research show promise in using ECC to improve the long-term fibers is coated with 1.2% (by weight) proprietary hydrophobic
durability performance of concrete structures including hydraulic agent to modify the interaction between the fiber and cement
structures. matrix for better performance. Mortar specimens adopted from
Among all durability related issues, sulfate attack is considered previous study [19] were also prepared as a control to compare
one of the most critical environmental deterioration mechanisms with ECC. The mortar mix consists of Type I Portland cement, water
affecting the durability of concrete hydraulic structures [25]. When and natural river sands. The detailed mix proportion is also shown
in contact with sulfate solution, the chemical reactions between in Table 1.
 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181 173

Table 1
Mix proportions of ECC and Mortar.

Material Cement Aggregates Fly ash Water HRWRA Fiber

ECC 393 457 (silica sand) 865 311 5 26
Mortar 614 1534 (river sand) – 215 – –

2.2. Durability test procedures

A typical ECC mixing procedure [31] was followed to prepare all

ECC mixtures. After mixing, fresh ECC were cast into cube speci-
mens measuring 50.8  50.8  50.8 mm for compression tests,
and dogbone-shaped specimens (dimensions as shown in Fig. 1)
for uniaxial tension tests. Cube mortar specimens with the same
aforementioned size were also prepared for control purpose. All

LVDT-2
specimens were demolded after 24 h. Afterwards, all cube speci-

LVDT-1
mens were covered with wet paper towel and stored in a plastic
bag at a room temperature of 23 ± 3 °C until the age of 28 days.
Dogbone specimens were directly cured in laboratory air at a tem-
perature of 23 ± 3 °C until 28 days prior to being exposed to
sulfate-chloride environments.
In order to investigate the durability of ECC under sulfate and
chloride conditions, the following aggressive solutions suggested
by former studies [19,20] were prepared:

(1) 5% (by mass) Na2SO4 solution.

(2) 5% (by mass) Na2SO4 + 3% (by mass) NaCl solution.

At the age of 28 days, cube specimens (ECC and mortar) were

put into and kept in the above solutions for 30, 60, 90, 120, 200
and 420 days under laboratory condition, while dogbone speci-
mens were kept for 30, 60, 90, 120 and 200 days. Meanwhile, con-
trol cube and dogbone specimens were immersed in water for the Fig. 2. Uniaxial tensile test setup.
same period of time. A set of three specimens was tested for each
mixture/exposure condition.
After environmental exposure, the specimens were air-dried for strain-hardening behavior under tension, which can be expressed
24 h and tested for their compressive and tensile properties. Com- in energy and strength terms, as shown in Eqs. (1) and (2):
pression tests on ECC and mortar specimens were conducted using
a compression test system at a loading rate of 1300 ± 300 N/s in Z d0
K 2m
accordance with ASTM C109 [32]. Uniaxial tension tests were con- Energy-based criterion : J 0b  r0 d0  rðdÞdd P Jtip  ð1Þ
0 Em
ducted in accordance to the recommendations for direct tension
testing of High Performance Fiber Reinforced Cementitious Com-
posites by the Japan Society of Civil Engineers (JSCE) [33]. Uniaxial Strength-based criterion : minðr0 Þ > rc ð2Þ
tensile loading was imposed on the specimens under displacement
control at a rate of 0.5 mm/min. Two external linear variable differ- where J 0b is the fiber bridging complementary energy; Jtip is the frac-
ential transducers (LVDTs) were attached on each side of the spec- ture energy of the matrix; r0 and d0 are the maximum bridging
imen, with a gage length of approximately 80 mm, to measure the stress of the crack plane and corresponding crack opening; rðdÞ is
tensile strain. The tensile test setup is shown in Fig. 2. the fiber bridging stress versus crack opening relationship of the
crack plane; rc is the cracking strength of the matrix; Km and Em
2.3. Micromechanical investigation are matrix fracture toughness and matrix Young’s modulus,
respectively.
The core of the micromechanics-based ECC theory are the Energy based criterion describes the energy balance during
two conditions that must be met for the composite to show crack extension. Eq. (1) requires that the complementary energy
J0b (associated with fiber/matrix interfacial bond) to be greater than
the fracture energy of the matrix Jtip (associated with matrix tough-
330
ness), as illustrated in Fig. 3. This guarantees the formation of flat
cracks under tension instead of the normal Griffith cracks
[34–36], and formation of flat cracks allows more cracks with tight
12,7

Gage length
crack width to form without extensive fiber rupture/pullout
[15,16]. Strength based criterion requires the maximum fiber
bridging capacity to be higher than the matrix cracking strength.
60

It is obvious that the strength based criterion also need to be

met to allow formation of more than one cracks. Therefore, satis-
85 40 80 faction of both criteria is necessary to achieve strain-hardening
behavior of ECC. Otherwise, a single localized fracture, and normal
Fig. 1. Dimensions of dogbone specimens for uniaxial tension test (unit: mm). tension-softening behavior as in the case of FRC will result.
174 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181

the test is shown in Fig. 4. The beam specimens measure 305 mm

long, 76 mm wide and 38 mm deep and were cast from ECC mix-
tures without fibers. The specimens were cured in laboratory air
under 23 ± 3 °C and 20 ± 5% RH for 28 days before exposed to the
sulfate and sulfate-chloride solutions for 30, 60, 90 and 200 days.
After designated environmental exposure, the specimens were
air-dried for 24 h and tested for matrix fracture toughness. Prior
to the test, a thin notch (0.6 mm thick) was made on the tension
face of the beam specimen at mid-span (as shown in Fig. 4) to ini-
tiate the fracture. The notch depth to beam height ratio is 0.4 as
recommended by ASTM E399. Four specimens were tested for each
case.
Based on the measured fracture load during the test, the matrix
fracture toughness Km was determined. Assuming the Young’s
modulus of the matrix is the same as that of the composite (since
the fiber volume fraction is low), using the modulus of ECC com-
Fig. 3. Typical fiber bridging stress versus crack opening rðdÞ curve [14]. Hatched posite calculated from the tensile stress-strain relationship, Jtip
area represents maximum complementary energy J0b . Cross area represents the was calculated following the right hand side expression of Eq. (1).
fracture energy of matrix J tip .
In order to obtain J0b , single fiber pullout tests were conducted to
determine the interfacial parameters including chemical bond
Based on above strain-hardening criteria, two pseudo strain- strength Gd, frictional bond strength s0 and slip hardening coeffi-
hardening indices as defined in Eqs. (3) and (4) can be used to cient b. During the test, individual fibers with embedment lengths
predict the tensile performance of ECC composites based on of approximately 1 mm were pulled out of small ECC mortar
micromechanical properties. prisms, as shown in Fig. 5. The details of specimen preparation
and test configuration can be found in [39]. The preparation of sin-
J0b gle fiber specimens followed the same curing and conditioning
PSHE ¼ ð3Þ procedure as for the beam specimens. Considering that single fiber
J tip

minðr0 Þ rt
PSHS ¼ ¼ ð4Þ
rc rc
where rt is the (ultimate) tensile strength of ECC composite. As pre-
viously discussed, both indices need to be greater than 1 for ECC to
achieve strain-hardening behavior and high tensile ductility. Con-
sidering variability in material properties, in practice, PSHE and Single fiber
PSHS need to be greater than 3 and 1.2, respectively to achieve
robust strain-hardening behavior [37]. In fact, PSHs index typically
does not govern for PVA-ECC, while PSHE index is considered as
the dominant factor that determines the tensile behavior of ECC.
Therefore, only PSHE index was considered and discussed in this Specimen
study.
To quantitatively assess the influence of sulfate and sulfate- Load cell
chloride attack on micromechanical properties of ECC, matrix frac-
ture toughness test and single fiber pullout test were conducted.
Matrix fracture toughness test (in accordance with ASTM E399
[38]) uses a three-point bend loading on notched beam specimens
made of ECC matrix (without fibers). The detailed configuration of Fig. 5. Single-fiber pullout test setup.

Fig. 4. Matrix fracture toughness test setup.

 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181 175

test, by nature exhibits large scatter in the data, twenty specimens For mortar specimens subjected to Na2SO4 solution and
were prepared and tested for each case. Once the single fiber pull- Na2SO4 + NaCl solution, the strength continues to increase up to
out behavior was characterized, by considering the average pull- 120 days and to 200 days respectively, after which significant
out behavior of all the fibers on a cross section, the fiber bridging drops were observed for specimens under both exposure condi-
stress-crack opening (r(d)) relationship can be derived [14]. The tions. Specifically, the compressive strength decreased from
complementary energy J 0b was then calculated from r(d) relation- 71 MPa and 73 MPa to around 56 MPa respectively for specimens
ship. Together with the matrix fracture toughness from the Km exposed to Na2SO4 solution and Na2SO4 + NaCl solution. This was
measurement, PSHE index was determined. also accompanied by surface cracks after 420 days (Figs. 7 and 8).
Mortar specimen at 28 days was shown in Fig. 9. As can be seen,
3. Experimental results and discussion no surface cracks were observed before exposure. Therefore, the
surface cracking is likely due to internal pressure caused by the
3.1. Compressive behavior continued formation of expansive ettringite even when the pores
were completely filled, when sulfate ions reacted with cement
The measured compressive strength of ECC and mortar speci- hydrates. When such pressure exceeded the mortar’s tensile
mens exposed to Na2SO4 solution, Na2SO4 + NaCl solution, and strength, cracks formed. However, for ECC specimens in Na2SO4
water is shown in Fig. 6. As can be seen, the compressive strength solution and Na2SO4 + NaCl solution, neither compressive strength
of specimens cured in water showed a notable increasing trend at reduction nor surface cracking (as shown in Figs. 7 and 8) was
the beginning stage up to 60 days for ECC specimens and 30 days observed. This could be attributed to the fact that ECC has much
for mortar specimens, respectively. Then the rate of strength higher tensile performance that delayed the cracking. The delay
increase slowed down and became stabilized. The strength of surface cracking kept diffusivity of sulfate ions low and pre-
increase is attributed to the continuous hydration of cement and vented accelerated deterioration. In addition, even when ECC even-
the pozzolanic reaction of fly ash. Similar trend was observed tually cracks, the cracks will be much tighter compared to that of
for ECC specimens that were exposed to Na2SO4 solution and mortar specimens, therefore preventing dramatic increase of the
Na2SO4 + NaCl solution: the compressive strength experienced a diffusivity of aggressive ions and keeping the deterioration slow.
sharp increase in the initial stage, and it was followed by a rela- As a result, ECC provides better sulfate resistance than normal
tively sluggish rise until the end of exposure. For those specimens, mortar specimens, which is highly desirable for hydraulic struc-
in addition to continuous cement hydration and pozzolanic reac- tures in sulfate-rich environments.
tion (only in ECC), sulfate ions were diffused into the pores within With regards to the effect of chlorides on sulfate attack, Fig. 6
the mortar specimen and reacted with the cement hydration prod- clearly shows that the compressive strength of all ECC specimens
ucts to form ettringite and gypsum, which may also contribute to subjected to Na2SO4 + NaCl solution is lower than that of the
the strength gain. same-age specimens exposed to Na2SO4 solution during the whole

Fig. 6. Compressive strength of (a) ECC and (b) mortar specimens subjected to different environmental exposures.

Fig. 7. (a) ECC and (b) mortar specimen after 420-day Na2SO4 exposure.
176 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181

Fig. 8. (a) ECC and (b) mortar specimen after 420-day Na2SO4 + NaCl exposure.

According to prior studies [28,29,40–42], this is mainly attributed

to the following reasons: (1) The sulfate diffusion rate is lower
for Na2SO4 + NaCl composite solution than that for the Na2SO4
solution. (2) The rate of chloride diffusion is much higher than that
of sulfate diffusion, therefore C3A firstly reacts with chloride ions
and forms calcium chloroaluminate hydrate (Friedel’s salt), and
the quantity of C3A available for the sulfate ions to react with is
consequently reduced. (3) The solubility of ettringite in chloride
solution is three times greater than that in water, thereby resulting
in a lower ettringite precipitation. However, experiments that con-
sider longer exposure time are needed in the future to further char-
acterize the long-term performance of ECC and mortar specimens
in Na2SO4 + NaCl solution.

3.2. Tensile behavior

Table 2 presents the measured (ultimate) tensile strength,

strain capacity and the average crack width of ECC specimens sub-
jected to different exposure conditions. The average crack width
presented in the table was obtained based on the method
described in [43]. Typical tensile stress-strain curves of ECC speci-
Fig. 9. Mortar specimen before exposure to aggressive solutions at 28 days.
mens under different exposure conditions are also shown in
Fig. 10.
exposure period. For mortar specimens, a similar trend was From the test results, it can be seen that the tensile strength
observed at the initial 120 days, after which the compressive of ECC specimens exposed to water, Na2SO4 solution and
strength of specimens in Na2SO4 solution tended to decrease, while Na2SO4 + NaCl solution all witnessed an increase over time. Specif-
specimens in Na2SO4 + NaCl solution maintained the upward trend ically, the average tensile strength of ECC specimens at 28 days is
until 200 days. This indicates that the presence of chloride ions 5.02 MPa. After 200 days’ exposure to Na2SO4 and Na2SO4 + NaCl
slowed down the sulfate attack process and therefore the strength solutions, the value has risen to 6.29 MPa and 6.22 MPa respec-
increasing (and possible later decreasing) trend was shifted. tively. For specimens kept in water, relatively smaller increase

Table 2
Tensile properties of ECC under different exposure conditions.

Exposure condition Exposure time (days) Tensile strength (MPa) Strain capacity (%) Average crack width (lm)
0 5.02 ± 0.26 4.20 ± 0.14 55
Water 30 5.65 ± 0.27 2.93 ± 0.72 52
60 5.81 ± 0.10 2.53 ± 0.39 51
90 5.53 ± 0.19 2.36 ± 0.25 46
200 5.62 ± 0.02 2.65 ± 0.79 49
Na2SO4 30 5.52 ± 0.20 2.14 ± 0.03 49
60 5.80 ± 0.31 2.69 ± 0.09 48
90 6.24 ± 0.09 2.21 ± 0.28 50
200 6.29 ± 0.08 2.14 ± 0.12 47
Na2SO4 + NaCl 30 5.84 ± 0.21 3.30 ± 0.23 49
60 5.61 ± 0.09 2.11 ± 0.01 51
90 6.45 ± 0.18 3.11 ± 0.17 48
200 6.22 ± 0.12 2.12 ± 0.23 50
 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181 177

interface between fibers and matrix, resulting in higher fiber bridg-

ing strength.
With regards to tensile strain capacity, a slight declining ten-
dency with exposure age was noticed for all specimens, especially
from 0 to 30 days. After 30 days’ exposure, the strain capacity
became stabilized and all remained above 2%. It should be noted
that despite the observed decreasing trends, the tensile ductility
of ECC after environmental exposure is still two orders of magni-
tude higher than that of normal concrete.
The high tensile ductility of ECC is a result of the multiple-
cracking and strain-hardening behavior, which was well pre-
served after 30, 60, 90 and 200 days of exposure to both water
and aggressive solutions. However, the average crack width
dropped from 55 lm to 49 lm (water exposure), 47 lm (Na2SO4
solution exposure) and 50 lm (Na2SO4 + NaCl solution exposure)
respectively after 200 days’ exposure. The drop in crack width is
associated with change of fiber/matrix interfacial properties,
which will be discussed in details in Section 3.3. Such decrease
of average crack width, although led to a slight reduction in ten-
sile ductility, is helpful for limiting the diffusion of sulfate and
chloride ions into the interior of the structures, slowing down
the deterioration, and enhancing the durability of hydraulic
structures.

3.3. Micromechanical investigation

The composite behavior, especially tensile behavior of ECC

under sulfate and sulfate-chloride environments can be further
understood by looking into their micromechanical properties.
The measured micromechanical parameters (matrix fracture
toughness Km, chemical bond Gd, frictional bond s0 and slip hard-
ening coefficient b) as a function of exposure time were plotted
in Fig. 11.
The matrix fracture toughness of ECC specimens exposed to
Na2SO4 and Na2SO4 + NaCl solutions all showed increasing trends.
For ECC specimens subjected to Na2SO4 solution, the matrix frac-
ture toughness increased notably during the first 30 days followed
by a relatively stable growth until the end of the test. The rise of
matrix toughness is most likely due to the continuous hydration
of cement, the pozzolanic reaction of fly ash and the formation of
ettringite and gypsum in sulfate environments that densifies the
matrix. For specimens that were immersed in Na2SO4 + NaCl solu-
tion, a slightly more gradual increasing trend and slight lower
matrix fracture toughness were observed, as the presence of chlo-
ride ions might have affected the formation of ettringite and
gypsum.
Regarding the fiber/matrix interfacial properties, despite the
large scatter of the single-fiber pullout test results, a significant
reduction in chemical bond Gd, and a notable increase in frictional
bond s0 were observed as shown in Fig. 11. According to previous
studies [44,45], chemical bond Gd is governed by the metal cation
concentration at the interface, in particular Al3+ and Ca2+; frictional
bond s0 is controlled by the microstructure of interfacial transition
zone. In the presence of sulfate and chloride ions, the concentra-
tion of Al3+ and Ca2+ tended to decrease due to the reaction that
forms ettringite and gypsum. Besides, the existence of chloride
ions could increase the leaching of Ca2+ [46,47]. Therefore, chemi-
Fig. 10. Typical tensile stress-strain relationship of ECC specimens under different
cal bond Gd showed a decreasing trend. While for frictional bond
exposure conditions for (a) 30 days, (b) 60 days, (c) 90 days and (d) 200 days. s0, the densification of the fiber/matrix interfacial transition zone
due to the continuous formation of hydration product, ettringite
and gypsum mainly contributed to the increasing trend. With
(from 5.02 MPa to 5.62 MPa) was observed after 200 days of expo- regard to the slip hardening coefficient b, no significant change
sure. This is similar to the trend of compressive strength. The was observed.
strength gain is again associated with continuous hydration of It should also be noted that all the micromechanical parameters
cement, pozzolanic reaction of fly ash and the reaction between as presented in Fig. 11 showed sharper changes during the first
sulfate and hydration products. These reactions could densify the 30 days and became stabilized later on. This is also consistent with
178 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181

Fig. 11. Influence of Na2SO4 and Na2SO4 + NaCl exposure on ECC micromechanical parameters: (a) matrix fracture toughness Km, (b) chemical bond Gd, (c) frictional bond s0
and (d) slip hardening coefficient b.

the composite level compressive/tensile behavior. It indicates that relationships at 28 days and after environmental exposures are
the related reactions mainly happened during the initial exposure plotted in Fig. 12, which clearly show the curve shift.
and slowed down afterwards. As can be seen, despite the decreasing chemical bond Gd, the
The change of micromechanical parameters directly affected maximum fiber bridging stress increased after environmental
the fiber bridging rðdÞ curves, thereby altered the tensile behavior exposure due to the increasing frictional bond s0 . This explains
of ECC composite. Based on the calculated results, after exposure to the increase of the tensile strength at composite level. In addition,
Na2SO4 and Na2SO4 + NaCl solutions, the fiber bridging curves the curve shifting leftwards also reduced the crack width. This is
shifted upwards and leftwards. The computed fiber bridging rðdÞ also consistent with the observed crack width data. The value of
 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181 179

Fig. 12. Calculated rðdÞ curves for specimens before and after 200-day exposure to (a) Na2SO4 and (b) Na2SO4 + NaCl solution.

Fig. 13. Influence of (a) Na2SO4 and (b) Na2SO4 + NaCl exposure on J0b .

Fig. 14. Influence of (a) Na2SO4 and (b) Na2SO4 + NaCl exposure on PSHE index.

the complementary energy J 0b was calculated from the rðdÞ rela- before and after environmental conditioning. As can be seen, even
tionship and shown in Fig. 13. From the plot, it is seen that J 0b though there is a slightly decrease in crack number after environ-
dropped after the initial exposure to aggressive solutions but mental exposure, all specimens showed significant multiple crack-
remained stable within the remaining exposure period. ing behavior.
The calculated PSHE index (J0b /Jtip) is shown in Fig. 14. As can be Macroscopically, the tensile ductility is governed by both crack
seen, the index dropped from 5.44 (before exposure) to 3.00 (expo- width and crack number. Crack number is controlled by the strain-
sure to Na2SO4 solution) and 3.12 (exposure to Na2SO4 + NaCl solu- hardening potential as reflected in the PSHE index. Crack width is
tion) after 60 days of exposure, and then the value slightly essentially determined by the fiber bridging rðdÞ curves. The
increased to 3.26 and 3.54 for specimens subjected to 200-day decrease in tensile strain capacity of ECC after environmental
Na2SO4 solution and Na2SO4 + NaCl solution respectively. The exposure is a combined result of both slightly reduced crack num-
recovery is mainly attributed to the increase of Young’s modulus ber (caused by reduced PSHE ) in certain exposure cases and
and the associated reduced matrix fracture energy Jtip. Throughout reduced crack width. Nevertheless, ECC still showed significant
the entire environmental exposure, PSHE index was above 3 at multiple cracking behavior, and reduced crack width is actually
most of the time, which guaranteed saturated and robust multiple considered beneficial for enhancing the durability of ECC in
cracking behavior. Fig. 15 shows the tested tensile specimens hydraulic structure applications.
180 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181

(a) 0 day

(b) 200 days Na2SO4 exposure

(c) 200 days Na2SO4+NaCl exposure

Fig. 15. Multiple cracking behavior of tensile specimens (a) before exposure and after 200-day exposure to (b) Na2SO4 and (c) Na2SO4 + NaCl solution.

4. Conclusions It should be noted that the above conclusions were drawn based
on the current experimental results up to 420 days of exposure to
This paper experimentally investigated the durability of ECC Na2SO4 and Na2SO4 + NaCl solutions. Further studies with longer
under sulfate and sulfate-chloride environment. Specific conclu- exposure time are needed for better understanding of the long-
sions can be drawn as follows: term performance of ECC under those environments. The findings
and micromechanics-based investigation approach of the present
(1) ECC remains durable after 420 days of exposure to concen- study serve as a valuable guideline for future studies. In addition,
trated sulfate and sulfate-chloride environment. In contrast, although the coupled effect of sulfate and chloride attack has been
notable deterioration was observed for mortar specimens investigated under laboratory condition in this paper, field condi-
under the same exposure conditions. The research findings tions are more complicated. Future in-situ tests will be of value
demonstrated the feasibility of using ECC to enhance the for a more complete assessment of ECC’s potential in hydraulic
performance and extend the service life of hydraulic struc- structure applications.
tures under aggressive environments.
(2) Long-term exposure to aggressive (sulfate and sulfate-
Acknowledgments
chloride) solutions leads to increase of compressive
strength, tensile strength and a reduction of tensile strain
This research was partially funded by National Natural Science
capacity of ECC. Nevertheless, ECC retains multiple cracking
Foundation of China (Grant Nos. 51139001, 51479054, 51579083,
and strain-hardening behavior with desirable high tensile
41323001), the National Key Research and Development Program
ductility above 2% after 200 days of environmental
of China (Grand No. 2016YFC0401601), the Doctoral Program of
exposure.
Higher Education of China (Grant No. 20130094110010), the Fun-
(3) The observed composite level behavior of ECC is a result of
damental Research Funds for the Central Universities (Grant No.
increased matrix fracture toughness, decreased fiber/matrix
2015B25414), China Scholarship Council (CSC) and the China
interfacial chemical bond and increased interfacial fric-
Thousand Talent Program. This research was performed at the
tional bond upon both sulfate and sulfate-chloride
University of Michigan, Ann Arbor, while H. Liu served as a visiting
exposures.
scholar.
(4) The reduction in tensile ductility of ECC after exposure to
sulfate and sulfate-chloride conditions is caused by slight
decrease of crack number and reduced crack width. Never- References
theless, ECC still retain significant multiple cracking and
strain-hardening behavior. In addition, the crack width [1] A.J. Schleiss, R.M. Boes, Dams and Reservoirs Under Changing Challenges, CRC
Press, 2011.
reduction is considered beneficial for long-term durability [2] C. Liu, Z. Wang, Present situation of dam concrete’s lifetime in world and
performance. existing problems in China, J. Yangtze River Sci. Res. Inst. 17 (1) (2000) 17–20.
 H. Liu et al. / Construction and Building Materials 133 (2017) 171–181 181

[3] L. Xing, G. Nie, Analysis on durability of concrete structure of hydropower [26] T.P. Dolen, G.A. Scott, K.F. von Fay, B. Hamilton, Effects of Concrete
stations in China, J. Hydroelectric Eng. 29 (2) (2003) 27–31. Deterioration on Safety of Dams, Dam Safety Office, 2003.
[4] ASCE, 2013 Report Card for America’s Infrastructures, 2013. [27] M.T. Bassuoni, M.L. Nehdi, Durability of self-consolidating concrete to sulfate
[5] K. Li, M. Ma, X. Wang, Experimental study of water flow behaviour in narrow attack under combined cyclic environments and flexural loading, Cem. Concr.
fractures of cementitious materials, Cem. Concr. Compos. 33 (10) (2011) 1009– Res. 39 (3) (2009) 206–226.
1013. [28] O.S.B. Al-Amoudi, M. Maslehuddin, Y.A.B. Abdul-Al, Role of chloride ions on
[6] T.S. Han, P.H. Feenstra, S.L. Billington, Simulation of highly ductile fiber- expansion and strength reduction in plain and blended cements in sulfate
reinforced cement-based composite components under cyclic loading, ACI environments, Constr. Build. Mater. 9 (1) (1995) 25–33.
Struct. J. 100 (6) (2003) 749–757. [29] Z.Q. Jin, W. Sun, Y.S. Zhang, J.Y. Jiang, J.Z. Lai, Interaction between sulfate and
[7] ACI Committee 224, Control of cracking in concrete structures, ACI J. Proc. 69 chloride solution attack of concretes with and without fly ash, Cem. Concr. Res.
(12) (1972) 717–753. 37 (8) (2007) 1223–1232.
[8] M. Tsukamoto, Tightness of fiber concrete, Darmstadt Concr. 5 (1990) 215– [30] Z. Zhang, S. Qian, H. Ma, Investigating mechanical properties and self-healing
225. behavior of micro-cracked ECC with different volume of fly ash, Constr. Build.
[9] V.C. Li, H. Stang, Elevating FRC material ductility to infrastructure durability, Mater. 52 (2) (2014) 17–23.
in: 6th RILEM Symposium on Fiber-Reinforced Concretes (FRC) – BEFIB 2004, [31] X. Huang, R. Ranade, V.C. Li, Feasibility study of developing green ECC using
Varenna, Italy, 2004, pp. 171–186. iron ore tailings powder as cement replacement, J. Mater. Civil Eng. 25 (7)
[10] B.Y. Ding, G.B. Wang, S.P. Huang, Y. Yue, H.P. Zhen, A review on causes of (2012) 923–931.
cracking in domestic concrete dams and preventive measures, Water Resour. [32] Standard test method for compressive strength of hydraulic cement mortars
Hydropower Eng. 25 (4) (1994) 12–18. (using 2-in. or [50-mm] cube specimens), ASTM Int. (2016).
[11] Y. Ma, Y. Zhu, Y. Liu, Y. Ning, Feedback study of temperature control and crack [33] JSCE concrete committee, recommendations for design and construction of
prevention of Jiangtanghu pump concrete sluice during construction, Water high performance fiber reinforced cement composites with multiple fine
Power 32 (1) (2006) 33–35. cracks, Jpn. Soc. Civil Eng. (2008).
[12] B.F. Zhu, Current situation and prospect of temperature control and cracking [34] A.A. Griffith, The phenomena of rupture and flow in solids, Philos. Trans. R. Soc.
prevention technology for concrete dam, J. Hydraul. Eng. 37 (12) (2006) 1424– London Ser. A 221 (1921) 163–198.
1432. [35] V.C. Li, C.K. Leung, Steady-state and multiple cracking of short random fiber
[13] M. Li, V.C. Li, Cracking and healing of engineered cementitious composites composites, J. Eng. Mech. ASCE 118 (11) (1992) 2246–2264.
under chloride environment, ACI Mater. J. 108 (2011) 333–340. [36] V.C. Li, H.C. Wu, Conditions for pseudo strain-hardening in fiber reinforced
[14] V.C. Li, Engineered cementitious composites (ECC) - tailored composites brittle matrix composites, Appl. Mech. Rev. 45 (8) (1992) 390–398.
through micromechanical modeling, in: N. Banthia, A. Bentur, A. Mufti (Eds.), [37] T. Kanda, Design of Engineered Cementitious Composites for Ductile Seismic
Fiber Reinforced Concrete: Present and the Future, Canadian Society of Civil Resistant Elements, University of Michigan, 1998.
Engineers, 1997, pp. 1–38. [38] American Society for Testing and Materials, Standard Test Method for Linear-
[15] V.C. Li, Advances in ECC research, ACI Spec. Publ. Concr. 206 (2002) 373–400. Elastic Plane-Strain Fracture Toughness KIc of Metallic Materials, ASTM, West
[16] V.C. Li, On engineered cementitious composites (ECC) – a review of the Conshohocken, PA, 2013.
material and its applications, J. Adv. Concr. Technol. 1 (3) (2003) 215–230. [39] C. Redon, V.C. Li, C. Wu, H. Hoshiro, T. Saito, A. Ogawa, Measuring and
[17] V.C. Li, Engineered cementitious composites (ECC) – material, structural, and modifying interface properties of PVA fibers in ECC matrix, J. Mater. Civil Eng.
durability performance, Concrete Construction Engineering Handbook, vol. 78, 13 (6) (2001) 399–406.
2008. [40] D.G. Miller, Laboratory investigation of the influence of curing conditions
[18] V.C. Li, S. Wang, C. Wu, Tensile strain-hardening behavior or polyvinyl alcohol and various admixtures on the life of concrete stored in sulphate
engineered cementitious composite (PVA-ECC), ACI Mater. J. 98 (6) (2001) solutions as indicated by physical changes, Proc ASTM, 1924, pp. 847–
483–492. 863.
[19] M. Sahmaran, M. Li, V.C. Li, Transport properties of engineered cementitious [41] A. Yeginobali, Sulfate resistance of mortars mixed with sea waters, in: Proc 3rd
composites under chloride exposure, ACI Mater. J. 104 (6) (2007) 604–611. International Conference on Durability of Building Materials and Components,
[20] E. Özbay, O. Karahan, M. Lachemi, K.M. Hossain, C.D. Atis, Dual effectiveness of 1984, pp. 55–65.
freezing-thawing and sulfate attack on high-volume slag-incorporated ECC, [42] W. Harrison, Effect of chloride in mix ingredients on sulphate resistance of
Composites Part B 45 (1) (2013) 1384–1390. concrete, Mag. Concr. Res. 42 (152) (1990) 113–126.
[21] M. Lepech, V.C. Li, Long term durability performance of engineered [43] H. Liu, Q. Zhang, C. Gu, H. Su, V. Li, Influence of micro-cracking on the
cementitious composites, Restor. Build. Monuments 12 (2) (2006) 119–132. permeability of engineered cementitious composites, Cem. Concr. Compos. 72
[22] V.C. Li, T. Horikoshi, A. Ogawa, S. Torigoe, T. Saito, Micromechanics-based (2016) 104–113.
durability study of polyvinyl alcohol-engineered cementitious composite, ACI [44] S. Wang, V. Li, Engineered cementitious composites with high-volume fly ash,
Mater. J. 101 (3) (2004) 242–248. ACI Mater. J. 104 (3) (2007) 233–241.
[23] V.C. Li, E.N. Herbert, Self-healing of microcracks in engineered cementitious [45] H. Ma, S. Qian, Z. Zhang, Z. Lin, V.C. Li, Tailoring engineered cementitious
composites (ECC) under a natural environment, Materials 6 (7) (2013) 2831– composites with local ingredients, Constr. Build. Mater. 101 (2015)
2845. 584–595.
[24] M. Sahmaran, V.C. Li, Durability of mechanically loaded engineered [46] A. Delagrave, M. Pigeon, J. Marchand, É. Revertégat, Influence of chloride ions
cementitious composites under highly alkaline environments, Cem. Concr. and pH level on the durability of high performance cement pastes (part II),
Compos. 30 (2) (2008) 72–81. Cem. Concr. Res. 26 (5) (1996) 749–760.
[25] R. Charlwood, Predicting the long term behavior and service life of concrete [47] K. Wang, D.E. Nelsen, W.A. Nixon, Damaging effects of deicing chemicals on
dams, in: Conference on Long Term Behavior of Dams, 2009, pp. 1–11. concrete materials, Cem. Concr. Compos. 28 (2) (2006) 173–188.