Construction and Building Materials 69 (2014) 370–380

Contents lists available at ScienceDirect

Construction and Building Materials

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

Effect of elevated temperature on strain-hardening engineered

cementitious composites
Prakash S. Bhat, Vivian Chang, Mo Li ⇑
Department of Civil and Environmental Engineering, University of Houston, N132 Engineering Building 1, 4800 Calhoun Road, Houston, TX 77204, United States

h i g h l i g h t s

 Engineered cementitious composite is proposed for spent nuclear fuel storage.

 High temperature effect on ECC uniaxial tension properties is characterized.
 ECC has high spalling resistance after 6 h of exposure to 600 °C.
 ‘‘Spider web’’ nano-cracks are absent in ECC at temperatures up to 600 °C.
 The change in ECC microstructure explains its mechanical properties deterioration.

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

Article history: Strain-hardening engineered cementitious composite materials (ECC) is proposed to substitute quasi-
Received 29 May 2014 brittle concrete materials for building extended spent nuclear fuel (SNF) storage systems in nuclear
Received in revised form 18 July 2014 power plants. While most of ECC properties have been established under normal temperature, the study
Accepted 18 July 2014
aims at understanding ECC material behavior under elevated temperature that is expected in a SNF stor-
Available online 15 August 2014
age environment. On the composite level, ECC specimens were characterized at various temperature lev-
els up to 600 °C under both uniaxial tension and compression. The elevated temperature effect on tensile
Keywords:
strength and strain capacity, compressive strength and failure mode, moisture loss, and spalling behavior
Engineered cementitious composites
Spent nuclear fuel storage
was studied. On the microstructure level, optical microscopy and scanning electron microscopy were
Uniaxial tension conducted to probe the degradation of components, and the change of pore structures due to fiber melt-
Elevated temperature ing within ECC. The results will provide crucial data and insights for future studies of re-engineering ECC
Degradation with robust properties specifically desired for nuclear engineering applications.
Spalling Ó 2014 Elsevier Ltd. All rights reserved.
Tensile Properties
Nuclear infrastructure

1. Introduction extended storage of SNF [4,5], it is challenging to achieve this goal

using conventional concrete materials.
Concrete is a major material component for nuclear power With a quasi-brittle nature, concrete is highly susceptible to
plants and spent nuclear fuel (SNF) storage systems, which provide cracking and fracture failure under combined mechanical loads
radiation shielding in steel-lined concrete pools, concrete dry- and environmental effects. Cracking causes strength loss and
storage casks, and foundation pads. The concrete is constantly greatly impairs the transport properties of concrete. This further
subjected to aging and deterioration under combined thermo- leads to other common deterioration mechanisms such as chemi-
chemo-hygro-mechanical effects, which often cause chemical and cal attack, chloride diffusion and corrosion of embedded steel,
physical alteration of the concrete and result in excessive cracking, moisture penetration, radioactive water leakage, and increased
spalling and loss of strength [1,2]. SNF storage concrete is also sus- radiation levels [5,6]. The deterioration process is further acceler-
ceptible to severely elevated temperatures during accident condi- ated when concrete is exposed to elevated temperature, causing
tions and extreme events, which can lead to catastrophic fracture strength loss and brittle fracture failure modes such as spalling [7].
failure [3,4]. While the long-term durability and safety of concrete Engineered cementitious composites, or ECC, is a class of fiber
structures for spent fuel pools and dry casks are key factors for reinforced cementitious composite materials that provide great
potential for application in SNF storage systems. ECC features large
⇑ Corresponding author. Tel.: +1 713 743 2650. tensile ductility and intrinsic crack width control capacity [8].
E-mail address: moli@uh.edu (M. Li). While containing similar ingredients as concrete or normal fiber

http://dx.doi.org/10.1016/j.conbuildmat.2014.07.052
0950-0618/Ó 2014 Elsevier Ltd. All rights reserved.
 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380 371

reinforced concrete (FRC), the microstructure of ECC can be delib- experiences a 30–60% drop in compressive strength at 600 °C
erately tailored through the use of micromechanical models to and a 60–90% drop at 800 °C [24–26]. By 800 °C, many cracks
achieve tensile strain-hardening behavior and ductility levels are visible with a change in color of the concrete, and only
approximately 200–600 times that of concrete or FRC under ten- 5–18% of the original compressive strength is retained
sion, thereby leading to delayed fracture localization [9]. The [24,26,27]. Cracks continue to appear and enlarge until spalling
fiber/matrix interface is engineered to allow ECC to dissipate occurs at 1200 °C. At this point, only 0.9% of the compressive
energy through multiple micro-cracking with crack widths less strength is retained [24]. Such a reduction in strength can be
than 100 lm. The tensile strain-hardening behavior of ECC differ- attributed to the dehydration of the concrete, and the cracks
entiates it from FRCs that exhibit tension-softening behavior. ECC’s and voids created by the degradation and decomposition of
high tensile ductility, deformation compatibility with existing con- aggregates and C–S–H gels [7,16–21,24]. The heating and cooling
crete, and self-controlled micro-crack width lead to its superior rates played important roles on the reduction of compressive
resistance to restrained shrinkage cracking, freeze–thaw, water strength and elastic modulus [26,29], while the heating durations
permeation and chloride diffusion [10–14]. did not show a significant impact.
The behavior of ECC materials at elevated temperature is of High performance concrete (HPC) features high compressive
great interest to safe operation of structural components and sys- strength and a dense microstructure. While the compressive
tems in nuclear power plants and SNF systems. For SNF pools, strength is higher, the temperature threshold for spalling is
the temperatures are typically maintained below 50 °C. For abnor- much lower. Some HPC spalled between 300 °C and 600 °C
mal and severe environmental conditions, temperature of local hot [28,29] after one hour heating duration, while normal strength
spots can reach 191 °C; for extreme conditions, the temperature concrete (NSC) only showed a reduction in compressive strength
may go from 140 °C to 260 °C; and for accident conditions, temper- without spalling. The spalling is often explosive and can be
atures may reach or exceed 600 °C (e.g., a large sodium spill in the attributed to the combined effect of two factors: (i) the increased
inert and air-filled equipment cells of a liquid–metal fast breeder brittleness of HPC compared to NSC, and (ii) an increase in the
reactor) [7,15]. Elevated temperatures can cause two forms of deg- vapor pressure from the addition of silica fume in making HPC
radation in concrete [7,16–21]. One is the degradation in mechan- [28,29]. Since silica fume leads to low permeability and low
ical properties of concrete, such as strength and Young’s modulus. porosity in concrete, the vapors from the evaporation of water
This is due to the physical–chemical changes of the cement paste cannot escape and will increase the pressure and tensile stress
and aggregates, change of pore structure, and the thermal incom- inside the concrete, causing spalling. It was also found that sili-
patibility between the aggregate and the cement paste which ceous aggregates generate higher thermal conductivity and
causes internal micro-cracking. The other form of damage is spall- expansions than carbonate aggregates in a concrete mixture
ing, which results from the internal tensile stress induced by the [30]. Consequently, HPC subjected to high temperature is more
vapor pressure. Spalling can be explosive, or be a gradual reduction susceptible to spalling and strength loss, despite its higher com-
of concrete cross section. Experimental results have shown that pressive strength than NSC.
spalling can occur in concrete under rapid heating in the tempera- FRCs use short discontinuous fibers to control cracking during
ture range of 200–350 °C [7]. the post-cracking stage, and feature a tension-softening behavior.
Sahmaran et al. [22,23] studied the effect of elevated tempera- It has been shown that the addition of steel fibers into HPC
ture up to 800 °C on the compressive behavior of ECC, i.e. compres- increased specific heat capacity and decreased thermal expansion
sive strength, stiffness, and compressive stress–strain relation. [31]. For HPC, the steel fibers increased the lower bound of the
These studies found that the compressive strength of ECC dropped spalling temperature range from 300 °C to 450 °C, but concrete still
by 40–50% after one hour exposure to 600 °C. The tensile behavior spalled between 450 °C and 800 °C [28,32]. In contrast, the addi-
of ECC under elevated temperature, however, has not been studied. tion of polypropylene fibers into HPC did not show spalling at
The knowledge is important because cracking and spalling are fail- any temperature range tested [28,32]. Due to melting of the fibers,
ure modes under tension. Given that ECC’s large tensile ductility more pores were created in the concrete. This caused a decrease in
and self-controlled tight crack width are essential for ensuring the compressive strength, but allowed water vapors to escape,
structural durability, a question that could naturally be raised is lessening the vapor pressure and tensile stress. FRCs still experi-
whether ECC can still maintain these tensile properties under ele- ence a large reduction in their mechanical properties such as com-
vated temperature. pressive strength, with the critical temperature also around 400 °C
This paper first summarizes previous studies on high tempera- [33–35]. By 650 °C, FRC of various fibers such as polypropylene,
ture effect on cementitious materials. To bridge the knowledge polyvinyl alcohol, carbon, glass, and aramid fibers have shown a
gap, this paper studied the elevated temperature effects on ECC 60–70% decrease in its compressive strength with major cracks
material properties and microstructure, especially under uniaxial being visible [34–36].
tension. The results help understand ECC material behavior under While previous studies have provided significant insights on
a high temperature environment, and lay out the groundwork for high temperature effect on cementitious materials, the focus
future research to develop new and robust strain-hardening ECC has been on quasi-brittle and tension-softening cementitious
materials for high temperature applications. materials and their compressive properties. For ECC, previous
work has focused on compressive properties under elevated tem-
2. Elevated temperature effect on cementitious materials perature. There has been a lack of knowledge on the tensile
behavior of strain-hardening ECC under elevated temperature.
Published literature on high temperature effect on cementi- This information is crucial, considering that cracking and spalling
tious materials has focused on concrete, high performance con- under high temperature are both induced by internal tensile
crete, high strength concrete, and fiber reinforced concrete. The stresses. Understanding both tensile and compressive properties
results vary, depending on the material ingredient and composi- of ECC subjected to elevated temperature is the focus of this
tion, curing conditions, heating rate and time, and specimen paper. Furthermore, this paper studied a special version of ECC
types. It has been shown that under high temperature environ- designed to possess microcracks width below 40 lm for reduced
ments, such as a fire, concrete starts to deteriorate and lose transport properties, which is different from the most well-
its mechanical properties. After 400 °C, concrete typically studied version of ECC, M45.
372 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380

3. Material and methods

In this study, the ECC mixture was designed by integrating

micromechanics theory, rheology control during processing, and
micro-structure tailoring. The ECC design theory requires the
simultaneous satisfaction of steady state cracking criteria [37]
and maximized micro-cracking density [38]. The ingredient parti-

101.6 mm (4 in)
cle size distribution and the combined amount of water and
admixtures were first determined to achieve a homogeneous

228.6 mm (9 in)
cementitious composite material at fresh state, with plastic viscos-
ity and yield stress tailored to an optimal level [38] that favors uni-
form dispersion of micro-scale polyvinyl alcohol (PVA) fibers. Then,
the micro-parameters of the hardened material, including matrix
properties (e.g., fracture toughness, flaw size distribution, hydra-
tion products), the fiber/matrix interfacial properties (e.g., interfa-
cial chemical and frictional bonds, slip-hardening coefficient, fiber
debonding and pullout behavior), and fiber properties (e.g., aspect
ratio, strength, Young’s modulus) were tailored to ensure the
strain-hardening criteria [37] were satisfied. Furthermore, the
micro-parameters were tailored so that the tensile stress vs. crack
open relation (i.e. r–d relation) of each individual microcrack
allows the steady-state microcrack to maintain a tighter crack
width below 40 lm.
The designed ECC (Table 1) binder system contained water, a
76.2 mm (3 in) 12.7 mm (0.5 in)
polycarboxylate-based high range water reducer, a viscosity mod-
ifying agent, Ordinary Portland Cement (OPC) Type I cement, ASTM
Fig. 1. Direct uniaxial tension test setup.
standard Type F fly ash, and silica sand that served as fine aggre-
gates. The cementitious ingredients, silica sand, water and admix-
ture together formed the ECC matrix, with tailored toughness and 28 days. Control mortar specimens were prepared in the same
tensile cracking strength satisfying the strain-hardening criteria. way, with the same mix design except that they do not contain
Polyvinyl alcohol (PVA) fibers were incorporated into the compos- any PVA fibers.
ite system at a volume fraction of 1.8%. The PVA fibers were 8 mm At the age of 28 days, the specimens were heated in an electric
long and 39 lm in diameter, with nominal tensile strength of furnace with a heating rate of 23.5 °C/min, up to different maxi-
1620 MPa and density of 1300 kg/m3 (Table 2). The fiber/matrix mum temperatures of 100 °C, 200 °C, 300 °C, 400 °C, and 600 °C.
interfacial bond was tailored so that the complementary energy Some control specimens were stored at 20 °C and were not sub-
(determined from fiber bridging–crack opening relation) and the jected to heating. When the target temperature was achieved,
maximum fiber bridging stress satisfy the strain-hardening the specimens were maintained at this temperature in the furnace
criteria. for 2 h (for the tensile specimens) and 6 h (for the compressive
All mixtures were prepared under controlled room temperature specimens) to ensure a steady-state thermal condition. The speci-
20 ± 1 °C and relative humidity conditions 50 ± 5% RH. To prepare mens were then tested in direct uniaxial tension and uniaxial com-
tensile specimens, the mixtures were casted into tensile coupon pression at room temperature. It should be noted that the direct
molds in a direction that is perpendicular to the length (305 mm) uniaxial tensile test (Fig. 1) is considered the most convincing
and width (76.2 mm) of each specimen. The specimen thickness method for validating strain-hardening behavior of ECC, because
was 12.7 mm. No external vibration was applied. To prepare com- some tension-softening materials can show ‘‘pseudo deflection
pressive specimens, the mixtures were casted into 76.2 mm (diam- hardening behavior’’ under bending [39,40]. The specimen dimen-
eter) by 152.4 mm (height) cylinder molds. The tensile coupon and sions were 229  76.2  12.7 mm, with the middle gage length of
compressive cylinder molds were then covered with plastic sheets 102 mm. Both ends of the specimen were strengthened so that
and demolded after 24 h. Finally, the specimens were moisture- multiple cracking only occurred within the gage length, allowing
cured in plastic bags at 95 ± 5% RH and 25 ± 1 °C for 7 days, and for more accurate measurement of strain. Tests were conducted
air cured at 65 ± 5% RH and 25 ± 1 °C for 21 days until the age of under a displacement control rate of 0.0025 mm/s to simulate a

Table 1
ECC and control mortar mixing proportion.

Material Cement Silica sand Fly ash Water Fiber, (Vf) (%) SP (%) VMA (%)
ECC 1.16 1.16 2.2 0.66 2.0 0.002 0.001
Control mortar 1.16 1.16 2.2 0.66 – 0.002 0.001

Table 2
Properties of PVA fiber.

Fiber type Nominal strength (MPa) Apparent strength (MPa) Diameter (lm) Length (mm) Young’s modulus (GPa) Elongation (%)
Poly-Vinyl Alcohol 1620 1092 39 12 42.8 6.0
 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380 373

Table 3 quasi-static loading condition. Two external linear variable differ-

ECC tensile specimen details and moisture loss. ential transformers (LVDTs) were attached to the specimen surface
Sample Temperature (°C) Specific Moisture Avg. moisture at a gage length of 102 mm to measure displacement, which was
label # for 2 h gravity loss (%) loss (%) used to calculate strain. Three specimens were tested for each tem-
1 20 2.09 0.00 0.00 perature exposure scenario. The uniaxial compression test was
2 2.12 0.00 conducted on sulfur-capped cylinder specimens to measure the
3 2.11 0.00 load–displacement curves, under displacement-control rate of
4 100 1.97 2.54 2.69 0.0025 mm/s. The test procedure followed ASTM C39 [41]. Three
5 1.98 3.19 specimens were tested for each temperature exposure scenario.
6 2.05 2.33
The microstructure of ECC after exposure to elevated temperature
7 200 1.96 7.10 7.34 was studied through optical microscopy and scanning electron
8 1.95 7.57
9 1.95 7.34
microscopy.

10 300 1.87 9.75 8.92

11 1.82 8.39
4. Results and discussion
12 1.84 8.62
13 400 1.84 12.38 12.68
4.1. Elevated temperature effect on tensile properties
14 1.82 13.33
15 1.85 12.34
The ECC tensile specimen details and moisture loss after heating
16 600 1.82 13.48 13.35
17 1.82 13.62 are summarized in Table 3 and Fig. 2. Test results revealed 2.69%,
18 1.83 12.96 7.34%, 8.92%, 12.68% and 13.35% of moisture loss of the ECC spec-
imen at 100 °C, 200 °C, 300 °C, 400 °C, and 600 °C, respectively. The
moisture loss was faster between 20 and 400 °C, and approached a
16% plateau between 400 and 600 °C. The moisture loss is associated
with the liberation of free water in large pores as well as small cap-
illary pores, followed by the loss of water adsorbed on the surface
12% of C–S–H and physically bound water from the decomposition of
calcium hydrates and other formed cement hydrates. During this
Moisture Loss

process, the high vapor pressure within the cementitious material

8% can lead to surface cracking or spalling. However, despite a high
amount of moisture loss, no spalling was observed in the ECC ten-
sile specimens, including those subjected to a maximum tempera-
4% ture of 600 °C. The specimens without heating were compared to
the specimens subjected to 2 h of heating at different temperatures
(Fig. 3). No spalling or degradation was observed from the speci-
0% men surface.
0 100 200 300 400 500 600 700 Uniaxial tensile test on ECC specimens without heating (20 °C)
Temperature (°C) was the sole method to validate that strain-hardening ECC, instead
of tension-softening FRC, was produced (Fig. 4(a)). Despite the var-
Fig. 2. Effect of temperature on moisture loss in ECC tensile specimens.
iation in the experimental data among specimens, the tensile

(a) 20 °C (b) 200 °C

(c) 400 °C (d) 600 °C

Fig. 3. Effect of temperature on ECC tensile specimens before mechanical testing – no degradation was observed from the surface.
374 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380

6 6

5 5
Tensile Stress (MPa)

Tensile Stress (MPa)

4 4

3 3

2 #01 2 #04

#02 #05
1 1
#03 #06

0 0
0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0% 0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0%
Tensile Strain Tensile Strain
(a) 20 °C (b) 100 °C

6 6

5 5
Tensile Stress (MPa)

Tensile Stress (MPa)

4 4

3 3

2 #07 2 #10

#08 #11
1 1
#09 #12

0 0
0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0% 0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0%

Tensile Strain Tensile Strain

(c) 200 °C (d) 300 °C

6 6

5 5
Tensile Stress (MPa)

Tensile Stress (MPa)

4 4

3 3

2 #13 2 #16

#14 #17
1 1
#15 #18

0 0
0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0% 0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0%

Tensile Strain Tensile Strain

(e) 400 °C (f) 600 °C
Fig. 4. Tensile stress–strain response of ECC after 2-h of exposure at different temperatures.

7 1.4
Normalized Strength (ft,T/ft)

6 1.2
Tensile Strength (MPa)

5 1.0

4 0.8

3 0.6

2 0.4

1 0.2

0 0.0
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700
Temperature (°C) Temperature (°C)
(a) Absolute tensile strength (b) Relative tensile strength
Fig. 5. Effect of elevated temperature on ECC tensile strength.
 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380 375

100 mm

(a) ECC, 20 °C
100 mm

(a) ECC, 20 °C
100 mm

(b) ECC, 200 °C for 2 hours

100 mm
100 mm
(c) ECC, 400 °C for 2 hours
(b) ECC, 200 °C for 2 hours

100 mm

(d) ECC, 600 °C for 2 hours

Fig. 6. Cross section view of ECC tensile specimens subjected to (a) 20 °C, (b) 200 °C,
(c) 400 °C and (d) 600 °C for 2 h. At 400 °C and 600 °C, the micro-pores were created
after fiber melting.
100 mm

(c) ECC, 400 °C for 2 hours

stress–strain curve followed the same pattern that contained three
stages: (1) The initial elastic stage, characterized by Young’s mod-
ulus. (2) The strain-hardening stage, accompanied by multiple
micro-cracking formations. During this stage, the width of each
micro-crack remained nearly constant after formation, while the
increasing applied strain increased the number of the micro-
cracks. This stage is characterized by first cracking strength (i.e.
the stress when the first micro-crack occurs), ultimate tensile
strength (i.e. the peak stress), and the strain capacity of the mate-
rial (i.e. the strain corresponding to peak stress). As this test was 100 mm
conducted under displacement control, each tiny load ‘‘drop’’ on
the curve corresponded to the released energy during the forma- (d) ECC, 600 °C for 2 hours
tion of each micro-crack. (3) The tension-softening stage, accompa-
Fig. 7. Side view of final failure cross section of ECC tensile specimens subjected to
nied by the formation of a localized fracture at one of the micro- (a) 20 °C, (b) 200 °C, (c) 400 °C and (d) 600 °C for 2 h. No fibers were observed at
cracks and the continuous drop of the ambient load. At this stage, 400 °C and 600 °C.
the ECC material behaves the same as tension-softening FRC mate-
rials. The results showed that at 20 °C the ECC specimens exhibited
strain-hardening behavior with tensile strain capacity 2.3% in aver- strain-hardening behavior still existed in ECC. The specimens ten-
age, which is approximately 230 times that of concrete adopted in sile strain capacity was further reduced to 0.5–0.8%, 50–80 times
current SNF systems. Such strain-hardening behavior will lead to that of concrete;. At 300 °C, 400 °C and 600 °C, the specimens lost
greatly higher fracture energy for providing damage tolerance to strain-hardening behavior; they did not exhibit tension-softening
safely store SNF. behavior as normal fiber reinforced concrete either, but a brittle
The specimens exposed to various temperature levels were behavior just as concrete.
then subjected to direct uniaxial tension test. The effect of ele- The effect of elevated temperature on ECC tensile strength is
vated temperature on ECC tensile stress–strain relation is plotted shown in Fig. 5. From 20 °C to 200 °C, ECC tensile strength
in Fig. 4. There was an obvious trend of reduction in tensile strain increased by 10% on average. As the ECC mixture contained a large
capacity of ECC. At 20 °C, all three specimens showed tensile amount of fly ash, the hydration process and pozzolanic reaction
strain-hardening behavior with strain capacity between 2 and continued after 28 days, and was accelerated due to the high tem-
2.5%. At 100 °C, the specimens still exhibited strain-hardening perature curing effect up to 200 °C. At 300 °C, 400 °C and 600 °C
behavior; however, the tensile strain capacity was reduced to compared to 20 °C, tensile strength decreased by 54%, 29% and
0.5–2.0%, with large variation indicating the change in matrix 41% respectively. This was explained by the rapid decomposition
flaw distribution and fiber bridging capacity. At 200 °C, and pyrolysis of PVA fibers followed by loss of all fibers, the
376 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380

Table 4 increased amount of pores after fiber melting, as well as the degra-
ECC compressive specimen details and moisture loss. dation of the cement paste and aggregates.
Sample Temperature (°C) Specific Moisture Avg. moisture The final failure section of each tensile specimen was studied
label # for 6 h gravity loss (%) loss (%) under a high-resolution optical microscope (Figs. 6 and 7). It was
1 20 2.00 0.00 0.00 clearly shown that the PVA fibers were present after exposure up
2 2.00 0.00 to 200 °C, but were melted and lost after 6 h of heating at 400 °C
3 1.99 0.00 or 600 °C. The melting of the fibers created a large quantity of
4 100 1.95 0.19 0.30 40–50 lm micro-pores in the cementitious matrix, which changed
5 1.98 0.18 the material’s pore structure and consequently influenced material
6 1.99 0.54
mechanical and transport properties. This explains the reduction in
7 200 1.90 4.71 5.48 tensile strength and the loss of strain-hardening behavior of ECC at
8 1.89 5.62
9 1.87 6.10
temperatures of 300 – 600 °C. The elevated temperature up to
200 °C did not seem to have an effect on PVA fibers. This can be fur-
10 300 1.79 9.97 9.96
11 1.77 9.49
ther evidenced by the un-decreased ultimate tensile strength of
12 1.79 10.42 ECC (Fig. 5), which was dominated by fiber bridging capacity. How-
13 400 1.73 12.24 12.42
ever, the decrease of tensile strain capacity from 20 °C to 200 °C
14 1.72 12.36 (Fig. 4) indicated that there was a combined effect of increase in
15 1.72 12.66 matrix toughness and degradation of fiber/matrix interfacial prop-
16 600 1.70 14.41 14.43 erties due to elevated temperature up to 200 °C, which negatively
17 1.69 14.47 impacted on the strain-hardening criteria described in Section 3.
18 1.68 14.40

4.2. Elevated temperature effect on compressive properties

The moisture loss of ECC compressive specimens compared

Table 5
Control mortar compressive specimen details and moisture loss.
with control mortar specimens, were summarized in Table 4 and
5. The results revealed 0.3%, 5.48%, 9.96%, 12.42% and 14.43% of
Sample Temperature (°C) Specific Moisture Avg. moisture
moisture loss of the ECC compressive specimens after 6 h of expo-
label # for 6 h gravity loss (%) loss (%)
sure to 100 °C, 200 °C, 300 °C, 400 °C, and 600 °C. The moisture loss
1 20 2.04 0.00 0.00 trend and level were also similar for control mortar specimens
2 2.07 0.00
3 2.05 0.00
(Fig. 8). The slightly higher moisture loss of ECC compared to con-
trol mortar was attributed to the higher initial porosity of ECC,
4 100 1.96 0.16 0.13
5 1.99 0.11
which resulted from the higher plastic viscosity at fresh state dur-
6 1.98 0.13 ing processing. Consequently, more air pores were entrapped dur-
7 200 1.89 5.14 4.32
ing mixing.
8 1.86 5.33 Spalling of the three mortar control specimens occurred after
9 1.91 2.48 30–75 min of exposure to the maximum temperature of 600 °C.
10 300 1.86 9.71 9.58 The spalling was explosive, as shown in Fig. 9. In contrast, the
11 1.84 9.89 ECC specimen did not exhibit any spalling or visible degradation
12 1.85 9.13 observed from the surface, despite the fact that moisture loss of
13 400 1.80 11.61 11.64 ECC was higher than mortar. The spalling resistance of ECC under
14 1.80 11.62 high temperature can be attributed to two mechanisms: First,
15 1.80 11.70
the melting of fibers created a large amount of micro-pores to
16 600 – – Spalling reduce the internal thermal stress and vapor pressure; Second,
17 – –
the tensile strain capacity of the ECC material suppressed brittle
18 – –
failure mode commonplace in cementitious materials, thus pre-
venting spalling caused by internal tensile stress.

16% 16%

12% 12%
Moisture Loss

Moisture Loss

8% 8%
SPALLING

4% 4%

0% 0%
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700
Temperature (°C) Temperature (°C)
(a) ECC (b) Control mortar
Fig. 8. Effect of elevated temperature on moisture loss of (a) ECC and (b) control mortar compressive specimens.
 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380 377

(a) ECC (b) Control mortar

Fig. 9. ECC vs. control mortar compressive specimens after 6-h exposure at 600 °C.

ECC, 20 °C Control Mortar, 20 °C

ECC, 100 °C for 6 hours Control Mortar, 100 °C for 6 hours

ECC, 200 °C for 6 hours Control Mortar, 200 °C for 6 hours

ECC, 400 °C for 6 hours Control Mortar, 400 °C for 6 hours

Fig. 10. Elevated temperature effect on compressive failure mode of ECC and control mortar specimens.
378 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380

60 60

Compressive Strength (MPa)

Compressive Strength (MPa)

50 50

40 40

30 30

SPALLING
20 20

10 10

0 0
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700
Temperature (°C) Temperature (°C)
(a) ECC (b) Control mortar
Fig. 11. Effect of elevated temperature on compressive strength of (a) ECC and (b) control mortar specimens.

(a) ECC, 20 °C (b) ECC, 200 °C

(c) ECC, 400 °C (d) ECC, 600 °C

Fig. 12. Lower-magnification scanning electron microscopy of ECC.

(a) ECC, 400 °C (b) Control mortar, 400 °C

Fig. 13. Higher-magnification scanning electron microscopy of ECC vs. Control Mortar at 400 °C.
 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380 379

The compressive failure modes are shown in Fig. 10. For speci- and dry casks. The ECC mixture investigated in this study exhibited
mens that had been subjected to 6 h heating at 100 °C and 200 °C, strain-hardening behavior under direct uniaxial tension at normal
and for specimens stored at 20 °C, different compressive failure temperature conditions of 20 °C, as well as self-controlled micro-
modes were observed in ECC compared to control mortar crack width below 40 lm (which is different from ECC M45). Such
(Fig. 12). The ECC specimens exhibited a ductile failure mode with behavior offers great advantage over normal concrete materials, in
an approximately 45-deg final fracture plane. The control mortar terms of improved damage tolerance and reduced transport prop-
specimens had a sudden and loud failure with large splitting cracks erties for SNF storage.
and large pieces of specimens spalled off. Obviously, the ductile This study found that the tensile strength and strain capacity of
strain-hardening behavior of ECC when the exposure temperature ECC deteriorated under elevated temperatures. At temperatures
was up to 200 °C, in contrast to the brittle behavior of the control below or equal to 200 °C, the increase in matrix toughness as well
mortar, led to the different compressive failure modes in the two as the degradation in fiber/matrix interfacial properties led to the
materials. Furthermore, after 6 h of exposure to 300 °C or 400 °C, reduced tensile strain capacity by one order, while the tensile
the ECC compressive specimens exhibited the same failure mode strength slightly increased by around 1%. At temperatures above
as control mortar, with splitting cracks and brittle failure. When 200 °C, ECC lost its tensile strain-hardening behavior, while its ten-
the maximum temperature was further increased to 600 °C, the sile strength was reduced by 40% approximately.
ECC compressive specimens failed with splitting cracks under uni- Interestingly, our work showed no reduction in ECC compres-
axial compression, while the control mortar specimens already sive strength even after 6 h of exposure to temperatures up to
experienced severe spalling before compressive testing. 600 °C, which differs from the observations of previous studies
Compressive strength of ECC compared to control specimens is [23,24]. However, when the temperature was above 200 °C, ECC
shown in Fig. 11. In contrast to the significant reduction in tensile compressive failure mode changed from ductile 45 deg plane fail-
strength, the compressive strength of ECC was not significantly ure to brittle splitting crack failure.
affected by elevated temperature. Interestingly, the compressive No spalling was observed in both ECC tensile and compressive
strength for the control mortar specimens increased with increas- specimens, even after 6 h of constant exposure to 600 °C. In con-
ing temperature up to 400 °C, and spalling suddenly occurred in trast, the control mortar specimens experienced severe explosive
all three specimens subjected to 600 °C. For mortar, the results spalling after 30–75 min of exposure to 600 °C. The spalling resis-
are different with the findings in previous research in literature that tance of ECC was attributed to its tensile strain capacity as well as
showed reduction in concrete compressive strength at high tem- the increase in porosity due to fiber melting. Scanning electron
peratures. The discrepancy can be due to two reasons: (1) the high microscopy of ECC microstructure clearly revealed the increased
amount of fly ash in the mortar mix design, and (2) the absence of amount of micro-pores at temperatures above 200 °C, and the
coarse aggregates; only fine aggregates with 275 lm mean size was absence of ‘‘spider web’’ nano-cracks that are common within con-
used. ECC compressive strength was affected by elevated tempera- trol mortar material at high temperatures.
ture in a similar trend as control mortar, except that ECC compres- The experimental results will guide future studies to re-engi-
sive strength was higher than mortar at 400 °C. This is due to the neer ECC materials with robust properties at elevated tempera-
less internal nano-cracking damage in ECC, which was later vali- tures for SNF storage applications.
dated by SEM results. Furthermore, ECC did not experience spalling
at 600 °C as mortar did. The results further indicated that compres-
sive strength and tensile strength of ECC are influenced by elevated Acknowledgment
temperature in different manners, which will be studied as an inter-
esting fracture mechanics problem in details in future studies. The authors would like to thank the US Department of Energy
Scanning electron microscopy (SEM) was adopted to probe the Nuclear Engineering Research Program (DE-AC07-05ID14517 No.
microstructure degradation of ECC and mortar composite materi- 00127468) for supporting this research.
als. The lower-magnification SEM of ECC subjected to different
temperatures is shown in Fig. 12. It was observed that the fibers
have melted when the temperature was 400 °C and 600 °C, creat- References
ing micro-pores and channels within the ECC microstructure. The
[1] Morton SL, Winson PL, Saegusa T, Shirai K, Sasahara A, Hattori T. Concrete
higher-magnification SEM of ECC subjected to 400 °C, compared shield performance of the VSC-17 spent nuclear fuel cask. International high
with control mortar, was shown in Fig. 13. A large amount of cracks level radioactive waste management conference; April 2006.
with width ranging from 200 nm to 500 nm was found in control [2] Fillmore DL. Literature review of the effects of radiation and temperature on
the aging of concrete, Idaho National Engineering and Environmental
mortar (Fig. 13(b)). These nano-cracks have a ‘‘spider web’’ pattern, Laboratory, Idaho Falls, Idaho 83415; 2004.
similar as the concrete micro-cracks caused by alkali silica reaction [3] Aquaro D, Castrataro P, Forasassi G. Impact resistance of a concrete bed for a
(ASR), suggesting that the nano-cracks were formed due to the spent nuclear fuel cask repository. 16th International conference on structural
mechanics in reactor technology; 2001. p. 1–8.
expansion of fine aggregates as well as internal vapor pressure. [4] Alvarez R. Spent nuclear fuel pools in the U.S.: Reducing the Deadly Risks of
In contrast, few nano-cracks were found in ECC subjected to 6 h Storage; May 2011.
of exposure to temperatures up to 600 °C (Fig. 13(a)), despite the [5] Douglas BR. Evaluation of the technical basis for extended dry storage and
transportation of used nuclear fuel; December 2010.
fact that ECC has lost its tensile strain-capacity at this temperature [6] Bangert F, Grasberger S, Kuhl D, Meschke G. Environmentally induced
level. This evidence strongly supports the hypothesis that the melt- deterioration of concrete: physical motivation and numerical modeling. Eng
ing of PVA fibers within ECC matrix created more pores and chan- Fract Mech 2003;70:891–910.
[7] Naus D. The effect of elevated temperature on concrete materials and
nels for abating the vapor pressure and the tensile stress caused by structures – a literature review. Technical Report, Oak Ridge National
aggregate expansion, leading to a much less microstructure degra- Laboratory; March 2006.
dation within ECC than normal mortar or concrete materials. [8] Li VC, Wu C, Wang S, Ogawa A, Saito T. Interface tailoring for strain-hardening
polyvinyl alcohol-engineered cementitious composites (PVA-ECC). ACI Mater J
2002;99(5):463–72.
[9] Li VC. Reflections on the research and development of engineered cementitious
5. Conclusions composites (ECC). Proceedings of the JCI international workshop – DFRCC,
Takayama, Japan; 2002. p. 1–21.
[10] Li M, Li VC. Non-brittle concrete for durable infrastructure in coastal regions.
Strain-hardening cementitious composite materials, ECC, was The international conference on future concrete, Doha, Qatar, November 1–3;
proposed for extended storage of SNF systems such as fuel pools 2010.
380 P.S. Bhat et al. / Construction and Building Materials 69 (2014) 370–380

[11] Li VC, Lepech M. Crack resistant concrete material for transportation [26] Lee J, Xi Y, Willam K. Properties of concrete after high-temperature heating
construction. In: TRB 83rd annual meeting, Washington, D.C., CD ROM, Paper and cooling. ACI Mater J 2008;15(4):334–41.
04-4680; 2004. [27] Huo JS, He YM, Xiao LP, Chen BS. Experimental study on dynamic behaviours of
[12] Li M. Multi-scale design for durable repair of concrete structures, Ph.D. concrete after exposure to high temperatures up to 700 °C. Mater Struct
Dissertation, University of Michigan; 2009. 2013;46(1–2):255–65.
[13] Lepech M, Li VC. Water permeability of cracked cementitious composites. [28] Sideris K, Manita P, Chaniotakis E. Performance of thermally damaged fibre
Proceedings of eleventh international conference on fracture, Turin, Italy. reinforced concretes. Construct Build Mater 2009;23(3):1232–9.
March 20–25; 2005. [29] Husem M. The effects of high temperature on compressive and flexural
[14] Li M, Li VC. High-early-strength engineered cementitious composites for fast, strengths of ordinary and high-performance concrete. Fire Saf J
durable concrete repair – material properties. ACI Mater J 2011;108(1):3–12. 2006;41(2):155–63.
[15] Willam K, Xi Y, Lee K, Kim B. Thermal response of reinforced concrete [30] Kodur V, Sultan M. Effect of temperature on thermal properties of high-
structures in nuclear power plants. SESM No. 02-2009, Research Report for strength concrete. J Mater Civ Eng 2003;15(2):101–7.
Oak Ridge National Laboratory; 2009. [31] Kodur V, Khaliq W. Effect of temperature on thermal properties of different
[16] Bazant Z. Analysis of pore pressure, thermal stress and fracture in rapidly types of high-strength concrete. J Mater Civ Eng 2011;23(6):793–801.
heated concrete. Proc. of int. workshop on fire performance of high-strength [32] Chen B, Liu J. Residual strength of hybrid-fiber-reinforced high-strength
concrete, Gaitherburg, MD, USA; February 13–14, 1997. concrete after exposure to high temperatures. Cem Concr Res
[17] Bažant Z, Kaplan MF. Concrete at high temperatures, material properties, and 2004;34:1065–9.
mathematical models. Burnt Mill: Longman; 1996. [33] Balazs GL, Lubloy E. Post-heating strength of fiber-reinforced concretes. Fire
[18] Bažant Z, Thonguthai W. Pore pressure and drying of concrete at high Saf J 2012;49:100–6.
temperature. Mag Concr Res 1979;31(107):67–75. [34] Cavdar A. The effects of high temperature on mechanical properties of
[19] Castillo C, Durrani AJ. Effect of transient high temperature on high-strength cementitious composites reinforced with polymeric fibers. Compos Part B: Eng
concrete. ACI Mater J 1990;97(1):47–53. 2013;45(1):78–88.
[20] Cruz CR, Gillen M. Thermal expansion of Portland cement paste, mortar and [35] Cavdar A. A study on the effects of high temperature on mechanical properties
concrete at high temperatures. Fire Mater 1980;4(2):66–70. of fiber reinforced cementitious composites. Compos Part B: Eng
[21] Komonen J, Vesa P. Effects of high temperature on the pore structure and 2012;43(5):2452–63.
strength of plain and polypropylene fiber reinforced cement pastes. Fire [36] Khaliq W, Kodur V. High temperature mechanical properties of high-strength
Technol 2003;39(1):23–34. fly ash concrete with and without fibers. ACI Mater J 2012;109(6):665–74.
[22] Sahmaran M, Lachemi M, Li VC. Assessing mechanical properties and [37] Li VC, Wu C, Wang S, Ogawa A, Saito T. Interface tailoring for strain-hardening
microstructure of fire-damaged engineered cementitious composites. ACI PVA-ECC. ACI Mater J 2002;99(5):463–72.
Mater J 2010;107(3):297–304. [38] Li M, Li VC. Rheology, fiber dispersion, and robust properties of engineered
[23] Sahmaran M, Ozbay E, Yucel HE, Lachemi M, Li VC. Effect of fly ash and PVA cementitious composites. J Mater Struct 2012;46(3):405–20.
fiber on microstructural damage and residual properties of engineered [39] Stang H. Scale effects in FRC and HPRFCC Structural elements. In: Naaman AE,
cementitious composites exposed to high temperatures. ASCE J Mater Civ Reinhardt HW, editors. High performance fiber reinforced cementitious
Eng 2011;23:1735–45. composites, RILEM proceedings pro 30; 2003. p. 245–58.
[24] Arioz O. Retained properties of concrete exposed to high temperatures: size [40] Li M, Li VC. Cracking and healing of engineered cementitious composites under
effect. Fire Mater 2009;33(5):211–22. chloride environment. ACI Mater J 2011;108(3):333–40.
[25] Li Z, Xu J, Bai E. Static and dynamic mechanical properties of concrete after [41] ASTM C39/C39M – 12a Standard Test Method for Compressive Strength of
high temperature exposure. Mater Sci Eng: A 2012;554:27–32. Cylindrical Concrete Specimens.