polymers

Article

Monotonic and Cyclic Bond Behavior of Deformed

CFRP Bars in High Strength Concrete
T. Tibet Akbas 1, *, Oguz C. Celik 2 , Cem Yalcin 3 and Alper Ilki 4
1
2
3
4

Institute of Science and Technologie, Istanbul Technical University, 34469 Istanbul, Turkey
Department of Architecture, Istanbul Technical University, 34437 Istanbul, Turkey; celikoguz@itu.edu.tr
Department of Civil Engineering, Bogazici University, 34342 Istanbul, Turkey; yalcince@boun.edu.tr
Department of Civil Engineering, Istanbul Technical University, 34469 Istanbul, Turkey; ailki@itu.edu.tr
Correspondence: tunc.akbas@gmail.com; Tel.: +90-212-318-6356

Academic Editor: Jorge de Brito

Received: 31 March 2016; Accepted: 19 May 2016; Published: 31 May 2016

Abstract: Composite reinforcing bars (rebars) that are used in concrete members with high
performance (strength and durability) properties could have beneficial effects on the behavior of these
members. This is especially vital when a building is constructed in an aggressive environment, for
instance a corrosive environment. Although tension capacity/weight (or volume) ratios in composite
rebars (carbon fiber reinforced polymer (CFRP), glass fiber reinforced polymer (GFRP), etc.) are very
high when compared to steel rebars, major weaknesses in concrete members reinforced with these
composite rebars may be the potential consequences of relatively poor bonding capacity. This may
even be more crucial when the member is subjected to cyclic loading. Although monotonic bond tests
are available in the literature, only limited experimental studies exist on bond characteristics under
cyclic loading conditions. In order to fill this gap and propose preliminary design recommendations,
10 specimens of 10-mm-diameter ribbed CFRP rebars embedded in specially designed high strength
concrete (f c = 70 MPa) blocks were subjected to monotonic and cyclic pullout tests. The experimental
results showed that cyclically loaded CFRP rebars had less bond strength than those companion
specimens loaded monotonically.
Keywords: bond strength; CFRP; cyclic; pullout; high strength concrete

1. Introduction
Reinforcements made of carbon fiber are widely used as retrofitting materials to increase the
capacity of existing buildings [1]. Recent studies have shown the usage of carbon fiber reinforcing
bars in concrete members effectively as an alternative material to steel rebars. Thus, fiber reinforced
polymer (FRP) reinforced concrete members such as beams, columns and slabs could be designed and
constructed. Carbon rebars are mostly preferred in harsh environments due to their chemical resistivity,
thermal conductivity, electrical conductivity, dimensional stability and durability [2,3]. High strength
and light weight characteristics are other remarkable features of these rebars. Some construction
applications with carbon rebars in marine structures, bridge concrete slabs and concrete pavements
have already proved their efficiency and suitability under corrosive exposure conditions. Research in
this field has also gained some momentum in terms of assessing their potential in the future.
Recently conducted experimental studies by various researchers showed similar behavioral
properties of carbon rebars in concrete sections when compared to steel rebars. These studies also
revealed that required embedment lengths of these rebars should be taken into consideration correctly
and carefully. Unlike steel rebars, less information is available in the literature on the bond behavior of
carbon rebars in normal and high strength concrete members, particularly for cyclic loading conditions.

Polymers 2016, 8, 211; doi:10.3390/polym8060211

www.mdpi.com/journal/polymers

Polymers 2016, 8, 211

2 of 11

In the past decades researchers have carried out numerous experimental studies and proposed
bond models to describe the behavior of steel rebars in terms of their embedment and anchorage
lengths. Alsiwat and Saatcioglu [4] formulated a bond-slip model using existing test results of
distribution and transfer of forces between steel reinforcement and concrete. Other researchers worked
on the impact of reinforced concrete (RC), detailing the bond-slip relationship where they evaluated
the bond behavior for different conditions of concrete cover and confinement [5]. According to detailed
experimental studies conducted using direct tension pullout tests, it has been observed that the force
distribution was not uniform along the rebar embedment length [6].
Similar to steel rebars, composite rebars also do not uniformly transfer forces to concrete [7].
Steel rebar surface properties are standardized as a result of past experimental studies. However,
composite rebars are produced with various base materials such as carbon, glass or aramid, and
may have different surface conditions such as ribbed or roughened. Therefore, there is still a lack of
sufficient experimental background and standardization. It should also be noted that plain composite
bars have only negligible bond resistance even in high strength concrete. American Concrete Institute
ACI 440 [8] recommends basic formulations for the calculation of the required embedment length of
composite rebars. Harajli and Abouniaj [9] have stated that these formulations are too conservative
and not realistic. Other researchers studies also support this conclusion about the bond requirements
of ACI 440 [10].
This experimental study mainly focuses on comparing the bond behavior of carbon fiber reinforced
polymer (CFRP) rebars in a high strength concrete block under monotonic and cyclic loading conditions.
In addition to the loading procedure, various embedment lengths were also taken as test parameters.
Obtained results showed that the sequence of loading has a significant impact on the behavior of CFRP
reinforced concrete members.
2. Specimen Preparation
All rebars had the same diameter (d) of 10 mm and were placed concentrically in the concrete
blocks. According to the predetermined specimen sizes, 10 plywood molds have been manufactured
with 350 mm 350 mm 300 mm dimensions. In this research, the concrete cover was kept constant
and taken as two times the diameter of the rebar to conform to the ACI 440 [8] guideline. The concrete
blocks were designed in a U-shaped plan configurationby using polystyrene foam elements with
150 mm 150 mm 300 mm dimensions) to provide the CFRP bars with the desired concrete cover.
In order to provide full fixity of the concrete specimen at the base of the experimental setup, four steel
pipes were placed at each corner of the specimens mold. The specimens were then fixed to the base
via four threaded steel bars. In order to determine the variation of bond stress distribution along the
CFRP rebar embedded inside the concrete block, five different embedment lengths were considered
as multiples of the bar diameters (5 d, 10 d, 15 d, 20 d and 25 d). To obtain the intended embedment
length, an unbonded surface is provided by means of wrapping the rebars with soft plastic pipes at two
ends of the embedded parts. Although the prepared test setup was suitable for concentric and eccentric
loading of the CFRP rebars, the specimens have been tested concentrically in this experimental work.
3. Materials
Unconfined specimens were designed and constructed. As stated in the preceding section, the
concrete cover was kept constant in order to observe the effect of 2 d cover spacing as per the ACI
440 guideline. It was thought that carbon rebars would be more appropriate to use in high strength
concrete members [11]. As shown in Figure 1, an appropriate mix design was developed to target
the C60 (cylindrical strength of 60 MPa) class of concrete as seen in Table 1. The standard cylinder
specimens (150 mm 300 mm) were tested under compression on the 28th day after pouring the
concrete in place. Experimentally obtained average cylinder compressive strength was found to be
68.8 MPa.

Table 1. Mixture proportions used for the test series.

Mixture

Water
/Cement

Water
(kg/m3)

Cement
(kg/m3)

135

300

Coarse
aggregate,
1225 mm
(kg/m3)
522

Polymers 2016, 8, 211

HSC

0.30

Coarse
aggregate,
612 mm
(kg/m3)
468

Fine
aggregate,
06 mm
(kg/m3)
787

Silica fume
(kg/m3)

Plasticizer
(kg/m3)

150

5.85

3 of 11

Polymers 2016, 8, 211

3 of 11

Table 1. Mixture proportions used for the test series.

Mixture

Water
/Cement

Water
(kg/m3)

Cement
(kg/m3)

HSC

0.30

135

300

Coarse
aggregate,
1225 mm
(kg/m3)
522

Coarse
aggregate,
612 mm
(kg/m3)
468

Fine
aggregate,
06 mm
(kg/m3)
787

Silica fume
(kg/m3)

Plasticizer
(kg/m3)

150

5.85

Figure 1. Mixture design.

Figure 1. Mixture design.

Sika CarboDur BC10 deformed rebars were selected in this research, as shown in Figure 2. The
1. Mixtureand
proportions
for the
test series. area can resist 110 kN
carbon rebars with d = 10Table
mm diameter
a 79 mm2 used
nominal
cross-sectional
tension force under 1% elongation [12].
Mixture
HSC

Water/Cement
0.30

Water
(kg/m3 )

Cement
(kg/m3 )

Coarse
Aggregate,
1225 mm
(kg/m3 )

Coarse
Aggregate,
612 mm
(kg/m3 )

Fine
Aggregate,
06 mm
(kg/m3 )

Silica
Fume
(kg/m3 )

Plasticizer
(kg/m3 )

135

300

522

468

787

150

5.85

Figure 1. Mixture design.

Sika CarboDur BC10 deformed rebars were selected in this research, as shown in Figure 2. The
Sika CarboDur BC10 deformed rebars were selected in this research, as shown in Figure 2. The
carbon rebars with d = 10 mm diameter and a 79 mm22 nominal cross-sectional area can resist 110 kN
carbon rebars with d = 10 mm diameter and a 79 mm nominal cross-sectional area can resist 110 kN
tension
forceforce
under
1% 1%
elongation
[12].
tension
under
elongation
[12].Figure 2. Rebar surface.
Three CFRP rebar specimens were tested at the Materials Laboratory of Bogazici University
(Figure 3). According to the test results, mechanical properties of CFRP rebars have been compared
with the manufacturers suggested values as seen in Table 2. The measured average modulus of
elasticity was 153,250 N/mm2. Measured rupture strengths for the specimens were 146.9, 104.4 and
126.6 kN, respectively. By using the average rupture value and also the experimentally obtained
strength and modulus of elasticity values, it was found out that the fracture strain was slightly over
1.0%, which was different than the catalogue value.
Figure 2. Rebar surface.
Figure 2. Rebar surface.
Table 2. Mechanical properties of CFRP rebars according to catalogue.

Three CFRP rebar specimens were tested at the Materials Laboratory of Bogazici University
Valuestested
in the longitudinal
direction
of the rebars
fibers [13]
(FigureCFRP
3). According
to the test results,
mechanical
properties
of CFRP
have
been compared
Three
rebar specimens
were
at the
Materials
Laboratory
of Bogazici
University
Elastic modulus
Mean
valuein Table 2. The
148,000
N/mm2 average modulus of
with
the
manufacturers
suggested
values
as
seen
measured
(Figure 3). According to the test results, mechanical properties of CFRP rebars
have been compared
2
value
>140,000
N/mmwere
2. MeasuredMinimum
was 153,250 N/mm
rupture
strengths
for2.the
specimens
146.9, 104.4
and of
with elasticity
the manufacturers
suggested
values
as
seen
in
Table
The
measured
average
modulus
in rupture
longitudinal
direction
of fibers
126.6 kN, respectively. By 2using the Values
average
value
and also
the [13]
experimentally obtained
elasticity was 153,250 N/mm . Measured rupture
strengths for the
specimens
were 146.9, 104.4 and
Mean value
3,100
N/mm2
strength and modulus of elasticity values, it was found out that the fracture strain
was slightly over
2
Tensile
strengththe average
Minimum
value value and
>2,800
126.6 kN, respectively.
By using
rupture
alsoN/mm
the experimentally obtained
1.0%, which was different than the catalogue value.
2
5%itFractile
Value out that the
2,900fracture
N/mm strain was slightly over
strength and modulus of elasticity values,
was found
95% Fractile Value
3,250 N/mm2
1.0%, which was different
the catalogue
value.
Tablethan
2. Mechanical
properties
of CFRP rebars according to catalogue.
Values in longitudinal direction of fibers [13]
value
ValuesMinimum
in the longitudinal
direction of>1.70%
the fibers [13]
Table
Mechanical properties
of CFRPforce
rebars according
catalogue.
2
1%2.elongation
Minimum
>110to
kN
Elastic
modulus
Meantensile
value
148,000
N/mm
Minimum value
>140,000 N/mm2
Values in the longitudinal direction of the fibers [13]
Values
in longitudinal direction of fibers
[13] N/mm2
Elastic modulus
Mean value
148,000
2
Mean
3,100 N/mm
Minimum
valuevalue
>140,000
N/mm2
2
Tensile strength
Minimum
>2,800
N/mm
Values invalue
longitudinal direction
of fibers
[13]
Mean
3,1002 N/mm2
5% value
Fractile Value
2,900 N/mm
Tensile strength
2
Minimum value
>2,800
2 N/mm
95%
Fractile
Value
3,250
N/mm
5% Fractile Value
2,900 N/mm2
longitudinal direction of fibers3,250
[13] N/mm2
95%Values
Fractilein
Value
Strain at break
Minimum
value
>1.70%
Values in longitudinal direction of fibers [13]
Strain at break
Minimum
value
1% elongation
Minimum
tensile force
>110 kN >1.70%
Strain at break

1% elongation

Minimum tensile force

>110 kN

Polymers 2016, 8, 211

Polymers 2016, 8, 211

4 of 11
4 of 11

Figure
test; (c)
(c) Location
Location of
of rupture.
rupture.
Figure 3.
3. Rebar
Rebar rupture
rupture test.
test. (a)
(a) Test
Test setup;
setup; (b)
(b) End
End of
of test;

4. Test Setup and Experimental Procedure

4. Test Setup and Experimental Procedure
All specimens were tested at the Structures Laboratory of Bogazici University. The existing
All specimens were tested at the Structures Laboratory of Bogazici University. The existing loading
loading frame, which was used previously for steel rebar tests, was modified according to the
frame, which was used previously for steel rebar tests, was modified according to the requirements
requirements of this research. The test setup was equipped with special steel mounting parts to
of this research. The test setup was equipped with special steel mounting parts to provide a reliable
provide a reliable connection between the rebar and hydraulic jack (200 kN capacity with 200 mm
connection between the rebar and hydraulic jack (200 kN capacity with 200 mm of stroke). A special
of stroke). A special conical steel jaw attachment was designed to carry 250 kN of load without any
conical steel jaw attachment was designed to carry 250 kN of load without any slipping occurrences.
slipping occurrences. The steel jaw assembly is the most crucial part of the experimental setup.
The steel jaw assembly is the most crucial part of the experimental setup.
In order to obtain proper and accurate force distribution between the rebar and surrounding
In order to obtain proper and accurate force distribution between the rebar and surrounding
concrete, compatible strain gauges were chosen for the CFRP rebar surfaces and attached to various
concrete, compatible strain gauges were chosen for the CFRP rebar surfaces and attached to various
locations before concrete pouring.
locations before concrete pouring.
Although CFRP rebars exhibit high tensile strengths in the longitudinal direction, they are weak
Although CFRP rebars exhibit high tensile strengths in the longitudinal direction, they are weak in
in the transverse direction because of their orthotropic material properties. Without using any
the transverse direction because of their orthotropic material properties. Without using any protecting
protecting cover for the rebar, the jaw would crush the rebar laterally, leading to erroneous results.
cover for the rebar, the jaw would crush the rebar laterally, leading to erroneous results. For this reason,
For this reason, to prevent any material crushing, the tested rebars were surrounded with a 250-mmto prevent any material crushing, the tested rebars were surrounded with a 250-mm-long steel pipe
long steel pipe sleeve with a diameter slightly larger than the rebars. The spacing between the rebar
sleeve with a diameter slightly larger than the rebars. The spacing between the rebar and steel sleeve
and steel sleeve was filled with a high strength epoxy having more than 3 MPa of bonding capacity
was filled with a high strength epoxy having more than 3 MPa of bonding capacity to the steel.
to the steel.
The schematic arrangement of the test setup is shown in Figure 4. A hydraulic jack was suspended
The schematic arrangement of the test setup is shown in Figure 4. A hydraulic jack was
from the top of the closed loading frame in the downward position with the conical jaw attached at the
suspended from the top of the closed loading frame in the downward position with the conical jaw
head of the actuator. The tip of the CFRP rebar was attached to the jaw. Specimens were seated on a
attached at the head of the actuator. The tip of the CFRP rebar was attached to the jaw. Specimens
steel plate having a thickness of 25 mm, which was welded to a concrete-filled steel foundation beam.
were seated on a steel plate having a thickness of 25 mm, which was welded to a concrete-filled steel
The concrete block was then fixed with four threaded rods at the edges. There were two steel plates
foundation beam. The concrete block was then fixed with four threaded rods at the edges. There were
with a 25 mm thickness connecting two of these rods together at the top of the specimen, providing a
two steel plates with a 25 mm thickness connecting two of these rods together at the top of the
full tension area between these steel plates.
specimen, providing a full tension area between these steel plates.
As opposed to other proposed test setups (e.g., ASTM D7913) that produce compression on the
As opposed to other proposed test setups (e.g., ASTM D7913) that produce compression on the
concrete surface, the test setup developed in this work allows us to measure the bond capacity of CFRP
concrete surface, the test setup developed in this work allows us to measure the bond capacity of
rebars with the pure tension area in a concentric or eccentric loading configuration. Also, various
CFRP rebars with the pure tension area in a concentric or eccentric loading configuration. Also,
embedment lengths from 5 d to 30 d with different concrete covers can be tested. Note that many test
various embedment lengths from 5 d to 30 d with different concrete covers can be tested. Note that
many test setups allow a 5 d embedment length. It is believed that the real behavior of the bond
mechanism can be better achieved with the testing method developed in this work.

Polymers 2016, 8, 211

5 of 11

setups allow a 5 d embedment length. It is believed that the real behavior of the bond mechanism can
be better
achieved
Polymers
2016,
8, 211 with the testing method developed in this work.
5 of 11

Figure 4. Test Setup. (a) Loading frame; (b) Concrete block dimensions; (c) LVDT location.
Figure 4. Test Setup. (a) Loading frame; (b) Concrete block dimensions; (c) LVDT location.

The specimens were tested monotonically and cyclically. In cyclic tests, only specimens with 25
specimens
were tested
monotonically
and
cyclically.
In cyclic tests, test
onlymethod
specimens
with
25of
d,
d, 20 The
d and
15 d embedment
depths
were tested
using
the force-controlled
with
5 kN
20 d and
15 d embedment
were tested
using the force-controlled
test method
with 5 kNfor
of
force
increments
and threedepths
repetitions.
Displacement-controlled
cyclic testing
was performed
force
increments
and
three
repetitions.
Displacement-controlled
cyclic
testing
was
performed
for
other
other specimens. Displacements were increased by 0.01 mm at a time.
specimens. Displacements were increased by 0.01 mm at a time.
5. Instrumentation
5. Instrumentation
One Linear Variable Differential Transformer (LVDT) (100 mm) was connected to the rebar with
One Linear Variable Differential Transformer (LVDT) (100 mm) was connected to the rebar with a
a manacle to measure the relative slip (Figure 4c). Another LVDT was connected on the concrete block
manacle to measure the relative slip (Figure 4c). Another LVDT was connected on the concrete block
surface and also to the hydraulic jack with a wire.
surface and also to the hydraulic jack with a wire.
In order to monitor the strain distribution inside the concrete block on rebar surfaces, three
In order to monitor the strain distribution inside the concrete block on rebar surfaces, three
BFLA-5 type strain gauges (SG1, SG2 and SG3) were attached on the rebar. Strain gauge
BFLA-5 type strain gauges (SG1, SG2 and SG3) were attached on the rebar. Strain gauge measurements
measurements were taken at the beginning of the bonded area (SG1), in the middle of the bonded
were taken at the beginning of the bonded area (SG1), in the middle of the bonded area (SG2) and at
area (SG2) and at the end of the bonded area (SG3) (Figure 4a).
the end of the bonded area (SG3) (Figure 4a).
6. Expected Failure Modes
6. Expected Failure Modes
Expected failure modes are explained in this section by comparing the experimental and
Expected failure modes are explained in this section by comparing the experimental and expected
expected failure modes. Figure 5a shows the failure mode regarding a rebar rupture. This happens
failure modes. Figure 5a shows the failure mode regarding a rebar rupture. This happens when the bar
when the bar is adequately embedded in the concrete block. On the other hand, a typical pullout
is adequately embedded in the concrete block. On the other hand, a typical pullout failure is shown in
failure is shown in Figure 5b, which corresponds to an inadequate embedment length.
Figure 5b, which corresponds to an inadequate embedment length.
If an adequate embedment length is provided in an unconfined concrete, the pullout force could
If an adequate embedment length is provided in an unconfined concrete, the pullout force could
exceed the tension capacity of the concrete and a shear failure would occur. Splitting failure occurs
exceed the tension capacity of the concrete and a shear failure would occur. Splitting failure occurs
when the surrounding concrete cannot resist the circumferential (radial) tensile stresses. Splitting
when the surrounding concrete cannot resist the circumferential (radial) tensile stresses. Splitting
generally occurs on the surface of the member as flexural cracks start forming and the rebar is stressed
generally occurs on the surface of the member as flexural cracks start forming and the rebar is stressed
and elongated towards failure with a continuous longitudinal crack along the rebar. Splitting cracks
generally occur on the concrete surface (along the bonded length) and propagate longitudinally and
transversely as the rebar is axially stressed.
In general, two types of splitting modes were expected prior to testing: concrete rupture and Vnotch failures. Concrete rupture is most likely to occur when the force exceeds the tension capacity

Polymers 2016, 8, 211

6 of 11

and elongated towards failure with a continuous longitudinal crack along the rebar. Splitting cracks
generally occur on the concrete surface (along the bonded length) and propagate longitudinally and
transversely as the rebar is axially stressed.
In 2016,
general,
and
Polymers
8, 211 two types of splitting modes were expected prior to testing: concrete rupture6 of
11
V-notch
failures.
Concrete
rupture
is
most
likely
to
occur
when
the
force
exceeds
the
tension
capacity
Polymers 2016, 8, 211
6 of 11
shown in
in Figure
Figure 5d,
5d, the
the V-notch
V-notch failure (spalling
of the unconfined concrete specimen (Figure 5c). As shown
cover is
relatively
small.
of the
the cover)
unconfined
specimen
5c).
As shown
in Figure 5d, the V-notch failure (spalling
of
occursconcrete
when the
concrete(Figure
of the cover) occurs when the concrete cover is relatively small.

Figure 5. Expected failure modes. (a) Rebar rupture side view; (b) Pullout side view; (c) Splitting
Figure 5. Expected failure modes. (a) Rebar rupture side view; (b) Pullout side view; (c) Splitting
failure
type) side
andmodes.
top views;
(d) Splitting
failure
top side
views.
Figure(shear
5. Expected
failure
(a) Rebar
rupture
side (V-notch
view; (b)type)
Pullout
view; (c) Splitting
failure (shear type) side and top views; (d) Splitting failure (V-notch type) top views.
failure (shear type) side and top views; (d) Splitting failure (V-notch type) top views.

The pullout mechanism may be defined in three different failure modes: rib crushing, failure of
The
mechanism
may
be
in
three
rib
crushing,
failure
of
The pullout
pullout
mechanism
may Rib
be defined
defined
inoccurs
three different
different
failure
modes:
ribas
crushing,
failure
of
the surface
and concrete
crushing.
crushing
in raised failure
types ofmodes:
ribs such
in the steel
rebar.
the
surface
and
concrete
crushing.
Rib
crushing
occurs
in
raised
types
of
ribs
such
as
in
the
steel
rebar.
the
surface
and concrete
crushing.
Rib rib,
crushing
occurs
in raised
types ofThe
ribsCFRP
such as
in theused
steelin
rebar.
If
the
load exceeds
the capacity
of the
it starts
bending
or yielding.
rebars
this
If
the
capacity
of
starts
The
CFRP
rebars
used
this
If the
the load
load
exceeds
theribs.
capacity
of the
the rib,
rib,
starts bending
bending
or
yielding.
The
CFRPmode
rebarsin
used
in
this
work
haveexceeds
indented
Therefore,
rib ititcrushing
is not or
anyielding.
expected
failure
thein
tests.
work
have
indented
ribs.
Therefore,
rib
crushing
is
not
an
expected
failure
mode
in
the
tests.
However,
work havefailure
indented
Therefore,
rib crushing
crushingmay
is not
anaccording
expected to
failure
mode
in rib
theheight
tests.
However,
of theribs.
surface
or concrete
occur
the ratio
of the
failure
of gap.
the
surface
orsurface
concrete
may occur
according
to the ratio
of ratio
the rib
height
to the
However,
failure
of the
or crushing
concrete crushing
may
occur according
to the
of the
rib height
to
the rib
rib
gap.
to the
rib gap.
When
the ribs are high and spaced too closely, the shear stress may govern the behavior and the
are
and
too
closely,
the
shear
stress
may
govern
the
and
the
When
the
ribs
are high
high
and spaced
spaced
tooshown
closely,in
the
shear6b,
stress
maythe
govern
the behavior
behavior
and
the
rebarWhen
wouldthe
beribs
pulled
out (Figure
6a). As
Figure
when
rib spacing
is larger
than
rebar
would
be
pulled
out
(Figure
6a).
As
shown
in
Figure
6b,
when
the
rib
spacing
is
larger
than
rebar would be10pulled
out ribs
(Figure
6a). the
As partly
showncrushed
in Figure
6b, when
ribaspacing
is front
largerofthan
approximately
times the
height,
concrete
maythe
form
wedge in
the
approximately
10
times
ribs
the
partly
may
wedge
front
approximately
10normally
times the
thebrought
ribs height,
height,
the
partly
crushed
concrete
may form
form aaconcrete
wedge in
in[14].
front of
of the
the
rib
and failure is
about
with
the crushed
splittingconcrete
of the surrounding
rib
and
failure
is
normally
brought
about
with
the
splitting
of
the
surrounding
concrete
[14].
rib and failure is normally brought about with the splitting of the surrounding concrete [14].

Figure 6. Failure mechanisms. (a) Failure of surface; (b) Concrete crushing.

Figure 6. Failure mechanisms. (a) Failure of surface; (b) Concrete crushing.
Figure 6. Failure mechanisms. (a) Failure of surface; (b) Concrete crushing.

7. Prediction of Strength According to ACI 440.R06

7. Prediction of Strength According to ACI 440.R06
Figure 7 shows the equilibrium condition of a CFRP rebar of length le embedded in a concrete
concrete
7 shows
equilibrium
of a CFRP
lengthon
le embedded
block.Figure
The force
in thethe
rebar
is resistedcondition
by an average
bond rebar
stressof
u acting
the surface in
of athe
rebar.
block.
The
force
in
the
rebar
is
resisted
by
an
average
bond
stress
u
acting
on
the
surface
of
the
rebar.
Equilibrium of forces can be written as follows;
Equilibrium of forces can be written as follows;
le d b u = Af,bar f f
(1)
le d b u = Af,bar f f
(1)
where Af,bar is the cross-section area of one rebar, db is the rebar diameter and ff is the stress developed
where
Af,bar is
area of onelength.
rebar, db is the rebar diameter and ff is the stress developed
in
the rebar
at the
the cross-section
end of the embedment
in the rebar at the end of the embedment length.

Polymers 2016, 8, 211

7 of 11

7. Prediction of Strength According to ACI 440.R06

Figure 7 shows the equilibrium condition of a CFRP rebar of length le embedded in a concrete
block. The force in the rebar is resisted by an average bond stress u acting on the surface of the rebar.
Equilibrium of forces can be written as follows;
le db u Af,bar f f

(1)

where Af,bar is the cross-section area of one rebar, db is the rebar diameter and f f is the stress developed
in
the rebar
the end of the embedment length.
Polymers
2016, at
8, 211
7 of 11

Figure 7.
7. Transfer
of force
force through
through bond.
bond.
Figure
Transfer of

The development length equation for steel reinforcing bars found in ACI 318-08 [15] is based on
The development length equation for steel reinforcing bars found in ACI 318-08 [15] is based
the work done by Orangun [16]. A similar methodology was used by Wambeke and Shield for
on the work done by Orangun [16]. A similar methodology was used by Wambeke and Shield for
representing the bond performances of GFRP rebars [17]. According to the test results, they have
representing the bond performances of GFRP rebars [17]. According to the test results, they have
derived an empirical bond formula. Based on their study, Equation (2) is proposed by ACI 440 [8]
derived an empirical bond formula. Based on their study, Equation (2) is proposed by ACI 440 [8] and
and recommended for bond capacity calculations of composite rebars.
recommended for bond capacity calculations of composite rebars.

db

= 4.0 + 0.3 C + 100d

u
a f ' 4.0 ` 0.3 d b` 100 ble
0.083
db
le
0.083 fcc 1

(2)
(2)

where C is the cover to the center of the rebar, u is the average bond stress acting on the surface of
where C is the cover to the center of the rebar, u is the average bond stress acting on the surface of the
the GFRP rebar, db is the diameter of the reinforcing bar, fc is the specified compressive strength of
GFRP rebar, db is the diameter of the reinforcing bar, f c is the specified compressive strength of the
the concrete and le is the embedded length of the reinforcing bar.
concrete and le is the embedded length of the reinforcing bar.
8. Experimental
8.
ExperimentalResults
Results
Experimentally obtained
obtained test
test results
results are
are summarized
summarized in
in Table
Table 3.
3. This
This table
table includes
includes specimen
specimen
Experimentally
labels,
embedment
lengths,
loading
types,
expected
bond
strengths
and
average
bond
stresses
labels, embedment lengths, loading types, expected bond strengths and average bond stresses
according to
and
average
bond
stresses
according
to the
according
to ACI
ACI 440
440[8],
[8],measured
measuredbond
bondstrengths
strengths
and
average
bond
stresses
according
to test
the
results,
and
observed
failure
modes.
test results, and observed failure modes.
Expected bond
bond capacities
capacities given
given in
Table 33 are
are calculated
calculated from
from Equation
Equation (2)
(2) which
which is
is proposed
proposed
Expected
in Table
by ACI
ACI 440
440 [8].
[8]. As
seen from
from these
these results,
results, measured
measured bond
bond stresses
stresses are
are generally
generally higher
higher than
than the
the
by
As seen
calculated
bond
capacities
according
to
ACI
440
[8].
For
the
embedment
lengths
of
25
d,
20
d
and
calculated bond capacities according to ACI 440 [8]. For the embedment lengths of 25 d, 20 d and 15 15
d,
d, measured
bond
strengths
stresses
are almost
twice
the code-calculated
However,
measured
bond
strengths
and and
stresses
are almost
twice the
code-calculated
values. values.
However,
similar
similar
bond capacities
were obtained
5 d embedment
length specimens.
bond
capacities
were obtained
for 10 d for
and105 d
d and
embedment
length specimens.
Under
monotonic
testing,
maximum
axial
loads
of
28.3
and
86.4
kN were
were obtained
obtained for
for 55 d
d and
and
Under monotonic testing, maximum axial loads of 28.3 and 86.4 kN
25 d
cyclic
(not
reversed)
loading,
16.216.2
and and
70.0 70.0
kN load
values
were
25
d specimens,
specimens,respectively.
respectively.AsAsforfor
cyclic
(not
reversed)
loading,
kN load
values
reached
for
5
d
and
25
d
specimens,
respectively.
Developed
bond
stresses
corresponding
to
these
were reached for 5 d and 25 d specimens, respectively. Developed bond stresses corresponding to
load values
were 18.0
11.0
MPa
monotonic
loading,loading,
and 10.3and
and10.3
8.9 MPa
for MPa
cyclicfor
loading.
these
load values
wereand
18.0
and
11.0for
MPa
for monotonic
and 8.9
cyclic
The experimental
bond stresses
could be
significantly
reduced
when cyclic
loads loads
are applied.
This
loading.
The experimental
bond stresses
could
be significantly
reduced
when cyclic
are applied.
results
in
reduced
capacities
for
the
specimens
under
cyclic
loading.
Reductions
in
shortand
longThis results in reduced capacities for the specimens under cyclic loading. Reductions in short- and
development-length specimens
are are
40%40%
and
20%~25%,
respectively.
showed
long-development-length
specimens
and
20%~25%,
respectively.Experimental
Experimentalresults
results showed
that the
the attained
attained minimum
minimum and
and maximum
maximum loads
loads of
of 16.2
16.2 and
were far
far below
below the
the
that
and 86.4
86.4 kN,
kN, respectively,
respectively, were
maximum
capacity
(104.4
~146.9
kN)
of
the
deformed
CFRP
bars
used.
maximum capacity (104.4 ~146.9 kN) of the deformed CFRP bars used.
Bond stress (in MPa) versus slip (in mm) graphs are plotted and illustrated in Figure 8. It is
recognized that the initial stiffness of the specimens changes depending on the loading procedure. In
monotonically loaded specimens, measured bond stresses for shorter embedment lengths (for
example for the 5 d specimen) are considerably greater than the stresses (max. 63.4% higher) that
were obtained in specimens with longer embedment lengths.
Specimens S21, S22, S23 and S26 failed from the splitting mode while specimens S24, S25, S27,
S28, S29 and S30 failed from the pullout failure mode. It was observed that splitting failure occurred

Polymers 2016, 8, 211

8 of 11

Polymers 2016, 8, 211

8 of 11

Table3.3. Test
Test results
results
Table
Specimen 1
S21-25D-T-U
S22-25D-M-U
S23-20D-M-U
S24-20D-T-U
S25-15D-T-U
S26-15D-M-U
S27-10D-T-U
S28-10D-M-U
S29-5D-M-U
S30-5D-T-U
1
6

ACI
Test
Specimen 1 le 2 (mm) Loading type
Failure mode
ACI
Test
u 3 (MPa) f 4 (kN) u (MPa) f (kN)
Loading type
le 2 (mm)
Failure Mode
3
4
5
u
(MPa)
f (kN)
u (MPa)
f (kN) 8.93
S21-25D-T-U
250
C
70.1
Splitting
6.02
47.31
S22-25D-M-U
250 C 5
M6
11.00
86.4 Splitting
Rupture
250
8.93
70.1
Splitting
6.02
47.31
Splitting
12.80 11.00
80.4
S23-20D-M-U
200
M
Splitting
6
250
86.4
M
6.71
42.18
Rupture
S24-20D-T-U
200
C
11.52
72.4
Pull-out
200
M
12.80
80.4
Splitting
6.71
42.18
13.84 11.52
65.2
S25-15D-T-U
150
C
Pull-out
200
C
72.4
Pull-out
7.86
37.04
150
C
65.2
Pull-out
S26-15D-M-U
150
M
17.74 13.84
83.6
Splitting
7.86
37.04
150
83.6
Splitting
S27-10D-T-U
100 M
C
11.01 17.74
34.6
Pull-out
100
C
11.01
34.6
Pull-out
10.15
31.90
10.15
31.90 12.32 12.32
S28-10D-M-U
100 M
M
38.7
Pull-out
100
38.7
Pull-out
50
M
28.3
Pull-out
S29-5D-M-U
50
M
18.02 18.02
28.3
Pull-out
17.04
26.77
17.04
26.77
50
C
16.2
Pull-out
S30-5D-T-U
50
C
10.31 10.31
16.2
Pull-out

2
3 average bond stress,
4 bond strength,
5 cyclic loading,
1 ribbed
ribbed
10 10
mm
rebar,
mm
rebar, 2embedment
embedment length,
length, 3 average
bond stress, 4 bond
strength, 5 cyclic
loading, 6
monotonic loading.
monotonic loading.

Figure 8. Experimental bond stress versus slip curves. (a) 25 d; (b) 20 d; (c) 15 d; (d) 10 d; (e) 5 d.

Figure 8. Experimental bond stress versus slip curves. (a) 25 d; (b) 20 d; (c) 15 d; (d) 10 d; (e) 5 d.

Bond stress (in MPa) versus slip (in mm) graphs are plotted and illustrated in Figure 8. It is
recognized that the initial stiffness of the specimens changes depending on the loading procedure. In
monotonically loaded specimens, measured bond stresses for shorter embedment lengths (for example

Polymers 2016, 8, 211

9 of 11

for the 5 d specimen) are considerably greater than the stresses (max. 63.4% higher) that were obtained
in specimens with longer embedment lengths.
Specimens S21, S22, S23 and S26 failed from the splitting mode while specimens S24, S25, S27, S28,
S29 and S30 failed from the pullout failure mode. It was observed that splitting failure occurred under
relatively high bond strengths, showing that the tensile strength of unconfined concrete controlled
the behavior at this load level. For the splitting-type failures, loud sounds were heard, whereas for
the pullout (or slip) failure no sound was heard. Since the concrete blocks were not confined, the
splitting
failures in some tests were quite brittle and caused complete wide cracks throughout
the
Polymers 2016, 8, 211
9 of 11
Polymers 2016,
8, 211
9 of 11
concrete
block.
Figure
behavioral
variation
between
monotonic
andand
cyclic
test
Figure 88 reveals
revealsthat
thatthere
thereisisa significant
a significant
behavioral
variation
between
monotonic
cyclic
Figure 8 reveals that there is a significant behavioral variation between monotonic and cyclic
results.
As aAs
general
observation
from from
these these
figures,
for specimens
with 10
d or10over,
values
test results.
a general
observation
figures,
for specimens
with
d orstiffness
over, stiffness
test results. As a general observation from these figures, for specimens with 10 d or over, stiffness
are
similar
at
early
displacement
levels.
However,
under
cyclic
loading,
stiffness
of
the
specimens
values are similar at early displacement levels. However, under cyclic loading, stiffness of the
values are similar at early displacement levels. However, under cyclic loading, stiffness of the
tends
to soften
under
larger
displacement
levels.
specimens
tends
to soften
under
larger displacement
levels.
specimens tends to soften under larger displacement levels.
The general shapes of the behavioral curves obtained from this study were similar to the curves
The general shapes of the behavioral curves obtained from this study were similar to the curves
given in the existing literature. Typical failure modes for splitting and pullout failures are given in
given in the existing literature. Typical failure modes for splitting and pullout failures are given in
Figures 9 and 10
10,respectively.
respectively.
Figures 9 and 10, respectively.

Figure
9. (a)
Splitting
failure;
(b) Section of
of bonded and
and unbonded regions.
regions.
Figure 9.
(a) Splitting
Splitting failure;
failure; (b)
Figure
9. (a)
(b) Section
Section of bonded
bonded and unbonded
unbonded regions.

Figure 10. (a) Pullout failure; (b) No damage observed on concrete surface.
Figure 10. (a) Pullout failure; (b) No damage observed on concrete surface.
Figure 10. (a) Pullout failure; (b) No damage observed on concrete surface.

Achillides and Pilakoutas [18] tested CFRP rebars monotonically and calculated the average
Achillides and Pilakoutas [18] tested CFRP rebars monotonically and calculated the average
bond strengths between 912 MPa according to their experiments. Note that concrete cube
bond strengths between 912 MPa according to their experiments. Note that concrete cube
compressive strength values of 4050 MPa were used in their work. According to the obtained
compressive strength values of 4050 MPa were used in their work. According to the obtained
experimental results in the current study, the bond strength values changed between 1018 MPa
experimental results in the current study, the bond strength values changed between 1018 MPa
under monotonic loading and 914 MPa under cyclic loading. These results show that obtained
under monotonic loading and 914 MPa under cyclic loading. These results show that obtained
monotonic loading capacities are remarkably higher than those obtained by the previous researchers
monotonic loading capacities are remarkably higher than those obtained by the previous researchers
[9,18]. This could be due to the better bond characteristics of the high strength concrete used in this

Polymers 2016, 8, 211

10 of 11

Achillides and Pilakoutas [18] tested CFRP rebars monotonically and calculated the average bond
strengths between 912 MPa according to their experiments. Note that concrete cube compressive
strength values of 4050 MPa were used in their work. According to the obtained experimental
results in the current study, the bond strength values changed between 1018 MPa under monotonic
loading and 914 MPa under cyclic loading. These results show that obtained monotonic loading
capacities are remarkably higher than those obtained by the previous researchers [9,18]. This could
be due to the better bond characteristics of the high strength concrete used in this study. Also, to the
knowledge of the authors, although there are some experimental studies conducted with 3050 MPa
concrete strengths, limited experimental work exists for tests performed with high strength concrete
(C60 and above).
9. Conclusions
Monotonic and cyclic experimental bond behavior of CFRP rebars in high strength (68.8 MPa)
concrete was investigated. Various embedment lengths were studied. All (10) specimens had
a constant concrete cover of 2 d. All monotonically-loaded specimens were loaded under a
displacement-controlled testing procedure as well as the 5 d and 10 d embedded cyclic loading
specimens. Other cyclic loaded specimens (25 d, 20 d and 15 d) were tested under the force-controlled
testing procedure. The following conclusions could be drawn from this experimental work:

In general, rebars with 5 d and 10 d embedment lengths failed due to pullout of the bar while
rebars with 15 d, 20 d and 25 d embedment lengths failed due to splitting of the concrete. These
failures were more brittle since the tension capacity of the unconfined concrete block was exceeded.
Therefore, a thicker cover concrete (i.e., >2 d) is more suitable for relatively deeper embedment
lengths when high strength concrete is used.
Stiffness of specimens varied significantly according to the selected loading protocols.
Monotonically-loaded specimens had a higher stiffness when compared to the stiffness obtained
for larger cycles of the cyclically-loaded specimens. However, generally, the stiffness values are
similar (especially for embedment lengths of 10 d or longer) at early displacement levels for both
cyclic and monotonic loading.
Cyclic loading reduced the bond capacity of the specimens. Reductions in short- and
long-development-length specimens are within the range of 40% and 20%~25%, respectively.
For tested specimens, the bond strength values changed between 10~18 MPa under monotonic
loading and 9~14 MPa under cyclic loading. These results showed that obtained monotonic
loading capacities were considerably higher than those obtained by the previous research.
Measured bond stresses are generally higher than the calculated bond capacities according to ACI
440 [8]. For the embedment lengths of 25 d, 20 d and 15 d, measured bond strengths and stresses
are almost twice the code-calculated values. However, the calculated bond capacities were similar
to obtained measurements for 10 d and 5 d embedment length specimens.

Acknowledgments: This research was supported by Istanbul Technical University Research Projects Unit
(ITU-BAP 34207). Technical assistance from the staff at Bogazici University, Civil Engineering Department,
Structures Laboratory are gratefully acknowledged and appreciated. However, any opinions, findings, conclusions,
and recommendations presented in this paper are those of the authors and do not necessarily reflect the views of
the supports.
Author Contributions: T. Tibet Akbas, Oguz C. Celik and Cem Yalcin designed the test setup and performed
the experiments; T. Tibet Akbas, Oguz C. Celik, Cem Yalcin and Alper Ilki analyzed the experimental data and
contributed in writing the paper.
Conflicts of Interest: The authors declare no conflict of interest.

Polymers 2016, 8, 211

11 of 11

References
1.
2.

3.
4.
5.
6.
7.
8.
9.
10.
11.

12.

13.
14.
15.
16.
17.
18.

Teng, J.G.; Chen, J.F.; Smith, S.T.; Lam, L. FRP Strengthened RC Structures, 1st ed.; Wiley: West Sussex,
UK, 2002.
Song, K.; Zhang, Y.; Meng, J.; Green, C.E.; Tajaddod, N.; Li, H.; Minus, M.L. Structural polymer-based carbon
nanotube composite fibers: Understanding the processing-structure-performance relationship. Materials
2013, 6, 25432577. [CrossRef]
Zhang, Y.; Tajaddod, N.; Song, K.; Minus, L.M. Low temperature graphitization of interphase
polyacrylonitrile (PAN). Carbon 2015, 91, 479493. [CrossRef]
Alsiwat, J.M.; Saatcioglu, M. Reinforcement anchorage slip under monotonic loading. ASCE J. Struct. Eng.
1992, 118, 24212438. [CrossRef]
Harajli, M.H. Bond stress-slip model for steel bars in unconfined or steel, FRC, or FRP confined concrete
under cyclic loading. ASCE J. Struct. Eng. 2009, 135, 509518. [CrossRef]
Tastani, S.P.; Pantazopoulou, S.J. Direct tension pullout test: Experimental results. ASCE J. Struct. Eng. 2010,
136, 731743. [CrossRef]
Sharaky, I.A.; Torres, L.; Baena, M.; Vilanova, I. Effect of different material and construction details on the
bond behavior of NSM FRP bars in concrete. Constr. Build. Mater. 2012, 38, 890902. [CrossRef]
ACI 440.1R06. In Guide for the Design and Construction of Structural Concrete Reinforced with FRP Bars; American
Concrete Institute: Farmington Hills, MI, USA, 2006.
Harajli, M.; Abouniaj, M. Bond performance of GFRP bars in tension: Experimental evaluation and
assessment of ACI 440 guidelines. ASCE J. Compos. Constr. 2010, 14, 659668. [CrossRef]
Hao, Q.; Wang, Y.; Ou, J. Design recommendations for bond between GFRP/steel wire composite rebars and
concrete. Eng. Struct. 2008, 30, 32393246. [CrossRef]
Akbas, T.T.; Celik, O.C.; Yalcin, C. Experimental bond behavior of deformed CFRP rebars in high strength
concrete. In proceeding of the ConcreteInnovation and Design, FIB Symposium, Copenhagen, Denmark,
1820 May 2015.
Sika CarboDur BC Rods Product Data Sheet. Available online: https://www.google.com/url?sa=t&
rct=j&q=&esrc=s&source=web&cd=1&ved=0ahUKEwi3-N32hujMAhVkK6YKHVClD1sQFggcMAA&url=
https%3A%2F%2Fgcc.sika.com%2Fdms%2Fgetdocument.get%2Fc64ae491-e7fa-3c19-bf32-0eb7ff401a47%
2FSika%2520Carbodur%2520BC%2520Rods%2520-%2520PDS.pdf&usg=AFQjCNFkYvG3C1Wg3ee2oH5Yy
Sei5YfO6g&sig2=NjTmA0ea0LkWHsaZcqZC3g&bvm=bv.122448493,d.dGY&cad=rja
(accessed
on
3 July 2013).
EN 2561. In Carbon Fiber Reinforced Plastics, Unidirectional Laminates, Tensile Test Parallel to Fibre Direction;
CEN: Brussels, Belgium, 1995.
Park, R.; Paulay, T. Reinforced Concrete Structures, 1st ed.; Wiley: West Sussex, UK, 1975.
ACI 318-08. In Building Code Requirements for Structural Concrete and Commentary; American Concrete Institute:
Farmington Hills, MI, USA, 2008.
Orangun, C.O.; Jirsa, J.O.; Breen, J.E. A reevaluation of test data on development length and splices. ACI J.
1977, 77, 114122.
Wambeke, B.; Shield, C. Development length of glass fiber reinforced polymer 21 bars in concrete. ACI Struct.
J. 2006, 103, 1117.
Achillides, Z.; Pilakoutas, K. Bond behavior of fiber reinforced polymer bars under direct pullout conditions.
ASCE J. Compos. Constr. 2004, 8, 173181. [CrossRef]
2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).