Engineering Structures 58 (2014) 1–11

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Analytical and experimental investigation for bond behaviour of newly

developed polystyrene foam particles’ lightweight concrete
A. Farghal Maree ⇑, K. Hilal Riad
Department of Structural Engineering, Ain Shams University, Cairo, Egypt

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

Article history: Structural lightweight concrete (LWC) is of high importance to the construction industry, as it is cost
Received 5 March 2013 effective and highly advantageous. A new kind of LWC was produced at the Department of Structural
Revised 5 October 2013 Engineering of Ain Shams University in 2005, which combines the advantages of normal density concrete,
Accepted 8 October 2013
cellular concrete and high workability concrete through partially replacing the normal weight aggregates
Available online 11 November 2013
with polystyrene foam particles. This leads to concrete’s unit weight reduction while maintaining ade-
quate strength. The latter material can therefore be produced using standard methods familiar to the con-
Keywords:
struction industry with a dry unit weight of 18.50 kN/m3, which in turn leads to self weight reduction of
Bond behaviour
Lightweight concrete
15–20% and the associated decrease in the structure’s overall cost, hence, providing a feasible challenge
Polystyrene foam concrete to normal weight concrete (NWC). The bond behaviour of structural polystyrene foam lightweight con-
Pull-out crete (PF-LWC) was investigated analytically and experimentally. The experimental program incorpo-
Bond–slip rated two phases. The first phase was performed on the standard pull-out specimens to compare its
results with the commonly conducted bond testing and to abstract the bond slip curve of the standard
pull-out test specimens. The second phase deal with a deduced beam-end specimen to assess the behav-
iour of bond between reinforcing bars and concrete in flexure members. Then analytical investigation of
the obtained experimental results was performed to develop a model capable of assessing the structural
bond behaviour of PF-LWC flexural members. The defining parameters of the bond stress–slip curve were
modified for NWC and PF-LWC using the best fit technique to the experimental results in order to add the
bar diameter as a variable in the bond stress–slip relationship. The defining parameters of the CEB-FIP
1990 bond stress–slip curve [1] were modified for NWC and PF-LWC using the best fitting technique.
Published by Elsevier Ltd.

1. Introduction
Most of the current concrete research focuses on high perfor-
mance concrete, which is a cost effective material that satisfies
demanding performance requirements, including durability. LWC
is of utmost importance to the construction industry that is cost
effective and highly advantageous. The primary advantage of
LWC is the reduction of the structural dead load, foundation loads,
the sections of all structural elements, and consequently the over-
all cost of the structure. Furthermore, the reduced mass will reduce
the lateral load that will be imposed on the structure during earth-
quakes, hence simplifying and reducing the lateral load carrying
system. That is why the global cost of the structure may be reduced
in case of using LWC rather than using NWC. The high workability
of the proposed PF-LWC reduces extensively the cost of compac-
tion. It possesses high ability to flow under its own weight, filling
formwork even in the presence of congested reinforcement or
complicated structure (as in most of bridge types) yet producing
⇑ Corresponding author. Tel.: +20 1227162032, +1 775 688 9801.
E-mail address: ahfarghal@hotmail.com (A. Farghal Maree).
a dense and homogenous hardened concrete. The fluidity and seg-
regation resistance of such type ensures a high level of homogene-
ity, minimal concrete voids and uniform concrete strength,
providing the potential for a superior level of finish and durability
to the structure.
The bond mechanism between deformed bars and concrete is
highly influenced by multiple parameters such as the strength of
surrounding material structure and its properties, the thickness
of concrete cover, amount of confining reinforcement, develop-
ment and splice length, the occurrence of splitting cracks in the
concrete cover, the reinforcement steel stress, lateral pressure,
direction of casting, reinforcing bar diameter and bar spacing [1].
LWC is considered more brittle than NWC, which might increase
the risk of splitting cracks and de-lamination of the concrete cover.
Since 2005, an extensive experimental and theoretical research
program, aiming to developing structurally and economically
efficient PF-LWC as well as establishing design guidelines for all
types of structural elements made using this material, is being
undertaken at the Department of Structural Engineering of Ain
Shams University.

0141-0296/$ - see front matter Published by Elsevier Ltd.

http://dx.doi.org/10.1016/j.engstruct.2013.10.015
2 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11
The first phase of this program resulted in a new kind of light-
weight concrete, which combines the advantages of normal den-
sity concrete, cellular concrete and high workability concrete,
through partially replacing the normal weight aggregates with
polystyrene foam particles, hence, leading to concrete’s unit
weight reduction while maintaining adequate strength.
For any reinforced concrete element, there should exist a suffi-
cient bond between the concrete and the reinforcing bars to allow
for force flow between the two materials. The significance of this
research appears from the ability to evaluate the bond properties
of LWC with ordinary deformed steel rebars [2]. The importance
of conducting an experimental and analytical investigation to iden-
tify the bond characteristic of PE-LWC is to allow for safe, reliable
and efficient utilisation of PF-LWC. The full understanding of the
behaviour of PF-LWC with steel rebars will encourage the confi-
dent implementation in the field by engineering community.
The objective of this experimental investigation is to evaluate
the bond properties, including the bond strength, of PF-LWC with
steel rebars using pullout specimens [3] and beam-end specimens
[4]. The investigation focused on studying the bond of deformed,
straight, unanchored steel rebars with PF-LWC in flexural members
and examining the applicability of the available codes design
expressions, available international bond models and some other
reliable and accurate descriptive equations to PF-LWC.
2. Constituent materials and mix proportioning
Two concrete mixes were used for the specimens, the first mix
was for the NWC specimens and the second mix was for the LWC
specimens. In order to achieve the high workability LWC; polysty-
rene foam, silica fume and super plasticizer were added to the mix.
The concrete mix consisted of natural sand as fine aggregate, fine
crushed stone of nominal maximum size of 10 mm as coarse aggre-
gate, fresh ordinary Portland cement and tab water. Super plasti-
cizer was added to the LWC mix in order to achieve a common
target compressive strength as that of NWC. The mix proportions
for NWC and PF-LWC in addition to the target compressive
strength and the resulted density are shown in Table 1.
Steel rebars of grade 240/350 and grade 400/600 were used. The
mild steel smooth rebars of grade 240/350 having minimum yield
strength of 240 MPa and ultimate tensile strength of 350 MPa were
used for bar size of 6 and 8 mm. The high tensile steel deformed re-
bars of grade 400/600 having minimum proof strength of 400 MPa
and ultimate tensile strength of 600 MPa were used for bar sizes of
12, 16 and 22 mm.
3. Experimental work program
The experimental program is carried out to investigate the bond
behaviour of PF-LWC. The program incorporated two phases in
which specimens were designed and cast to study the desired
properties. The first phase is performed on the standard pullout
specimens to compare the results with commonly conducted bond
testing and to abstract the bond slip curve of the standard pullout
test specimens. The second phase deals with a reduced beam-end
specimen to assess the behaviour of bond between reinforcement
Table 1
Concrete mixes’ proportions/m3 and obtained properties.
Concrete Cement Silica fume Coarse aggregate
mix (kg) (kg) (kg)
NWC 400 – 1200
PF-LWC 450 40 630
bars and LWC in flexure members in both cases; confined with stir-
rups and unconfined [2].
3.1. Description of test specimens
3.1.1. Pullout specimens
This group consisted of 9 sets of cubes having side length of
150 mm a steel rebar embedded during casting as a standard pull-
out specimen according to ASTM C234 [3]. These specimens were
tested in a direct pullout test to determine their bond strength
and their bond stress versus slippage responses. The dimensions
of the specimens were chosen according to ASTM C234 [3] speci-
men so that comparison with various researches is applicable.
Three parameters were assessed in this investigation; the concrete
type (NWC and LWC), the rebar diameter (12 mm, 16 mm and
22 mm) and the bonded length (2T, 3T and 4T). Table 2 and
Fig. 1 summarise the number, details, coding and specifications
of these specimens.
3.1.2. Beam-end specimens
This phase consisted of eight medium scale beam-end speci-
mens as shown in Fig. 2. The figure first shows that the beam-
end specimen had a rectangular cross-section of dimensions
150 mm  450 mm over a total length of 700 mm representing a
beam-end anchorage zone as shown in Fig. 3. These reduced spec-
imens totally reflect the behaviour of bond between rebars and
concrete in flexure members which was abstracted from ASTM
A944 [4]. In this respect, specimens EL-T12 and EN-T12 had the
same test bar T12 but with different concrete mixes. In addition
to specimens EL-T16 and EN-T16 had the same test bar T16. Simi-
larly specimens EL-T12CON, EN-T12CON, EL-T16CON and EN-
T16CON had additional confining reinforcement. EL-T12CON and
EN-T12CON had confining reinforcement of T6 with 120 mm spac-
ing all over the specimens’ length and T8 with 120 mm spacing for
specimens EL-T16CON and EN-T16CON. The previous illustration
clearly pronounces the parameters of the study; the concrete type
through having NWC and PF-LWC mixtures, bar size through using
T12 and T16, finally, the confining reinforcement through using
stirrups T6 and T8 with 120 mm spacing. The coding, concrete
dimensions and reinforcement details of the aforementioned spec-
imens are also summarised in Table 3.
The steel test bars were prepared using watering hose as a bond
breaker also end of test bars were threaded using lathing machine.
The threaded end of the test bar was fixed to the test setup using
washer and nuts.
3.2. Instrumentation
The accuracy of the measurement devices determines the reli-
ability of the experimental program results. In this respect, accu-
rate measurement devices were used to monitor and record the
test outputs. The used LVDTs were directly connected to data
acquisition; where the data is automatically collected every time
interval. The time interval chosen during the tests was 0.1 s.
Sand Polystyrene Super plasticizer Water Obtained Target fcu
(kg) foam (l) (l) (l) density (kN/m3) (MPa)
600 – – 230 22.5 27
630 330 13.5 139 18.5 24
 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11 3

Table 2 3.2.1. Test setup and loading scheme

Specifications of the tested pull-out specimens. The used test setup developed pulling of the bar by applying
Specimen ID Concrete mix Rebar Bond length (mm) Number compression on the beam-end concrete block and fixing the bar in-
PL-4T12 LWC T12 4T = 48 3 stead of applying direct tension to the bar and fixing the block. The
PL-4T16 LWC T16 4T = 64 3 test setup was designed to maintain safety during loading and up
PL-4T22 LWC T22 4T = 88 3 to failure of specimens. Schematic drawing of the test setup is
PN-4T12 NWC T12 4T = 48 3 shown in Fig. 4. The test setup consisted of a rigid horizontal steel
PN-4T16 NWC T16 4T = 64 3
PN-4T22 NWC T22 4T = 88 3
beam (2 UPN 120) used to fix the threaded part of the test bar
PL-3T12 LWC T12 3T = 36 2 using washer and nut. A gusset plate of 20 mm thickness is rested
PL-3T16 LWC T16 3T = 48 2 on the edge of the beam-end concrete specimen. This plate is used
PL-2T16 LWC T16 2T = 32 2 to transfer the load from the hydraulic jack to the concrete speci-
men through a steel column (2 UPN 160). A bearing cylinder is
used to support the specimen from the lower side laterally. In addi-
tion a cylindrical screw is used to support the specimen from the
upper side. Teflon pads were used to reduce the friction between
the specimen and lateral supporting elements. Using both supports
the loading scheme of the reduced beam-end specimen is provided
as shown in Fig. 4. Also two cylindrical safety screws were used to
maintain safety of the test setup at sudden failure. The relative
slippage was measured from the difference of LVDT1 and LVDT2
readings.

3.2.2. Determination of design bond strength

Park and Paulay explained the determination of usable bond
strength that traditionally the bond performance of various
Fig. 1. Details of the tested cubes. reinforcing bars, embedded in concrete of different strengths, has

Fig. 2. Details of the tested beam-end specimens.

Fig. 3. Beam-end specimen reduced from full scale bond beam specimen.
4 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11

Table 3
Specifications of the tested beam-end specimens.

Specimen ID Conc. mix b (mm) t (mm) Rebar Bond length (mm) Confining rebars
EL-T12 LWC 150 450 T12 48 –
EL-T12-CON LWC 150 450 T12 48 T6@120
EL-T16 LWC 150 450 T16 64 –
EL-T16-CON LWC 150 450 T16 64 T8@120
EN-T12 NWC 150 450 T12 48 –
EN-T12-CON NWC 150 450 T12 48 T6@120
EN-T16 NWC 150 450 T16 64 –
EN-T16-CON NWC 150 450 T16 64 T8@120

Fig. 4. Schematic drawing of test setup for beam-end specimens.

been determined from pullout tests. Generally the bars were pulled
from the surrounding concrete in such a way that transverse com-
pression against the bar was also included. This transverse com-
pression had a beneficial effect on the bond strength and was not
therefore typical of situation encountered in structures [5]. For this
reason various forms of test specimen have been proposed to elim-
inate transverse compression. The usable bond strength is seldom
a given fraction of the ultimate bond strength developed in a par-
ticular pullout test. Generally, the slip at the loaded or unloaded
end of the bar governs the ‘‘critical’’ bond intensity that can be
developed under working load conditions, since this slip will sig-
nificantly affect crack widths. It is important, therefore, that the
full load–slip history be determined when such a test is carried
out. According to Mathey and Watstein, the ‘‘critical’’ bond stress
may be defined as the smaller of the bond stresses associated with
either a free-end slip of 0.05 mm or a loaded-end slip of 0.25 mm in
beam tests. This slip can be considerably affected by the position of
the bars when cast [6].
In pullout tests some slip at the free end of top-cast bars has
been observed before the development of significant strength or
of any cracking. The slip at the loaded end of an embedded bar
(i.e., at the face of a crack) is largely governed by the concentration
of bond stresses in its immediate vicinity. An increase of embed-
ment length and a consequent lowering of the average bond stress
has little effect on the slip at the loaded end before it exceeds
0.25 mm. The limitation on slip at the loaded end, taken as one-half
an acceptable crack width of 0.5 mm, may set an upper limit to the
usable strength of large bars. Also Leonhardt defined the maximum
allowable slippage to be 0.1 mm under service loads in order to
minimise generated cracking at concrete surface [7]. The limitation
of slippage as one-half an acceptable crack width of 0.2 mm was
set as an upper limit to the usable strength. This crack width is
the limiting value in several international codes such as Building
Code Requirements for Structural Concrete (ACI 318-05) [8] and
Egyptian Code of Practice [9] for most of the environmental catego-
ries. As the maximum slippage is not the governing design value,
the bond stress corresponding to 0.1 mm slippage was determined
for each experimented specimen, representing the allowable work-
ing slippage between reinforcing bars and concrete.
4. Experimental results and discussions
4.1. Results of pullout test specimens
This section presents the results of phase one of the experimen-
tal program. The bond stress–slip relation for standard pullout test
specimens for various bar diameters having different bond length
and using two types of concrete; LWC and NWC was obtained.
The bond slip relationship for the three test specimens PL-4T12
is illustrated in Fig. 5 where, the first ascending part refers to the
stage in which the ribs penetrate into the mortar matrix, character-
 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11
Fig. 5. Bond stress–slip curves for specimens PL-4T12.
ised by local crushing and microcracking. The decreasing branch
refers to the reduction of bond resistance due to the occurrence
of splitting cracks along the bar. For each specimen the failure load
and corresponding slippage and the failure mode was determined.
As the maximum slippage is not the governing design value [7], the
bond stress corresponding to 0.1 mm slippage was determined for
each specimen. The common failure mode of most specimens was
splitting failure as shown in the physical photos of tested specimen
PL-4T16 in Fig. 6.
Results of experimental program phase one are summarised in
Table 4. Specimen code, bond strength and corresponding slippage,
mode of failure and bond stress corresponding to 0.1 mm slippage
are presented in the table.
The measured slippage for NWC specimens is lower than the
slippage for LWC specimens for all test bar diameters T12, T16
and T22. This is due to the higher bond strength for LWC than
NWC and the lower modulus of elasticity for LWC than NWC as
presented in the current British Concrete Code [10] and Building
Code Requirements for Structural Concrete (ACI 318-05) and Com-
mentary (ACI 318R-05) [8].
In order to compare the test results, the bond stress ratio
according to Eq. (1) has been calculated for all specimens. This ratio
is driven from the current British Concrete Code [10] and Egyptian
Code [9] equations for the concrete bond strength. It is also inde-
pendent of all investigated variables so it simplifies the comparison
between investigated variables affecting the bond strength to a
common base.
Fig. 6. Photo for mode of failure of specimen PL-4T16.
5
s
Bond Stress Ratio ðbÞ ¼ pffiffiffiffiffi ð1Þ
fcu
In which s is the measured ultimate bond stress and fcu is the con-
crete cube compressive strength. This ratio was selected both to al-
low for variations in the concrete strength between NWC and LWC
and to simplify the comparisons down to a common base. Further-
more, this ratio is considered constant and independent of all vari-
ables. This value of bond stress ratio has been plotted against the
other independent variables such as bar diameter and concrete
type.
4.1.1. Effect of bar diameter on bond strength
Test results shown in the following figures illustrate the effect
of the investigated variables on the bond strength. Bar diameter
is one of the major variables that affect the bond strength. Fig. 7
shows the bond stress ratio plotted against bar diameter for the
two types of concrete mixes NWC and LWC with bond length equal
to four times the test bar diameter.
A reduction in the bond stress ratio can be noticed with the
increase in bar diameter. Also, the bond stress ratios of LWC
specimens are higher than those of NWC. This result is due to
the higher tensile strength of LWC than that of NWC having the
same compressive strength as illustrated by Okail in [11], due to
lower water/cement ratio in addition to using super plasticizer
admixtures in LWC that the cement paste in case of LWC possesses
higher tensile strength. This result was obtained by Okail through
testing LWC cylinders in the splitting tensile strength test. As the
mode of failure for all specimens with equal bond length of four
times the bar diameter was splitting mode of failure, the increase
in tensile strength for LWC caused an increase in the bond
strength.
Fig. 8 shows the bond stress ratio corresponding to 0.1 mm slip-
page against bar diameter for the two types of concrete mixes NWC
and LWC with bond length equal to four times the test bar diame-
ter. A reduction in the bond stress ratio corresponding to 0.1 mm
slippage is observed with the increase in bar diameter. But the
bond stress ratio corresponding to 0.1 mm slippage of LWC is low-
er than those of NWC. This is due to the higher slippage and lower
elastic modulus of LWC than that of NWC.
4.1.2. Effect of bond length on specimens’ failure mode
The failure load of specimens with bond length less than four
times the bar diameter is nearly equal to their opposites having
four times the bar diameter bond length. On the other hand, the
obtained bond stress of the specimens having bond length less
than four times the bar diameter is higher than their opposites
having four times the bar diameter bond length. The average bond
strength of specimens PL-3T12 is 26.85 MPa, whereas for speci-
mens PL-4T12 having the same concrete mixture (LWC) and the
same test bar (T12) but with longer bond length equal to 48 mm
instead of 36 mm have an average bond strength of 22.01 MPa.
This presents an increase of 22% when using shorter bond length
in specimens PL-3T12. Also, specimens PL-3T16 and PL-2T16 hav-
ing an average bond strength of 19.31 MPa, whereas for specimens
PL-4T16 is 15.67 MPa which presents an increase in bond strength
of 23.23%. This increase in both cases when using T12 and T16 is
due to the mode of failure. As specimens PL-3T12, Pl-3T16 and
PL-2T16 reached their ultimate bond capacity in the pull out fail-
ure mode, where the test bar slides along a shearing cylindrical
plane of failure after crushing and shearing off the concrete keys
enclosed between the bar ribs.
However, for the other specimens the splitting mode of failure
occurred. This results from the splitting of the concrete cover sur-
rounding the bar. This phenomenon of cover splitting happens
when the tangential tensile stress in the concrete cover, caused
6 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11

Table 4
Results of pull-out test specimens.

Specimen ID Strength (so) (MPa) Slippage (So) (mm) Stress at 0.1 mm (MPa) Mode of failure
PL-4T12 19.80 0.207 10.24 Splitting
23.62 0.227 10.60 Splitting
22.31 0.211 7.70 Splitting
PL-4T16 15.60 0.235 6.67 Splitting
16.15 0.255 6.70 Splitting
16.15 0.237 12.21 Splitting
PL-4T22 17.26 0.426 4.80 Splitting
15.53 0.367 4.06 Splitting
16.40 0.476 5.83 Splitting
PN-4T12 19.60 0.240 13.60 Splitting
20.60 0.195 14.85 Splitting
19.20 0.186 17.23 Splitting
PN-4T16 14.30 0.190 10.25 Splitting
13.26 0.160 9.18 Splitting
18.03 0.258 9.88 Splitting
PN-4T22 16.93 0.350 6.22 Splitting
16.10 0.423 7.13 Splitting
14.82 0.310 5.62 Splitting
PL-3T12 27.20 0.230 18.48 Pull out
26.50 0.210 18.78 Pull out
PL-3T16 19.64 0.410 14.97 Pull out
19.40 0.467 12.63 Pull out
PL-2T16 19.54 0.313 14.48 Pull out
18.64 0.260 15.66 Pull out

4.2. Results of beam-end specimens

This section presents the results of phase two of the experimen-

tal program. The tested beam-end specimens were casted using
two types of concrete mixes (LWC and NWC) also various test bars
were used T12 and T16. Finally, half of the specimens were con-
fined using stirrups and the second half were unconfined.
The results of experimental program phase two are summarised
in Table 5 where specimen code, bond strength and corresponding
slippage, mode of failure and bond stress corresponding to 0.1 mm
slippage are shown.
For all test specimens, the bond stress ratio (Eq. (1)) has been
plotted against the other independent variables such as bar diam-
eter, concrete types and presence of confining stirrups.

4.2.1. Effect of bar diameter on bond strength

Fig. 7. Obtained bond stress ratio for various bar diameters. A reduction in the bond stress ratio can be noticed with the in-
crease in bar diameter. Also, the observed bond stress ratios of
LWC specimens are higher than those of NWC. This result is due
to the higher tensile strength of LWC than that of NWC as illus-
trated in [11]. As the mode of failure for all specimens was splitting
mode of failure, this increase in tensile strength for LWC will cause
an increase in the bond strength.

4.2.2. Effect of confining stirrups on bond strength

The confined specimens – LWC or NWC – possessed higher
bond strength rather than the unconfined ones as shown in
Fig. 9. An increase of about 4% was observed.

5. Analytical model

5.1. Proposed modified bond model

Fig. 8. Obtained bond stress ratio corresponding to 0.1 mm slippage for various bar
diameters.
by the radial outward pressure, exceed the tensile strength of the
concrete. This results in lower bond strength which is more critical
in design [12].
In the bond law according to CEB-FIP 1990 model code [1], bond
slip relationship is presented by a set of equations. The laws pre-
sented in the model are generated based on several parameters;
the concrete compressive strength, reinforcement bar diameter,
reinforcement type (smooth and deformed bars), confinement con-
ditions and the quality of concrete casting. The bond law according
CEB-FIP 1990 model code shown in Fig. 10 is described by the
 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11 7

Table 5
Results of beam-end specimens.

Specimen ID Strength (so) (MPa) Slippage (So) (mm) Stress at 0.1 mm (MPa) Mode of failure
EL-T12 23.51 0.964 7.29 Splitting
EL-T12-CON 24.62 0.975 8.02 Splitting
EL-T16 20.55 0.974 4.01 Splitting
EL-T16-CON 21.21 0.972 4.61 Splitting
EN-T12 19.06 0.785 8.97 Splitting
EN-T12-CON 21.94 0.815 9.41 Splitting
EN-T16 18.35 0.697 7.27 Splitting
EN-T16-CON 19.26 0.811 7.35 Splitting

occurs only for confined concrete, referring to advanced crushing

and shearing off of the concrete between the ribs, where the maxi-
mum slippage of this constant part is (S2). The decreasing branch re-
fers to the reduction of bond resistance due to the occurrence of
splitting cracks along the bar, where the maximum slippage of this
decreasing branch is (S3). The horizontal part represents a residual
bond capacity, which is maintained by virtue of a minimum trans-
verse reinforcement, keeping a certain degree of integrity intact.
Haskett et al. [13] developed a new technique, which was used
to derive a general local bond stress–slip relationship from test re-
sults. Their new technique coincides with the developed bond
stress–slip relationship by Eligehausen et al. [14]. Importantly, it
suggests that the CEB Mode Code 1990 local bond stress–slip rela-
tionship is relatively accurate. The magnitude of the peak bond
stress suggested by CEB Mode Code 1990 accurately models the
peak loads attained experimentally.
Fig. 9. Comparison between bond stress ratio of different confined and unconfined
beam-end specimen.
This model has several application of the bond stress–slip
relation such as reinforced concrete tie and anchorage of bars. As
for bars with relatively large rib area the bond strength smax
increases and the characteristic slip value S1 decreases, which is
neglected in CEB-FIP 1990 bond model [1]. In order to count for
the observed effect of bar diameter and concrete type on the bond
stress–slip relationship, the defining parameters of the bond
stress–slip curve were modified for NWC and LWC. The defining
parameters of the CEB-FIP 1990 bond stress–slip curve were
modified for NWC and LWC using the best fitting technique to
the experimental results and are summarised in Table 6. The char-
acteristic slippage (S1), bond strength (smax) and the power (a) are
the modified parameters for various types of concrete, NWC and
LWC.
For unconfined concrete with good bond condition, several
modified parameters defining the mean bond stress–slip relation-
ship are introduced in Table 6. First the characteristic slip value
(S1), which is modified from 0.6 mm to 0.11 e0.054/ depending on
Fig. 10. Bond slip law by CEB-FIP model code 1990 [1]. bar diameter /, as the characteristic slip value increases with
increasing the bar diameter. Another increase in the characteristic
slip value is obtained in case of using LWC due to its lower modu-
following set of equations. In order to analyse the bond behaviour
lus of elasticity causing higher deformation where characteristic
of concrete, some modifications are suggested on the defining
slip (S1) becomes 1.11  0.11e0.054/. The second parameter illus-
parameters of CEB-FIP 1990 bond model for both NWC and LWC.
trated is the power (a) where it is increased to 0.55 in case of using
 a LWC instead of 0.40 in case of using NWC, that is also due to the
s
s ¼ smax ; 0 6 S 6 S1 ð2Þ lower elasticity of LWC than NWC result in higher curvature in case
s1
of NWC. The third parameter defined in the modified proposed pffiffiffiffi
s ¼ smax ; S1 < S 6 S2 ð3Þ model is the bond pffiffiffiffi strength which is increased from 2 fc to
0
0:35
10:4  kL  / fc depending also on bar diameter, as the bond
0

  strength increases with decreasing the bar diameter. This increase

s  s2
s ¼ smax  ðsmax  sf Þ ; S 2 < S 6 S3 ð4Þ in bond strength is obviously due to the relative large rib area for
s3  s2
used bars [12]. Another increase in the bond strength is obtained
in case of using LWC due to its higher tensile strength as illustrated
s ¼ s j ; S3 < S ð5Þ
previously, where the bond strength increases by a magnification
In this figure the first ascending non-linear part refers to the stage factor (kL) equals to 1.11.
in which the ribs penetrate into the mortar matrix, characterised The proposed bond stress–slip relationship possesses a higher
by local crushing and microcracking. The maximum slippage of this bond strength smax and lower characteristic slippage S1 than that
ascending part is the characteristic slippage (S1). The constant part presented by CEB-FIP bond model [1], that is due to using
8 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11

Table 6
Modified parameters for defining the mean bond stress–slip relationship according to Eqs. (2)–(5) for both NWC and LWC.

CEB-FIP Model Code 1990 Modified for NWC Modified for LWC
0:054/
S1 (mm) 0.6 mm 0:11  kL  e 0:11  kL  e0:054/
S2 (mm) 0.6 mm 0:11  kL  e0:054/ 0:11  kL  e0:054/
S3 (mm) 1.0 mm 1.0 mm 1.0 mm
a 0.4 0.40 0.55
pffiffiffiffi pffiffiffiffi pffiffiffiffi
smax (MPa) 2 fc0 10:4  kL  /0:35 fc0 10:4  kL  /0:35 fc0
sf (MPa) 0.15 smax 0.15 smax 0.15 smax

deformed bars with relatively large rib area. One of the major
parameters affecting the bond stress–slip relationship is the geo-
metrical dimensions of the deformed bars’ ribs, which can be pre-
sented by the relative rib area of reinforcing bars. Relative rib area
(fR) is defined as area of the projection of all ribs on a plane perpen-
dicular to the longitudinal axis of the bar, rod or wire, divided by
the rib spacing and the nominal circumference [15]. The relative
large ribs size in the available deformed bars in Egypt causes an in-
crease in bond stress and reduction in slippage. The properties of
the deformed bars used in the experimental work are summarised
in Table 7, where bar diameter, rib height, rib spacing and relative
rib area are illustrated in the table. A photo of the ribbed bars (T12,
T16 and T22) used in the experimental work is shown in Fig. 11.
The international codes determine a minimum value for relative
rib area of deformed bars such as Euro code [16]. It states in Annex
(C) that the minimum relative rib area is 0.04 for bars with diam-
eter less than 12 mm and 0.056 for bigger bar diameters. This min-
Fig. 11. Ribbed bars used in experimental work.
imum values for relative rib area according to [16] are shown in
Table 7. These values illustrate the difference in the relative rib
area of the used bars, where it is greater by 160%, 80% and 89%
for bars T12, T16 and T22, respectively.
Fig. 12 shows the proposed modified bond models for NWC and
LWC compared with the original CEB-FIP Model Code 1990.

5.2. Verification of proposed modified bond model

A comparison between the experimental results and the pro-

posed model is conducted to verify the proposed modified bond
model illustrated previously. These experimental results include
both obtained ones from self-performed experimental work and
other available literature experimental results in the same scope.
The verification includes comparison between the proposed bond
slip relationship with the experimental results of pullout NWC
specimens and pullout LWC specimens. Figs. 13–15 show the com-
parison between bond slip relationship of experimental pullout
specimens and modified proposed model for NWC specimens of
bar diameters 12 mm, 16 mm and 22 mm, respectively. In addition
the proposed model is compared to the experimental work for LWC
specimens of bar diameters 12 mm, 16 mm and 22 mm in Figs. 16–
18 respectively.
Xiao and Falkner studied bond behaviour between recycled
aggregate concrete (RAC) and steel rebars where three recycled
coarse aggregate (RCA) replacement percentages (i.e., 0%, 50%
and 100%) and two types of steel rebar (i.e., plain and deformed)
were considered in this study. Based on the test results obtained Fig. 12. Proposed modified bond models for NWC and LWC and the original CEB-FIP
from pullout test, the influences of both RCA replacement Model Code 1990.

Table 7
Properties of the deformed bars used in the experimental work.

Test bar Rib height (mm) Rib spacing (mm) Relative rib area (fR) Minimum allowable relative rib area [14]
T12 1.20 7.8 0.104 0.040
T16 1.60 10.7 0.101 0.056
T22 2.20 14.2 0.106 0.056
 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11 9

Fig. 13. T12 NWC specimens. Fig. 16. T12 LWC specimens.

Fig. 14. T16 NWC specimens.

Fig. 17. T16 LWC specimens.

Fig. 15. T22 NWC specimens. Fig. 18. T22 LWC specimens.

percentages and the rebar surface type on the bond strength concrete type used in this specimen was NWC with 0% replace-
between the RAC and steel rebars were investigated. The experi- ment of recycled coarse aggregate (RCA).This pullout specimen
mental results of specimen RAC-II-0 presented in their work were was a cube of side length 100 mm with an embedded deformed
used for comparison with the proposed modified bond model. The test bar of 10 mm diameter. The bond length of the test bar was
10 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11
Fig. 19. Comparison between bond slip relationship of RAC-II-0 presented in [17]
and modified proposed bond model.
Fig. 20. Comparison between bond slip relationship of RAC-II-0 presented in [17]
and experimental work of NWC specimens.
five times the bar diameter. The cube compressive strength of the
pullout specimen was 34.4 MPa. Four pullout specimens with the
previously described specifications were presented in their work,
which were used in verification of the proposed modified bond
model. The bond slip relationship of the four specimens RAC-II-0
presented in [17] is plotted against the proposed modified model
as shown in Fig. 19. This relation shows a good correlation in the
pre-failure zone between proposed bond slip relationship and
experimental work results of specimens RAC-II-0. Also Xiao and
Falkner work were compared to the experimental work of NWC
pullout specimens as shown in Fig. 20 and its shows also a good
correlation in the pre-failure zone.
6. Conclusions
1. Modified bond–slip relationship similar to CEB-FIP model
[1] for unconfined concrete with good bond conditions,
but considering the effects of bar diameter and concrete
type was proposed. This model was verified to yield predic-
tions in good correlation with experimental results con-
ducted in this study and others available in the relevant
literature.
2. The proposed model successfully incorporates the above
mentioned effects in a single combined expression for the
characteristic slip value (S1) to be 0:11  kL  e0:054/ lieu
of the constant value of 0.6 mm in the p original
ffiffiffiffi expression
pffiffiffiffi
and bond strength of 10:4  kL  /0:35 fc0 in lieu of 2 fc0
of the original CEB-FIP model. Where kL is a magnification
factor for LWC that equals 1.11 instead of being unity for
NWC.
3. The curvature of bond–slip relationship in LWC is found to
be lower than the corresponding one of NWC. This is
reflected in the power (a) of the proposed bond–slip rela-
tionship, to have value of 0.55 in case of LWC in lieu of
0.40 in case of NWC, which could be attributed to the lower
elastic modulus of LWC.
4. The major parameter affecting the bond stress–slip rela-
tionship is the geometric characteristics of the deformed
bars’ ribs, which is evaluated through the relative rib area
of reinforcing bars. The proposed bond stress–slip relation-
ship possesses a higher bond strength smax and lower char-
acteristic slippage S1 than that presented by CEB-FIP bond
model, due to the use of deformed bars with relatively large
rib area in Egypt.
5. The slippage of LWC is higher than that of NWC, which also
increases slightly with the increase in bar diameter.
6. The bond stress ratios of LWC specimens are found to be
higher than those of NWC due to the higher tensile strength
of LWC. However, the bond stress ratio corresponding to
0.1 mm slippage of LWC is lower than that of NWC. This
stress reduction could be attributed to the lower elastic
modulus of LWC than that of NWC.
7. Mode of failure of bond specimens significantly affects the
bond strength. Specimens of the pull-out failure mode pos-
sess higher bond strength than those of splitting mode of
failure.
8. Specimens having bond length less than four times the bar
diameter consistently yielded pull-out failure, whereas
specimens with bond length of four times the bar diameter
usually showed splitting failure.
9. The confined beam-end specimens with stirrups made of
LWC or NWC yielded higher bond strength rather than
the unconfined ones.
Acknowledgements
Authors would like to express deep appreciation to Prof. Ahmed
Sherif Essawy, Professor of Concrete Structures, Faculty of Engi-
neering, Ain Shams University, for his guidance and valuable sug-
gestions throughout the research. As well as his experienced
advice, continuous support and deep encouragement through all
phases toward completion of M.Sc. thesis of the first author [2].
References
[1] Comite EURO – International du Beton. CEB-FIP model code 1990 for concrete
structures. Thomas Telford, London; 1993. 437p.
[2] Farghal A. Bond behavior of structural lightweight concrete. M.Sc. Thesis, Ain
Shams University; 2012. 247p.
[3] American Society for Testing and Materials. ASTM C 234: standard test method
for comparing concrete on the basis of bond development with reinforcing
steel. In: Annual book of ASTM; 2002.
[4] American Society for Testing and Materials. ASTM A 944: standard test method
for comparing bond strength of steel reinforcing bars to concrete using beam-
end specimens. In: Annual book of ASTM; 2005. p. 566–9.
[5] Park R, Paulay T. Reinforced concrete structures. New Zealand: John Wiley and
Sons, Chritchurch; 1975. p. 769.
[6] Mathey R, Watstein D. Investigation of bond in beam pull-out specimens with
high yield strength deformed bars. ACI Struct J 1961;57(9).
 A. Farghal Maree, K. Hilal Riad / Engineering Structures 58 (2014) 1–11
[7] Leonhardt F, Mönnig E. Vorlesungen über Massivbau: Teil 3: Grundlagen zum
Bewehren im Stahlbetonbau. German edition; 1977.
[8] ACI Committee 318. Buiding code requirements for structural concrete
(ACI318-11) and commentary (ACI 318R-11). American Concrete Institute.
MI, USA; 2011. 479p.
[9] Egyptian Code of Practice Committee ECP, 203. Egyptian code for design and
construction of concrete structures. HBRC, Cairo; 2007. 454p.
[10] British Standards BSI. BS 8110: structural use of concrete, Part 1: Code of
practice for design and construction. BSI, GB; 1997. 168p.
[11] Okail H. Flexural behavior of reinforced self-compacting lightweight concrete
beams. M.Sc. Thesis, Ain Shams University; 2008. 124p.
[12] RILEM/CEB/FIP. Bond test for reinforcing steel: 2. Pullout test.
Recommendation RC 6; 1978.
11
[13] Haskett M, Ochlers D, Ali M. Local and global bond characteristics of steel
reinforcing bars. Eng Struct J 2008;30:376–83.
[14] Eligehausen R, Popov EP, Bertero VV. Local bond stress–slip relationship of
deformed bars under generalized excitations. UCB/EERC-83/23. Berkeley:
Earthquake Engineering Research Center, University of California; 1983.
[15] Sistonen E, Tukiainen P, Huovinen S. Bonding of hot dip galvanised
reinforcement in concrete. Finland; 2001. <http://www.tekna.no/ikbViewer/
Content/225411>.
[16] European Committee for Standardization. EN 1992-1-1: Design of concrete
structures, Part 1–1: General rules and rules for buildings. European Standard,
ICS 91.010.30; 91.080.40; 2004.
[17] Xiao J, Falkner H. Bond behaviour between recycled aggregate concrete and
steel bars. Constr Build Mater J 2007;21:395–401.