Materials and Structures (2013) 46:145159
DOI 10.1617/s11527-012-9890-6
ORIGINAL ARTICLE
A comparison of engineered cementitious composites
versus normal concrete in beam-column joints
under reversed cyclic loading
Fang Yuan Jinlong Pan Zhun Xu
C. K. Y. Leung
Received: 28 November 2011 / Accepted: 11 June 2012 / Published online: 3 July 2012
RILEM 2012
Abstract Engineered cementitious composites (ECC)
are a class of high-performance fiber reinforced cemen-
titious composite with strain hardening and multiple
cracking properties. For a reinforced concrete member,
substitution of conventional concrete with ECC can
significantly improve the deformation characteristics in
terms of reinforced composite tensile or shear strength
and energy dissipation ability. In this paper, a number of
RC/ECC composite beam-column joints have been
tested under reversed cyclic loading to study the effect
of substitution of concrete with ECC in the joint zone on
the seismic behaviors of composite members. The
experimental parameters include shear reinforcement
ratio in the joint zone, axial load level on the column and
substitution of concrete with ECC or not. According to
the test results, for the specimens without shear rein-
forcement in the joint zone, substitution of concrete with
ECC in the joint zone cannot change the brittle shear
failure in the joint zone, but can significantly increase
the load capacity and ductility of the beam-column joint
F. Yuan  J. Pan (&)  Z. Xu
Key Laboratory of Concrete and Prestressed Concrete
Structures of Ministry of Education, School of Civil
Engineering, Southeast University, Nanjing 210096,
China
e-mail: jinlongp@gmail.com
C. K. Y. Leung
Department of Civil and Environmental Engineering,
Hong Kong University of Science and Technology,
Hong Kong, China
specimens, as well as the energy dissipation ability due
to high ductility and shear strength of ECC material. For
the specimens with insufficient or proper shear rein-
forcement ratio, substitution of concrete with ECC in
the joint zone can lead to failure mode change from
brittle shear failure in the joint zone to a more ductile
failure mode, i.e. flexural failure at the base of the beam,
with increased load capacity, ductility and energy
dissipation ability. Increase of axial load on column and
shear reinforcement in the joint zone have little effect on
seismic behaviors of the members when they failed by
flexural failure at the base of beam. In a word, the
substitution of concrete with ECC in the joint zone was
experimentally proved to be an effective method to
increase the seismic resistance of beam-column joint
specimens.
Keywords ECC  Ductility  Multiple cracking 
Beam-column joints  Seismic resistance
1 Introduction
For conventional reinforced concrete frame structures,
the seismic performance mostly depends on the
deformation ability of key components such as beams,
columns and their joint zones. Under earthquake
actions, these members are expected to maintain
substantial inelastic deformations without a significant
loss of load carrying capacity. Among these structural
146
components, beam-column joints are designed to
sustain vertical live or dead loads transferred from
beams and slabs, horizontal loads from earthquake
actions and wind, leading to complicated stresses in the
joint zone, as shown in Fig. 1. Evidence from recent
earthquakes showed beam-column joints with insuffi-
cient transverse steel reinforcement often failed by
brittle shear failure with x shape cracks under reversed
cyclic loading during the earthquake actions. Once
brittle shear failure in the joint zones occurs, the joint
cannot sustain any external and internal loading and
maintain integrity of the frame structure, indicating that
final failure of the frame structure is reached [1]. To
increase seismic performance of frame structures, the
joint zones are reinforced with additional transverse
steel reinforcement, which serve as confinement of the
concrete core and lead to enhancement of shear capacity
in the joint zone. Meanwhile, premature buckling of
longitudinal steel reinforcement can be avoided by the
confining effect of the transverse steel reinforcement
[2]. With additional transverse steel reinforcement,
brittle shear failure can be avoided in the joint zone with
a significant increase of structural ductility and seismic
resistance. However, on the other hand, an increased
amount of transverse steel reinforcement in the joint
zone will also bring forth two aspects of problems.
Dense shear reinforcement ratio may lead to difficulties
in placing steel bars because of space limitation, and the
compactness of concrete cannot be guaranteed, leading
to more defects in the joint zone. For reinforced
concrete structures, another intrinsic deficiency is the
brittleness of concrete especially in tension or shear.
For concrete frame structures, seismic cyclic loading
always leads to concrete spalling, bond splitting, brittle
shear failure in the joint zone. Though transverse steel
reinforcement can provide composite action with
concrete and achieve a virtually ductile deformation
behavior, the inherent brittleness of concrete cannot be
modified and the deficiency with respect to steel/
concrete interaction, interfacial bond deterioration, and
composite integrity are still challenges for conventional
reinforced concrete. Incompatible deformation between
concrete and steel reinforcement can decrease interfa-
cial slip, bond deterioration, resulting in decreased
deformation ability and load capacity of concrete
members.
To increase interfacial bond and composite integ-
rity between concrete and steel reinforcement, it is
proposed to use fiber reinforced concrete (FRC) to
Materials and Structures (2013) 46:145159
M1 N1
V1
Beam Joint zone
M2
M2
N2 N2
V2 V2
Column
V1
M1
N1
Fig. 1 External forces applied on a beam-column joint
substitute concrete in the reinforced concrete mem-
bers. FRC is a kind of high performance cementitious
material with good crack controlling properties. For
FRC material, after first crack is formed in the
material, the fibers cannot sustain the load at cracking,
so further deformation occurs under continuously
decreasing applied stress. The post-peak stress versus
crack opening indicates the energy absorption capac-
ity of the material and hence its resistance to crack
propagation. Experimental results from Shannag et al.
[3] indicated that using steel FRC to replace conven-
tional concrete in the beam-column connection zone
can significantly increase seismic behaviors of interior
beam-column joints, including higher load capacity,
larger displacement and curvature ductility, slower
stiffness degradation and higher energy dissipation.
Khuntia and Goel [4] developed a new type concrete-
encased steel member by substituting regular concrete
with FRC to eliminate the need for longitudinal and
transverse steel reinforcement in normal RC members.
The use of FRC can maintain the integrity under
earthquake actions, while the use of open web steel
member enhances the interaction between the steel
and FRC due to interlocking effect. Parra-Montesinos
et al. [5] proposed precast FRC-encased steel truss
beam-RC column connections, which relied on the use
of steel rods external to the beam for moment transfer
and diagonal steel plates for force transfer between the
truss chords and the external rods. Under earthquake
actions, it was designed that all connection members
behaved elastically, while inelastic rotations occurred
in the beam region near the anchorage location of the
external rods. Under larger horizontal displacement,
although the FRC was not reliable to provide shear
Materials and Structures (2013) 46:145159
resistance in the plastic region, it was effective in
providing confinement to the embedded truss and
increasing the beam plastic hinge rotation.
In recent years, a class of high performance fiber
reinforced cementitious composites (called engi-
neered cementitious composites ECC) with ultra
ductility, has been developed for applications in
construction industry [68]. Substitution of conven-
tional concrete with ECC strategically in concrete
frame structures may provide a method to solve the
deficiencies resulting from brittleness of concrete.
ECC and concrete have similar range of tensile
(46 MPa) and compressive strengths (3080 MPa),
while they have distinct difference in tensile defor-
mation behaviors. For conventional concrete, it fails in
a brittle manner once its tensile strength is reached.
However, for an ECC plate under uniaxial tension,
after first cracking, tensile load capacity continues to
increase with strain hardening behavior accompanied
by multiple cracks along the plate. For each individual
crack, the crack tends to open steadily up to a certain
crack width, and increasing loading will result in
formation of an additional crack. With the same
cracking mechanism, cracking of the ECC member
can reach a saturated state with small crack spacing,
which is determined by the stress transfer capacity of
the fibers in the matrix. With increasing loading, a
random single crack localizes and softening behavior
is followed. Typically, mechanical softening starts at a
tensile strain of 35 %, with a crack spacing of
36 mm and crack width of about 60 lm [9]. In
compression, ECC has a similar strength as concrete
with increased strain at the ultimate strength, resulting
in a lower elastic stiffness compared with concrete due
to lack of coarse aggregate. After the peak stress, the
compressive stress drops to 0.5 fc and is followed by
descending stress with further increasing compression
deformation [10]. Existing research has indicated that
the mechanical properties of ECC material in shear are
similar to those in tension [11]. The enhanced shear
capacity and ductility of ECC provide an alternative
way to increase the shear resistance and ductility of
reinforced concrete members.
Previous study has indicated that the combination of
ECC and steel reinforcement can lead to compatible
deformation in uniaxial tension, resulting in decreased
interfacial bond stresses and elimination of bond
splitting cracks and cover splitting [12]. ECC beams
without transverse steel reinforcement demonstrated
147
superior mechanical performance to concrete beams
with closely spaced stirrups, indicating that elimination
of shear reinforcement is feasible when concrete was
replaced by ECC [13]. Experiments on the cyclic
response of steel reinforced ECC columns [14] and
frames [15] also confirmed that the structure integrity
could be maintained better when concrete was replaced
by ECC. For reinforced concrete frame structures, it is
vital to avoid brittle shear failure in critical components
such as columns and beam-column joints. Substitution
of concrete with ultra ductile ECC in the joint zone and
the end zones of beams and columns are also expected
to obtain compatible deformation between ECC and
longitudinal steel reinforcement especially in inelastic
deformation regime, resulting in increased load capac-
ity, ductility and energy dissipation of the frame
structure. With higher shear strength, application of
ECC in the joint zones or outside can decrease the
amount of transverse steel reinforcement in these
regimes. Moreover, the confinement effect of ECC
together with that from transverse shear reinforcement
can avoid buckling of longitudinal steel reinforcement
and maintain composite integrity, and a ductile failure
can be guaranteed. Though ECC itself is unable to
recover the energy dissipation capacity under reversed
cyclic loading, the stabilizing effect from strategically
application of ECC on the longitudinal reinforcement
and damage tolerance at large deformation can consid-
erably increase the seismic performance of reinforce-
ment concrete structures.
In this paper, structural behaviors of beam-column
joint specimens with application of ECC in the joint
zone were investigated and compared with those of
conventional reinforced concrete beam-column joint
specimens. The influence of different parameters,
including transverse steel reinforcement ratio in the
joint zone, axial load level on the column, usage of
ECC in the joint zone or not, on the ultimate strength,
rigidity, and energy dissipation ability, etc., are
evaluated.
2 Experimental program
2.1 Preparation of specimens
Due to higher cost of PVA-ECC compared with
normal commercial concrete [16], application of ECC
for a whole structure is essentially uneconomic. For a
148
concrete frame structure, ECC can be utilized in some
key positions by taking advantage of strain hardening
property of ECC, and the seismic resistance of the
structure can be significantly enhanced. In this exper-
imental study, ECC is only involved in the connection
zone of beam-column joint specimens, as shown in
Fig. 2.
In this experimental study, four RC/ECC composite
beam-column joints and two reinforced concrete
beam-column joints were tested. These joint speci-
mens are all T type joints for simulating the edge
beam-column joints in the frame structures. The
specimens were all designed based on the principal
of strong component/weak joint in order to investi-
gate seismic behaviors of the joints. For the joint
specimens, the experimental parameters include appli-
cation of ECC in the joint zone or not, transverse steel
reinforcement ratio (0, 0.69 and 1.04 %) in the joint
zone, axial load level on the column, etc. Totally, four
different specimen configurations were designed in
this experimental study. Specimen S1 and S2 are
normal reinforced concrete beam-column joint with-
out stirrup and with two stirrups in the joint zone, and
specimen S3 and S4 are ECC/RC composite beam-
column joints without stirrup and with two stirrups
in the joint zone respectively. The transverse steel
reinforcement in specimen S5 is the same as specimen
S4 but with higher axial load on the column, while the
specimen S6 is an ECC/RC composite beam-column
joint with increased transverse steel reinforcement
(three stirrups) in the joint zone. Table 1 gives the
details of each specimen. For all specimens, two levels
of axial loads (350 and 525 kN, corresponding to 20
and 30 % of the load carrying capacity of the column)
Fig. 2 Part made with ECC for RC/ECC joint specimens
Materials and Structures (2013) 46:145159
were applied on the column during the joint specimens
were loaded in horizontal direction. All the specimens
have the same longitudinal and transverse steel
reinforcement in the columns and beams. The details
of steel reinforcement configuration were shown in
Fig. 3.
2.2 Material properties
ECC is a kind of fiber reinforced cementitious com-
posites with ultra ductility and the properties of multiple
cracking and strain hardening [17]. In order to evaluate
the ductility behavior of ECC used for the beam-column
joints, direct tensile tests were conducted. Figure 4
shows the tensile stressstrain curves of ECC material
used for casting the beam-column joints. The test results
indicated that the tensile strength exceeded 5 MPa and
the ultimate tensile strain approached 4 %. As for the
fiber reinforced cementitious composite, the shear
property of ECC material is similar to that in tension,
which has been experimentally proved [18, 19]. For the
compressive strength (fcu) of ECC and concrete used in
the area except the joint zone, a number of cubic
specimens with dimension of 150 9 150 9 150 mm
were tested in compression. The compressive strength
of ECC and concrete are 49.6 and 52.4 MPa respec-
tively, and the modulus of elasticity (Ec) of ECC and
concrete are 34.49 and 18.50 GPa respectively. For the
steel reinforcement, tension tests were carried out to
determine the material properties. Table 2 shows the
measured average yield strength (fy), tensile strength
(fu) and modulus of elasticity (Es) for the steel
reinforcement.
2.3 Loading configuration
To investigate the seismic behaviors of beam-column
joints with different configurations, the tests were
designed as shown in Fig. 5. For a beam-column joint
specimen, the column was horizontally and simply
supported on the ground with the left end of the
column leaned against the rigid reaction wall. A
hydraulic jack was installed between the other end of
the column and a steel frame anchored on the ground.
To avoid significant displacement of the steel frame,
two steel strands with high strength were tensioned
and fixed on the reaction wall and the steel frame. For
each specimen, the axial load can be applied on the
column with the hydraulic jack. For the beam-column
Materials and Structures (2013) 46:145159 149

Table 1 Summary of specimen information

Specimen Composite qatransverse Axial load (kN) Yielding load Ultimate load Failure load Ductility
coefficient
(%) Py Dy Pmax Dmax Pu Du
(kN) (mm) (kN) (mm) (kN) (mm)

S1 RC 0 350 75.3 16.2 102.4 37.8 87.0 47.9 2.96

S2 R/C 0.69 350 80.2 15.3 107.0 47.3 90.9 64.8 4.25
S3 RC/ECC 0 350 83.1 13.5 119.2 26.9 101.9 64.2 4.76
S4 RC/ECC 0.69 350 99.7 12.7 128.5 26.7 109.2 83.8 6.26
S5 RC/ECC 0.69 525 97.6 11.9 125.6 19.98 106.7 84.5 7.10
S6 RC/ECC 1.04 350 96.5 13.3 119.7 26.6 101.7 74.6 5.63
a
The transverse reinforcement ratio of joints

Fig. 3 Details of the test

joints (unit mm) (a) for
Specimens S-1, S-3 (b) for
Specimens S-2, S-4, S-5, S-6

6
joint specimen, to apply horizontal load on the end of
Tensile stress (MPa)

5
the beam, a hydraulic actuator was horizontally
4 installed between the reaction wall and the end of
3 the beam. The whole loading system is shown in
2
Fig. 5. For each specimen, the loading history
included elastic and inelastic cycles. The elastic cycles
1
were conducted under load control at load levels of
0
0 1 2 3 4 5
0.25, 0.5 and 0.75 Py, where Py is the estimated lateral
Strain (%) yielding load corresponding to the lateral yielding
displacement Dy. The load was increased at intervals
Fig. 4 Tensile stressstrain relationship of ECC of 0.05 Py when the specimen is approaching yield
150 Materials and Structures (2013) 46:145159

Table 2 Material properties of steel reinforcement

Steel type Diameter Yield Ultimate Modulus of
strength strength elasticity
(mm) fy fu (MPa) Es (GPa)
(MPa)

Mild stirrup 6 407.5 454.8 181

Deformed 20 359.4 541.6 187
longitudinal
bar

Fig. 6 Cyclic loading history for each specimen

strains from the strain gauges were automatically

collected by a data logger.

3 Results and discussion

3.1 Failure characteristics and crack patterns

Fig. 5 Test setup for all specimens
strength. After yielding of the specimen occurred,
inelastic cycles were conducted under displacement
control at displacement levels of Dy, 2Dy, 3Dy, 4Dy,
5Dy and so on. Three cycles were imposed at each
inelastic displacement level described above. The
loading history is shown in Fig. 6. For each specimen,
the test was terminated when the residual load capacity
of the specimen decreased to 85 % of the peak load
capacity.
During the loading process, a displacement trans-
ducer was installed to obtain the displacement at the
top of beam. The other two displacement transducers
(LVDT) were installed to measure the shear deforma-
tion of the joint zone, as shown in Fig. 3. To measure
the strain variation of the steel reinforcement, a
number of strain gauges were attached on the longi-
tudinal steel bars of the beam at a space of 80 mm
within the joint zone and near the beam end, and two
strain gauges were used at each side of the stirrup in
the joint zone for the specimens with stirrups in the
joint zone. The displacement from the LVDTs and
For specimen S1, which is a normal reinforced
concrete beam-column joint specimen without stirrups
in the joint zone, a number of flexural cracks appeared
in the height of 800 mm from the base of the beam
before yielding of longitudinal steel reinforcement
was reached. The cracks spacing was approximated to
be 100 mm, and the cracks extended to near the center
line of the beam. Some of the cracks formed connected
cracks through the thickness of the beam. Diagonal
cracks were observed in the joint zone. For specimen
S1, steel yielding occurred at the displacement of
16.2 mm corresponding to a yield load of 75.3 kN.
Beyond yielding, the cracks in the beam region kept
constant while the diagonal cracks in joint zone
became wider and wider. The maximum crack width
observed in joint zone increased to 5 mm at the
displacement of 2Dy. Ultimate load capacity was
obtained at the displacement of 37.8 mm with a lateral
load of 102.4 kN. Longitudinal splitting cracks and
concrete spalling were observed at the displacement of
3Dy and the residual strength decline to 85 % of the
ultimate strength, indicating final failure of the
specimen was reached. The specimen S1 finally failed
Materials and Structures (2013) 46:145159
by brittle shear failure of concrete in joint zone, final
crack pattern of specimen S1 is shown in Fig. 7.
For specimen S2 is a normal RC beam-column joint
with two stirrups in the joint zone. Before yielding
occurred, a number of flexural cracks occurred in the
height of 850 mm from the base of the beam. The
opening and spacing of the cracks in the beam were
similar to S1 while the crack width within the joint
zone was much smaller than S1 due to addition of steel
stirrups. With increasing external loading, yielding
occurred at the displacement of 15.3 mm with corre-
sponding load of 80.2 kN. Beyond yielding, several
cracks propagated towards the centerline of the beam
with angles of 45, and connected with the cracks from
the other side, forming intersectional crack patterns in
the joint zone. With further increase of the external
load, the crack arising from the base of the beam
increased to 7 mm at the displacement of 2Dy, and
more intersectional shear cracks occurred in the joint
zone. When the displacement reached 3Dy, localiza-
tion of cracks at the height of 100 mm and 150 mm
from the base of the beam occurred and connected
with the cracks at the height of 300 and 450 mm from
the other side. For these localized cracks, the crack
width approached 5 mm. With increasing displace-
ment, more shear cracks occurred in the joint zone and
localized cracks tended to open significantly to 9 mm.
Meanwhile, splitting of concrete occurred at the base
of the beam and within the joint zone. Finally,
specimen S2 failed at a displacement of 64.8 mm
Fig. 7 Crack patterns of specimens after failure
151
(5Dy)with a hybrid failure mode from flexural yielding
at the base of the beam to shear crushing of concrete in
the joint zone, and final crack pattern of specimen S2
was shown in Fig. 7.
For specimen S3, which is a RC/ECC composite
beam-column joint without stirrups in the joint zone,
the initial tiny crack occurred at a load of 40 kN at the
interface between concrete and ECC. Prior to yielding,
some tiny cracks that occurred in the ECC zone of the
beam, and extended to approximately 40 mm far from
the tension side, while larger flexural cracks formed in
the beam of concrete part (4001,000 mm from the
base of the beam). In this stage, no cracks appeared in
the joint zone. With increasing external loading, steel
yielding occurred at the displacement of 13.5 mm with
corresponding load of 83.1 kN. For the beam of
concrete part, the flexural cracks extended from two
sides of the beam and formed intersectional cracks
due to reversed horizontal cyclic loading. Meanwhile,
multiple tiny cracks occurred in the ECC zone around
the base of the beam. When the displacement reached
2Dy, a major flexural crack appeared at the base of the
beam, and some secondary cracks occurred in the joint
zone. When the displacement reached 4Dy, the crack at
the base of the beam increased to 10 mm, and more
tiny cracks occurred and formed a few intersectional
shear cracks in the ECC joint zone. When the
displacement reached 5Dy, a major shear crack in the
ECC joint zone suddenly increased to 15 mm, indi-
cating ultimate stage of the beam-column joint
152
specimen was reached. The corresponding ultimate
load capacity was 119.2 kN. With increasing dis-
placement, the external load decreased with the
displacement, and final failure occurred at the dis-
placement of 6Dy with corresponding failure load of
101.9 kN. The crack pattern after failure is shown in
Fig. 7. Compared with specimen S1 and S2, specimen
S3 showed much better ductility and higher ultimate
load capacity.
Specimen S4 is a RC/ECC composite beam-column
joint specimen with two stirrups in the joint zone. Prior
to yielding of steel reinforcement, a number of tiny
flexural cracks occurred within the height of 850 mm
from the base of the beam. The extension of the cracks
was around 50 mm. In this stage, no cracks occurred in
the joint zone. With increasing external loading,
specimen S4 reached yielding of steel reinforcement
at the load value of 99.7 kN with corresponding
displacement of 12.7 mm. After yielding, the joint
specimen is loaded by displacement control. When the
displacement reached 2Dy, a major flexural crack
formed at the base of the beam, and three groups of
connected shear cracks formed along the beam. In this
stage, two small shear cracks occurred within the joint
zone, but they didnt seem to open significantly. When
the displacement reached 3Dy, the crack at the base of
the beam continued to open and reached about 8 mm,
but the crack could not extend further along the depth
of the beam and multiple tiny cracks formed near the
crack tip. During increase of displacement from 4 to
6Dy, the beam tended to slide along the cracked
section at the base of the beam, and further displace-
ment loading had no effect on the cracks along the
beam and within the joint zone. Final failure of
specimen S4 was caused by the tensile rupture of
longitudinal reinforcement at the displacement of 7Dy
(Fig. 7) with corresponding load of 109.2 kN. Com-
pared with specimen S2, specimen S4 failed by full
development of plastic hinge at the end of the beam,
but not the shear failure in the joint zone.
Specimen S5 is a RC/ECC composite beam-column
joint with two stirrups in the joint zone and constant
axial load of 525 kN on the column during the loading
process. For specimen S5, the deformation and cracking
behaviors were very similar to that of specimen S4,
while yielding of the specimen occurred at the load of
97.6 kN with corresponding displacement of 11.9 mm,
which was smaller than that of specimen S4. After
yielding, the flexural cracks concentrated near the base
Materials and Structures (2013) 46:145159
of the beam where yielding of steel reinforcement
occurred. Shear sliding occurred along the section of the
beam base at the displacement of 7Dy, and steel rupture
occurred at the same section at the displacement of 7Dy
due to reversed horizontal loading. The failure load of
specimen S5 was 106.7 kN with corresponding dis-
placement of 84.5 mm. The failure mode of specimen
S5 was the same as that of specimen S4, i.e. fully
development of plastic hinge at the base of the beam for
joint specimen. The crack pattern of specimen S5 is
shown in Fig. 7.
Specimen S6 is a RC/ECC composite beam-column
joint with three stirrups in the joint zone and a constant
axial load of 350 kN on the column. The deformation
and cracking behaviors of S6 were definitely similar
to S4 and S5, while yielding of S6 occurred at the
load of 96.5 kN with corresponding displacement of
13.3 mm. With increasing displacement loading,
specimen S6 finally failed by rupture of reinforcement
due to fully development of the plastic hinge at the
base of the beam (Fig. 7). The failure load and
displacement were 101.7 kN and 74.6 mm respec-
tively. The detailed test results for each specimen can
be obtained in Table 1.
3.2 Lateral load (P) versus lateral displacement
(D) curves
Figure 8 shows the curves of lateral load (P) versus
lateral displacement (D) for each specimen. For
specimen S1, the shear force in the joint zone was
only undertaken by concrete and longitudinal steel
reinforcement, leading to premature cracking of
concrete under shear and compression stresses. When
the lateral displacement reached 2Dy, shear cracks
opened significantly and spalling of concrete occurred,
indicating final failure was reached. For specimen S1,
pinching effect of hysteresis loops was apparent as
shown in Fig. 8a, indicating a brittle failure charac-
teristic of specimen S1. Compared with S1, the
hysteresis loops of S2 were relatively full and stable,
and no apparent pinch effect was observed. This is due
to the fact that application of steel stirrups in the
connection zone enhanced the resistance to shear force
and provided confinement of concrete core, resulting
in higher compressive strength of concrete.
ECC is a kind of composite material with superior
high ductility and damage tolerance, and deforms
compatibly with steel reinforcement due to the same
Materials and Structures (2013) 46:145159 153

Fig. 8 Cyclic load 150 150

(P) versus lateral
100 100
displacement (D) curves
(a) for S1 (b) for S2 (c) for 50 50

P(kN)

P(kN)
S3 (d) for S4 (e) for S5
0 0
(f) for S6
-50 -50

-100 -100

-150 -150
-100 -50 0 50 100 -100 -50 0 50 100
(mm) (mm)
(a) (b)
150 150

100 100

50 50
P(kN)

P(kN)
0 0

-50 -50

-100 -100

-150 -150
-100 -50 0 50 100 -150 -100 -50 0 50 100 150
(mm) (mm)
(c) (d)
150 150

100 100

50 50
P(kN)

P(kN)

0 0

-50 -50

-100 -100

-150 -150
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
(mm) (mm)
(e) (f)

tensile properties. For steel reinforced ECC member,
bond splitting or ECC spalling can be avoided under
external loading, which can also be observed from the
test results of specimen S3. Meanwhile, ECC had
much better shear strength than normal concrete with
the same compressive strength [8]. For the specimens
without stirrups in the joint zone, specimen S3 showed
much higher load capacity and ductility than specimen
S1, although they failed in the same failure mode, i.e.
brittle shear failure in the joint zone. According to
Fig. 8, the area within the hysteresis loops of S3 was
much higher than that of S1 and S2, indicating that
substitution of ECC in the joint zone can significantly
increase the energy dissipation ability under reversed
cyclic loading. Even compared with specimen S2
which had two stirrups in the joint zone, the ultimate
load capacity of specimen S3 is 11.4 % higher than
that of S2, which means that application of ECC in the
joint zone can improve the shear strength significantly
and even can substitute the steel stirrups with the same
structural performance.
Compared with specimen S2, the ultimate load
capacity of specimen S4 is 16.7 % higher than that of S2
due to the dual enhancement from ECC and stirrups in
the joint zone. For specimen S2 and S4, the failure mode
also transferred from the shear failure in the joint zone
to flexural failure at the base of the beam with full
development of plastic hinge. For specimen S5 and S6,
increase of the axial load in the column and addition of
three stirrups within the joint zone seemed to have little
154
Fig. 9 P-D envelope 150
curves of the specimens
(a) for specimens without 100
stirrups in the joint (b) for
50
Load(kN)
specimens with stirrups in
the joint
0 0
-50
-100
-150
-100 -50
Displacement(mm)
(a)
effect on the seismic behavior of specimens, which may
be due to that they both failed by flexural failure at the
base of the beam. The hysteresis loops of specimens S-5
and S-6 were generally similar to those of specimen S3
and S4.
3.3 P-D envelope curves
Figure 9a shows the P-D envelope curves of each
specimen without stirrup. The curves of RC specimen
S1 tend to drop more quickly due to premature failure
of the concrete in joint. The curves of RC/ECC
composite specimen S3 drops more gently and the area
encompassed by the curve is much larger than S1. This
is attributed to the fact that substitution of concrete
with ECC improves the shear strength of the joint,
resulting in much higher load capacity and ductility
even S3 final failed by shear failure in the joint zone.
For the specimens with stirrups in the joint zone, the
P-D envelope curves are shown in Fig. 9b. The curves
of RC/ECC composite specimens (S4, S5 and S6) were
more stable than RC specimen S2 especially in inelastic
deformation regime. According to the test results,
specimen S2 failed from flexural failure at the base of
the beam to shear failure of concrete in the joint zone,
while specimen S4 to S6 all changed to flexural failure
at the base of the beam due to enhancement of shear
strength in the joint zone with the application of ECC.
The application of ECC in the joint can increase the
ultimate load capacity and displacement of the speci-
men. Meanwhile, with the same failure mode of flexural
failure at the base of the beam, further increase of shear
reinforcement in the joint zone or increase of axial load
in the column can not improve the ultimate load and
displacement of the joint specimen, which can be
proved from Fig. 9b.
Materials and Structures (2013) 46:145159
150
100
50
Load(kN)
S-2
S-4
-50
S-1 S-5
-100
S-6
S-3
-150
0 50 100 -150 -100 -50 0 50 100 150
Displacement(mm)
(b)
3.4 Strain analysis
For each specimen, the strain variations in the stirrups
have been collected during the loading process.
Figure 10 shows the maximum strains of the stirrups
in the joint zone at the yield load and peak load. When
the specimens reached yielding load, the maximum
strain in the stirrups of specimen S2 is 1,243 le, which
is much larger than that of the other three RC/ECC
composite joint specimens due to wide opening of
shear cracks in RC joint. In the ultimate stage, the
strain in the stirrups of specimen S2 is 2,037 le, which
is beyond the yielding strain (1,800 le) and is more
than two times of the other three specimens. It is
attributed to the fact that the crack width in joint of RC
specimen (S2) is much larger than that of specimens
S4, S5 and S6, in which the fibers have bridged and
restrained the cracks.
To analyze the strains distributions along the
longitudinal reinforcement, the results of specimen
S2 and S4 are used for examples. Figure 11 shows the
strain distributions along the longitudinal reinforce-
ment of specimen S2 and S4 at different load levels.
For specimen S2, the premature flexural cracks
occurred along the RC beam, which led to the
fluctuated distribution of strains along the longitudinal
reinforcement in the beam. In contrast, for specimen
S4, the strains along the longitudinal reinforcement in
the beam distribute uniformly until the load increased
to 100 kN due to opening of multiple and tiny cracks
along the beam. In this stage, the longitudinal steel
reinforcement had compatible deformation with ECC
material and showed good bond with ECC. For each
load value, the strains along the longitudinal rein-
forcement in specimen S4 are much smaller than those
in specimen S2 due to strain hardening of ECC and
Materials and Structures (2013) 46:145159 155

1600 2400
2037
1243
1200 1800

()
1205

()
800 620 1200
585 882 870

400 600
166

0 0
S-2 S-4 S-5 S-6 S-2 S-4 S-5 S-6
(a) (b)
Fig. 10 Strain values for each specimen (a) at the yielding load (b) at the peak load

Fig. 11 The strain distributions along the longitudinal steel reinforcement (a) for specimen S2 (b) for specimen S4

compatible deformation between ECC and longitudi-
nal steel reinforcement.
3.5 Deformation analysis
In this experimental program, the total displacement at
the top of the beam can be divided into the flexural and
shear deformations of the beam and column (Dbc) and
the shear deformation of the joint core (Dpz).The
control specimens designed in the tests were based on
the principal of strong component and weak joint,
and hence the failure mechanism in the joint zone can
be analyzed. The diagonal shear deformation in the
joint zone was measured by two displacement trans-
ducers located at the joint core as shown in Fig. 3. The
mean value of the deformation along the diagonal lines
in joint can be calculated as:
    in the joint zone. The shear deformation ability was
d1 d0  d2 d0 
1 2
X 1
2
where d1, d10 , d2 and d20 are the deformations along the
diagonal lines in the joint zone, as shown in Fig. 12.
Then the absolute value of the shear drift angle can
be determined as follows:
p
a2 b2
jcshear j a1 a2 X 2
2ab
where a1 and a2 are the shear drift angle corresponding
to the directions along joint height and width respec-
tively, h and b are the height and width of the joint
core, as shown in Fig. 12.
Figure 13 shows the shear drift angle in the joint
zone of specimens at different load levels. For S1, the
shear drift angle came up to 0.017 rad at the
displacement of 3Dy, and shear failure occurred
subsequently due to absence of steel stirrups in the
joint zone. For specimen S2, the longitudinal steel
reinforcement at the base of the beam first reached
yielding with increasing external loading, and final
failure also occurred by excessive shear deformation
much better than that of S1 due to increase of shear
reinforcement in the joint zone. Specimen S3 is a RC/
ECC composite joint specimen without stirrups in the
joint zone, which has the same failure mode and shear
156
Fig. 12 Shear deformation in the joint core
0.04
0.03
/rad
0.02
0.01
0
0 2 4 6 8
/ y
Fig. 13 Shear deformation angle in different stages
deformation as specimen S2 due to high shear strength
and ductility of ECC. Specimen S4, S5 and S6 all
failed by flexural failure at the base of the beam, and
the shear drift angle are much smaller than 0.006 rad,
indicating that the ECC/RC joint zones keep in elastic
stage up to final failure.
For each specimen, image analysis was used to obtain
the displacement distribution of the joint zone. Before
testing, a grid with a spacing of 20 mm was drawn on the
surface of the specimen to assess the deformation of
specimens. Images of the specimens were taken by a
digital camera at different load levels. These images
were analyzed by software to obtain the coordinates of
intersectional dots of the grid painted on the specimen
surface. The deflection curves for each specimen were
shown in Fig. 14. In Fig. 14, position 0 of the vertical
axis denotes the intersection of the column and beam,
and negative values denote the beam-column connection
zone. For the beam part of the joint specimen, the
deflection curves were fitted by a polynomial expression
from the point displacement values from image analysis.
Materials and Structures (2013) 46:145159
With the fitted deflection curves, the flexural curvature
of the beam can also be obtained. At the displacement of
3Dy, the maximum curvature and plastic hinge length
are 0.004 l and 320 mm respectively. However, the
curvatures along the beam were very small, which meant
that specimen S1 had failed in the joint zone before
flexural hinge could occur in the beam part. For
specimen S2, the plastic hinge length is 460 mm with
180 mm in the beam part, indicating that plastic hinges
occurred at both the joint zone and the base of the beam.
For Specimen S2, the beam rotated to some extent
before the shear failure in the joint zone occurred. For
specimen S3, the maximum curvature (0.00351/mm)
was smaller than S1, while the plastic hinge length
(410 mm) was larger than S1 at the displacement of 3Dy,
indicating that the plastic hinge zone located both at the
base of the beam and in the joint zone. For specimens S4,
S5 and S6, the curvature in the joint zone was relatively
small, and the plastic hinges developed sufficiently at the
base of the beam, indicating flexural failure of the beam
for these joint specimens. In the ultimate stage, the final
deflection at the base of the beam reached 18.1, 12.4 and
11.1 mm for S4, S5 and S6.
3.6 Ductility and energy dissipation
The ductility coefficient (l) is an important parameter
for evaluating the ductility performance of beam-
column joints. The ductility coefficient (l) is defined
asDu/Dy, where Dy is the lateral displacement at yield
load and Du is the displacement when the applied load
declines to 85 % of the maximum load. The value of l
for each specimen is listed in Table 1. For specimens
without stirrups in the joint zone, the ductility
coefficient of S3 is 1.61 times of that of S1, which is
due to substitution of concrete with ECC in the joint
zone. For the specimens with stirrups in the joint zone,
the ductility coefficients of S4, S5 and S6 are 1.47,
1.67 and 1.32 times of that of S2. The enhanced shear
strength and confinement effect of ECC are respon-
sible for the improvement of ductility coefficient of
RC/ECC composite joint specimens. The ductility
coefficient of specimen S5 is larger than S4, which is
due to the fact that improvement of axial load on the
column contributed to restrain propagation of cracks
in the joint zone and near the base of the beam.
Redundant amount of stirrup may lead to difficulties in
arranging transversal reinforcement and proper place-
ment of concrete in the joint zone. That is why the
Materials and Structures (2013) 46:145159 157

Fig. 14 Displacement 800 800 800

distributions of each 600 600 600

Position(mm)

Position(mm)

Position(mm)
specimen (a) for S1 (b) for 400 400 400
S2 (c) for S3 (d) for S4 200 200 200
(e) for S4 (f) for S5 0
1y
0
1y
0
1y
2y 3y 3y
-200 3y -200 4y -200 4y
-400 -400 -400
-20 0 20 40 -20 0 20 40 -20 0 20 40
Displacement(mm) Displacement(mm) Displacement(mm)
(a) (b) (c)
800 800 800
600 600 600
Position(mm)

Position(mm)

Position(mm)
400 400 400
200 200 200
1y 1y 1y
0 0 0
2y 2y 2y
-200 6y -200 6y -200 6y
-400 -400 -400
0 20 40 60 80 0 20 40 60 80 0 20 40 60 80
Displacement(mm) Displacement(mm) Displacement(mm)
(d) (e) (f)

ductility coefficient of specimen S6 is smaller than that

of specimen S4.
Equivalent damping coefficient (neq) [20] is another
important parameter for evaluating the energy dissi-
pation capacity of beam-column joint specimens.
Equivalent damping coefficient (neq) can be calculated
according to the hysteresis loops in Fig. 15, and can be
expressed as:
1 SABC SCDA
ne q
2p SOBE SODF
where SABC and SCDA are the areas enclosed by the
curves ABC and CDA respectively, which denote the
inelastic dissipating energy in one complete hysteresis
loop. Similar definitions were used for SOBE and SODF
which denote the inelastic strain energy at a given
displacement amplitude.
For each specimen, the cumulative energy dissipa-
tion is defined as the sum of the areas of each hysteresis
loop before the considered load level or displacement
step. Figures 16 and 17 show the equivalent damping
coefficient neq-D/Dy curves and cumulative energy
dissipation -D/Dy curves for each specimen. Speci-
men S1 and S3 are the beam-column joint specimens
without stirrups in the joint zone. Specimen S1 reached
the ultimate load at displacement of 3Dy, and showed
little energy dissipation capacity beyond ultimate load
due to brittle shear failure in the joint zone, while S3
showed steadily energy dissipation capacity after
ultimate load and failed at the displacement of
3
Fig. 15 Hysteresis loop and energy dissipation
64.2 mm, resulting from strain hardening property of
ECC in the joint zone. The cumulative energy
dissipation of S3 is 3.7 times of that of S1.
For specimens with stirrups in the joint zone,
specimen S2 showed the same equivalent damping
and cumulative dissipated energy with S4 when the
displacement was smaller than 4Dy. After that, spall-
ing of concrete occurred in the joint zone and the
energy dissipation ability of S2 decreased sharply with
further loading. While, specimen S4 showed steadily
energy dissipation capacity until lateral displacement
reached 7Dy (83.8 mm), which is attributed to superior
ductile behavior of ECC and the abundant inelastic
deformation of steel reinforcement along with ECC.
During the loading process, no spalling of ECC and no
buckling of longitudinal steel reinforcement occurred
158 Materials and Structures (2013) 46:145159

0.3 on the seismic behavior of the beam-column joint

Equivalent damping coefficient specimens. For the specimens without stirrups in the
0.25
joint zone, addition of ECC in the joint zone can
0.2 significantly increase the load capacity and ductility of
0.15
the beam-column joint specimens, as well as the
energy dissipation due to high ductility and shear
0.1 strength of ECC material. For the specimens with
S-1 S-2
0.05 reduced or proper shear reinforcement in the joint
S-3 S-4
zone, replacement of concrete with ECC in the joint
S-5 S-6
0 zone can lead to failure mode change from brittle shear
0 2 4 6 8
/ y failure in the joint zone to flexural failure due to
yielding of longitudinal steel reinforcement at the base
Fig. 16 Equivalent damping coefficient neq-D/Dy curves of of the beam. The RC/ECC composite beam-column
each specimen
joint showed higher load capacity, ductility and energy
dissipation when compared with normal RC beam-
6 column joint specimen. Increase of the axial load on
S-1 S-2 the column cannot increase the ultimate load capacity
5
S-3 S-4 and ductility since they all failed by flexural failure at
S-5 S-6 the base of the beam, but can result in increased
4
ductility coefficient because the additional axial load
3 can restrain propagation of cracks in the joint spec-
imen. Experimental results showed that increase of
2 shear reinforcement in the joint zone may lead to
difficulty in concrete casting, and the ultimate load
1 capacity and ductility showed a small decrease with
increasing the shear reinforcement ratio. In a word,
0
0 2 4 6 8 substitution of concrete with ECC in the joint zone can
significantly increase the seismic performance of
beam-column joints even with decreased shear rein-
Fig. 17 Cumulative energy dissipation -D/Dy curves of each forcement in the joint zone compared with proper
specimen
designed RC members.

in the joint zone. It is found that the cumulative Acknowledgments Financial support of the work by National
Natural Science Foundation of China under 50808043, by the
dissipated energy of specimen S5 is larger than
National Basic Research Program of China (973 Program)
specimen S4, indicating that higher axial load on the under 2009CB623200 and the Priority Academic Program
column is beneficial to prevent propagation of cracks Development of Jiangsu Higher Education Institutions, is
in the joint zone and can lead to higher energy gratefully acknowledged.
dissipation ability. For specimen S6, which had three
stirrups in the joint zone, the energy dissipation
capacity is little smaller than that of S4, which may be References
caused by the defects in the joint zone when concrete
casting was conducted. 1. Ghobarah A, Said A (2002) Shear strengthening of beam-
column joints. Eng Struct 24(7):881888
2. Waston S, Zahn FA, Park R (1994) Confining reinforcement
of concrete columns. ACI J Struct Eng 120(6):17981823
4 Conclusions 3. Shannag MJ, Abu-Dyya N, Abu-Farsakh G (2005) Lateral
load response of high performance fiber reinforced concrete
beam-column joints. Constr Build Mater 19(7):500508
In the present paper, a number of beam-column joint 4. Khuntia M, Goel SC (1998) FRC encased steel joist com-
specimens with different configurations have been posite beams under reversed cyclic loading. ASCE J Struct
tested to investigate the effect of ECC in the joint zone Eng 124(10):11151124
Materials and Structures (2013) 46:145159
5. Parra-Montesinos G, Dasgupta P, Goel SC (2005) Devel-
opment of connections between hybrid steel truss-FRC
beams and RC columns for precast earthquake-resistant
framed construction. Eng Struct 27(13):19311941
6. Kim YY, Fischer G, Li VC (2004) Performance of bridge
deck link slabs designed with ductile ECC. ACI Struct J
101(6):792801
7. Lepech MD, Li VC (2009) Application of ECC for bridge
deck link slabs. RILEM J Mater Struct 42(9):11851195
8. Lepech MD, Li VC (2010) Sustainable pavement overlays
using engineered cementitious composites. J Pavement Res
Technol 3(5):241250
9. Zhang J, Leung CKY, Gao Y (2009) Simulation of crack
propagation of fiber reinforced cementitious composite
under direct tension. Eng Fract Mech 78(12):24392454
10. Li VC, Mishra DK, Wu HC (1995) Matrix design for pseudo
strain-hardening fiber reinforced cementitious composites.
RILEM J Mater Struct 28(183):586595
11. Li VC, Mishra DK (1996) Structural applications of engi-
neered cementitious composites. Indian Concr J 70(10):
561574
12. Fischer G, Li VC (2002) Influence of matrix ductility on
tension-stiffening behavior of steel reinforced engineered
cementitious composites. ACI Struct J 99(1):104111
159
13. Li VC, Wang S (2002) Flexural behaviors of glass fiber-
reinforced polymer (GFRP) reinforced engineered cemen-
titious composite beams. ACI Mater J 99(1):1121
14. Fisher G, Li VC (2002) Effect of matrix ductility on
deformation behavior of steel reinforced ECC flexural
members under reversed cyclic loading condition. ACI
Struct J 99(6):781790
15. Fisher G, Li VC (2003) Intrinsic response control of moment
resisting frames utilizing advanced composite materials and
structural elements. ACI Struct J 100(2):166176
16. Cheung YN (2004) Investigation of concrete components
with a pseudo-ductile layer. Doctoral Dissertation. Hong
Kong University of Science and Technology
17. Li VC (2003) On engineering cementitious composites
(ECC). J Adv Concr Technol 1(3):215229
18. Fukuyama H, Matsuzaki Y, Nakano K, Sato Y (1999)
Structural performance of beam elements with PVA-ECC.
Proceedings HPFRCC-3, pp 531542
19. Kanda T, Watanabe S, Li VC (1998) Application of pseudo
strain hardening cementitious composites to shear resistant
structural elements. Proceedings FRAMCOS-3, pp 14771490
20. Tang JR (1989) Earthquake resistance design of reinforced
concrete frame connection. Nanjing: Southeast University
Press [in Chinese]