ACI MATERIALS JOURNAL

TECHNICAL PAPER

Title no. 108-M04

Influence of Surface Crack Width on Bond Strength of

Reinforced Concrete
by Denglei Tang, Thomas K. C. Molyneaux, David W. Law, and Rebecca Gravina
This paper reports the results of an experimental program to
investigate the changes in the bond characteristics of deformed
mild steel reinforcing due to chloride-induced corrosion. The
principal parameters investigated are the cover depth, bar
diameter, degree of corrosion, and the surface crack width. The
results show a strong relationship between the average surface
crack width and the average bond strength; the degree of corrosion
does not demonstrate such a clear relationship. The bars with
a 1 C/ (cover/diameter) show an initial increase in bond strength
at the first visible crack; no similar initial increase is observed for
the bars with 3 C/. The bottom-cast bars display a higher bond
strength with no corrosion, but a similar bond strength is observed
for both top- and bottom-cast bars once cracking is observed.
Keywords: bond; corrosion; cover; deformed reinforcing bar; surface
crack width.

INTRODUCTION
In reinforced concrete (RC), it is the bond behavior
between the reinforcement and the concrete that ensures
effective composite behavior; however, reinforcement
corrosion affects composite performance because the
corrosion products formed around the steel bars ultimately
result in a deterioration of the bond. The deterioration of
structures due to the corrosion of the reinforcement is mainly
characterized by a general or local section loss of the reinforcing
bars, cracks in the concrete, and loss of bond strength.
RESEARCH SIGNIFICANCE
Previous research on corrosion and bond has focused on the
relationship between the corrosion level (corrosion
penetration or mass loss of the steel) and cracking,1-8 current
density versus surface crack width,1,2,9 and bond strength
versus corrosion level (mass loss or corrosion penetration).10-17
Relatively little research has studied the relationship between
bond strength and surface crack width.11,13 This paper reports
a study on the relationship between the change of bond strength
and the surface crack width, a parameter that can readily be
measured on a structure.
EXPERIMENTAL INVESTIGATION
Specimens
Beam-end specimens11,18-20 were selected for this study.
This type of eccentric pullout or beam end-type specimen
uses a bonded length representative of the anchorage zone of
a typical simply supported beam. Specimens of rectangular
cross section were cast with a longitudinal reinforcing bar in
each corner and without stirrups (Fig. 1).
A deformed reinforcing bar of 12 and 16 mm (0.4724 and
0.6299 in.) diameter with covers of one and three times bar
diameter were investigated. This gave four groups of
specimens with a combination of different bar diameter and
ACI Materials Journal/January-February 2011

Fig. 1Sketch of adopted beam end specimen.

C/ ratios. These were selected to investigate size effects

and the influence of C/ on the loss of bond strength.
Materials
The concrete mixture design comprised normal portland
cement, fine aggregate (medium-sized natural sand), and
crushed coarse aggregate. The ratio by weight of
cement/sand/coarse aggregate was 182:247:442, with a
water-cement ratio (w/c) of 0.49 for a target strength of 40 MPa
(5.8 ksi). The mixture design is given in Table 1. In
accordance with AS1012,21 100 mm (0.3937 in.) diameter
cylinders were cast to assess concrete compressive strength
and 150 mm (5.9 in.) diameter cylinders were cast to assess
indirect tensile strength after 28 days. Curing was
undertaken by spraying with water for 5 days following
casting. Specimens were then stored in the lab. Accelerated
corrosion testing was undertaken 28 days after casting.
The specimens were cast in different groups; the mean
compressive and tensile strengths are given in Table 2. To
compare bond strength for the different concrete
compressive strengths, Eq. (1) is used to normalize bond
strength for noncorroded specimens, as has been used by
other researchers.22
ACI Materials Journal, V. 108, No. 1, January-February 2011.
MS No. M-2009-058.R4 received April 12, 2010, and reviewed under Institute publication
policies. Copyright 2011, American Concrete Institute. All rights reserved, including the
making of copies unless permission is obtained from the copyright proprietors. Pertinent
discussion including authors closure, if any, will be published in the November-December
2011 ACI Materials Journal if the discussion is received by August 1, 2011.

29

accuracy of 0.02 mm (0.00079 in.). Measurements were

taken normal to the bar direction. When the average surface
crack width was equal to the target crack width, that bar was
removed from the circuit. Pullout testing was undertaken
when all four bars exhibited the target crack width.
Average surface crack widths of 0.05, 0.5, 1, and 1.5 mm
(0.002, 0.02, 0.039, and 0.059 in.) were adopted as corrosion
stages. In addition, control specimens were cast that did not
undergo accelerated corrosion.
Prior to casting, the bars were cleaned with nitric acid23 to
remove any weathered corrosion products, so as to enable
the mass loss of corrosion products to be assessed by
comparing the initial and final weight. Following the pullout
test, all bars were cleaned in the same way and weighed
again. The difference in mass was then used to determine the
extent of corrosion.
Bond strength tests were conducted by means of a handoperated hydraulic jack and a custom-built test rig, as shown
in Fig. 4. The loading scheme is illustrated in Fig. 5.11 A
plastic tube with a length of 80 mm (3.15 in.), as bond
breaker, was provided at the end of the concrete section
underneath the transverse reaction to ensure that the bond
strength was not enhanced by this reactive force.

Denglei Tang is a Bridge Design Engineer at VicRoads, Melbourne, Australia. He

received his MSc from the Royal Melbourne Institute of Technology (RMIT
University), Melbourne, Australia, in 2008. His interests include the structural
analysis and durability of reinforced concrete structures.
Thomas K. C. Molyneaux is an Associate Professor at RMIT University. He
received his BSc and PhD from the University of Leeds, Leeds, UK. His research
interests include the nonlinear analysis of masonry structures and durability of
reinforced concrete structures.
David W. Law is a Lecturer at Heriot-Watt University, Edinburgh, UK. He received
his BSc and MSc from Salford University, Salford, UK, in 1984 and 1985, respectively,
and received his PhD from the University of Reading, Reading, UK, in 1989. His
research interests include the durability of reinforced and service-life modeling of
concrete structures.
Rebecca Gravina is a Senior Lecturer at RMIT University in the School of Civil,
Environmental and Chemical Engineering. She received her PhD from the University
of Adelaide, Adelaide, Australia, in 2002. Her research interests include the nonlinear
analysis of reinforced concrete structures and ductility and bond mechanics of the
steel-to-concrete and fiber-reinforced polymer-to-concrete interface.

40
= exp tl -----fc

(1)

where is the bond strength for Grade 40 concrete, exptl is

the experimental bond strength, and fc is the experimental
compressive strength.
The tensile strength of the 12 and 16 mm (0.4724 and
0.6299 in.) steel bars used was nominally 500 MPa (72.52 ksi),
which equates to a failure load of 56.5 and 100.5 kN (12.70 and
22.59 kips), respectively.

ANALYTICAL INVESTIGATION
The amount of steel lost due to corrosion can be calculated
by applying Faradays law.24 This relates the corrosion
current to the loss of steel mass by accounting for the charge
on each ion that is transported during the process.

Experiment procedure
To accelerate the corrosion process, a constant current was
applied. The steel bars served as the anode and four metal
plates were fixed on the surface to serve as cathodes. Sponges
sprayed with saltwater were placed between the metal plates
and concrete to provide an adequate contact (Fig. 2).
A current density of 200 A/cm2 (186.92 A/ft2) was
applied to the bars in pairs (top pair and bottom pair), with
each pair being wired in series with a power supply
maintaining the constant current flow. When the required
surface crack width was achieved for a particular bar, that
bar was removed from the circuit and the current was
adjusted accordingly to maintain the required current
density. A 3% saltwater spray was applied under a 1-day
wet and 2 days dry cyclic regime. The surface crack
width was measured at 20 mm (0.7874 in.) intervals along
the length of the crack, giving a maximum of 14 measuring
points, (Fig. 3), depending on the length of the crack.
Measurements were taken by using microscopy with an

m = MIt/zF

(2)

where m is the mass of steel consumed (g), M is the atomic

weight of the metal (56 g for iron), t is the time in seconds, z is
the ionic charge (2 for iron), and F is a constant (96,500 amp.s).
The resulting equation in terms of penetration is
x = 11.6Icorrt

(3)

where x is the corrosion penetration in microns, t is the time

in years elapsed since the onset of corrosion, and Icorr is the
corrosion current in A/cm2 throughout time t.
The method of applying an impressed current to accelerate
corrosion has been used by many researchers1,2,5,7-14,1618,21,25-30 to obtain results in a reasonable time. Maximum
corrosion rates reported on actual structures are generally
less than 10 A/cm2 (9.346 A/ft2).31 The maximum natural
corrosion rate (without impressed current) reported from

Table 1Concrete mixture design

Material

Cement

Quantity

381 kg/m3
(23.774 lb/ft3)

w/c

Sand

10 mm (0.394 in.)
washed aggregate

7 mm (0.276 in.)
washed aggregate

Salt

Slump

0.49

517 kg/m3
(32.261 lb/ft3)

463 kg/m3
(28.891 lb/ft3)

463 kg/m3
(28.891 lb/ft3)

18.84 kg/m3
(1.176 lb/ft3)

140 25 mm
(5.51 0.95 in.)

Table 2Mean compressive and tensile strengths

Group no.

Specification

No. of specimens

Mean compressive strength, MPa

Mean tensile strength, MPa

1
2

12 mm (0.4724 in.), 1 C/
12 mm (0.4724 in.), 1 C/

5
4

37
36

2.8
2.8

3
4

16 mm (0.6299 in.), 1 C/
16 mm (0.6299 in.), 3 C/

5
7

39
40

3.3
3.2

Note: 1 MPa = 144.93 psi.

30

ACI Materials Journal/January-February 2011

Fig. 2Accelerating system.

Fig. 3Sketch of surface crack width.

measurements taken in laboratory studies, however, was

100 A/cm2 (93.46 A/ft2). The majority of previous studies
using impressed currents have used current densities that are
three to 100 times greater than this.9 The current density
levels are known to affect the surface crack width, probably
because a low current density level gives corrosion products
more time to dissipate through the concrete pores.2,9
Research suggests that current densities up to 200 A/cm2
(186.92 A/ft2) result in similar stresses during the early
stages of corrosion when compared to 100 A/cm2
(9.346 A/ft2).9 Consequently, the work reported herein has
adopted a current density of 200 A/cm2 (186.92 A/ft2).
Surface cracking due to the corrosion of RC structures has
been investigated by other researchers.1,3-6,8,11-13,32-35
When the production of rust begins, it gradually builds
pressure around the reinforcing steel. This buildup of
pressure eventually cracks the concrete around the steel, and
the crack or cracks propagate with further increase of
pressure. If the cracks propagate to the surface, the concrete
will eventually spall. Andrade et al.36 assessed cracking
conditions in corroded structures. They proposed the following
relationship between crack width and corrosion penetration
W = 0.05 + ( x x 0 ) [ W 1.0 mm (0.039 in.) ]

(4)

where W is the estimated surface crack width (mm), x is the

corrosion penetration (microns), x0 is the corrosion
penetration (microns) at crack initiation, and is a
coefficient that depends on the position of the bar in the
section ( = 0.01 for top-cast bars and 0.0125 for bottom-cast
bars). Andrade et al.36 suggested that the x0 needed for the
cracking initiation can be estimated by the following expression
x 0 = 83.8 + 7.4c 22.6f c, sp

(5)

where c is the concrete cover (mm), is the bar diameter

(mm), and fc,sp is the tensile strength (MPa).
Alternatively, Eq. (6) is recommended by Alonso et al.,1
based on a study of the influence of the cover/diameter ratio.
x 0 = 7.53 + 9.32c
ACI Materials Journal/January-February 2011

Fig. 4Pullout test.

(6)

Fig. 5Loading scheme.

Previous work comparing these two approaches by Vidal

et al.3 concluded that their experimental results are more in
agreement with the predictions calculated using Eq. (6) than
Eq. (5). As such, another equation, Eq. (7), was proposed to
relate the surface cracks and steel loss of section.
W = k ( A s A s0 )

(7)

where W is the surface crack width (mm), As is the steel

loss of cross section in mm2, As0 is the steel cross-section
loss needed for cracking initiation, and k = 0.0575.
31

Table 3Comparison of measured and predicted surface crack width and section loss at initial cracking
Specification

12 mm (0.4724 in.), 1 C/

16 mm (0.6299 in.), 1 C/

16 mm (0.6299 in.), 3 C/

Measured W, mm (in.)
0.0480 (0.0019)

Measured X0, m (in.)

X0(Andrade), m (in.)

X0(Alonso), m (in.)

81.3116 (0.0032)

30.406 (0.0012)

16.85 (0.00066)

0.0593 (0.0023)
0.0423 (0.0017)

79.9516 (0.0031)
71.3437 (0.0028)

30.406 (0.0012)
25.886 (0.0010)

16.85 (0.00066)
16.85 (0.00066)

0.0492 (0.0019)

64.5521 (0.0025)

25.886 (0.0010)

16.85 (0.00066)

0.078 (0.0031)

80.7331 (0.0032)

27.242 (0.0011)

16.85 (0.00066)

0.068 (0.0027)
0.048 (0.0019)

66.8255 (0.0026)
75.3042 (0.0030)

27.242 (0.0011)
27.242 (00011)

16.85 (0.00066)
16.85 (0.00066)

0.044 (0.0017)

90.4079 (0.00356)

27.242 (0.0011)

16.85 (0.00066)

0.058 (0.0023)

180.137 (0.0071)

23.058 (0.00091)

35.49 (0.0014)

0.096 (0.0038)

216.5482 (0.0085)

23.058 (0.00091)

35.49 (0.0014)

Fig. 6Average bond strength versus average surface crack

width for 12 mm (0.4724 in.) bars with 1 C/. (Note: 1 MPa
= 144.93 psi; 1 mm = 0.039 in.)

Fig. 7Average bond strength versus extent of corrosion for

12 mm (0.4724 in.) bars with 1 C/. (Note: 1 MPa =
144.93 psi; 1 mm = 0.039 in.)

COMPARISON OF PREDICTIONS AND

EXPERIMENTAL RESULTS
Table 3 gives the comparison of experimental results
with results from Eq. (4) and (5) at the initial cracking stage
(0.05 mm [0.002 in.]). It is found that experimental values x0
are greater than the predicted values.
Vidals module (Eq. (6)) is used to compare the predicted
surface crack width with the observed average surface crack
width at various corrosion degrees (Table 4). The data show
a reasonable correlation for the 12 mm (0.4724 in.), 1 C/
bars; however, the correlation for the other bars is
appreciably worse. For the 16 mm (0.6299 in.), 1 C/ bars,
the predicted values are consistently between two to three
times lower than those observed. For the 3 C/ data, a
number of bars displayed very large variationsboth higher
and lower surface crack widthsthan those observed.

bond strength was reduced by 27 and 29% for the top and
bottom bars, respectively. Comparing Fig. 6 and 7, the bond
strength shows a straight line relationship for average bond
strength against average surface crack width for crack
widths over 1 mm (0.039 in.) (Fig. 6). Extrapolating the data
would indicate that the fall in bond strength commences at a
crack width of approximately 0.5 mm (0.020 in.). The data
for the extent of corrosion against bond strength show a fall
in bond strength with an increase in the extent of corrosion,
but the data are considerably more scattered.
Figures 8 and 9 for 12 mm (0.4724 in.) bars with 3 C/
show an initial fall in bond strength for both the top- and
bottom-cast bars. Again, there is a clear trend for bond
strength and average surface crack width. The bond strength
gradually drops with an increase of surface crack width,
although the shape is logarithmic rather than the straight line
observed at 1 C/; however, results are more scattered for
extent of corrosion and bond strength, particularly for more
highly corroded bars. No difference is observed between
top-cast and bottom-cast bars.
Figures 10 and 11 present the loss of bond strength with
the change of surface crack width and corrosion level for
16 mm (0.6299 in.) bars with 1 C/. The data show a trend
of decreasing bond strength, although a slight increase of
bond strength for low corrosion levels (less than 2%) is
observed for bars with a relative crack width of 0.05 mm
(0.002 in.). The data become more scattered at the higher
crack widths. The bond strength of both top and bottom bars

EXPERIMENTAL RESULTS AND DISCUSSION

Effect of surface crack width and extent of
corrosion on bond strength
Plots of average bond strength against average surface
crack width and extent of corrosionfor top and bottom
barsare given in Fig. 6 to 13.
For 12 mm (0.4724 in.) bars with 1 C/, the bond strength
shows a slight increase in bond strength for the top-cast bars
but remains constant for bottom-cast bars at the appearance of
the first visible crack (0.05 mm [0.002 in.]). The bond strength
then significantly decreases for all bars at a crack width of 1 mm
(0.039 in.) and above. For crack widths of 1 mm (0.039 in.), the
32

ACI Materials Journal/January-February 2011

Table 4Comparison of measured and predicted surface crack widths

Specification

12 mm (0.4724 in.), 1 C/

16 mm (0.6299 in.), 1 C/

12 mm (0.4724 in.), 3 C/

16 mm (0.6299 in.), 3 C/

Measured W, mm (in.)
0.983 (0.0387)
1.248 (0.049)
1.022 (0.04)
1.119 (0.044)
1.360 (0.054)
1.130 (2.34)
1.207 (0.0475)
1.160 (0.046)
0.420 (0.017)
0.422 (0.017)
0.668 (0.026)
0.650 (0.026)
0.583 (0.023)
0.932 (0.037)
0.868 (0.034)
0.897 (0.035)
1.703 (0.067)
1.434 (0.056)
1.450 (0.057)
1.770 (0.07)
0.190 (0.0075)
0.360 (0.014)
0.220 (0.009)
0.739 (0.029)
1.164 (0.046)
0.370 (0.0146)
0.648 (0.0255)
1.758 (0.069)
1.426 (0.056)
1.436 (0.057)
0.409 (0.016)
0.489 (0.019)
0.439 (0.017)
0.379 (0.0149)
0.096 (0.0038)
0.230 (0.009)
0.322 (0.013)
0.136 (0.0054)
0.454 (0.018)
0.205 (0.0081)
0.714 (0.028)
0.580 (0.023)
0.804 (0.032)
1.316 (0.052)

was reduced by 51% at crack widths of 0.9 mm (0.035 in.)

with corresponding corrosion levels of 3.5%.
Results for 16 mm (0.6299 in.) bars with 3 C/ are plotted
in Fig. 12 and 13. Again, a clear trend is observed for the
average surface crack width and bond strength, as with the
12 mm (0.4724 in.) 1 C/ bars. No difference between the
top-cast and bottom-cast bars is observed.
The effect of surface crack width on bond strength loss can
be explained as loss of confinement. Research by Cairns et
al.37 shows that corrosion products do not impair friction
characteristics of a bar/concrete interface where the surface
crack width does not exceed 1.0 mm (0.039 in.), indicating
that the extent of corrosion has less effect on the change of
bond strength.
ACI Materials Journal/January-February 2011

Eq. (6) W, mm (in.)

1.111 (0.044)
1.181 (0.046)
1.504 (0.059)
1.006 (0.040)
1.663 (0.065)
2.340 (0.092)
1.023 (0.040)
1.947 (0.077)
0.374 (0.015)
0.398 (0.016)
0.309 (0.012)
0.377 (0.015)
0.409 (0.016)
0.369 (0.015)
0.267 (0.011)
0.316 (0.012)
0.720 (0.028)
0.559 (0.022)
0.527 (0.021)
0.552 (0.022)
0.011 (0.0004)
0.025 (0.0010)
0.003 (0.0001)
0.624 (0.025)
0.572 (0.023)
0.517 (0.020)
0.524 (0.021)
0.294 (0.012)
0.427 (0.017)
0.452 (0.018)
0.241 (0.009)
0.234 (0.009)
0.266 (0.010)
0.196 (0.0077)
0.107 (0.004)
0.102 (0.0041)
0.041 (0.0016)
0.403 (0.016)
0.436 (0.017)
0.335 (0.0132)
0.292 (0.0115)
3.337 (0.1314)
4.070 (0.16)
1.043 (0.041)

W (measured)/W (Eq. (6))

0.88
1.06
0.68
1.11
0.92
0.48
1.18
0.60
1.12
1.06
2.16
1.72
1.42
2.53
3.25
2.84
2.37
2.57
2.75
3.21
17.27
14.4
73.33
1.18
2.03
0.71
1.23
5.98
3.34
3.18
1.70
2.09
1.65
1.93
0.90
2.25
7.85
0.34
1.04
0.61
2.44
0.17
0.20
1.26

Bond strength
A linear regression analysis was undertaken for average
crack width and extent of corrosion versus bond strength.
The best fit data are presented in Table 5, linear regression,
and Table 6, logarithmic regression.
The data show that a best fit is obtained for the mean
surface crack width for seven out of the eight plots. Only the
16 mm (0.6299 in.) 1 C/ bars gave a better fit with the
extent of corrosion. This would indicate that the average surface
crack width gives a better fit than the extent of corrosion.
The data show that for the 12 mm (0.4724 in.) 3 C/ bars,
the 16 mm (0.6299 in.) 3 C/ bars, and the 16 mm (0.6299 in.)
1 C/ bars, a natural log fit gives a better correlation than a
straight linear fit. Only for the 12 mm (0.4724 in.) 1 C/ bars
33

Fig. 8Average bond strength versus average surface crack

width for 12 mm (0.4724 in.) bars with 3 C/. (Note: 1 MPa
= 144.93 psi; 1 mm = 0.039 in.)

Fig. 11Average bond strength versus extent of corrosion

for 16 mm (0.6299 in.) bars with 1 C/. (Note: 1 MPa =
144.93 psi; 1 mm = 0.039 in.)

Fig. 9Average bond strength versus extent of corrosion for

12 mm (0.4724 in.) bars with 3 C/. (Note: 1 MPa =
144.93 psi; 1 mm = 0.039 in.)

Fig. 12Average bond strength versus average surface

crack width for 16 mm (0.6299 in.) bars with 3 C/. (Note:
1 MPa = 144.93 psi; 1 mm = 0.039 in.)

Fig. 10Average bond strength versus average surface

crack width for 16 mm (0.6299 in.) bars with 1 C/. (Note:
1 MPa = 144.93 psi; 1 mm = 0.039 in.)

Fig. 13Average bond strength versus extent of corrosion

for 16 mm (0.6299 in.) bars with 3 C/. (Note: 1 MPa =
144.93 psi; 1 mm = 0.039 in.)

Table 5Best fit parameters, crack width versus

bond strength, linear analysis, 1 mm = 0.039 in.

Table 6Best fit parameters, crack width versus

bond strength, logarithmic analysis, 1 mm = 0.039 in.

Mean surface crack width

Linear analysis 12 mm 1 C/ 12 mm 3 C/ 16 mm 1 C/ 16 mm 3 C/

Mean surface crack width

Logarithmic
analysis
12 mm 1 C/ 12 mm 3 C/ 16 mm 1 C/ 16 mm 3 C/

R squared
Slope (m)

0.8565
1.6418

0.9331
4.0958

0.458
0.8461

0.6718
2.998

R squared
Slope (m)

0.7595
0.5315

Intercept (b)

3.7187

7.6772
3.0058
Extent of corrosion

6.4958

Intercept (b)

1.9217

R squared
Slope (m)

0.7521
0.072

0.6322
0.4474

0.4313
0.221

0.5763
0.1559

R squared
Slope (m)

0.7445
0.5602

0.7396
1.5784

0.5085
0.6394

0.8362
1.3402

Intercept (b)

3.6739

7.7465

3.2053

6.4846

Intercept (b)

3.6126

6.9422

3.0918

6.972

34

0.9503
2.4284

0.5896
0.3728

0.7507
1.0345

2.944
2.1093
Extent of corrosion

3.1578

ACI Materials Journal/January-February 2011

did the linear fit give a higher R2 value. The data analysis
does not take into account any data points prior to corrosion
being initiated. This is due to the initial increase in bond
strength observed in the 1 C/ bars. For consistency, the 3 C/
bars have been analyzed in the same manner. The data for
the 1 C/ bars show a higher degree of correlation than
the 3 C/ bars. For those bars with 1 C/, a correlation of
greater than 0.856 was observed, indicating that the mean
crack width may be a potential method for measuring the
reduction in bond strength. For the 3 C/ bars, the R2 was
over 0.75 for the 16 mm (0.6299 in.) bars and 0.58 for the
12 mm (0.4724 in.) bars.
The suggested fitting equations for the four sets of
specimens are:
Mean bond strength = 1.6418 (mean crack width) +
3.7187 (12 mm 1 C/ bars);
Mean bond strength = 2.4284Ln (mean crack width) +
2.944 (12 mm 3 C/ bars);
Mean bond strength = 0.3728Ln (mean crack width) +
2.1093 (16 mm 1 C/ bars); and
Mean bond strength = 1.0345Ln (mean crack width) +
3.1578 (16 mm 3 C/ bars).
Influence of C/ ratio
The generally accepted concept of the change of bond
strength with corrosion is that the bond strength will increase
before corrosion-induced cracking is observed. The data
show that the bond strength for 12 mm (0.4724 in.) bars with
1 C/ initially increases for top bars and remains constant for
bottom bars when the first initial cracks occur. For 16 mm
(0.6299 in.) bars with 1 C/, both top and bottom bars show
an increase of bond strength at initial cracking; however, for
12 and 16 mm (0.4724 and 0.6299 in.) bars with 3 C/, the
bond strength reduces at the initial cracking stage. The
difference is probably because of the different range of
corrosion level required for initial cracking for different
C/. It was found that an approximately 1% mass loss was
required for initial cracking for bars with 1 C/, whereas a
2.5% mass loss was needed for initial cracking for bars with
3 C/. Another reason is probably due to the quality of the
surrounding concrete. For bars with 1 C/, the composition
of the concrete around the bar may be less homogeneous
than for bars with 3 C/. This would be analogous with the
increase in top-cast bars compared to bottom-cast bars. The
corrosion products formed in the 1 C/ specimens would
result in an increase in bond strength due to the increase in
internal pressures, caused by the corrosion products
increasing the confinement and mechanical interlocking
around the bar due to the poor-quality concrete coupled with
increased roughness of the bar, resulting in a greater friction
between the bar and the surrounding concrete. The additional
cover in the 3 C/ would nullify this effect; thus, there would
be minimal increase in the mechanical interlock and, once
cracking occurred, these pressures would be relieved,
resulting in the reduction in bond strengths observed.
For cracks greater than 1.0 mm (0.039 in.), the bond
strength has a similar residual strength for 16 mm (0.6299 in.)
bars for the range of C/ tested. The bond strength decreased
more sharply for 3 C/ than for 1 C/ at similar crack widths.
Again, this would indicate that with the increase of C/, the
effect of surface crack width becomes more pronounced.
ACI Materials Journal/January-February 2011

Fig. 14Average surface crack width versus corrosion level

for 12 mm (0.4724 in.) bars. (Note: 1 mm = 0.039 in.)

Fig. 15Average surface crack width versus corrosion level

for 16 mm (0.6299 in.) bars. (Note: 1 mm = 0.039 in.)
Extent of corrosion versus surface crack width
Studies on corrosion cracks have been reported by many
researchers. Several have found that there is a weak
relationship between surface crack width and corrosion
rates7,38; however, others conclude that there is no
relationship between surface crack width and corrosion
rates.4,33,34 Research by Mohammed et al.32 reported that
the surface crack widths were observed to correlate with the
corrosion rate only at the very beginning of the exposure
period. Consequently, there is no accepted clear relationship
between the corrosion level and surface crack width. This is
probably due to nonuniform corrosion at the bar surface.
Localized corrosion significantly affects the extent of
corrosion; however, it has less effect on the average surface
crack width. Another possibility is that once surface cracks
occur, the growth of the surface crack width is not fully
dependent on the growth of corrosion product due to other
cracks occurring inside the concrete.
Figures 14 and 15 show the average surface crack width
versus change of corrosion level. Although there is a trend,
it can be seen that there is a spread of results at the greater
crack widths (Fig. 14). Similar behavior was observed by
Alonso et al.1 and Vidal et al.3 This can be explained due to
localized corrosion having less effect on corrosion cracking.
Figure 15 shows that for 16 mm (0.6299 in.) bars, a higher
corrosion level was required for a larger C/ to achieve a
similar crack width. Corresponding data for 12 mm (0.4724 in.)
bars is inconclusive and additional data would be required to
determine if similar trends occur.
35

In several cases, it was observed that initial crack width

increased at a steady (near constant) (time) rate up to a point
(at approximately 1 mm [0.039 in.] or even larger); after this
point, the rate of crack width growth dramatically increased,
eventually resulting in spalling of the concrete. This suggests
that the surface crack width is initially related to the volume
of corrosion product, but after a critical width is reached,
other factors contribute to crack propagation and the onset of
spalling. Spalling was observed to occur at different crack
widths, between 1 and 2 mm (0.039 and 0.079 in.) for
different bars.
Crack patterns
Three stages in the cracking process were observed. The
initial cracks occurred in a very short period, usually
generated within several days. After that, most cracks grew
gradually until they reached 1 mm (0.039 in.). In the final
stage, it was difficult to predict the surface crack width based
on the corrosion level because it was difficult to ascertain the
condition of the surrounding concrete. The results between
the corrosion level and surface crack width show scatter,
indicating that it is not feasible to reliably relate the extent of
corrosion and the surface crack width when the corrosion is
above a certain degree. The results were found to diverge
when the extent of corrosion is greater than 10%.
The first crack observed was a hairline crack with a width
of approximately 0.05 mm (0.002 in.). This was termed
crack initiation. These continuous longitudinal cracks
appeared on the concrete surface parallel to the bars. Short,
unconnected, longitudinal cracks were also occasionally
observed parallel to the bars at lower corrosion levels.
For specimens with 1 C/, the main longitudinal
corrosion cracks appeared on the vertical sides, and some of
them appeared on either the top or bottom surface. For
specimens with 3 C/, all the main longitudinal cracks were
observed parallel with the bars on the vertical sides of the
specimens. It was also observed that cracks progressed
between two adjacent bars at higher corrosion levels, and
eventually connected. For some specimens, it was found that
there were radial cracks inclined to the main cracks passing
to another face.
CONCLUSIONS
Based on the results reported in this study and
observations during the experimental investigation, the
following conclusions can be drawn:
1. The data indicate that there is a relationship between
bond strength and average surface crack width. This is
clearly evident for specimens with 12 mm (0.4724 in.) bars
with 1 C/ and 3 C/ at all crack widths. For specimens
with 16 mm (0.6299 in.) bars with 1 C/, the relationship is
less apparent at crack width greater than 1 mm (0.039 in.).
2. The relationship between surface crack width and bond
strength is dependent on the C/. The crack appearance is
delayed (requiring greater corrosion) for larger covers. The
C/ should be taken into account when using surface crack
width to assess bond- strength loss.
3. The bond strength has a better relationship with the surface
crack width than the extent of corrosion. This is probably
because once cracks reach a certain width, corrosion has less
impact on crack progression. The data show that a best fit is
obtained for the mean surface crack width than for extent of
corrosion for seven out of the eight plots.
36

4. Pullout tests on unconfined bars with 1 C/ show an

initial increase of bond strength at the first visible external
crack for top-cast bars and no decrease for bottom-cast bars.
Specimens with 3 C/ display a fall in bond strength at
initial cracking.
5. The data show that for the 12 mm (0.4724 in.) 3 C/ bars,
the 16 mm (0.6299 in.) 3 C/ bars, and the 16 mm (0.6299 in.)
1 C/ bars, a natural log fit gives a better correlation.
The 12 mm (0.4724 in.) 1 C/ bars give a better fit with a
linear plot. The suggested fitting equations for the four sets
of specimens are:
Mean bond strength = 1.6418 (mean crack width) +
3.7187 (12 mm [0.4724 in.] 1 C/ bars);
Mean bond strength = 2.4284Ln (mean crack width) +
2.944 (12 mm [0.4724 in.] 3 C/ bars);
Mean bond strength = 0.3728Ln (mean crack width) +
2.1093 (16 mm [0.6299 in.] 1 C/ bars); and
Mean bond strength = 1.0345Ln (mean crack width) +
3.1578 (16 mm [0.6299 in.] 3 C/ bars).
6. The corrosion penetration measured is larger than that
predicted for initial crack appearance. The predicted crack
widths are comparable for 1 C/ but display significant
variation at 3 C/.
REFERENCES
1. Alonso, C.; Andrade, C.; Rodriguez, J.; and Diez, J. M., Factors
Controlling Cracking of Concrete Affected by Reinforcement Corrosion,
Materials and Structures, V. 31, 1998, pp. 435-441.
2. Andrade, C.; Alonso, C.; Rodriguez, J.; and Garcia, M., Cover
Cracking and Amount of Rebar Corrosion: Importance of the Current
Applied Accelerated Tests, Concrete Repair, Rehabilitation and
Protection, R. K. Dhir and M. R. Jones, eds., E&FN Spon, London, UK,
1996, pp. 263-273.
3. Vidal, T.; Castel, A.; and Franois, R., Analyzing Crack Width to
Predict Corrosion in Reinforced Concrete, Cement and Concrete
Research, V. 34, No. 1, 2004, pp. 165-174.
4. Arya, C., and Ofori-Darko, F. K., Influence of Crack Frequency on
Reinforcement Corrosion in Concrete, Cement and Concrete Composites,
V. 26, No. 3, 1996, pp. 345-353.
5. Maaddawy, E. T.; Soudki, K.; and Topper, T., Long-Term
Performance of Corrosion-Damaged Reinforced Concrete Beams, ACI
Structural Journal, V. 102, No. 5, Sept.-Oct. 2005, p. 649-656.
6. Molina, F. J.; Alonso, C.; and Andrade, C., Cover Cracking as a
Function of Rebar Corrosion: Part 2Numerical Model, Materials and
Structures, V. 26, No. 9, 1993, pp. 532-548.
7. Vu, K.; Stewart, G. M.; and Mullard, J., Corrosion-Induced
Cracking: Experimental Data and Predictive Models, ACI Structural
Journal, V. 102, No. 5, Sept.-Oct. 2005, pp. 719-726.
8. Andrade, C.; Alonso, C.; and Molina, F. J., Cover Cracking as a
Function of Rebar Corrosion: Part 1 Experimental Test, Materials and
Structures, V. 26, No. 9, 1993, pp. 453-464.
9. Maaddawy, A. E. T., and Khaled, A. S., Effectiveness of Impressed
Current Technique to Simulate Corrosion of Steel Reinforcement in
Concrete, Journal of Materials in Civil Engineering, V. 15, No. 1, 2003,
pp. 41-47.
10. Amleh, L., and Mirza, S., Corrosion Influence on Bond between
Steel and Concrete, ACI Structural Journal, V. 96, No. 3, May-June 1999,
pp. 415-423.
11. Cairns, J. J.; Du, Y.; and Law, D. W., Residual Bond Strength of
Corroded Plain Round Bars, Magazine of Concrete Research, V. 58, No. 4,
2006, pp. 221-231.
12. Cabrera, J. G., Deterioration of Concrete Due to Reinforcement
Steel Corrosion, Cement and Concrete Composites, V. 18, No. 1, 1996,
pp. 47-59.
13. Almusallam, A. A.; Al-Gahtani, A. S.; Aziz, A. R.; and
Rasheeduzzafar, Effect of Reinforcement Corrosion on Bond Strength,
Construction and Building Materials, V. 10, No. 2, 1996, pp. 123-129.
14. Stanish, K.; Hooton, R. D.; and Pantazopoulou, S. J., Corrosion
Effects on Bond Strength in Reinforced Concrete, ACI Materials Journal,
V. 96, No. 6, Nov.-Dec. 1999, pp. 915-921.
15. Mangat, P. S., and Elgarf, M. S., Bond Characteristics of Corroding
Reinforcement in Concrete Beams, Materials and Structures, V. 32, 1999,
pp. 89-97.

ACI Materials Journal/January-February 2011

16. Fang, C.; Lundgren, K.; Plos, M.; and Gylltoft, K., Bond Behaviour
of Corroded Reinforcement Steel Bars in Concrete, Cement and Concrete
Research, V. 36, 2006, pp. 1931-1938.
17. Lee, H.-S.; Noguchi, T.; and Tomosawa, F., Evaluation of the Bond
Properties between Concrete and Reinforcement as a Function of the
Degree of Reinforcement Corrosion, Cement and Concrete Research, V. 32,
No. 8, 2002, pp. 1313-1318.
18. Rodriguez, J.; Ortega, L. M.; and Casal, J., Corrosion of
Reinforcing Bars and Service Life of Reinforced Concrete Structures:
Corrosion and Bond Deterioration, International Conference on Concrete
across Borders, Odense, Denmark, 1994, pp. 315-326.
19. Chana, P. S., A Test Method to Establish Realistic Bond Stresses,
Magazine of Concrete Research, V. 42, No. 151, 1990, pp. 83-90.
20. Rodriguez, J.; Ortega, L. M.; Casal, J.; and Diez, J. M., Corrosion
of Reinforcement and Service Life of Concrete Structures, Durability of
Building Materials and Components 7, C. Sjostrom, ed., E&FN Spon,
London, UK, 1996, pp. 117-126.
21. Standards Australia Committee AS 1012.11993, Methods of
Testing ConcreteMethod 1: Sampling of Fresh Concrete, Australian
Standard, 1993, 8 pp.
22. Baldwin, M. I., and Clark, L. A., The Assessment of Reinforcing
Bars with Inadequate Anchorage, Magazine of Concrete Research, V. 47,
No. 171, 1995, pp. 95-102.
23. Fang, C.; Lundgren, K.; Chen, L.; and Zhu, C., Corrosion Influence
on Bond in Reinforced Concrete, Cement and Concrete Research, V. 34,
No. 11, 2004, pp. 2156-2167.
24. ASTM G102-89 (2010), Standard Practice for Calculation of
Corrosion Rates and Related Information from Electrochemical
Measurements, ASTM International, West Conshohocken, PA, 2010, 7 pp.
25. Gonzalez, J. A.; Andrade, C.; Alonso, C.; and Feliu, S.,
Comparison of Rates of General Corrosion and Maximum Pitting
Penetration on Concrete Embedded Steel Reinforcement, Cement and
Concrete Research, V. 25, No. 2, 1995, pp. 257-264.
26. Broomfield, J. P., Corrosion of Steel in Concrete: Understanding,
Investigation, and Repair, E&FN Spon, London, UK, 1997, 264 pp.
27. Al-Sulaimani, G. J.; Kaleemullah, M.; Basunbul, I. A.; and
Rasheeduzzafar, Influence of Corrosion and Cracking on Bond Behavior

ACI Materials Journal/January-February 2011

and Strength of Reinforced Concrete Member, ACI Structural Journal,

V. 87, No. 2, Mar.-Apr. 1990, pp. 220-231.
28. Rasheeduzzafar; Al-Saadoun, S. S.; and Al-Gahtani, A. S., Corrosion
Cracking in Relation to Bar Diameter, Cover and Concrete Quality, Journal
of Materials in Civil Engineering, V. 4, No. 4, 1992, pp. 327-342.
29. Mangat, P. S., and Elgarf, M. S., Flexural Strength of Concrete
Beams with Corroding Reinforcement, ACI Structural Journal, V. 96,
No. 1, Jan.-Feb. 1999, pp. 149-158.
30. Anyeung, Y.; Balaguru, P.; and Chung, L., Bond Behavior of
Corroded Reinforcement Bars, ACI Materials Journal, V. 97, No. 2, Mar.Apr. 2000, pp. 214-220.
31. Law, D. W.; Blin, F.; and Wilson, K., Interpretation and Modelling
of Corrosion Rate Measurements on Reinforced Concrete Structures,
Proceedings of the 4th International Conference on Non Destructive
Monitoring, Harrogate, UK, June 2007, pp. 2276-2287.
32. Mohammed, T. U.; Otsuki, N.; Hisada, M.; and Shibata, T., Effect
of Crack Width and Bar Types on Corrosion of Steel in Concrete, Journal
of Materials in Civil Engineering, V. 13, No. 3, 2001, pp. 194-201.
33. Franois, R., and Arliguie, G., Effect of Microcracking and Cracking
on the Development of Corrosion in Reinforced Concrete Members,
Magazine of Concrete Research, V. 51, No. 2, 1999, pp. 143-150.
34. Schiel, P., and Raupach, M., Laboratory Studies and Calculations
on the Influence of Crack Width on Chloride-Induced Corrosion of Steel in
Concrete, ACI Materials Journal, V. 94, No. 1, Jan.-Feb. 1997, pp. 56-62.
35. Ohtsu, M., and Yosimura, S., Analysis of Crack Propagation and
Crack Initiation Due to Corrosion of Reinforcement, Construction and
Building Materials, V. 11, 1997, pp. 437-442.
36. Andrade, C.; Alonso, C.; Rodriguez, J.; and Casal, J. M., Relation
between Corrosion and Cracking, Internal Report of Brite/Euram Project
BE-4062. DG XII, C.E.C, 1995.
37. Cairns, J.; Du, Y.; and Law, D. W., Influence of Corrosion on the
Friction Characteristics of the Steel/Concrete Interface, Construction and
Building Materials, V. 21, No. 1, 2005, pp. 190-197.
38. Okada, K., and Miyagawa, T., Chloride Corrosion of Reinforcing
Steel in Cracked Concrete, Performance of Concrete in Marine
Environment, SP-65, V. M. Malhotra, ed., American Concrete Institute,
Farmington Hills, MI, 1980, pp. 237-254.

37