Magazine of Concrete Research, 2004, 56, No.

9, November, 545–557

Bond of high-strength concrete under

monotonic pull-out loading
M. Alavi-Fard* and H. Marzouk†

Grand Canada Engineering; Memorial University of Newfoundland, Canada

An experimental investigation was conducted on high-strength concrete specimens to examine the bond strength
characteristics under monotonic loading. The range of compressive strengths tested was between 70 and 95 MPa.
The influences of load history, confining reinforcement, bar diameter, concrete strength, reinforcement spacing and
rate of pull out were investigated experimentally. The internal concrete strains close to the contact surface and also
the steel strain were measured. The test set-up, load application, instrumentation and measurement, and test
procedure were designed to measure strains and deformations. Several specimens with reinforcement bar diameters
of 20, 25 and 35 mm were tested. The test results revealed that the bond strength of high-strength concrete is higher
than the corresponding normal strength concrete. However, the bond behaviour of high-strength concrete is more
brittle in comparison with normal strength concrete. The concrete strains were measured around the steel reinforce-
ment. Concrete strain measurements are useful to identify the internal crack pattern and to predict possible failure
modes. The area under the curve of the bond stress–slip curve can define the bond energy. The bond energy is
related to ductility and can be used along with the bond strength in evaluating the bond behaviour of high-strength
concrete.

vations of secondary cracks are reported, as well as the

Introduction
High-strength concrete is used mostly in the con-
struction of bridges, high-rise buildings, marine and
offshore structures. Bond strength between high-
strength concrete and reinforcement is an important
factor in designing any reinforced concrete structure
under various kinds of loading. Therefore, this study
was conducted to investigate bond behaviour between
high-strength concrete and steel reinforcement, and to
determine the internal stress and strain along the rein-
forcement interface with high-strength concrete. The
experimental data are valuable to understand the bond
behaviour of high-strength concrete.
Several experimental and theoretical investigations
1
were conducted by Somayaji and Shah and Jiang et
2
al., on the behaviour of bond for normal strength con-
crete. Improved tools for measurement of local bond
and local slip were introduced and applied. The obser-
* President, Grand Canada Engineering, Ottawa, Canada K1P 5G4.
† Faculty of Engineering and Applied Science, Memorial University
of Newfoundland, St. John’s, NF, Canada A1B 3X5.
(MCR1185) Paper received 26 September 2003; last revised 21 May
2004; accepted 9 August 2004
distribution of strain in concrete in the vicinity of the
3
reinforcing bar. Darwin et al. studied development
length criteria for conventional and high relative rib
area of reinforcement. On the basis of a statistically
based expression, the development reinforcement
length and splice strength of reinforcement for concrete
with strengths between 17 and 110 MPa, with and with-
out confining reinforcement, was investigated. The
effects of cover, spacing, development/splice length,
geometric properties of the development and spliced
reinforcement were incorporated into their design equa-
tion.
4,5
Azizinamini et al. examined bond performance of
reinforcing bars and tension development length of
reinforcing bars embedded in high-strength concrete.
The effects of concrete compressive strength, splice
length, and casting position on the bond strength of
reinforcing bars have been studied. It was concluded
that in the case of high-strength concrete, increasing
the tension development length (or equivalent tension
splices) was not an efficient way of increasing the bond
capacity of deformed reinforcing bars, especially when
6
the concrete cover is small. Azizinamini et al. indi-
cated that when calculating the tension development
length of high-strength concrete and tension splice, a
545

0024-9831 # 2004 Thomas Telford Ltd

Alavi-Fard and Marzouk

minimum number of stirrups should be provided over

the splice region. Based on the 70-beam specimens an
expression is recommended to calculate the extra num-
7
ber of stirrups required. De Larrard et al. have investi-
gated the effect of bar diameter on bond strength in
high-performance concrete. It was concluded that bond
capacity increases with the tensile strength of the con-
crete and it is at a higher rate with smaller reinforce-
ment. It was also found that the bond is greater for
smaller bar diameters than for larger bar diameters.
8
Eligehausen et al. conducted one of the main com-
prehensive investigations on the effect of the bar dia-
meter embedded in normal-strength concrete. It was
concluded that the maximum bond capacity decreased
slightly with the increasing bar diameter. The frictional
bond resistance was not influenced significantly by the
different bar diameter, lug spacing, or the relative rib
area.
Experimental investigation
This investigation was designed to test the confined
region of a joint in a high-strength concrete structure in
order to study the behaviour of the bond between re-
inforcement and high-strength concrete. The load his-
tory, confining reinforcement, bar diameter, concrete
strength, bar spacing, and rate of pull out were consid-
ered as the main study parameters. The study para-
meters were evaluated under monotonic loading in
tension and compression. The internal strain in concrete
close to the contact surface area was measured using
strain gauges.
A summary of the test programme is presented in
Tables 1 to 3. The test results are subdivided into six
series (M1–M6) as shown in Tables 4 to 6. Only one
parameter was changed at a time, and all other para-
meters were kept constant. The influence of each study

Table 1. Experimental programme for monotonic test (M1)

Series Investigation Loading history Specimen notation Rebar Concrete strength: Vertical confining Slip rate: Type of
parameter diameter: f9c rebar: mm/min deformation
mm MPa mm pattern
M1 Load history Monotonic in 1HNM-19-1 35 85.00 10 1.51 Canadian
tension 1HNM-19-1A 35 81.00 10 1.51 standard
1HNM-19-1B 35 81.00 10 1.51
1HNM-19-1C 35 81.00 10 1.51
Monotonic in 1HNM-19-2A 35 81.00 10 1.51 Canadian
compression 1HNM-19-2B 35 85.12 10 1.51 standard
1HNM-19-2C 35 83.16 10 1.51
Monotonic in 1HNM-19-1 25 81.70 10 1.51 Canadian
tension 1HNM-19-1A 25 86.00 10 1.51 standard
1HNM-19-1B 25 85.12 10 1.51
1HNM-19-1C 25 85.12 10 1.51
Monotonic in 1HNM-19-2 25 83.60 10 1.51 Canadian
compression 1HNM-19-2A 25 83.67 10 1.51 standard
1HNM-19-2B 25 83.67 10 1.51
1HNM-19-2C 25 83.67 10 1.51

Table 2. Experimental programme for monotonic test (M2 and M3)

Series Investigation Loading history Specimen notation Rebar Concrete strength Vertical confining Slip rate: Type of
parameter diameter: f9c : rebar: mm/min deformation
mm MPa mm pattern
M2 Confining Monotonic in 2HNM-8-1 35 85.00 No conf. 1.51 Canadian
reinforcement tension 2HNM-8-2 35 85.00 10 1.51 standard
2HNM-8-3 35 85.00 20 1.51
2HNM-8-4 35 85.00 25 1.51
Monotonic in 2HNM-8-1 25 81.70 No conf. 1.51 Canadian
tension 2HNM-8-2 25 81.70 10 1.51 standard
2HNM-8-3 25 85.00 20 1.51
2HNM-8-4 25 85.00 25 1.51
Monotonic in 2HNM-8-1 25 81.70 No conf. 1.51 Canadian
compression 2HNM-8-2 25 81.70 10 1.51 standard
2HNM-8-3 25 85.00 20 1.51
2HNM-8-4 25 85.00 25 1.51
M3 Rebar Monotonic in 3HNM-12-1 20 78.00 10 1.51 Canadian
diameter tension 3HNM-12-2 25 90.30 10 1.51 standard
3HNM-12-3 35 87.45 10 1.51

546 Magazine of Concrete Research, 2004, 56, No. 9

 Bond of high-strength concrete under monotonic pull-out loading

Table 3. Experimental programme for monotonic test (M4 and M6)

Series Investigation Loading history Specimen notation Rebar Concrete strength Vertical confining Slip rate: Type of
parameter diameter: f9c : rebar: mm/min deformation
mm MPa mm pattern
M4 Concrete Monotonic in 4HNM-12-1 35 50.61 10 1.51 Canadian
strength tension 4HNM-12-2 35 87.45 10 1.51 standard
4HNM-12-3 35 91.45 10 1.51
4HNM-12-4 35 92.70 10 1.51
4HNM-12-1 25 50.61 10 1.51 Canadian
4HNM-12-2 25 86.00 10 1.51 standard
4HNM-12-3 25 92.61 10 1.51
4HNM-12-4 25 94.96 10 1.51
M5 Bar spacing: Monotonic in
1db tension 5HNM-4-1 35 86.59 10 1.51 Canadian
2db 5HNM-4-2 35 94.64 10 1.51 standard
1db 5HNM-4-1 25 90.30 10 1.51
2db 5HNM-4-2 25 92.61 10 1.51
M6 Rate of Monotonic in 6HNM-6-1 35 89.00 10 1.5100 Canadian
loading tension 6HNM-6-2 35 87.45 10 0.0151 standard
6HNM-6-3 35 78.00 10 75.0000
6HNM-6-1 25 83.30 10 1.5100 Canadian
6HNM-6-2 25 83.67 10 0.0151 standard
6HNM-6-3 25 83.30 10 75.0000

parameter on the bond behaviour under monotonic
loading was examined.
Test specimens
The test specimen represented the confined region of
a joint in high-strength concrete structures. The test
set-up, specimen size, number of specimens and testing
parameters were similar to that conducted on normal-
8
strength concrete by Eligehausen et al. The reinforced
high-strength concrete specimen was confined by extra
reinforcement representing the joint reinforcement. The
confinement index is calculated as defined by the
9 10
Canadian Building Code CSA.
The confinement index for the tested specimen was
greater than 2.5 without the contribution of the extra
steel confining reinforcement. Extra top and bottom
10 mm stirrups were added to the specimens to ensure
good confinement of reinforcement. The dimensions of
tested specimen were 250 mm in length, 15db (tested
bar diameter) in width and 5db to 7db in thickness as
shown in Fig. 1. The bond length is located at the
middle of the specimen and the rest of the specimen is
de-bonded by the use of two small PVC pipes at the
end of each specimen. The embedded length of rebar
diameters of 25 and 35 mm in the high-strength con-
crete block were taken as 75 and 100 mm, respectively,
as well as extra reinforcement provided with top and
bottom stirrups. These embedment lengths were short
enough to result in a fairly uniform bond stress when
the rebar is pulled out, but not long enough to reduce
the scatter usually observed in test results when a very
Magazine of Concrete Research, 2004, 56, No. 9
short-bonded length was used. Three concrete strain
gauges were used to measure the concrete strain at the
middle and edge of the specimen. In each testing series
a minimum of three specimens were fully instrumented
to measure steel and concrete strain gauges. The posi-
tion of the concrete strain gauges is shown schemati-
cally in Fig. 2 for a typical test specimen that was
considered in this investigation.
Concrete mix design
The high-strength concrete mixture contained a nor-
mal Portland cement, type 10 in accordance with Cana-
9
dian Standards Association CSA, and 10% of silica
fume used on the basis of weight as detailed in Alavi-
Fard. The cement content ranged between 450 and
470 kg/m3 and a water : cement ratio of 0.29 was used.
A non-chloride water-reducing agent, retarder and
superplasticiser were used in the mix. The target com-
pressive strengths for the high-strength concrete speci-
mens ranged between 70 and 95 MPa.
Local aggregates were used for the concrete mix
design. The coarse aggregate was mostly crushed quart-
zite sandstone with a maximum nominal size of
20 mm. The fine aggregate was identical in composi-
tion to the coarse aggregate with a minor percentage of
sandstone and shale. The tested specimens were cured
for 28 days at the concrete laboratory, and then were
transferred for the structural laboratory for testing.
Steel reinforcement
The reinforcing bars were steel conforming to Cana-
dian Standard Association CSA-G40.20-M94. Two
samples of each bar size with diameter of 10, 25 and
547
548

Alavi-Fard and Marzouk

Table 4. Test results (M1)
Group Investigation Loading history Specimen notation Embedded length: Slip at peak S: Peak load Pmax Bond stress ìt : Normalised test Bond stress CSA94 ìCSA /(f9c )1=3
parameter mm mm KN MPa result* ìt /(f9c )1=3 ìCSA
MPa
M1 Load history Monotonic in 1HNM-19-1 100 7.71 278.00 24.800 5.458 5.126 1.100
tension 1HNM-19-1A 100 6.53 218.00 19.470 4.499 5.004 0.899
1HNM-19-1B 100 5.24 167.80 14.980 3.462 5.004 0.692
1HNM-19-1C 100 6.06 210.35 18.780 4.341 5.004 0.868
Monotonic in 1HNM-19-2A 100 6.71 218.65 19.520 4.511 5.004 0.901
compression 1HNM-19-2B 100 6.69 217.54 19.420 4.415 5.129 0.861
1HNM-19-2C 100 5.72 191.08 17.060 3.908 5.129 0.762
Monotonic in 1HNM-19-1 73 8.19 180.14 30.190 6.958 5.025 1.385
tension 1HNM-19-1A 73 9.67 157.65 27.340 6.194 5.126 1.201
1HNM-19-1B 73 5.92 137.43 23.820 5.415 5.129 1.056
1HNM-19-1C 73 5.08 146.23 25.350 5.763 5.129 1.124
Monotonic in 1HNM-19-2 73 5.52 135.91 23.420 5.356 5.086 1.053
compression 1HNM-19-2A 73 5.12 139.08 24.050 5.498 5.086 1.081
1HNM-19-2B 73 3.86 150.67 25.890 5.919 5.086 1.164
1HNM-19-2C 73 5.29 178.38 30.930 7.072 5.086 1.390
* ìt ¼ peak load/(ðdL); d ¼ rebar diameter; L ¼ bond length.

Table 5. Test results (M2 to M3)

Group Investigation Loading history Specimen notation Embedded length: Slip at peak S: Peak load Pmax Bond stress ìt : Normalised test Bond stress CSA94 ìCSA /(f9c )1=3
parameter mm mm KN MPa result* ìt /(f9c )1=3 ìCSA
Magazine of Concrete Research, 2004, 56, No. 9

MPa
M2 Confining Monotonic 2HNM-8-1 100 7.01 270.64 23.84 5.422 5.126 1.058
reinforcement in tension 2HNM-8-2 100 6.43 253.23 22.61 5.142 5.126 1.003
2HNM-8-3 100 17.54 272.54 24.34 5.536 5.126 1.080
2HNM-8-4 100 5.14 190.16 16.98 3.862 5.126 0.753
2HNM-8-1 73 8.71 168.79 29.14 6.716 5.025 1.337
2HNM-8-2 73 8.41 171.23 29.65 6.833 5.025 1.360
2HNM-8-3 73 8.48 168.23 29.14 6.628 5.126 1.293
2HNM-8-4 73 - - - - - -
Monotonic in 2HNM-8-1 73 6.83 193.84 33.45 7.709 5.025 1.534
compression 2HNM-8-2 73 3.23 185.54 32.05 7.386 5.025 1.470
. 2HNM-8-3 73 5.94 186.37 31.93 7.261 5.126 1.417
2HNM-8-4 73 2.32 115.29 19.91 4.528 5.126 0.883
M3 Rebar Monotonic in 3HNM-12-1 73 7.29 112.67 24.89 5.825 4.909 1.187
diameter tension 3HNM-12-2 73 6.84 143.55 24.84 5.537 5.283 1.048
3HNM-12-3 100 6.82 220.66 19.69 4.436 5.198 0.853
* ìt ¼ peak load/(ðdL); d ¼ rebar diameter; L ¼ bond length.
Magazine of Concrete Research, 2004, 56, No. 9

Table 6. Test results (M4 to M6)

Group Investigation Loading history Specimen notation Embedded length: Slip at peak S: Peak load Pmax Bond stress ìt : Normalised test Bond stress CSA94 ìCSA /(f9c )1=3
parameter mm mm KN MPa result* ìt /(f9c )1=3 ìCSA
MPa
M4 Concrete Monotonic in 4HNM-12-1 100 6.89 68.53 6.12 4.069 3.954 1.029
strength tension 4HNM-12-2 100 6.95 248.95 22.23 5.008 5.198 0.963
4HNM-12-3 100 6.38 248.85 22.22 4.934 5.316 0.928
4HNM-12-4 100 8.06 273.18 24.39 5.389 5.352 1.007

Bond of high-strength concrete under monotonic pull-out loading

4HNM-12-1 73 6.72 141.19 24.45 6.609 3.954 1.671
4HNM-12-2 73 9.38 157.65 27.34 6.194 5.156 1.201
4HNM-12-3 73 6.92 167.24 28.96 6.401 5.351 1.196
4HNM-12-4 73 8.83 141.21 24.45 5.359 5.418 0.989
M5 Bar spacing Monotonic in 5HNM-4-1 100 5.80 203.650 18.180 4.109 5.174 0.794
tension 5HNM-4-2 100 8.78 244.810 21.856 4.796 5.408 0.887
5HNM-4-1 73 6.83 143.540 24.850 5.539 5.283 1.048
5HNM-4-2 73 7.19 167.240 28.960 6.401 5.351 1.196
M6 Rate of Monotonic in 6HNM-6-1 100 6.85 236.730 21.030 4.711 5.244 0.898
pull-out tension 6HNM-6-2 100 6.96 248.950 22.230 5.008 5.198 0.963
6HNM-6-3 100 5.98 227.670 20.320 4.756 4.909 0.969
6HNM-6-1 73 5.54 28.250 4.880 1.117 5.074 0.220
6HNM-6-2 73 5.29 150.360 25.890 5.919 5.085 1.164
6HNM-6-3 73 7.19 106.890 18.530 4.243 5.074 0.836
* ìt ¼ peak load/(ðdL); d ¼ rebar diameter; L ¼ bond length.
549
Alavi-Fard and Marzouk
Direction of casting

5·5db
Stirrups

120 mm
250 mm
250 mm

120 mm
Test bar (d
10 mm)

5db 5db 5db 5db–7db

Elevation Side view
Vertical bars

PVC pipe
Bond free length
5db–7db

3·5db

Bonded length of the test bar

4·25db 6·5db 4·25db

15 db
Plan

Fig. 1. Dimensions of bond test specimen

used to apply monotonic tensile and compression load.

Typical concrete strain gauge
35 mm
100 mm
250 mm
525 mm 175 mm
Fig. 2. Typical test specimen including concrete strain gauges
35 mm were tested for tensile strength. A hydraulic
machine was used to apply load and electrical resis-
tance strain gauges were used to measure the strain up
to yield. The recorded yield stress of the steel bars was
found between 450 and 475 MPa.
Test set-up
A test frame was designed to carry out the experi-
mental programme. Some additional accessory parts
were designed and fitted to facilitate the examination
of bond strength investigation. Fig. 3 shows the test set-
up with a specimen mounted. The vertical loading
frame consisted of two main vertical W-shape columns
connected by two horizontal cross channels as detailed
11
by Alavi-Fard and Marzouk. An electro-hydraulically
controlled testing actuator with capacity 700 kN was
550
A load cell was attached to the actuator to measure the
load. The load cell was connected to an internal ampli-
fier via the controller. The output voltage from the
controller was fed into the input channel of the data
acquisition system. All tests were run under displace-
ment control. The displacement was measured at the
end of the loaded steel bar by the use of the built in
linear variable differential transducer (LVDT) in the
actuator and confirmed by the external linear potential
differential transducer (LPDT) at the unloaded end of
the bar. The typical displacement rate was at 1.51 mm/
min for the steel bar. The slip is the relative displace-
ment between the steel bar and the high-strength con-
crete block. The slip is calculated as the difference
between the LVDT reading minus the elongation of the
steel bar outside of the concrete block. Therefore, the
maximum measured slip represents the maximum local
slip at the middle of the embedded length with a
relatively acceptable accuracy. Since the bond location
is situated at the middle of the specimen, the bond
length is short compared to the specimen size, there-
fore, it can be assumed that the stress distribution is
uniform along the tested bond length. The main objec-
tive was to record the bond stress versus the slip dis-
placement during the duration of the test. The area
under the bond stress–slip displacement curve is de-
fined as the bond energy. The specimens were tested
over a three-month testing period at the structural
laboratory of Memorial University of Newfoundland.
Fig. 4 shows a tested specimen after failure.
Magazine of Concrete Research, 2004, 56, No. 9
 Bond of high-strength concrete under monotonic pull-out loading
Fig. 3. Specimen during testing: (a) test set-up; (b) specimen close-up
Test results
Loading history
The details of the test specimen under monotonic
loading can be found in Tables 1 to 3. The maximum
magnitudes of slip, load and bond strength for each
specimen are shown in Tables 4 to 6. A typical bond
stress–slip curve for high-strength concrete specimen is
shown in Fig. 5. The comparison of results for four
specimens under pull-out test with an embedded bar
Fig. 4. Specimen after testing
Magazine of Concrete Research, 2004, 56, No. 9
diameter of 35 and 25 mm each are plotted in Figs 6
and 7, respectively. The bond stress and slip curve for
push-in tests have been plotted on the graphs (Fig. 8).
The comparison between the pull-out test of Fig. 7 and
the push-in test of Fig. 8 indicates that there was no
major difference between the two types of loading. In
all of the tested specimens, the main cracks developed
in the longitudinal direction, while some minor cracks
also developed in the transverse direction. The test
results indicated a nonlinearity of the bond stress–slip
behaviour in the ascending portion of the curve, espe-
cially close to maximum stress. All results confirmed
a
Bond stress
c
b
e
Slip
Fig. 5. Typical bond stress–displacement for high-strength
concrete
551
Alavi-Fard and Marzouk
1·2
1HNM-19-1 (fc′
85 MPa)

1HNM-19-1A (fc′
81 MPa)

1
1HNM-19-1B (fc′
81 MPa)

1HNM-19-1C (fc′
81 MPa)

0·8

Normalised bond stress

0·6
A
C

0·4

1
0·2
B

0
0 0·2 0·4 0·6 0·8 1 1·2

Normalised displacement

Fig. 6. Bond stress–displacement for 35 mm bar (pull-out) test

1·2
1HNM-19-1 (fc′
81·70 MPa pull-out)
1HNM-19-1A (fc′
86·00 MPa pull-out)
1 1HNM-19-1B (fc′
85·12 MPa pull-out)
1 1HNM-19-1C (fc′
85·12 MPa pull-out)
Normalised bond stress

0·8 1A

1B
0·6

1C
0·4

0·2

0
0 0·2 0·4 0·6 0·8 1 1·2

Normalised displacement

Fig. 7. Bond stress–displacement for 25 mm bar (pull-out) test

the sudden drop of the stress level at the beginning of the result of the value of tested bond stress divided by
the descending section of the curve followed by the cube root of compression strength of concrete.
gradual decrease of stress. The slope of the curves in
the ascending section in case of the push-in was higher Concrete and steel strains
than that of the pull-out test. Hence, the total bond The steel strain and internal concrete strain for a
energy in compression is slightly higher than expected typical specimen with embedded bar diameter 35 mm
for the pull-out test. However, this difference is not is plotted in Figs 9 and 10. The specimen is tested
large enough to provide a distinct difference between under push-in load. The two strain gauges STC1 and
the two behaviours. STC2 are installed in the bond area and the strain
The average equivalent bond stress for experimental gauge STC3 is far from the bond area, as shown sche-
phase of this investigation is calculated as the result of matically in Fig. 2. There are differences between the
tension or compression force divided by contact surface magnitudes of concrete strain reading of STC1 and
area (bond area). The curves are normalised based on STC2 due to the location of the strain gauges and the
552 Magazine of Concrete Research, 2004, 56, No. 9
 Bond of high-strength concrete under monotonic pull-out loading
1·2
1HNM-19-2A (fc′
83·67 MPa push-in)
1HNM-19-2B (fc′
83·67 MPa push-in)
1 1HNM-19-2C (fc′
85·12 MPa push-in)
Normalised bond stress

0·8
2C

0·6
2B

0·4
2A

0·2

0
0 0·2 0·4 0·6 0·8 1 1·2

Normalised displacement

Fig. 8. Bond stress–displacement for 25 mm bar (push-in) test

250 indicated that when secondary cracks started to open,

the concrete strain gauges were damaged. Fig. 10
200 shows the measured steel strain for a typical bond
Compression load: kN

specimen. The steel strain gauge was glued to the steel

150 bar at the middle of the specimen. The steel stress–
strain curve demonstrates the increase of steel strain
100 proportional to the load and the decrease of the strain
Concrete strain gauge (STC1) at the beginning of the descending portion of the bond
50 Concrete strain gauge (STC2) stress–slip curve.
Concrete strain gauge (STC3)

0 Confining reinforcement
1500 1000 500 0 500 1000 1500
Concrete strain: micro strain The influence of different confinement bar diameters
on local bond behaviour of deformed bars was investi-
Fig. 9. Concrete strains (gauges 1 and 2 at bond area and 3 gated in series M2 of Tables 2 and 5. The vertical and
at the edge) horizontal steel reinforcement shown in Fig.1 repre-
sents the extra confining reinforcement for tested speci-
men. The selected different vertical bar diameters were
250 ranged from 0.3 to 0.6 of the tested bar diameter. Table
5 presents the measured values of slip, load and bond
200 strength at maximum load. All of the tested concrete
Compression load: kN

specimens have adequate concrete cover, edge distance

150
f c′
85·12 MPa
and bar spacing requirement. The normalised bond-
stress–displacement curves for different size of confin-
100 ing reinforcement for a 35 mm diameter bar with re-
spect to the maximum load and maximum deflection
50 are presented in Fig. 11.
STS1 When no extra reinforcement was provided as con-
0 finement, the bond stresses vanished as soon as the
50 0 50 100 150 200 250 300 350 400 450 500
longitudinal crack developed through the cover. The
Steel strain: micro strain
failure mode was of a sudden splitting type. In addition,
Fig. 10. Steel strains a large amount of energy was released due to sudden
splitting of the specimen. However, when extra con-
finement was provided the total area of the bond en-
ergy curve for specimens with different confinement
distance from the bond surface. The concrete strain bars of 10 and 20 mm diameters was higher than that
reading of STC3 is not very significant, since it is very for a confinement bar diameter of 25 mm. A
far from the location of the bond failure. The result 10 mm diameter stirrup was used for confinement of
Magazine of Concrete Research, 2004, 56, No. 9 553
Alavi-Fard and Marzouk
1·2
2HNM-8-1 (without Confinement)
2HNM-8-2 (with Confinement, bar dia.
10 mm)
1
2HNM-8-3 (with Confinement, bar dia.
20 mm)
2HNM-8-4 (with Confinement, bar dia.
25 mm)

Normalised bond stress

3
0·8
2

0·6
1
4
0·4
f′c
85 MPa

0·2

0
0 0·2 0·4 0·6 0·8 1 1·2
Normalised displacement

Fig. 11. Bond stress–displacement for different confining bars

25 and 35 mm diameter bars efficiently. In addition, the bar. The bond strength of the 20 mm diameter bar is
failure mode for specimens with extra confinement was greater than the tensile capacity of the bar cross-section
characterised by pulling out the bar from concrete area. Therefore, the bar is broken during testing and
rather than the sudden splitting behaviour for speci- this curve does not demonstrate the entire bond behav-
mens with no confinement. Therefore, high-strength iour. The ascending slope of the curve for a 20 mm
concrete needs extra steel to reach an adequate level of diameter bar is much steeper than for the 25 and
bond energy and ductility. The confinement reinforce- 35 mm diameter bars. Further comparison between the
ment of high-strength concrete is considered adequate, area under the curve shows that the area under the
when the diameter of the confinement bar is from 0.3 curve for the bar with a diameter of 35 mm is less than
to 0.6 of the tested bar. for the bar diameter of 25 mm. Hence, the experiment
results show that the bond resistance for the smaller
Bar diameter diameter is higher than that for the larger diameter bar.
The bond test results of the 20, 25 and 35 mm In general, the results of this series agree with the find-
8
diameter bars are shown in series M3 of Tables 2 and ings of Eligehausen et al. , using a similar test set-up for
5. The effect of varying bar diameter on the bond normal-strength concrete. However, the bond strength
strength is illustrated in Fig. 12. All curves in this values are different for each concrete strength level. The
figure are normalised with respect to maximum bond sudden drop of the stress level at the beginning of the
stress and maximum displacement of a 25 mm diameter descending section of the bond stress–displacement

1·2
Bar dia. 20 mm 3HNM-12-3 (Bar dia. 35 mm, f′c
87·45 MPa

3HNM-12-2 (Bar dia. 25 mm, f′c
90·30 MPa

1
3HNM-12-1 (Bar dia. 20 mm, f′c
78·00 MPa
Normalised bond stress

0·8

0·6
Bar dia. 25 mm

0·4
Bar dia. 35 mm

0·2

0
0 0·2 0·4 0·6 0·8 1 1·2
Normalised displacement

Fig. 12. Bond stress–displacement for different bar diameters

554 Magazine of Concrete Research, 2004, 56, No. 9

curve is attributed to the characteristics of the high-
strength concrete.
Bar spacing
The influence of different lateral bar spacing on local
bond behaviour was investigated in series M5 of Tables
3 and 6. The lateral bar spacing was taken as 3.5 db for
all the tested specimens as shown in the plan view of
Fig. 1. However, this spacing was changed in four
specimens during this investigation to examine the ef-
fect of the lateral spacing on the bond behaviour. Two
specimens were tested with bar diameter of 25 mm and
lateral spacing of 25 and 50 mm and the other two
specimens were tested with bar diameter of 35 mm and
bar spacing of 35 and 70 mm.
In the case of bar diameter of 25 mm there was no
significant difference between the two in the ascending
portion, while in the descending portion there was a
small difference. The bond strength for specimen with
bar spacing of 25 mm was about 20% less than for a
specimen with 50 mm bar spacing.
The test results for the specimen with bar diameter
of 35 mm show that the increase in bar spacing had
more influence on the bond resistance of the initial part
of the bond stress–slip relationship than on the bond
strength. The bond strength is improved with the in-
crease of lateral bar spacing. The result of observation
and comparison between bond strength for this series
of tests showed a possible difference of 20% in the
bond strength due to the lateral bar spacing.
Rate of loading
This section presents the experimental results from
the testing of three specimens tested under different
displacement rates of 75, 1.51 and 0.0151 mm/min as
shown in series M6 of Tables 3 and 6. Fig. 13 shows a
1·2
Curve 1: Specimen 6HNM-6-1 with rate of loading 75·0000 mm/min & f′c
89·00 MPa
Curve 2: Specimen 6HNM-6-2 with rate of loading 0·0151 mm/min & f′c
78·00 MPa
Curve 3: Specimen 6HNM-6-3 with rate of loading 1·5100 mm/min & f′c
87·45 MPa
1
3 2
Normalised bond stress
0·8
0·6
0·4
0·2
0
0 0·2
Fig. 13. Bond stress–displacement for different rate of loading
Magazine of Concrete Research, 2004, 56, No. 9
Bond of high-strength concrete under monotonic pull-out loading
comparison of the normalised bond stress–displace-
ment response for the three specimens tested under
different rates of loading for a tested bar diameter of
35 mm. Although at the initial stage of loading there
was a noticeable difference for the specimen with rate
of loading of 1.51 mm/min the bond–displacement
curves were similar. The result indicated that the bond
strength is approximately the same under different rates
of loading for a tested bar of 35 mm as presented in
series M6 of Table 6.
Concrete strength
The influence of the concrete compressive strength
on the bond behaviour was investigated in series M4 of
Tables 3 and 6. In principle, the results of this series of
tests agree with the previous study for normal concrete
8
strengths of 30 and 55 MPa. The main difference be-
tween high-strength concrete and normal-strength con-
crete is the instantaneous drop of the curve at the
beginning of the descending branch of the curves, as a
result of losing adhesion. This brittle behaviour is well
illustrated in series M4 of tests especially for concrete
with higher strength. The results of all tests in this
series confirmed the nonlinear brittle behaviour of the
bond for high-strength concrete. In the case of high-
strength concrete the capacity of bond strength is high-
er than the normal; however, the impact of the instanta-
neous drop in the bond stress for high-strength concrete
must be recognised. The bond strength is strongly de-
pendent on concrete strength and this parameter has a
direct effect on bond behaviour. A comparison of nor-
malised bond stress with regard to the maximum load
of specimen with compressive strength of 94.96 MPa to
other specimens, with different concrete strengths for a
25 mm diameter bar is shown in Fig. 14.
1
0·4 0·6 0·8 1 1·2
Normalised displacement
555
Alavi-Fard and Marzouk
1·2
4
1 f′c
92·61 MPa, Bar dia. 25 mm, pull-out
3 f′c
86·00 MPa, Bar dia. 25 mm, pull-out
Normalised bond stress
0·8
0·6
0·4
0·2
0
0 0·2
Fig. 14. Bond stress–displacement for different concrete strengths
Test results and code of practice
The load required to pull a reinforcing bar out of a
concrete block will obviously increase as the length of
the bar cast into the block increases. When the em-
bedded length becomes long enough, the bar will yield
in tension before it pulls out of the block. The mini-
mum embedded length required to develop the yield
force of the bar is called the minimum required
development length. The development length require-
ments are used by all North American codes to indicate
the bond strength of concrete. Both of the Canadian
9 12
Standard CSA and the American ACI-318, express
the bond strength of concrete as a function of the
square root of concrete compressive strength. The test
results of the current investigation indicated that the
bond strength of high-strength concrete, from 70 to
95 MPa is more appropriately proportional to the cube
root rather than the square root based on the statistical
10
sensitivity analysis conducted by Alavi-Fard. There-
fore, the test results were normalised with regard to the
cube root of the compressive strength as given in
Tables 4 to 6. The ratio between the normalised test
9
results versus the CSA Standard prediction is given in
Tables 4 to 6. This ratio indicates that the square root
of the compressive strength approach adopted by the
Canadian Standard does not provide a good prediction
of the bond strength for the tested high-strength con-
crete specimens. The bond strength of concrete is pro-
portional to the compressive strength, to the power of
13
1/3 in the British Code, BS 8110, and to the power of
14
2/3 in both of the CEB-FIP and Norwegian NS-
15
3474 Codes.
6
More recently, Azizinamini et al. recommended a
new expression for confinement of the tension develop-
ment length and lap splice for high-strength concrete to
ensure adequate ductility. A modification to the ACI-
12
318 expression has been recommended based on the
assumption that the bond strength of high-strength con-
crete of 69 MPa and over is proportional to the fourth
root of the compressive strength.
556
f′c
94·96 MPa, Bar dia. 25 mm, pull-out
f′c
50·61 MPa, Bar dia. 25 mm, pull-out
4HNM-12-1 (f′c
56·61 MPa, pull-out)
4HNM-12-2 (f′c
80·00 MPa, pull-out)
4HNM-12-3 (f′c
92·61 MPa, pull-out)
4HNM-12-4 (f′c
94·96 MPa, pull-out)
1
2
0·4 0·6 0·8 1 1·2
Normalised displacement
Failure mechanism
There are two types of failure mechanisms that are
known for the pull-out test. The first type is splitting of
the concrete cover and the second type is pull-out of
the bar. There are several parameters that govern the
mode of failure such as: type of loading, rate of load-
ing, thickness of the cover, bar diameter, confinement
reinforcement and bar deformation patterns. The test
results of the current investigation revealed that the
behaviour of the bond stress–slip response was brittle
and nonlinear for high-strength concrete. The magni-
tude of the maximum slip at failure was estimated to be
five times the slip at the maximum bond stress. There-
fore, the primary cracks will be developed and ob-
served before failure of the specimen. The longitudinal
crack at the surface of the specimen appeared approxi-
mately at the maximum load and the bar failed due to
pull-out from the concrete block. The rib of the de-
formed bar can be a major cause of longitudinal split-
ting along the bar for a confined specimen. The face
angle, height and spacing of the ribs of the deformed
bar have an important effect on the bond energy on
high-strength concrete.
Splitting type failure occurred when the cracks flow-
ing from the contact area of the bar reached the surface
of the high-strength concrete prism. The splitting fail-
ure was initiated by the wedging action of the ribs as
the reinforcement bar moved with respect to the con-
crete. Splitting was characterised by planar-like cracks
in those planes radial to the axis of the rebar. In a
16
recent study by Alavi-Fard and Marzouk on five de-
formation patterns of steel reinforcement for high-
strength concrete, it was found that the Canadian Stan-
dard deformation is one of the most efficient patterns
to be used for high-strength concrete reinforcement.
Furthermore, the wedging action of the rib had a higher
percentage of failure to the bond specimens. In the case
of the unconfined specimens, failure occurred only due
to longitudinal splitting along the bar and at the same
time it was accompanied with a large amount of energy.
Magazine of Concrete Research, 2004, 56, No. 9
 Bond of high-strength concrete under monotonic pull-out loading
Summary and conclusions
The bond strength of high-strength concrete is sub-
jected to the effects of several factors. In this investiga-
tion, the effect of the load history, confinement, bar
diameter, concrete strength, bar spacing and pulls out
rate were examined. The range of the tested concrete
compressive strengths was between 70 and 95 MPa.
The bond stress–slip curve of high-strength concrete
is characterised by a sharp drop of the stress at the
beginning of the descending portion of the bond
stress–slip curve. The area under the curve of the bond
stress–slip curve can be defined as bond energy. The
bond energy is recommended to be used to evaluate the
bond behaviour rather than the maximum bond stress.
The influence of extra confinement on bond is sig-
nificant on improving the ductility and the bond energy,
especially after reaching the maximum bond strength.
The test results revealed that a confinement bar dia-
meter from 0.3 to 0.6 of the confined bar is adequate
for high-strength concrete under monotonic loading.
The result of tests examining the effect of varying
bar diameter embedded in high-strength concrete indi-
cates that the bond is higher for the smaller bar dia-
meter than for the larger one. The bond strength for a
25 mm diameter bar is approximately 15% higher than
that of a 35 mm diameter bar. A sharp drop of the bond
stress–slip curve at the beginning of the descending
portion was confirmed for all bar diameters. The level
of bond stress drops by about 30% of maximum bond
stress at the beginning of the descending branch of the
bond stress–slip curve.
The results of the investigation regarding the influ-
ence of the later bar spacing revealed that the bond
strength could be improved by selecting a proper lateral
bar spacing and adequate concrete cover. An investiga-
tion into the bond strength subjected to the effect of the
concrete strength concluded that the bond strength for
high-strength concrete is higher than the corresponding
one for normal-strength concrete. However, the behav-
iour of high-strength concrete is more brittle and it
must be reflected in the bond model.
9 British Standard Institute, London, UK, 1985, 126 pp.
The current Canadian Standard CSA must be mod-
ified to reflect the reality of the bond strength results
for high-strength concrete of compressive strength
70 MPa and over. The use of the square root of the
compressive strength to predict the bond strength will
lead to unsafe bond design. The use of the square root
expression for the bond strength of high-strength con-
crete of compressive strength, 70 MPa and over was
abandoned in the latest proposed revision of the ACI-
12
318. It is important at this point to publish all the
available test results to guide the Canadian Code com-
mittee to adopt the appropriate expression, whether it
uses cube root like the European codes or a fourth root
Magazine of Concrete Research, 2004, 56, No. 9
like the American Code. On the basis of the limited test
results, the bond strength is more appropriately propor-
tional to the cube root rather than the current square
root.
References
1. Somayaji S. and Shah S. P. Bond stress versus slip relationship
and cracking response of tension members. ACI Journal, 1981,
78, No. 3, 217–225.
2. Jiang D. H., Shah S. P. and Andonian A. Study of the transfer
of tensile forces by bond. ACI Journal, 1984, 81, No. 2,
251–258.
3. Darwin D, Zuo J, Tholen M. L. and Idun E. K. Development
length criteria for conventional and high relative rib area rein-
forcing bars. ACI Structural Journal ,1996, 93, No. 3, 347–359.
4. Azizinamini A., Stark M., Roller J. J. and Ghosk S. K.
Bond performance of reinforcing bars embedded in high-
strength concrete. ACI Structural Journal, 1993, 90, No. 5,
554–561.
5. Azizinamini A., Pavel R., Hatfield E. and Ghosh S. K.
Behavior of spliced reinforcing bars embedded in high-strength
concrete. ACI Structural Journal, 1999, 96, No. 5, 826–835.
6. Azizinamini A., Darwin D, Eligehausen R, Pavel R and
Ghosh S. K. Proposed modification to ACI 318-95 tension
development and lap splice for high strength concrete. ACI
Structural Journal, 1999, 96, No. 6, 922–926.
7. De Larrard F., Schaller D. and Fuchs J. Effect of bar
diameter on the bond strength of passive reinforcement in high-
performance concrete. ACI Materials Journal , 1993, 90, No. 4,
333–339.
8. Eligehausen R., Popov E. P. and Betero V. Local bond
stress-slip relationships of deformed bar under generalized ex-
citations. In Report No. UCB/EERC -83/23. Earthquake Engi-
neering, Research Center, Berkeley, CA, USA, 1983, p. 185.
9. CSA. Design of Concrete Structures. Standard CSA-A23.3.
Canadian Standards Association, Rexdale, Ont, Canada, 1994.
10. Alavi-Fard M. Bond Characteristics of High Strength Con-
crete. PhD thesis, Memorial University of Newfoundland, St.
John’s, NF, Canada, 1999.
11. Alavi-Fard M. and Marzouk H. M. Bond behavior of high
strength concrete under reversed pull-out cyclic loading. Cana-
dian Journal of Civil Engineering, 2002, 29, 191–200.
12. ACI Committee 318. Building code requirements for structural
concrete and commentary. American Concrete Institute, Farm-
ington Hills, MI, 2002, pp. 369.
13. British Standards Institute BS 8110. Structural Use of
Concrete: Part 1, Code of Practice for Design and Construction.
14. CEB-FIP. CEB-FIP model code for concrete structures. Gmi-
te’Euro-International du Be’ton el Federation International de
al Pre-Contrainte, Lausanne, Switzerland, 1990.
15. Norwegian Standard NS 3473.Concrete structures. Design
rules. Norwegian Council for Building Standardization [NBR],
1992, 4th edition.
16. Alavi-Fard M. and Marzouk H. M. Influence of concrete
strength, size and deformation pattern on bond capacity of high
strength concrete. International Conference on HPHSC, Perth,
Australia, 1998.
Discussion contributions on this paper should reach the editor by
1 May 2005
557