Construction and Building Materials 223 (2019) 12–27

Contents lists available at ScienceDirect

Construction and Building Materials

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

Investigation of interface shear properties and mechanical model

between ECC and concrete
Jun Tian a,b, Xiaowei Wu a,b,⇑, Yu Zheng a, Shaowei Hu a,c, Yinfei Du d, Wenwei Wang e, Can Sun a,
Lifei Zhang b,c
a
School of Environment and Civil Engineering, Dongguan University of Technology, Dongguan 523808, China
b
Key Laboratory of Transport Industry of Bridge Detection Reinforcement Technology (Chang’an University), Xi’an 710064, China
c
Department of Materials and Structural Engineering, Nanjing Hydraulic Research Institute, Nanjing 210029, China
d
School of Civil Engineering, Central South University, Changsha 410075, China
e
Department of Bridge Engineering, School of Transportation, Southeast University, Nanjing 210096, China

h i g h l i g h t s

The effect of ECC construction method on shear properties of ECC-to-concrete interface was studied.

The construction method of spraying ECC had an obvious decreasing effect on interface tensile strength.

A shear strength mechanical model and a shear stress-slip model of ECC-to-concrete interface were proposed.

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

Article history: This paper reveals an investigation of the influences of ECC mixture construction method (i.e., common
Received 5 February 2019 casting and spraying ECC), ECC strength grade, interface roughness of concrete substrate, and polyvinyl
Received in revised form 19 June 2019 alcohol (PVA) fiber type on shear properties of ECC-to-concrete interface. The test results show that
Accepted 23 June 2019
two failure modes of interface shear failure and ECC failure were observed. The roughness degree of con-
crete substrate had a strong effect on the failure mode, while ECC strength grade, ECC mixture construc-
tion method, and PVA fiber type had almost no influence on the failure mode. The shear load-slip curves
Keywords:
of all specimens were basically similar. Compared to the casting ECC, the spraying ECC had an obvious
Engineered cementitious composites (ECC)
Interface
decreasing effect on interface shear strength. The ECC strength grade and roughness degree of concrete
Shear strength substrate had a significant influence on interface shear strength, and higher ECC compressive strength
Interface bond and coarser interface could cause higher interface shear strength. On the contrary, PVA fiber type had
Shear strength model a slight influence on interface shear strength. Based on the micro-bonding mechanism and test results
Shear stress-slip model of shear load-slip curves, a shear strength mechanical model and a shear stress-slip model of ECC-to-
concrete interface were proposed to predict interface shear strength and shear stress-slip curve,
respectively.
Ó 2019 Published by Elsevier Ltd.

1. Introduction In recent years, a novel of cement-based material named as

Engineered Cementitious Composites (ECC) draws attention in aca-
After several years of service, concrete structures are faced with demic and engineering industries, because of prominent advan-
many serious problems, such as crack width more than 0.2 mm [1], tages in terms of ultra-high toughness [7–12], multiple micro-
concrete cover spalling [2], rebar corrosion [3], degradation of flex- cracking behavior [13–21], self-healing characterization [22–27],
ural stiffness and bearing capacity [4,5], and concrete durability better fatigue resistance [28–31], and good durability [32–35].
[6]. As a result, it is evident and urgent to repair and strengthen ECC has been now introduced to repair concrete structures [36–
the damaged concrete structures. 40], such as bridge decks, girders, and tunnel linings. For ECC
repaired concrete structures, the interface bonding performance
⇑ Corresponding author at: School of Environment and Civil Engineering, between ECC and concrete is the weakest compared to the concrete
Dongguan University of Technology, No. 1, University Road, Dongguan City, and ECC, due to a lot of micro-cracks, defects, and gaps in the inter-
Guangdong Province P.C. 523808, China. face. Moreover, the interface bonding performance between ECC
E-mail address: [email protected] (X. Wu).

https://doi.org/10.1016/j.conbuildmat.2019.06.188
0950-0618/Ó 2019 Published by Elsevier Ltd.
 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27 13

and concrete is the basic premise of ECC repairing concrete struc- on the micro-bonding mechanism and test results of shear load-
tures. Therefore, the interface properties between ECC and con- slip curves, a shear strength mechanical model and a shear
crete become urgent issues to be coped with [41–43]. stress-slip model of ECC-to-concrete interface were proposed to
Many factors, including ECC strength, interface roughness, con- predict interface shear strength and shear stress-slip curve,
crete substrate strength grade, etc., can influence the mechanical respectively.
properties of ECC-to-concrete interfaces, such as axial tensile prop-
erties, shear properties, fracture properties, and fatigue properties
2. Experimental study
of ECC-to-concrete interface. Therefore, the influences of ECC
strength grade, concrete substrate strength grade, interface rough-
2.1. Materials
ness of concrete substrate, interface adhesive types, interface cur-
ing ages (i.e., 28 days, 120 days, and 365 days), ECC reinforcement
Four ECC mixtures and a concrete mixture were prepared. The
methods (i.e., cast-in-place ECC and epoxy resin bond precast ECC
ECC mixtures consisted of ordinary Portland cement (OPC 42.5N),
components) on the axial tensile properties of ECC-to-concrete
fly ash, silica sand, water, polyvinyl alcohol (PVA) fibers, and some
interface have been studied [44,45] by researchers. This investiga-
concrete admixtures, including water-reducing and accelerating
tion reveals that ECC strength grade, interface roughness of con-
agents. The effect of accelerating agents is to accelerate the con-
crete substrate, interface adhesive type, interface curing age, and
densation and hardening of ECC. The polycarboxylic water-
ECC reinforcement method had a significant influence on interface
reducing agent (PCA-I) with reduction rate of 30% was manufac-
tensile strength, while concrete substrate strength grade had a
tured by SuBoTe New Material Co., LTD of China. The accelerating
slight impact on interface tensile strength.
agent (Red star I type) with initial setting time of 1–5 min and final
Besides, the influences of ECC strength grade, concrete substrate
setting time of 5–10 min was manufactured by Xi’an QinBoTe
strength grade, interface roughness of concrete substrate, interface
Material Co., LTD of China. Two kinds of PVA fibers were manufac-
adhesive type, interface moisture condition (i.e., wet saturated
tured by Kuraray Corporation of Japanese (i.e., Type J) and SuBoTe
state and dry state), interface orientation (i.e., horizontal plane
New Material Co., LTD of China (i.e., Type G), respectively, and their
and vertical plane), interface curing ages (i.e., 28 days, 120 days,
physical and mechanical properties are listed in Table 1. The
and 365 days), sand-cement ratio of ECC (i.e., 0.5, 0.6, and 0.8),
amount of PVA fibers added to the matrix was 2% of the total mate-
and content of slag and fly ash in ECC on the shear properties of
rial volume. The particle sizes of silica sand ranged approximately
ECC-to-concrete interface were also studied [46–51] by research-
from 0.15 mm to 0.32 mm. The measured 28-day average com-
ers. This investigation reveals that ECC strength grade, concrete
pressive strengths of each group of ECC-1, ECC-2, ECC-3, and
substrate strength grade, interface roughness of concrete substrate,
ECC-4 specimens (70.7 mm  70.7 mm  70.7 mm in dimensions)
interface adhesive type, interface moisture condition, interface
were 21.7 MPa, 32.6 MPa, 32.6 MPa, and 40.8 MPa, respectively,
curing age, and sand-cement ratio of ECC had a significant influ-
with standard deviations of 0.858, 1.349, 1.409 and 1.470, respec-
ence on interface shear strength, while the interface orientation,
tively. Ordinary Portland cement (OPC 52.5) was applied to manu-
and content of slag and fly ash in ECC had a slight impact on inter-
facture the concrete and the measured 28-day average
face shear strength.
compressive strength and splitting tensile strength of cubic speci-
Until now, the influences of roughness degree of concrete sub-
mens (150 mm  150 mm  150 mm in dimensions) were
strate on the fracture properties of ECC-to-concrete interface
43.2 MPa and 1.43 MPa, respectively, with standard deviations of
[52,53] and the fatigue properties [54,55] of ECC-to-concrete inter-
1.071 and 0.118, respectively. The detailed mixture proportions
face have also been studied by researchers. This investigation
of ECC and concrete are shown in Table 2.
reveals that the roughness degree of concrete substrate had a sig-
In order to ensure that each group of specimens has basically
nificant influence on interface fracture energy and fatigue proper-
identical interface roughness, the Sikadur-330 epoxy resin only
ties of ECC-to-concrete interface, and coarser interface could cause
was used to bond aggregates and not for an interface adhesive
higher interface fracture energy and better fatigue properties. And
between ECC and concrete, which was produced by Sika Corpora-
the relationship between interface fracture energy and interface
tion of Switzerland. Material parameters of Sikadur-330 epoxy
roughness is a quadratic function.
resin were provided by the manufacturer. The ultimate tensile
Due to the conventional construction, low cost and high con-
strength, elastic modulus, and elongation of Sikadur-330 epoxy
struction efficiency of spraying ECC [56,57], it has been widely
resin were 30 MPa, 3.8 GPa, and 1.6%, respectively.
applied in many ECC repaired concrete structures [58]. Moreover,
the absence of vibrating construction of spraying ECC leads to a sig-
nificant difference in the internal microscopic pore structure of 2.2. Specimen preparation
ECC-to-concrete interface compared to casting ECC. However, a lit-
erature review indicates that there are no studies on the shear One group of dumbbell-shaped specimens of each ECC mixture
properties of spraying ECC-to-concrete interface, shear strength was prepared to measure tensile properties and each group had
mechanical model, and shear stress-slip model of ECC-to- three identical specimens. The geometric dimensions of all
concrete interface up to date. dumbbell-shape ECC specimens were 280 mm in length, 15 mm
Therefore, this paper focuses on investigating the influence of in thickness, and the maximum and minimum width of 60 mm
ECC mixture construction method (i.e., common casting and spray- and 30 mm, respectively, as shown in Fig. 1.
ing ECC) on shear properties of ECC-to-concrete interface. Also, the A total of 45 shear specimens were prepared to investigate the
ECC strength grade, interface roughness of concrete substrate, and influences of ECC mixture construction method, ECC strength
PVA fiber type were considered in the experimental design. Based grade, interface roughness of concrete substrate, and PVA fiber

Table 1
Physical and mechanical properties of PVA fibers.

PVA fiber types Diameter (lm) Length (mm) Density (kg/m3) Ultimate tensile strength (MPa) Tensile elastic modulus (GPa) Elongation (%)
Type J 39 12 1300 1560 38 7
Type G 20 15 1300 1300 37 7
14 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27

Table 2
Mixture proportions of concrete and ECC (kg/m3).

Mixture Water Cement Fly Silica Fine Coarse Water-reducing Accelerating Fiber volume Compressive strength
types ash fume aggregate aggregate agent agent fraction (MPa)
Type Type
J G
Concrete 162 455 66 - 5320 1213 4 – – – 49.7
ECC-1 397 248 992 40 248 – – – 2% – 21.7
ECC-2 473 957 287 40 250 – 1 20 – 2% 32.6
ECC-3 473 957 287 40 250 – 1 20 2% – 32.6
ECC-4 353 379 954 40 309 – 4 – 2% – 40.8

vibrating machine, while the later was sprayed directly on the con-
crete substrate without vibrating.
In order to ensure that each group of specimens has basically
identical interface roughness, the interface processing technology
is adopted as follows. Three types of interfaces were generated
by treating the concrete surface bonding with ECC. Type A pro-
cessed by removing cement paste attached to the concrete sub-
strate. On the base of Type A, a 2 mm thick epoxy resin layer was
coated on the interface, and 1.75–2.0 mm and 2.5–5.0 mm aggre-
gates were then adhered to this interface to form the interfaces
of type B and C, respectively, as shown in Fig. 6. The average value
of vertical height difference between protruding and concave por-
tions of the interface is used to quantitatively assess the interface
roughness [59], which is named interface roughness value in this
paper. A set of the device, which is consisted of a dial indicator
and a flat base, is designed to measure vertical height difference
between protruding and concave of the interface, as shown in
Fig. 3. The measured interface roughness values of type A, type B,
and type C were 0.57 mm, 1.32 mm, and 3.33 mm, respectively.
All the 45 shear specimens were divided into 15 groups and
each group had three identical coupons, as shown in Table 3. The
Fig. 1. Details of ECC specimens: (a) front view; (b) side view (mm). shear specimens’ notation is defined as follows: the first letters
(N or S) indicates the construction method of ECC (i.e., N is referred
type on the shear properties of ECC-to-concrete interface. All shear as the casting ECC and S is referred as the spraying ECC); the sec-
specimens had a prism shape with the designed geometric dimen- ond numbers (1, 2 or 3) indicates the ECC strength grade (i.e., 1
sions of 100 mm in length, 100 mm in depth, and 40 mm in width. is 21.7 MPa, 2 is 32.6 MPa, and 3 is 40.8 MPa). The third letters
The geometric dimensions of the interface between ECC and con- (A, B or C) represents the interface type (i.e., A, B or C). The fourth
crete were 100 mm in depth and 40 mm in width, as shown in letters (G or J) represents the PVA fiber type (i.e., G is PVA fiber
Fig. 2. made from SuBoTe New Material Co., LTD and J is PVA fiber made
Three designed strength grades of ECC (i.e., 21.7 MPa, 32.6 MPa, from Kuraray Corporation). For example, N-1-A-J refers to the spec-
and 40.8 MPa) and one designed strength grade of concrete (i.e., imen which was bonded by type A interface with casting ECC
49.7 MPa) were prepared. Two kinds of PVA fibers, manufactured (strength of 21.7 MPa) incorporating with type J fiber, while S-2-
by Kuraray Corporation of Japanese (i.e., Type J) and SuBoTe New C-J represented the specimen which was bonded by type C inter-
Material Co., LTD of China (i.e., Type G), respectively, were selected face with spraying ECC (strength of 32.6 MPa) incorporating with
to mix into the cement-based composite. Two construction type J fiber.
methods (i.e., casting and spraying) were determined to construct Before preparing the shear specimens, several steel moulds
ECC mixture. The significant difference between casting ECC and were manufactured, as shown in Fig. 4. The preparation process
spraying ECC was that the former has been vibrated using the of shear specimens is as follows.

Fig. 2. Details of shear specimens (mm): (a) side view; (b) top view.
 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27 15

2.3.2. Single-sided shear test of ECC-to-concrete interface

The single-sided shear test method is shown in Fig. 10. The
shear force acting on the bonding interface was applied by a uni-
versal testing machine. The whole loading process was controlled
by at a speed of 0.2 mm/min. The real-time interface slip and load
were collected by an automatic data acquisition system.

3. Results and discussion

3.1. Tensile properties of ECC specimens

Fig. 3. Measurement principle of interface roughness value.
The stress-strain curves of all ECC specimens under axial pull-
out load are shown in Fig. 11. It can be found that the stress-
1) A steel mould and a plate were installed together to form a strain curves of all ECC specimens are basically similar and exhibits
50 mm  100 mm  40 mm chamber, as shown in Fig. 5. a strain hardening behavior, the tensile strain is up to 4.2%, ulti-
Concrete mixture was poured into the chamber to form mate tensile stress is up to 5.6 MPa, because of the bridging effect
the concrete substrate. After 24 hours, the steel plate was of fibers in the ECC mixture. And it was also found that the ECC
removed, and then the concrete substrate was cured 28 days specimens under axial pull-out load exhibit multiple micro-
under the temperature of 20 ± 2 °C and humidity of 95%. cracking behaviors, as shown in Fig. 12.
2) The three types of the interfaces were manufactured accord-
ing to the above-mentioned method, as shown in Fig. 6. After
cleaning the bonding surface of concrete substrate with 3.2. Failure modes and shear load-slip curves of single-sided shear test
water, the surface was naturally dried in laboratory.
3) A steel mould and a concrete substrate were installed Two failure modes of interface shear failure and ECC failure
together to form another 50 mm  100 mm  40 mm cham- were observed in the test, as shown in Figs. 13 and 14, respectively.
ber, as shown in Fig. 7. For specimens S-2-A-J, S-2-B-J, and S- The failure mode of specimens with interface types A and B was
2-C-J, ECC mixture was sprayed into the chamber by a mor- interface shear failure, as shown in Fig. 13. The failure mode of
tar spraying machine, without any vibration. For all other specimens with interface type C was ECC failure, as shown in
specimens, the ECC mixture was casted into the chamber, Fig. 14, because the better interface bonding strength leaded to
which was then vibrated for 60 seconds on a vibrating failure on the ECC side. According to the results, it can be found
machine with a vibration frequency of 50 ± 3 Hz and effec- that the roughness degree of concrete substrate had a significant
tive amplitude of 0.5 ± 0.1 mm. Before loading, all specimens influence on the failure mode, while ECC strength grade, ECC mix-
were cured 28 days in a chamber with the temperature of ture construction method, and PVA fiber type had no influence on
20 ± 2 °C and relative humidity of 95%. the failure mode.
The shear load-slip curves of each group specimens are shown
2.3. Test methods in Fig. 15. It can be found that the shear load-slip curves of all spec-
imens were basically similar. The shear load-slip curve responses
2.3.1. Axial tensile test of ECC specimens exhibited nearly a linear relationship before the applied loads
In order to ensure that the ECC specimen bear axial tensile load- reached to the maximum values. After the maximum loads, the
ings, a loading device with two gimbal joints was designed, as shear load sharply decreased with the increase of slip and the shear
shown in Fig. 8. The axial pull-out force acting on the ECC specimen load-slip curves presented some ductility. After shear force sharply
was applied by a universal tensile machine, as shown in Fig. 9. The decreased and crack penetration interface, the interface could still
whole loading process was controlled by at a speed of 0.2 mm/min. bear small shear loads, as shown in Fig. 16, because mechanical
The real-time strain and load were collected by an automatic data interlocking force and fibers bridging effect played an important
acquisition system. role in bearing the loads.

Table 3
Detailed characterizations of specimens.

Specimen types Construction methods of ECC mixture ECC strength (MPa) Interface types PVA fiber types Numbers of specimens
N-1-A-J Casting 21.7 Type A Type J 3
N-1-B-J 21.7 Type B 3
N-1-C-J 21.7 Type C 3
N-2-A-J Casting 32.6 Type A Type J 3
N-2-B-J 32.6 Type B 3
N-2-C-J 32.6 Type C 3
S-2-A-J Spraying 32.6 Type A Type J 3
S-2-B-J 32.6 Type B 3
S-2-C-J 32.6 Type C 3
N-2-A-G Casting 32.6 Type A Type G 3
N-2-B-G 32.6 Type B 3
N-2-C-G 32.6 Type C 3
N-3-A-J Casting 40.8 Type A Type J 3
N-3-B-J 40.8 Type B 3
N-3-C-J 40.8 Type C 3
16 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27

Fig. 7. Chamber of ECC.

Fig. 4. Details of steel moulds (mm): (a) front view; (b) side view.

Fig. 5. Chamber of concrete.

Fig. 8. Schematic diagram of axial tensile loadings.

Fig. 6. Interface types A, B, and C.

3.3.1. Influence of ECC mixture construction method
The influences of ECC mixture construction method (i.e., casting
ECC and spraying ECC) on interface shear strengths of specimens
3.3. Interface shear strength are shown in Fig. 17.
In Fig. 17, when spraying ECC instead of casting ECC, the inter-
The interface shear strength of all specimens can be calculated face shear strengths of specimens all decreased. The corresponding
according to Eq. (1): reduction rates of specimen groups with interface A, B, and C types
were 47%, 19%, and 13%, respectively, which indicated compared to
su ¼ Pu =A ð1Þ
the casting ECC, the spraying ECC had an obvious negative effect on
where su is interface shear strength, MPa; Pu is maximum shear interface shear strength. It might because the absence of vibrating
load, kN; A is interface area, mm2. In Figs. 17–20, the shear stress construction of spraying ECC leads to a significant difference in the
values are the average values of each group shear specimens. Exper- internal microscopic pore structure of ECC-to-concrete interface
imental results of all specimens are shown in Table 4. compared to casting ECC.
 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27 17

ECC compressive strength could cause higher interface shear

strength no matter which types of interfaces were applied.

3.3.3. Influence of PVA fiber type

The influences of PVA fiber type on interface shear strength of
specimens with type A, B, and C interfaces are shown in Fig. 19.
Two kinds of PVA fibers, manufactured by Kuraray Corporation
of Japanese and SuBoTe New Material Co., LTD of China, respec-
tively, were selected to mix into ECC, respectively. In Fig. 19, when
PVA fiber type of SuBoTe New Material Co., LTD instead of PVA
fiber type of Kuraray Corporation, the variations of interface shear
strength were very small. The corresponding difference values of
specimen groups with interfaces type A, B, and C were 0.01 MPa,
0.09 MPa, and 0.08 MPa, and the corresponding change rates
were 1%, 5%, and 4%, respectively.
According to the results, it can be found that no matter which
types of PVA fiber were applied, the variations of interface shear
strength were very small, which indicated the PVA fiber type had
a slight influence on interface shear strength. It might due to the
fact that, according to the shear failure mechanism of ECC-to-
concrete interface, interface shear strength is composed of chemi-
cal gluing force, van der Waals force, mechanical interlocking force,
and fiber bridging effect, of which the fiber bridging effect played
an insignificance role in interface shear strength.

3.3.4. Influence of interface roughness

The influences of interface roughness of concrete substrate on
interface shear strength are shown in Fig. 20.
In Fig. 20, it can be found that the interface shear strength
Fig. 9. Test setup of axial tensile. enhanced with increasing interface roughness value. Taking the
specimen groups N-3-J for an example, when interface roughness
values increased from 0.57 mm to 1.32 mm, the interface shear
strength increased rapidly and the corresponding increase rate
was 114%. And then interface roughness values increased from
1.32 mm to 3.33 mm, interface shear strength had relative small
growths and the corresponding increase rate was 40%. For the
specimen groups N-1-J, N-2-J, S-2-J, and N-2-G, the influences of
interface roughness on interface shear strength had the same
trends.
According to the results, it can be found that the interface
roughness of concrete substrate had a significant influence on
interface shear strength, and coarser interface could cause higher
interface shear strength. It can be inferred that mechanical inter-
locking force played an important role in interface shear strength.

4. Micro-bonding mechanism and shear mechanical model of

ECC-to-concrete interface

4.1. Micro-bonding mechanism of ECC-to-concrete interface under

shear loads
Fig. 10. Single-sided shear test setup.
The bonding force of ECC-to-concrete interface is composed of
3.3.2. Influence of ECC strength grade chemical gluing force, van der Waals force, and mechanical inter-
The concrete was designed such to have always higher com- locking force. Under shear loads, when the interface aggregate,
pressive strength than ECC in this paper. The influences of ECC and ECC did not slip, the interface bonding force was composed
strength grade on interface shear strength of specimens are shown of chemical adhesive force, van der Waals force, and mechanical
in Fig. 18. In Fig. 18, it can be found that the interface shear interlocking force; when the interface aggregate and ECC slipped,
strength enhanced with increasing ECC strength. Taking the speci- the adhesive force and van der Waals force would all disappear,
mens groups with type C (N-C-J) for an example, when ECC com- and the interface bond force was only composed of mechanical
pressive strength increased from 21.7 MPa to 40.8 MPa, the interlocking force.
interface shear strength increased by 3 times. For the specimen
groups with A and B types interface, the influences of ECC strength 4.2. Shear strength model of ECC-to-concrete interface
on interface shear strength had the same trend.
Therefore, it can be concluded that the ECC strength grade had a According to the micro-bonding mechanism and interface fail-
significant influence on the interface shear strength, and the higher ure modes of ECC-to-concrete interface under single-sided shear
18 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27

Fig. 11. Stress-strain curves of ECC specimens: (a) ECC-1; (b) ECC-2; (c) ECC-3 and (d) ECC-4.

Fig. 13. Representatives of interface shear failure.

Some assumptions were made for shear failure mechanism of

ECC-to-concrete interface as follows.

Fig. 12. Representatives of ECC multiple micro-cracking behaviours.

1) When ECC-to-concrete interface reaches ultimate limit state,
V 1 and V 4 are assumed to be rigid body, V 2 and V 3 are
assumed to be plastic body, respectively.
loads, a shear failure mechanism of ECC-to-concrete interface was 2) The r is velocity reduction coefficient of plastic body V 2 , and
proposed in this paper, which was consisted of concrete without 0 < r  1; The b is velocity reduction coefficient of plastic
shear deformation (i.e., V 1 ), the part of interface aggregate combin- body V 3 , and 0  b < 1.
:
ing with ECC (i.e., V 2 ), ECC with shear deformation (i.e., V 3 ), ECC 3) The strain rate field with motion permission eij (motion field
without shear deformation (i.e., V 4 ), and velocity discontinuity sur- for short in this paper) in shear failure mechanism of ECC-to-
face of SD , as shown in Fig. 21. concrete interface meets the following conditions:
 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27 19

a. The geometric condition of plastic area meets:

:
 :
eij ¼ 1=2 v i;j þ v i;j , eij can be derived from velocity field
v i ;

b. Velocity boundary conditions meets v i ¼ v i ;
c. The external power is greater than zero, namely,
R
T v  dS  0;
ST i i
:
d. The volume is incompressible, namely, eii ¼ 0.

Based on above basic assumptions, the shear force P v  was

solved by using discontinuous surface theory and upper limit the-
orem of limit analysis, as shown in Eq. (2).
Z XZ
:  
Pv  ¼ Dðeij ÞdV þ s0S v t dS ð2Þ
V SD

: :
where Dðeij Þ ¼ rij eij .
Fig. 14. Representatives of ECC failure.
The P v  was obtained by sorting Eq. (2), as shown in Eq. (3).

Fig. 15. Shear load-slip curves: (a) N-1-J; (b) N-2-J; (c) S-2-J; (d) N-2-G; (e) N-3-J.
20 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27

Fig. 16. Representatives of crack penetration interface.

Fig. 17. Influence of ECC mixture construction method on interface shear strength. Fig. 19. Influence of PVA fiber types on interface shear strength.

Fig. 18. Effect of ECC strength grade on interface shear strength. Fig. 20. Influence of interface roughness on interface shear strength.
 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27 21

Table 4
Experimental Results.

Specimen types Interface area (mm2) Maximum shear load (kN) Interface shear strength (MPa) Average values (MPa) Standard deviation
N-1-A-J-1 40  100 1.137 0.28 0.33 0.04
N-1-A-J-2 40  100 1.539 0.38
N-1-A-J-3 40  100 1.236 0.31
N-1-B-J-1 40  100 2.540 0.64 0.56 0.07
N-1-B-J-2 40  100 2.300 0.58
N-1-B-J-3 40  100 1.867 0.47
N-1-C-J-1 40  100 3.450 0.86 0.86 0.08
N-1-C-J-2 40  100 3.075 0.77
N-1-C-J-3 40  100 3.823 0.96
N-2-A-J-1 40  100 3.664 0.92 0.93 0.01
N-2-A-J-2 40  100 3.764 0.94
N-2-A-J-3 40  100 3.672 0.92
N-2-B-J-1 40  100 6.707 1.68 1.72 0.04
N-2-B-J-2 40  100 7.125 1.78
N-2-B-J-3 40  100 6.839 1.71
N-2-C-J-1 40  100 8.882 2.22 2.23 0.04
N-2-C-J-2 40  100 9.138 2.28
N-2-C-J-3 40  100 8.731 2.18
S-2-A-J-1 40  100 2.021 0.51 0.49 0.02
S-2-A-J-2 40  100 1.967 0.49
S-2-A-J-3 40  100 1.873 0.47
S-2-B-J-1 40  100 5.652 1.41 1.40 0.05
S-2-B-J-2 40  100 5.329 1.33
S-2-B-J-3 40  100 5.767 1.44
S-2-C-J-1 40  100 7.486 1.87 1.93 0.06
S-2-C-J-2 40  100 7.641 1.91
S-2-C-J-3 40  100 8.032 2.01
N-2-A-G-1 40  100 3.664 0.92 0.92 0.02
N-2-A-G-2 40  100 3.743 0.94
N-2-A-G-3 40  100 3.584 0.9
N-2-B-G-1 40  100 6.707 1.68 1.63 0.04
N-2-B-G-2 40  100 6.321 1.58
N-2-B-G-3 40  100 6.538 1.63
N-2-C-G-1 40  100 9.437 2.36 2.31 0.05
N-2-C-G-2 40  100 8.937 2.23
N-2-C-G-3 40  100 9.289 2.32
N-3-A-J-1 40  100 5.237 1.31 1.11 0.15
N-3-A-J-2 40  100 3.850 0.96
N-3-A-J-3 40  100 4.175 1.04
N-3-B-J-1 40  100 9.415 2.35 2.38 0.23
N-3-B-J-2 40  100 8.486 2.12
N-3-B-J-3 40  100 10.707 2.68
N-3-C-J-1 40  100 12.865 3.22 3.33 0.13
N-3-C-J-2 40  100 13.050 3.26
N-3-C-J-3 40  100 14.050 3.51

Z Z XZ
: :  
Pv  ¼ Dðeij ÞdV þ Dðeij ÞdV þ s0S v t dS ð3Þ
V2 V3 SD

R :
4.2.1. Calculation of V Dðeij ÞdV
The Ducker-Prager model [60] was selected as the yield condi-
tion, which was the extension form of Mises yield condition, as
shown in Eq. (4).
1
f ðI1 ; J 2 Þ ¼ aI1 þ ðJ 2 Þ2  k ¼ 0 ð4Þ
where a and k are all positive constant; I1 and J2 are the first and
second invariant of deviator stress tensor, respectively.
The associated flow law of Eq. (4) is shown in Eq. (5).
" #
p @f ðI1 ; J2 Þ Sij
de ¼ dk ¼ dk adij þ pffiffiffiffi ð5Þ
ij
@ rij 2 J2

1 ;i ¼ j
where dij ¼ ; Sij is deviator stress tensor.
0 ; i–j
p p
The deij deij is obtained according to Eq. (5) and plastic potential
theory, as shown in Eq. (6).
" #2
p p 2 Sij
de e
ij d ij ¼ ðdkÞ adij þ pffiffiffiffi ð6Þ
2 J2
Fig. 21. Shear failure mechanism of ECC-to-concrete interface.
22 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27

2
pffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
:
The dk is obtained by substituting depij depij ¼ 1:5ðdepk Þ ,  2 : : 2 : : 2 : : 2 : 2 : 2 : 2
e k ¼ ðex  ey Þ þ ðey  ez Þ þ ðez  ex Þ þ 6½ðexy Þ þ ðexz Þ þ ðeyz Þ 
dij Sij ¼ Sii ¼ 0, and Sij Sij ¼ 2J 2 into Eq. (6), as shown in Eq. (7). 3
pffiffiffi ð12Þ
6 depk :
dk ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð7Þ The equivalent plastic strain increment ek of plastic body V 2 and
2 3a2 þ 0:5
V 3 are obtained substituting strain rates of plastic body V 2 and V 3
where depk is equivalent plastic strain increment. into Eq. (12), as shown below.
: pffiffi
The depij is obtained by substituting Eq. (7) into Eq. (5), as shown For plastic body V 2 :ek ¼ 2 3 3 1d c v  ;
: pffiffi
in Eq. (8). For plastic body V 3 : ek ¼ 2 3 3 cdb v  ;
pffiffiffi " #
6 depk Sij Therefore, Eq. (13) can be obtained by combining Eq. (10), as
depij ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi adij þ pffiffiffiffi ð8Þ shown below.
2 3a2 þ 0:5 2 J2
R
: R : R :
Dðeij ÞdV ¼ V 2 Dðeij ÞdV þ V 3 Dðeij ÞdV
V
For rigid-perfectly plastic materials, when small deformation
pffiffi 1=2
ðJ2 ÞV pffiffi pffiffi 1=2
ðJ 2 ÞV pffiffi
occurs, Eq. (8) can be changed into Eq. (9), as shown below. ¼ 26 pffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi 2 3 1c
v  hb 2d þ 26 pffiffiffiffiffiffiffiffiffiffiffiffi
3 ffi 2 3 cb
v  hb 2d ð13Þ
pffiffiffi : " # 2 3 d
3a2 þ0:5 2 3 d 3a3 þ0:5
: ek Sij
e 
ij ¼
6
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi adij þ pffiffiffiffi ð9Þ ¼ pffiffiffiffiffiffiffiffiffiffi
1
2
ð1  cÞðJ 2 ÞV1=2
6a2 þ1 2
v  hb þ pffiffiffiffiffiffiffiffiffiffi
1
2 V 3 v hb
cbðJ2 Þ1=2 
6a3 þ1
2 3a2 þ 0:5 2 J2
R :
: : If you interchange the plastic body V 2 and V 3 , the V Dðeij ÞdV
where eij is strain rate in motion field; ek is equivalent plastic strain
should remain the same. Therefore, Eq. (14) can be obtained by
increment in motion field.
: swapping places with V 2 and V 3 for Eq. (13), as shown below.
The Dðeij Þ obtained according to volume of rigid-perfectly plas- R : R : R :
tic material is incompressible, as shown in Eq. (10). V
Dðeij ÞdV ¼ V2
Dðeij ÞdV þ V3
Dðeij ÞdV
pffiffiffi : "
S
# ¼ pffiffiffiffiffiffiffiffiffiffi
1
ð1  cÞðJ 2 ÞV1=2 v  hb þ pffiffiffiffiffiffiffiffiffiffi
1
V 2 v hb
cbðJ2 Þ1=2 
ð14Þ
: : : 6 ek 2 6a2 þ1 2 2 6a3 þ1
Dðe 
ij Þ ¼r  
e ¼ Sij e 
¼ ffi Sij adij þ pijffiffiffiffi
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ij k k
2 3a2 þ 0:5 2 J2 ¼ pffiffiffiffiffiffiffiffiffiffi
1
6a23 þ1
ð1 c ÞðJ 2 ÞV1=2
3
v 
hb þ pffiffiffiffiffiffiffiffiffiffi
1
6a22 þ1
cbðJ 2 Þ1=2
V2 v 
hb
pffiffiffi :
6 ek
¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi J21=2 ð10Þ The Eq. (15) can be obtained by rearranging the Eq. (14), as
2 3a2 þ 0:5 shown below.
The strain rates of V 2 and V 3 in motion field are obtained 1
V 2 v hbð1  c  cbÞ

according to Eq. (10) and Fig. 21, as shown in Eq. (11). qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðJ 2 Þ1=2
6a22 þ 1
: 1
 
eij ¼ v i;j þ v i;j ð11Þ 1
2 ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðJ 2 ÞV1=2
3
v  hbð1  c  cbÞ ð15Þ
:
 @ v x
:
 @ v y 6a23 þ 1
For plastic body V 2 : e ¼
x @x
¼ 0; e ¼
y @y
¼ 0;
If the Eq. (15) is true, there must be Eq. (16), as shown below.
:
@ v y

@ v x
¼ 12 v d=2
 cv 
exy ¼ 12 @y
þ @x
¼ 1d c v  ; 1
c¼ ð16Þ
1þb
: :
e ¼ @@xv ¼ 0; e ¼ @@yv ¼ 0;
 
For plastic body V 3 :  x  y R :
x y So the V
Dðeij ÞdV is obtained, as shown in Eq. (17).

 2 3
:
exy ¼ 12 @ v x
þ
@ v y
¼ 12 cbv 
¼ cdb v  ; Z :
6 1 1 1=2 7 b
@y @x d=2
Dðe 
ij ÞdV ¼ 4qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðJ 2 Þ1=2
V 2 þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðJ 2 ÞV 3 5 v  hb
V 6a2 þ 1
2
6a3 þ 1
2 1 þ b
:
In motion field, the equivalent plastic strain increment ek is
shown in Eq. (12). ð17Þ

Table 5
Fitting parameters.

Specimen types Interface roughness values (mm) ECC strength rs (MPa) a b c Correlation coefficient
N-1-A-J 0.57 21.7 0.04334 1.69877 0.04589 0.91997
N-1-B-J 1.32
N-1-C-J 3.33
N-2-A-J 0.57 32.6 0.08155 0.84199 0.06987 0.9959
N-2-B-J 1.32
N-2-C-J 3.33
S-2-A-J 0.57 32.6 0.09284 0.80271 0.06067 0.99447
S-2-B-J 1.32
S-2-C-J 3.33
N-2-A-G 0.57 32.6 0.07562 1.21648 0.07555 0.99492
N-2-B-G 1.32
N-2-C-G 3.33
N-3-A-J 0.57 40.8 0.10478 0.95897 0.08487 0.96456
N-3-B-J 1.32
N-3-C-J 3.33
 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27 23

Fig. 22. Comparisons of experimental values of interface shear strength with its prediction curves: (a) N-1-J; (b) N-2-J; (c) S-2-J; (d) N-2-G and (e) N-3-J.

where h and b are the height and width of plastic body, respec- where sS1 , rS1 , sS2 , and rS2 are shear strength and uniaxial com-
tively; a2 and a3 are the positive constant, respectively, which can pressive strength of plastic body V 2 and V 3 , respectively.
be determined by calculation as follows.
PR  
In the state of uniaxial normal stress and uniaxial shear stress, 4.2.2. Calculation of s0 v t dS
SD S
Eq. (18) can be obtained from Eq. (4), as shown below. The velocity discontinuity value in velocity discontinuity sur-
( face SD can be obtained according to Fig. 21, as shown in Eq. (19).
a rS þ prSffiffi3  k ¼ 0  
ð18Þ 1b 
a 0 þ sS  k ¼ 0 v t ¼ cv   cbv  ¼ ð1  bÞcv  ¼ v ð19Þ
1þb
k ¼ sS , a ¼ rsSS  p1ffiffi3 can be obtained by solving Eq. (18), therefore, The s0S takes the Coulomb shear failure condition, namely,
k2 ¼ sS1 ;a2 ¼ rsS1
S1
 p1ffiffi3; k3 ¼ sS2 ; a3 ¼ rsS2S2  p1ffiffi3 can be also obtained, s0S ¼ f ðC; rn ; uÞ, where C and u are combined cohesive force and
24 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27

friction angle of ECC-to-concrete failure interface, respectively; rn a parameter of interface roughness. Therefore, f ðC; uÞ can be con-
is normal stress of velocity discontinuity surface SD ; tensile stress sidered as a function with integrated cohesive force C and interface
is positive and compressive stress is negative. roughness of failure interface as two independent variables.
The Eq. (20) is obtained by substituting Eq. (19) into According to the micro-bonding mechanism of ECC-to-concrete
PR  
s0 v t dS, as shown below.
SD S
interface and shear load-slip curves, when shear loads reached ulti-
XZ   1b 
mate shear strength, the interface aggregate and ECC had been
s0S v t dS ¼ s0S v hb ð20Þ slipped, and the adhesive force and van der Waals force were all
SD 1þb disappear. Therefore, the interface bond force was only composed
of mechanical interlocking force. So, the f ðC; uÞ will become a func-
4.2.3. Upper limit solution of interface shear strength tion with only one independent variable of interface roughness.
The Eq. (21) is obtained by substituting Eq. (17) and Eq. (20) According to the above discussion and analysis, the Eq. (24)
into Eq. (2), as shown below. would be a function with ECC axial compressive strength and inter-
2 3 face roughness as two independent variables. It can be inferred that
1b the ultimate shear strength of ECC-to-concrete interface is mainly
6 1 1 1=2 7 b
su ¼ 4qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðJ 2 Þ1=2
V 2 þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðJ 2 ÞV 3 5 þ s0S ð21Þ affected by ECC axial compressive strength and interface roughness.
6a2 þ 1
2
6a3 þ 1
2 1 þ b 1þb
Meanwhile, the test results in this paper also shown that shear
strength of ECC-to-concrete interface was significantly affected by
Since Eq. (21) is more complex, the coefficients are difficult to ECC axial compressive strength and interface roughness.
determine, and it is difficult to apply, so it would be simplified.
According to Eq. (4), the Eq. (21) can be transformed into an 4.2.5. Determine relevant parameters of shear strength model
expression of first invariant of stress tensor ðI1 ÞV 2 and ðI1 ÞV 3 . In order to simplify the Eq. (24), let a ¼ ð1  bÞ=ð1 þ bÞ and
When the hydrostatic stress is relatively small, the yield condi- pffiffiffi
c ¼ 2 3b=3ð1 þ bÞ. The Eq. (25) is obtained by substituting
tion of Eq. (4) can be simplified as the yield criterion of Mises pffiffiffi
a ¼ ð1  bÞ=ð1 þ bÞ and c ¼ 2 3b=3ð1 þ bÞ into Eq. (24), as shown
[60]. Therefore, a2 and a3 in Eq. (21) are all approximately zero,
below.
as shown below.
su ¼ rS ða f ðC; uÞ þ cÞ ð25Þ
1=2 1 1=2 1
ðJ 2 ÞV 2 ¼ pffiffiffi rS1 ; ðJ 2 ÞV 3 ¼ pffiffiffi rS2
3 3 According to experimental data of interface shear strength, the
Eq. (25) is fitted, as shown below.
1=2 1 1=2 1
or ðJ 2 ÞV 2 ¼ pffiffiffi rS2 ; ðJ 2 ÞV 3 ¼ pffiffiffi rS1 su ¼ rS ða ex=b þ cÞ ð26Þ
3 3
where su is shear strength of ECC-to-concrete interface, MPa; rS is
So, the Eq. (21) can be simplified, as shown below.
uniaxial compressive strength of ECC; x is interface roughness val-
pffiffiffi
3 b 1b ues, mm; a, b, and c are fitting parameters, respectively. The fitting
su ¼ ðrS1 þ rS2 Þ þ s0S ð22Þ parameters in Eq. (26) and correlation coefficient are shown in
3 1þb 1þb
Table 5.
where rS1 and rS2 are uniaxial compressive strength of plastic body The comparisons of experimental values of interface shear
V 2 and V 3 , respectively. strength with model prediction curves are shown in Fig. 22. It
In this paper, the plastic body V 3 is ECC, and the plastic body V 2 can be found that the shear strength model had good consistencies
is the part of interface aggregate combining with ECC. Since com- with experimental results, although there were some discrepancies
pressive strength of interface aggregate is greater than ECC, the between the predicted values and test values. Accordingly, Eq. (26)
compressive strength of plastic body V 2 mainly depends on ECC, could be used to predict the variation laws of interface shear
so the compressive strength of plastic body V 2 would be approxi- strength with ECC compressive strength and interface roughness.
mately equal to compressive strength of plastic body V 3 . Therefore,
Eq. (22) can be simplified, as shown in Eq. (23). 4.3. Shear stress-slip model of ECC-to-concrete interface
pffiffiffi
2 3 b 1b
su ¼ rS þ s0S ð23Þ According to test results, the rising and descent sections of
3 1þb 1þb
shear load-slip curves were approximately straight lines in general.
where rS is uniaxial compressive strength of ECC.
The s0S ¼ f ðC; rS ; uÞ is a function with ECC compressive strength
rS , integrated cohesive force C, and friction angle u as three inde-
pendent variables. When the ECC compressive strength is known,
the s0S ¼ f ðC; rS ; uÞ can be changed into s0S ¼ rS f ðC; uÞ, so the
Eq. (24) is obtained by substituting s0S ¼ rS f ðC; uÞ into Eq. (23),
as shown below.
pffiffi
su ¼ 2 3 3 rS 1þb
b 1b
þ rS 1þb f ðC; uÞ
h pffiffi i ð24Þ
1b
¼ rS 1þb
f ðC; uÞ þ 2 3 3 1þb
b

4.2.4. Discussion of upper limit solution of interface shear strength

The f ðC; uÞ in Eq. (24) is a function with integrated cohesive
force C and friction angle u as two independent variables. The inte-
grated cohesive force C is composed of chemical gluing force and
van der Waals force. Since the tangent value of friction angle u is
the friction coefficient, the friction angle u can be considered as Fig. 23. Shear load-slip curves.
 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27 25

For simplification, they were simplified into straight lines. There- is calculated according to Eq. (26); s is interface slip, mm; s0 is
fore, an interface shear stress-slip curve in the form of double- the interface slip corresponding to ultimate shear strength, mm;
linear property was proposed in this paper, as shown in Fig. 23, su is ultimate interface slip, mm.
the equation of interface shear stress-slip curve as shown in Eq. (27). The s0 and u0 can be obtained by substituting ECC compressive
( strength and interface roughness into Eqs. (28) and (29), respec-
su ss0 0  s  s0 tively. The Eqs. (28) and (29) are derived from the regression of
s¼ ð27Þ experimental data in this paper.
su ssuuss0 s0 < s  su
s0 ¼ ð0:013 ex=0:89 þ 0:016ÞrS þ ð29310:74 ex=1084419:4
where s is shear stress of ECC-to-concrete interface, MPa; su is
þ 29310:8Þ ð28Þ
ultimate shear strength of ECC-to-concrete interface, MPa, which

Fig. 24. Comparisons of test shear load-slip curves with its prediction curves: (a) N-1-J; (b) N-2-J; (c) S-2-J; (d) N-2-G and (e) N-3-J.
26 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27

su ¼ ð0:0044 ex=0:99 þ 0:0092ÞrS þ ð0:76 ex=0:82 þ 0:44Þ ð29Þ Declaration of Competing Interest
where x is interface roughness values, mm; rS is uniaxial compres-
None.
sive strength of ECC, MPa; e is 2.71828; All the remaining data into
Eqs. (28) and (29) were obtained by regression of experimental
data.
The comparison shear stress-slip curves between experiment Acknowledgments
and prediction of each group of specimens are shown in Fig. 24.
It can be found that shear stress-slip curves had good consistencies This work was financially supported by the National Natural
with experimental curves before the maximum loads. After the Science Foundation of China (Grant number 51809046,
maximum loads, although there were some discrepancies between 51808562, 51878156, and 51578135), Key Program of National
the predicted curves and test curves for minority specimens (e.g., Natural Science of China (Grant number 51739008 and
specimens N-1-A-J, specimens N-1-B-J, and specimens N-2-C-G, 51527811), Open Fund of Key Laboratory of Transport Industry of
etc.), the curves variations of experiment and prediction were basi- Bridge Detection Reinforcement Technology (Chang’an University)
cally consistent for majority specimens. The main reasons for the (300102219527 and 300102219528), Start-up Funds of DGUT
large error between the predicted and the experimental results (GC300502-29 and GC300502-32), Higher Education Research Pro-
of minority specimens are the experiment error and approximate ject of DGUT (2018GJYB04 and 2017GJYB06), and Guangdong Nat-
simplified functions of theoretical model. Therefore, it can be con- ural Science Foundation (2018A030313864).
cluded that the interface shear stress-slip model proposed in this
paper can accurately simulate the shear stress-slip curves of ECC- Data availability statement
to-concrete interface.
All data included in this study are available upon request by
5. Conclusions contact with the corresponding author.

This paper experimentally investigated the influences of ECC References

mixture construction method (i.e., common casting and spraying
ECC), ECC strength grade, interface roughness of concrete sub- [1] S.J.E. Dias, J.A.O. Barros, W. Janwaen, Behavior of RC beams flexurally
strate, and PVA fiber type on the shear properties of ECC-to- strengthened with NSM CFRP laminates, Compos. Struct. 201 (2018) 363–376.
[2] D. Wang, X. Zhou, Y. Meng, et al., Durability of concrete containing fly ash and
concrete interface, and then a shear strength mechanical model silica fume against combined freezing-thawing and sulfate attack, Constr.
and a shear stress-slip model of ECC-to-concrete interface were Build. Mater. 147 (2017) 398–406.
proposed, respectively. The main conclusions are as follows. [3] M. Elghazy, A. El Refai, U. Ebead, et al., Effect of corrosion damage on the
flexural performance of RC beams strengthened with FRCM composites,
Compos. Struct. 180 (2017) 994–1006.
1) In the single-sided shear test, two failure modes of interface [4] W. Dodds, C. Goodier, C. Christodoulou, et al., Durability performance of
shear failure and ECC failure were observed. The roughness sustainable structural concrete: effect of coarse crushed concrete aggregate on
microstructure and water ingress, Constr. Build. Mater. 145 (2017) 183–195.
degree of concrete substrate had a significant influence on
[5] T. Lou, T.L. Karavasilis, Time-dependent assessment and deflection prediction
the failure mode, while ECC strength grade, ECC mixture of prestressed concrete beams with unbonded CFRP tendons, Compos. Struct.
construction method (i.e., common casting and spraying 194 (2018) 365–376.
ECC), and PVA fiber type had almost no influence on the fail- [6] J. Tian, W.W. Wang, Y.F. Du, Damage behaviors of self-compacting concrete
and prediction model under coupling effect of salt freeze-thaw and flexural
ure mode. load, Constr. Build. Mater. 119 (2016) 241–250.
2) The shear load-slip curves of all specimens were basically [7] S.F.U. Ahmed, M. Maalej, Tensile strain hardening behaviour of hybrid steel-
similar. The shear load-slip curve responses exhibited nearly polyethylene fibre reinforced cementitious composites, Constr. Build. Mater.
23 (1) (2009) 96–106.
a linear relationship before the applied loads reached to the [8] S.T. Kang, J.K. Kim, The relation between fiber orientation and tensile behavior
maximum values. After the maximum loads, the shear load in an ultra high performance fiber reinforced cementitious composites
sharply decreased with the increase of slip and the shear (UHPFRCC), Cem. Concr. Res. 41 (10) (2011) 1001–1014.
[9] D.J. Kim, S.H. Park, G.S. Ryu, et al., Comparative flexural behavior of hybrid ultra
load-slip curves presented some ductility. The ECC mixture high performance fiber reinforced concrete with different macro fibers, Constr.
construction method (i.e., common casting and spraying Build. Mater. 25 (11) (2011) 4144–4155.
ECC), ECC strength grade, roughness degree of concrete sub- [10] C. Lu, C.K.Y. Leung, Theoretical evaluation of fiber orientation and its effects on
mechanical properties in engineered cementitious composites (ECC) with
strate, and PVA fiber type had almost no influence on the various thicknesses, Cem. Concr. Res. 95 (2017) 240–246.
variation trend of shear load-slip curves. [11] Y.Z. Zheng, W.W. Wang, K.M. Mosalam, et al., Mechanical behavior of ultra-
3) Compared to the casting ECC, the spraying ECC had an obvi- high toughness cementitious composite strengthened with fiber reinforced
polymer grid, Compos. Struct. 184 (2018) 1–10.
ous decreasing effect on interface shear strength. The ECC
[12] Z.F. Zhu, W.W. Wang, K.A. Harries, et al., Uniaxial tensile stress-strain behavior
strength grade and roughness degree of concrete substrate of carbon-fiber grid-reinforced engineered cementitious composites, J.
had a significant influence on interface shear strength, and Compos. Constr. 22 (6) (2018) 04018057.
higher ECC compressive strength and coarser interface could [13] M. Sßahmaran, V.C. Li, Durability properties of micro-cracked ECC containing
high volumes fly ash, Cem. Concr. Res. 39 (11) (2009) 1033–1043.
cause higher interface shear strength. On the contrary, PVA [14] C. Lu, C.K.Y. Leung, V.C. Li, Numerical model on the stress field and multiple
fiber type had a slight influence on interface shear strength. cracking behavior of engineered cementitious composites (ECC), Constr. Build.
4) Based on the micro-bonding mechanism and test results of Mater. 133 (2017) 118–127.
[15] H. Liu, Q. Zhang, C. Gu, et al., Influence of micro-cracking on the permeability
shear load-slip curves, a shear strength mechanical model of engineered cementitious composites, Cem. Concr. Compos. 72 (2016) 104–
and a shear stress-slip model of ECC-to-concrete interface 113.
were proposed to predict interface shear strength and shear [16] K. Kobayashi, D. Le Ahn, K. Rokugo, Effects of crack properties and water-
cement ratio on the chloride proofing performance of cracked SHCC suffering
stress-slip curve, respectively. Although there was a certain from chloride attack, Cem. Concr. Compos. 69 (2016) 18–27.
deviation between the predicted values and the experimen- [17] M. Sß ahmaran, V.C. Li, Influence of microcracking on water absorption and
tal values of a minority of specimens, the predicted values of sorptivity of ECC, Mater. Struct. 42 (5) (2009) 593–603.
[18] M. Sahmaran, M. Lachemi, K.M.A. Hossain, et al., Influence of aggregate type
the majority of specimens were in good agreement with and size on ductility and mechanical properties of engineered cementitious
experimental values. composites, ACI Mater. J. 106 (3) (2009) 308.
 J. Tian et al. / Construction and Building Materials 223 (2019) 12–27 27

[19] M. Sß ahmaran, V.C. Li, Engineered cementitious composites: can composites be [40] H.E. Yucel, H. Jashami, M. Sahmaran, et al., Thin ECC overlay systems for
accepted as crack-free concrete?, Transport Res Rec. 2164 (1) (2010) 1–8. rehabilitation of rigid concrete pavements, Mag. Concr. Res. 65 (2) (2013) 108–
[20] M. Sahmaran, H.E. Yücel, S. Demirhan, et al., Combined effect of aggregate and 120.
mineral admixtures on tensile ductility of engineered cementitious [41] M.A. Carbonell Muñoz, D.K. Harris, T.M. Ahlborn, et al., Bond performance
composites, ACI Mater. J. 109 (6) (2012) 627. between ultrahigh-performance concrete and normal-strength concrete, J.
[21] M. Sßahmaran, M. Al-Emam, G. Yıldırım, et al., High-early-strength ductile Mater. Civil Eng. 26 (8) (2013) 04014031.
cementitious composites with characteristics of low early-age shrinkage for [42] L. Hou, H. Liu, S. Xu, et al., Effect of corrosion on bond behaviors of rebar
repair of infrastructures, Mater. Struct. 48 (5) (2015) 1389–1403. embedded in ultra-high toughness cementitious composite, Constr. Build.
[22] B. Suryanto, S.A. Wilson, W.J. McCarter, et al., Self-healing performance of Mater. 138 (2017) 141–150.
engineered cementitious composites under natural environmental exposure, [43] M.D. Lepech, V.C. Li, Application of ECC for bridge deck link slabs, Mater. Struct.
Adv. Cem. Res. 28 (4) (2015) 211–220. 42 (9) (2009) 1185–1195.
[23] Y. Zhu, Z.C. Zhang, Y. Yao, et al., Analysis of crack microstructure, self-healing [44] L. Bu, N. Zhou, C. Lu, et al., Research on the bond-performance of polyvinyl
products, and degree of self-healing in engineered cementitious composites, J. alcohol-engineered cementitious composite to concrete, J. Shandong Univ.
Mater. Civil Eng. 28 (6) (2016) 04016017. (Eng. Sci.) 42 (2) (2012) 45–51.
[24] C. Lu, V.C. Li, C.K.Y. Leung, Flaw characterization and correlation with cracking [45] B. Wang, S. Xu, F. Liu, Evaluation of tensile bonding strength between UHTCC
strength in engineered cementitious composites (ECC), Cem. Concr. Res. 107 repair materials and concrete substrate, Constr. Build. Mater. 112 (2016) 595–
(2018) 64–74. 606.
[25] M. S ß ahmaran, V.C. Li, Durability of mechanically loaded engineered [46] N. Wang, Experimental Research on Bonding Mechanical Performance
cementitious composites under highly alkaline environments, Cem. Concr. Between Ultra High Toughness Cementitious Composites (UHTCC) and
Compos. 30 (2) (2008) 72–81. Existing Concrete. Doctoral thesis, Dalian University of Technology, Dalian,
[26] M. Sahmaran, G. Yildirim, T.K. Erdem, Self-healing capability of cementitious 2011.
composites incorporating different supplementary cementitious materials, [47] Z.C. Deng, H.Q. Xue, Experiment and study on direct shear behavior for
Cem. Concr. Compos. 35 (1) (2013) 89–101. adhesion of fiber reinforced high ductility cementitious composites on existing
[27] G. Yıldırım, Ö.K. Keskin, S.B. Keskin, et al., A review of intrinsic self-healing concrete, Highw 2 (2011) 027.
capability of engineered cementitious composites: recovery of transport and [48] M. Sahmaran, H.E. Yücel, G. Yildirim, et al., Investigation of the bond between
mechanical properties, Constr. Build. Mater. 101 (2015) 10–21. concrete substrate and ECC overlays, J. Mater. Civil Eng. 26 (1) (2013) 167–174.
[28] C.K.Y. Leung, Y.N. Cheung, J. Zhang, Fatigue enhancement of concrete beam [49] B. Wang, Ultra High Toughness Cementitious Composite: Its Bond Behavior
with ECC layer, Cem. Concr. Res. 37 (5) (2007) 743–750. with Concrete and Structural Application for Flexural Strengthening of RC
[29] S.Z. Qian, V.C. Li, H. Zhang, et al., Life cycle analysis of pavement overlays made Beams. Doctoral thesis, Dalian University of Technology, Dalian, 2011.
with engineered cementitious composites, Cem. Concr. Compos. 35 (1) (2013) [50] N. Zhou, The Research on Interfacial Bond Behavior of Polyvinyl Alcohol Fiber
78–88. Cement Mortar to Concrete. Master thesis, Hunan University, Changsha, 2011.
[30] J. Qiu, E.H. Yang, Micromechanics-based investigation of fatigue deterioration [51] Z.H. Li, Q.F. Cui, K.L. Shi, Experimental research on the interface bonding
of engineered cementitious composite (ECC), Cem. Concr. Res. 95 (2017) 65– performance of ECC-RC functionally gradient concrete, Concr 11 (2016) 16–20.
74. [52] C.H. Hu, Research on the Interfacial Bonding Behavior of Existing Concrete
[31] S. Müller, V. Mechtcherine, Fatigue behaviour of strain-hardening cement- Members Repaired with SHCC. Doctoral thesis, Xi’an University of Architecture
based composites (SHCC), Cem. Concr. Res. 92 (2017) 75–83. and Technology, Xi’an, 2013.
[32] G. Yildirim, M. Sahmaran, M. Balcikanli, et al., Influence of cracking and healing [53] J.T. Yu, W.L. Xu, Y.M. Zhang, Experiment study on fracture property of ECC-
on the gas permeability of cementitious composites, Constr. Build. Mater. 85 concrete interface, J. Build. Mater. 18 (6) (2015) 958–963.
(2015) 217–226. [54] J. Zhang, V.C. Li, Monotonic and fatigue performance in bending of fiber-
[33] H. Liu, Q. Zhang, V. Li, et al., Durability study on engineered cementitious reinforced engineered cementitious composite in overlay system, Cem. Concr.
composites (ECC) under sulfate and chloride environment, Constr. Build. Res. 32 (3) (2002) 415–423.
Mater. 133 (2017) 171–181. [55] W. Liu, S. Xu, Q. Li, Flexural behaviour of UHTCC-layered concrete composite
[34] J. Nam, G. Kim, B. Lee, et al., Frost resistance of polyvinyl alcohol fiber and beam subjected to static and fatigue loads, Fatigue Fract. Eng. M. 36 (8) (2013)
polypropylene fiber reinforced cementitious composites under freeze thaw 738–749.
cycling, Compos. Part B-Eng. 90 (2016) 241–250. [56] Y.Y. Kim, H.J. Kong, V.C. Li, Design of engineered cementitious composite
[35] C. Wu, V.C. Li, Thermal-mechanical behaviors of CFRP-ECC hybrid under suitable for wet-mixture shotcreting, ACI Mater. J. 100 (6) (2003) 511–518.
elevated temperatures, Compos. Part B-Eng. 110 (2017) 255–266. [57] T. Kanda, T. Saito, N. Sakata, et al., Tensile and anti-spalling properties of direct
[36] V.C. Li, Durable overlay systems with engineered cementitious composites sprayed ECC, J. Adv. Concr. Technol. 1 (3) (2003) 269–282.
(ECC), Int. J. Restor. Build. Count 9 (3) (2003) 215–234. [58] Y. Lin, D. Lawley, L. Wotherspoon, et al., Out-of-plane testing of unreinforced
[37] Y.Z. Zheng, W.W. Wang, J.C. Brigham, Flexural behaviour of reinforced concrete masonry walls strengthened using ECC shotcrete, Struct 7 (2016) 33–42.
beams strengthened with a composite reinforcement layer: BFRP grid and ECC, [59] J. Tian, X.W. Wu, Y. Zheng, et al., Experimental study and mixed-dimensional
Constr. Build. Mater. 115 (2016) 424–437. FE analysis of T-rib GFRP plate-concrete composite bridge decks, Adv. Mater.
[38] A.P. Lampropoulos, S.A. Paschalis, O.T. Tsioulou, et al., Strengthening of Sci. Eng. (2018) 7531912.
reinforced concrete beams using ultra high performance fibre reinforced [60] F.Z. Zhao, H.G. Zhou, Q. Yuan, J.G. Ma, Study on Bonding Mechanism and
concrete (UHPFRC), Eng. Struct. 106 (2016) 370–384. Engineering Application of New and Old Concrete, China Water & Power Press,
[39] Y. Zheng, L.F. Zhang, L.P. Xia, Investigation of the behaviour of flexible and 2003.
ductile ECC link slab reinforced with FRP, Constr. Build. Mater. 166 (2018)
694–711.