Construction and Building Materials 54 (2014) 180189

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Effect of corrosion method of the reinforcing bar on bond characteristics

in reinforced concrete specimens
Yoon Suk Choi a,1, Seong-Tae Yi b, Myung Yu Kim a, Woo Young Jung a, Eun Ik Yang a,
a
b

Department of Civil Engineering, Gangneung-Wonju National University, 7, Jukheon-gil, Gangneung-si, Gangwon-do 210-702, South Korea
Department of Civil and Environmental Engineering, Inha Technical College, 100, Inha-road, Nam-gu, Inchon-si 402-752, South Korea

h i g h l i g h t s
 The bond capacity is affected by the corrosion method of rebar in RC specimens.
 The slip at failure was increased when the corrosion level was higher than 5%.
 The bond strength was increased when corrosion level was lower than 1%.
 The expansion pressure of corroded rebar ranged from 9.3 to 13.4 MPa.
 The brittle failure pattern was observed when the area of corrosion exceeded 50%.

a r t i c l e

i n f o

Article history:
Received 13 May 2013
Received in revised form 18 December 2013
Accepted 19 December 2013
Available online 14 January 2014
Keywords:
Bond characteristics
Corrosion method
Articial corrosion
Natural corrosion
Reinforced concrete

a b s t r a c t
Reinforcing bars embedded in concrete structures are corroded, for example, by the ingression of chlorides from sea sand and deicing materials. Such corrosion induces interior pressure which is increased
in the surrounding concrete due to expansion of rebar. As a result, the bond strength and stiffness of reinforcing bar are reduced and the structure eventually undergoes deterioration of concrete which lead to
drastic shortening of service life of concrete structures. Accordingly, many researchers have investigated
on relationship of bond characteristics of reinforced concrete (RC) members and corrosion of the reinforcing bar. One technique to dene the relationship is to articially induce rapid corrosion of the reinforcing
bar. However, this articial corrosion method failed to provide an accurate representation of real conditions and led to overestimation of the performance of RC members in real situations. The objective of this
paper is to investigate the differences of bond characteristics in RC members corroded by articial rapid
and natural corrosion methods. The evaluation on technique suitability was also performed.
The results of this investigation indicated that the failure pattern and the critical corrosion level at
which the bond capacity failed, varied depending on the RC corrosion method. The RC specimens were
deteriorated at a low corrosion stage under natural corrosion conditions. More attention is needed to
be given to the rapid articial RC member corrosion method. Non-destructive test (NDT) methods were
also applied to RC specimens for investigation of the naturally corroded members. A formula of corrosion
area prediction from NDT was proposed.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
When reinforcing bar is corroded, the durability of reinforced
concrete (RC) members will be severely fell and the service life will
be accordingly shortened. The big and long infrastructures become
more vulnerable to safety hazards as their size increase. Accurate
prediction of the service life of these members has become a considerable challenge [1,2]. Corrosion of reinforcing bars in concrete
Corresponding author. Tel.: +82 33 640 2418; fax: +82 33 646 1391.
E-mail addresses: yoons0305@gwnu.ac.kr (Y.S. Choi), eiyang@gwnu.ac.kr (E.I.
Yang).
1
Tel.: +82 33 640 2416; fax: +82 33 646 1391.
0950-0618/$ - see front matter 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2013.12.065

can be caused by sea water, sea sand, deicing salt, or carbonation of

concrete [3,4]. Among these causes, corrosion by chloride ions is a
particular concern since it causes the most serious damage to reinforcing bar.
Research on bond property and other behaviors of RC member
according to the corrosion of the reinforcing bar has been actively
conducted. The investigation on the quantication of corrosion
behaviors is being pursued [5,6]. Nonetheless, corrosion of the
reinforcing bar takes a considerable amount of time and is strongly
inuenced by environmental factors. Consequently, investigations
are difcult to perform.
The most common cause of corrosion initiation of reinforcing
bar in concrete is the ingress of chloride ions to the steel surface.

181

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

Corrosion of rebar damages the reinforced concrete structures in

two ways. First, it reduces the cross-sectional area of reinforcing
bar. Second, it produces corrosion products with a larger volume
than the rebar itself. This volume augmentation induces the tensile
stress in concrete which lead to cracking and structural failure.
In response to this problem, various evaluations have been carried out by using analysis techniques such as corrosion property
analysis of corroded rebar, accelerated corrosion, and simulation
of extreme environmental conditions [79]. However, these experiments do not represent the actual behavior of a structure. The possibility of false representation of corrosion behavior cannot be
excluded. Presently, there are limited studies on the comparison
of corrosion behavior of reinforcing bar in concrete according to
different corrosion methods [1013]. Moreover, corrosion behavior
studies are also limited for natural and articial environmental
conditions.
Therefore, the purpose of this study is to investigate the relationship between the level of reinforcement corrosion and the
bond properties in accordance with the corrosion methods.
In this study, the examination of the variation of bond behavior
of RC members with different corrosion methods is conducted. At
the end of this paper, an appropriate method for predicting service
life of the structure will be suggested. More specically, the differences between members corroded naturally during a long period
and members subjected to an articially accelerated process in
the bond property are investigated. Based on these research outcomes, data guideline for the evaluation of the corrosion property
of such structures is proposed. For naturally corroded case, the
possibility of corrosion prediction by a non-destructive test
(NDT) is also examined in parallel with electrochemical methods.
2. Experiments and measurement
2.1. Experimental program
2.1.1. Test variables
This test was carried out to evaluate the bond characteristics between concrete
and reinforcing bars corroded in RC members. Moreover, corrosion methods of rebar were divided into three patterns. To evaluate the effect of rebar cross-section
reduction (High level corrosion), the rst case, rebar corroded before concrete
placement (method A), was conducted. The corroded rebar was actually prepared
before concrete placement. To evaluate the effect of the tensile stress due to the volume increasing, the second case, rebar corroded articially after concrete placement (method B) in which the rebar was corroded by an electrical accelerating
method. To evaluate the effect of low level corrosion, the nal case (method C), rebar corroded naturally by mixing chlorides into specimens, was applied. Watercement ratio (W/C = 0.4, 0.6), corrosion level of the reinforcing bar, and the amount of
chlorides inside the specimens were considered as test variables. Additionally, with
the aim of developing NDT method for predicting corrosion level, the test results
from specimens corroded naturally were compared and evaluated by using half-cell
potential and polarization resistance methods. Experimental variables are summarized in Table 1. In this study, test data for corrosion were obtained and evaluated
over a period of two years since a long duration is required to induce natural
corrosion.

2.1.2. Specimen preparation

Fig. 1 showed dimension of molds for the pullout test specimens. Various types
of specimens were used for bond tests such as the popular ones from American
Society for Testing and Materials (ASTM) C 234 [14] and modied beam tests. In
ASTM, it was specied that the concentric pullout test can be used only for

comparison purposes. The details of other test setups can be found in a research
report by Chapman and Shah [15]. After evaluating various setups, a modied
version of the setup proposed in Danish Standard (DS) 2082 [16] was chosen to
be used in this study (Fig. 1).
The test specimens were fabricated in cubic with 150  150  150 mm cubes
dimension by considering the appropriate size and shape for the pullout test and
current corrosion measurement. The bond length was set to four times the diameter
(db  25 mm) of rebar (i.e., 4db  100 mm). Moreover, the rebar exposed to air was
treated with a rust inhibitor to prevent its corrosion. After removal from the mold,
the specimen was water-cured in a water chamber for four weeks at a temperature
of 20 3 C. For the case of articially induced corrosion of the rebar (method A),
the bond test was conducted at 28 days. On the other hand, the specimens with
articially corroded rebar after placement (method B) were subjected to accelerated
corrosion for 14 days and a bond test was performed at 28 days. In contrast, the naturally corroded specimens (method C) were air-cured in an outdoor environment
until the occurrence of corrosion. The usual cylinder molds with diameter of
100 mm and height of 200 mm were used to prepare specimens for the compression test.

2.2. Materials and concrete mixture proportion

In this test, D25 deformed bars and ordinary Portland cement were used for the
bond test. Natural sand and crushed stones of 25 mm maximum size were used as
ne and coarse aggregates, respectively. The physical properties of the aggregates
used in the concrete mixture are listed in Table 2.
Table 3 showed the mixture proportion which satised the target slump and
void content of 10 2 cm and 5 1%, respectively. Chemical admixtures were used
in order to get the same slump and void content for each W/C.

2.3. Articial corrosion method of the reinforcing bar

2.3.1. Corrosion of the rebar before concrete placement (method A)
The anode of direct current power supply was connected to the rebar and the
cathode was connected to a mesh wire placed inside a 5% NaCl solution as a method
to corrode the rebar prior to the concrete placement (Fig. 2(a)).
The corrosion level of reinforcing bar was represented by the weight reduction
ratio of corroded rebar. The level was determined by measuring the weight of the
rebar before and after the corrosion procedure. Fig. 3 illustrated the corrosion status
of the rebar.

Level of weight reduction %

weight before test  weight after test

 100
weight before test

2.3.2. Corrosion of the rebar after concrete placement (method B)

Accelerated corrosion test on rebar was carried out by building a direct current
circuit on it with anode and cathode of the direct current power supply connected
to reinforcement and NaCl solution, respectively [7]. The concrete specimen with
embedded rebar was then immersed in a 3% NaCl solution (Fig. 2(b)). The amount
of corrosion was indicated by the quantity of corrosion including the cumulative
current value and the weight reduction ratio. The expected corrosion value with
the current value is estimated by the following equation based on Faradays law.

Amount of corrosion mole

1
2  96; 500

qdt

In this equation, the amount of corrosion (mole) is the number of moles of corroded iron ion (Fe), and q (Ampere) is the amount of current passed through each
stage.
The articial corrosion accelerated test was carried out up to a 10% corrosion
level. However, the expansion pressure of the rebar caused corrosion cracks at only
above 2% of corrosion. When cracks occur, the durability and strength of the specimen were suddenly decreased [8,9]. Accordingly, the bond test was conducted on
the specimens with corrosion up to approximate 2%. Fig. 4 showed the cracks of
specimen due to excessive corrosion.

Table 1
Experimental variables.
W/C (%)

Corrosion method

40

Articial corrosion

Rebar corroded before placement (method A) (corrosion level: 0%, 2%, 5%, 7%, 10%)
Rebar corroded after placement (method B) (expected corrosion level: 0%, 2%, 5%, 7%, 10%)

Measurement
Bond capacities by pullout of rebar

60

Natural corrosion

Corrosion by chloride ion intrusion (method C) (0.6, 0.9, 1.2, 1.5, 1.8, 2.4, 3.0, 3.6 kg/m3)

Bond capacities by pullout of rebar

Half-potential
Polarization resistance

182

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

Fig. 1. Molds and dimensions for pullout test specimens.

Table 2
Physical properties of the aggregate.
Type

Specic
gravity

Absorptivity
(%)

Fineness modulus
(F.M.)

Fine aggregate
Coarse
aggregate

2.59
2.64

0.99
0.70

2.65
6.82

Table 3
Concrete mix proportion.
W/C (%)

40
60

S/a (%)

43
46

Unit weight (kg/m3)

Admixture (kg)

Water

Cement

Sand

Gravel

170
172.5

425
288

716
813

1022
978

2.125
0.864

2.4. Natural corrosion method of the reinforcing bar (method C)

When concrete structures are exposed to chloride contaminated environments,
chloride ions can penetrate into concrete cover and reach the reinforcing bar. Thus,
the penetration of chloride ions is the most important factor in the corrosion of
reinforcing bars embedded in concrete. Several laboratory test methods have been
developed to realize the simulation of corrosion on reinforcing bar in concrete.
However, it takes very long time for chloride ions to reach the reinforcing bar since
the penetration of chlorides into concrete is slow even under severe test condition
[1113]. Thus, in method C, chlorides were directly mixed in fresh concrete to

(a) before concrete placement (method A)

induce the natural corrosion surrounding reinforcing bar. The chloride contents
were 0.6, 0.9, 1.2, 1.5, 1.8, 2.4, 3.0, and 3.6 kg/m3. The specimens were put in a
chamber with 60% of RH and 20 C of temperature. Natural corrosion test (method
C) was performed within a period of two years. The corrosion progress of reinforcing bar in concrete was estimated by the NDT. Fig. 5 showed the method used to
compare the degree of corrosion measured by NDT methods and the actual amount
of corrosion of the rebar inside the concrete specimen. The surface of the rebar perpendicularly penetrated the cross-section of the reference electrode and was considered when measuring the polarization resistance. Then, the total cross-section
and the ratio of corrosion area were calculated as the corroded area of the total
cross-section.
When the half-cell potential method was applied, the concrete surface was
maintained in a condition of wetness for 30 min prior to the measurement. The
standard electrode (i.e., saturated copper sulfate electrode) was then connected
to the anode and the rebar inside the concrete was connected to the cathode. After
building a cell by combining these elements, the potential difference between cathode and anode of the assembly was measured [17].
In addition, the measurement of polarization resistance of the rebar was carried
out by using a PR-4500 model device from CC Technologies. The standard electrode
was a saturated copper sulfate electrode. A guard electrode was used to conne the
reference electrode and the polarization area. The rebar and electrodes were electrically conducted through a sponge wetted with liquid detergent. The shift range of
potential during the polarization resistance measurement was from 20 mV to
+20 mV based on the corrosion potential. The polarization area of the rebar was
then computed by measuring the surface area of rebar which was covered by the
vertical projection of the reference electrode cross-section [18].
2.5. Bond properties with corrosion of the reinforcing bar
The bond property evaluation of the corroded rebar was executed by a pullout
test by considering the corrosion degree of the rebar as the main variable. In this
test, a Universal Testing Machine (UTM) was used to apply the load. The experiment
was performed by using a load control method until the specimen failed. The

(b) after concrete placement (method B)

Fig. 2. Articial corrosion methods for the reinforcing bar.

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

183

development. In addition, the compressive strength at 28 days

was 21 MPa and 32 MPa, respectively, at W/C 0.6 and 0.4.
3.2. Bond properties of the rebar corroded articially

Fig. 3. Appearance after corrosion of the rebar (method A).

Fig. 4. Appearance of cracking for the specimen with 2% corrosion level (method B).

amount of relative slip by loading was measured by using two linear variable displacement transducers (LVDTs) which were installed between the rebar and concrete (Fig. 6). The displacement and loading value at each step were then
automatically recorded. The strain in the reinforcing bar was very low and its yielding was not observed in any tests.

3. Experimental results and evaluation

3.1. Compressive strength
Fig. 7 showed the development of the concrete compressive
strength used in this experiment. The experimental results indicated that compressive strength increased with age, and the
long-term strength of the low W/C concrete showed larger

3.2.1. Case of rebar corroded before concrete placement (method A)

For each corrosion level of 0%, 2%, 5%, 7%, and 10%, three specimens were prepared for each of W/C values (0.4 and 0.6). With
these samples, the bond property of the RC member with corrosion
of the reinforcing bar was evaluated and the load and free-end slip
were measured at 28 days.
The test results consisted of loadslip curve (Fig. 8) and maximum bond load (Fig. 9) according to the corrosion level. The slip
at maximum load was measured and found from 0.3 to 0.5 mm
for the un-corroded specimens. This slip tended to increase at a
high corrosion level. The loadslip curves did not show any clear
peak in this test since failure pattern showed pull-out failure (with
splitting of concrete) rather than slip failure (without splitting of
concrete). Specically, the specimens exhibited brittle failure pattern with small slip when the corrosion level lower than 5%. On
the other hand, a ductile failure pattern was manifested at 10% of
corrosion level and the slip at the failure was gradually increased
when the corrosion level higher than 5%.
Although the maximum bonding load varied with the strength
of concrete (Fig. 9), the load capacity of high-strength concrete
changed considerably as the corrosion level increased. It appeared
that as the W/C reduced, the bond strength diminution was manifested quicker due to the increase of corrosion. It means that high
strength concrete became more vulnerable after critical corrosion.
3.2.2. Case of rebar corroded by electrical accelerating method after
concrete placement
The corrosion of the reinforcing bar embedded in concrete differs by various factors (i.e., type of cement, quality of the concrete,
surrounding environmental conditions). More effort is needed for
accurate evaluation of such corrosion. In this study, a bond test
of rebar corroded by the accelerating method (method B) was carried out since it requires short time. In this test, D25 deformed bars
were used to prepare three specimens for each values of W/C (0.4
and 0.6) and for each corrosion level. The loadslip curves were
then measured and evaluated.
As mentioned in Section 2.3.2, although articially imposed corrosion was possibly attained 2% of corrosion level, the specimens
failed prior to reaching the desire level of corrosion due to the
expansion pressure when the corrosion level exceeded 2%. Accordingly, this study was carried out with a bond test on the specimens
up to 2% of corrosion level. When the corrosion level approached
approximately 2%, cracks started to form along the corroded bar.
Because the corrosion products make about 10 times of the original
volume, 2% mass loss translates to 20% volume increase [7]. The

Fig. 5. Measurement of the corrosion area (method C).

184

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

Fig. 6. Pullout test setup of the loading point.

low level corrosion. Although there was not signicant change on

reinforcement cross-section, high expansion pressure and cover
cracking were known to be occurred in early corrosion process of
corroded RC members. Therefore, it is necessary to determine the
expansion pressure of the corroded bars. Theoretically, the expansion pressure according to the corrosion of the rebar can be calculated in the following equation.

Compressive strength (MPa)

60

50

40

30

ft  A

20

W/C=0.4
W/C=0.6

10

0
0

300

600

900

1200

Age (days)
Fig. 7. Development of the compressive strength with age.

bond behavior of corroded RC members has been previously investigated through experimental and analytical studies and most of
them were focused on corroded bar experiencing severe corrosion.
However, not much attention was given to the bond behavior in

Lp

100

(b)

(a)
80

Load (kN)

80

Load (kN)

where ft is the tensile strength of concrete (MPa); D is the diameter

of the rebar (mm); A is the area of concrete excluding the cross-section of the rebar (mm2); L is the bond length of the rebar (mm); and
p is the expansion pressure of the rebar (MPa).
In this study, when corrosion cracks occurred in the RC specimens, the expansion pressure (p) of D25 deformed bars was found
from 9.3 to 13.4 MPa.
Meanwhile, Fig. 10 showed the test results consisted of load
slip curves according to the corrosion level. The relationship between maximum bond load and corrosion level was also showed
in Fig. 11. As expected, low corrosion level provided better bond
strength which conrmed other published results [19]. At initial
stages of corrosion (less than 1% corrosion level), the loadslip
curve was steeper compare to that at high corrosion level because
the corrosion products improved the friction at the interface.

100

60

40

0%
2%
5%
7%
10 %

20

0
0.0

pD

0.2

0.4

0.6

Slip (mm)

0.8

60

40

0%
2%
5%
7%
10 %

20

1.0

0
0.0

0.2

0.4

0.6

Slip (mm)

Fig. 8. Loadslip curve with corrosion level (method A): (a) W/C = 0.4 and (b) W/C = 0.6.

0.8

1.0

185

100

100

80

80

Maximum bond load (kN)

Maximum bond load (kN)

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

60

40

W/C=0.4
W/C=0.6

20

0
0

60

40

W/C=0.4
W/C=0.6

20

0
0.0

10

0.5

1.0

1.5

2.0

2.5

Corrosion level (%)

Corrosion level (%)

Fig. 9. Maximum bond load versus corrosion level (method A).

Fig. 11. Maximum bond load versus corrosion level (method B).

Moreover, it can be seen from Fig. 11 that there was a decrease

in bond strength when corrosion level exceeds 1%. Although the
bond stiffness and maximum load increased when the corrosion level was lower than 1%, they decreased as the corrosion went beyond this level which conrmed the optimal level of corrosion on
reinforcing bar.

researchs specimens were experienced severe condition (more

than 24% of corrosion rate in mass) and had longitudinal cracks
before the test. Furthermore, initial stage and low level of corrosion
were induced by the articial condition. This corrosion progress
may be faster than that of the natural condition. Thus, the level
of initial stage corrosion progressed in natural condition (such as
the time-to-corrosion initiation) is lower than that of the articial
condition.

3.3. Bond properties of the rebar corroded naturally by mixing

chlorides
In this study, different amount of chlorides were added to fresh
concrete and natural corrosion was induced by exposing those
specimens to an outdoor environment for over two years. These
specimens were prepared according to W/C values. The relationship of bond strength and the corrosion level was then evaluated
at each planned stage. Since the level of corrosion was not severe
in two years, the weight loss due to corrosion was not noticeable.
Consequently, the relationship between the corrosion area ratio
and the bond properties of the rebar was examined. The bond test
on the corroded reinforcing bar in high corrosion levels has been
conducted by other researchers [611]. However, most of these

3.3.1. Slip properties with corrosion of the rebar

The loadslip curves and slip properties according to the corrosion area ratio of the rebar are exhibited in Figs. 12 and 13, respectively. In these gures, it was noticed that the amount of slip
decreased as the area of corrosion increased regardless of W/C. Particularly, brittle failures were manifested even in the case of specimens with a corrosion area ratio of over 50%. This phenomenon
was observed as described in Section 3.2.1.
Since natural corrosion was produced, the weight of rebar reduced by the corrosion was not notable and only the corrosion area
of the rebar inside the concrete specimen could be measured. In
this case, the corrosion state of rebar with corrosion area exceeding

100

100

(a)

(b)
80

Load (kN)

Load (kN)

80

60

40

0%
0.5 %
0.9 %
1.1 %
1.7 %

20

0
0.0

0.2

0.4

0.6

Slip (mm)

0.8

60

40

0%
0.8 %
1.3 %
1.8 %
2.3 %

20

1.0

0
0.0

0.2

0.4

0.6

Slip (mm)

Fig. 10. Loadslip curve with corrosion level (method B): (a) W/C = 0.4 and (b) W/C = 0.6.

0.8

1.0

186

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

100

100

(b)

(a)
80

Load (kN)

Load (kN)

80

60

40

0%
15 %
35 %
50 %
70 %

20

0
0.0

0.2

0.4

0.6

0.8

60

40

0%
15 %
50 %
80 %
95 %

20

1.0

0
0.0

0.2

0.4

Slip (mm)

0.6

0.8

1.0

Slip (mm)

Fig. 12. Loadslip curve with corrosion area rate ratio (method C): (a) W/C = 0.4 and (b) W/C = 0.6.

100

W/C=0.4
W/C=0.6

Maximum bond load (kN)

Maximum slip (mm)

0.8

0.6

0.4

0.2

80

60

W/C=0.4
W/C=0.6

0.0
0

20

40

60

80

100

Corrosion area ratio (%)

40
0

20

40

60

80

100

Corrosion area ratio (%)

Fig. 13. Maximum slip versus corrosion area ratio (method C).
Fig. 14. Maximum bond load versus corrosion area ratio (method C).

50% was almost identical to that with 0% of corrosion area in

method A and method B. However, the research results revealed
that the specimen took the pattern of brittle failure as the area of
corrosion increased. Particularly, brittle failure was the predominant pattern even in the concrete specimen with the corroded
rebar that did not have a measurable reduction in weight.
3.3.2. Bond properties with corrosion of the rebar
Fig. 14 showed the relationship between maximum bond load
and corrosion area ratio of rebar with a mixture proportion which
was used to investigate the variation of bond strength according to
corrosion. From the test results, although the maximum load did
not change greatly when the corrosion area ratio reached 50%, that
of the specimens with W/C value of 0.6 decreased by over 10%
when the area of corrosion exceeded 80%.
However, according to the experimental results for the case of
electrical corrosion of the rebar before the concrete placement
(method A), corrosion typically had a positive effect on bond
strength when its level was lower than 5%. On the other hand, bond
strength increased when corrosion level was lower than 1% for the

case of rebar corroded after the concrete placement (method B).

Hence, electrical corrosion may lead to a high level of corrosion
and the corrosion products around the rebar increase the friction
which caused an increase in bond strength.
For the natural corrosion test, it was found that the bond
strength decreased as the area of corrosion on the surface of the rebar increased. The same manner was found in the case of an unnoticeable corrosion level, (i.e., close to 0% corrosion level) of the
rebar. When corrosion products are not generated and only rust
is formed on the surface of the rebar as in the case of natural corrosion, the bond between the surface of the reinforcing bar and
concrete decreases, causing the reduction of bond strength. The
test results indicated that natural corrosion should be used in order
to realistically evaluate the bond strength between corroded reinforcing bar and concrete at initial state of corrosion.
3.4. Non-destructive evaluation of naturally corroded specimens
The evaluation of natural corrosion progression of embedded
reinforcing bar was performed. The corrosion level was measured

187

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

3.4.1. Evaluation by half-cell potential method

Fig. 15 showed the relationship between the actual corrosion
area ratio and the distribution of measured half-cell potential.
The gure showed that the half-cell potential had lower distribution pattern as the corrosion area increased. According to the ASTM
C 876 specication [14], when the potential is over 200 mV and
below 350 mV, the probability of corrosion is below 10% and over
90%, respectively. However, when the ratio of the actual area of
corrosion was below 10% and over 60%, the potential was generally
distributed between 100 to 250 mV and below 300 mV in
average, respectively. It indicated that the test results obtained
from this study is different from the specication in ASTM C 876
[14]. Additionally, when the ratio of corrosion area was 60% or
above, the corrosion probability was distributed from 250 to
350 mV. It showed that there is difculty in quantifying the level
of corrosion. As a result, only qualitative judgment of the presence
or absence of corrosion is possible.
The equation in Fig. 15 was obtained from regression analyses
by using the relationship of actual corrosion area ratio (y, %) and
the measured half-cell potential (x, mV).
Fig. 16 showed the distribution of half-potential according to
the amount of chlorides inside the concrete. This gure indicated
that the distribution of the potential reduced as W/C and the quantity of added chlorides increased. It is attributed to the increase of
conductivity of micro-current as the concentration of salinity increased. Furthermore, the moisture absorbed by the voids in concrete and the chlorides acted as electrolytes which kept the
surface of concrete moist for the measurement of half-cell potential. Meanwhile, as the W/C ratio increased, the half-cell potential
showed a reduction of distribution which is attributed to the
amount of chloride and the water content inside the concrete affected the half-cell potential. However, when the quantity of chlorides inside the concrete is over 3.0 kg/m3 and the specimen is
subjected to corrosion, the inuence of water content is negligible
as compare with the uncertain level (250  350 mV) of potentials of the result with the specication .

Half-cell potential (mV vs CSE)

by non-destructive methods of half-cell potential and polarization

resistance. The area measured by the polarization resistance method was then considered as the total area. The ratio between the
corroded area and the total area of rebar was then computed and
a comparison between actual corrosion area and values measured
from non-destructive methods was carried out.

-200

-300

-400

-500
0.0

0.8

1.6

2.4

3.2

4.0

Chloride content (kg/m )

Fig. 16. Change of half-cell potential versus chloride content.

3.4.2. Evaluation by polarization resistance method

Fig. 17 showed the relationship o ratio of actual corrosion area
and polarization resistance. This gure indicated the value of
polarization resistance distribution which is below 100 kO cm2
and over 150 kO cm2 for the specimens with an 80% or higher
and less than 20% of actual corrosion area ratio, respectively. The
standard for the evaluation of corrosion due to polarization resistance (CEB Bulletin No. 243) [20] species that the speed of corrosion is very fast when the value of polarization resistance is below
26 kO cm2. According to the test results, if the standard is used, it is
expected that the corrosion phenomenon of the real reinforcing bar
can be underestimated. Thus, CEB Bulletin No. 243 [20] needs correction for considering the relationship between the value of polarization resistance and the corrosion of actual reinforcing bar. It
depends on the equipment setup used for the corrosion study.
A regression equation to obtain the relationship between the ratio of actual area of corrosion (y, %) and the value of polarization
resistance (x, kX cm2) is expressed in Fig. 17. This equation was
proposed from the results of natural corrosion. It is expected that
this relational equation can be used to reasonably predict the corrosion state.

100

100

y=-0.34x-54.0
2
R =0.56

80

Corrosion area ratio (%)

Corrosion area ratio (%)

W/C=0.4
W/C=0.6

-100

60

40

20

0
-500

y=-23.4ln(x)+145.6
2
R =0.79

80

60

40

20

0
-400

-300

-200

-100

Half-cell potential (mV vs CSE)

Fig. 15. Corrosion area ratio versus half-cell potential.

100

200

300

400

500
2

Polarization resistance (kcm )

Fig. 17. Corrosion area ratio versus polarization resistance.

600

188

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

polarization resistance showed a gradual decreasing trend as the

corroded area of the rebar increased. It was noted that the use of
quantitative analysis on the value of polarization to conrm the level of corrosion phase is.

Polarization resistance (kcm )

500

W/C=0.4
W/C=0.6

400

4. Conclusions
300

In this study, the bond property of corroded rebar at each corrosion stage due to different corroding methods of reinforcing
bar were compared and evaluated. The conclusions obtained from
this study are summarized below:

200

100

0
0

20

40

60

80

100

Corrosion area ratio (%)

Fig. 18. Change of polarization resistance with corrosion area ratio.

Polarization resistance (kcm )

600

500

W/C=0.4
W/C=0.6

400

300

200

100

0
0.0

0.8

1.6

2.4

3.2

4.0

Chloride content (kg/m )

1. Even if the bond load is increased when corrosion level

reaches 5% in the case of corrosion of the rebar prior to
the concrete placement (method A), the failure pattern is
gradually transferred to ductile mode as the corrosion
level grows higher than 5%.
2. The reduction in bond load due to corrosion was more
noticeable for high-strength concrete. Conversely, for the
case of corrosion of the rebar after concrete placement
(method B), the bond strength and maximum load
increased until the actual corrosion level reached 1%. However, cracks, which cause a signicant reduction in the
bond strength, were observed in concrete when the level
of corrosion exceeded 2%.
3. For the case of natural corrosion of the rebar with minimal
corrosion close to a corrosion level of 0%, the specimens
exhibited a brittle failure pattern as the area of corrosion
increased and the bond strength decreased. The trend for
brittle failure was observed even when the area of corrosion exceeded 50%. Moreover, the maximum load was
decreased by about 10% or more for the case of specimens
with an area of corrosion of 80% or above and 60% of W/C.
Thus, it is deemed that the deterioration of a concrete
member due to corrosion could be underestimated in
specimens corroded by an articially induced method.
Such evaluations should be conducted with naturally corroded specimens.
4. The concrete properties and concentration of chlorides
inside concrete did not greatly affect the polarization
resistance technique. Accordingly, this method made it
possible to quantitatively assess corrosion levels.

Fig. 19. Change of polarization resistance versus chloride content.

Acknowledgments
The data for Fig. 17 were used again in Fig. 18 to examine the
distribution of the polarization resistance according concrete properties. From the relationship, it is found that the value of polarization resistance is somewhat high when W/C is low for a corrosion
area ratio range from 10% to 60%. The difference in the polarization
resistance is attributed to the increase of concrete resistance since
concrete becomes water-tight when W/C is low or admixture is
added. Due to this resistance, the ow of micro-current also becomes faster. In the case of a corrosion area ratio of 80% or above,
corrosion was observed throughout the specimen. However, the effect of W/C was not signicant. Moreover, the effect of resistance
due to the corrosion products increased and was consistent with
the half-cell potential measurements.
Fig. 19 showed the changes in the polarization resistance
according to the amount of chlorides in concrete. The half-cell potential showed lower distribution as the amount of chlorides increased. On top of that, the value of polarization resistance was
also decreased with an increase of chloride content as the progression of corrosion due to the chlorides inside concrete.
In summary, the polarization resistance method was not greatly
affected by the physical properties of concrete. The value of

This research was supported by Basic Science Research Program

through the National Research Foundation of Korea (NRF) funded
by the Ministry of Education, Science and Technology (No. 20100023118).
References
[1] Khatri RP, Sirivivatnanon V. Characteristic service life for concrete exposed to
marine environments. Cem Concr Res 2004;34(5):74552.
[2] Schueremans L, Gemert DV, Giessler S. Chloride penetration in RC-structures in
marine environment long term assessment of a preventive hydrophobic
treatment. Constr Build Mater 2007;21(6):123849.
[3] Mohammed TU, Hamada H, Yamaji T. Performance of seawater-mixed
concrete in the tidal environment. Cem Concr Res 2004;34(5):593601.
[4] Dias WPS, Seneviratne GAPSN, Nanayakkara SMA. Offshore sand for reinforced
concrete. Constr Build Mater 2008;22(7):137784.
[5] Vidal T, Castel A, Franois R. Corrosion process and structural performance of a
17 year old reinforced concrete beam stored in chloride environment. Cem
Concr Res 2007;37(8):155161.
[6] Parande AK, Babu BR, Karthik MA, Kumaar D, Palaniswamy N. Study on
strength and corrosion performance for steel embedded in metakaolin blended
concrete/mortar. Constr Build Mater 2008;22(3):12734.
[7] Auyeung Y, Balaguru P, Chung L. Bond behavior of corroded reinforcement
bars. ACI Mater J 2000;91(2):21420.

Y.S. Choi et al. / Construction and Building Materials 54 (2014) 180189

[8] Lee HS, Noguchi T, Tomosawa F. Evaluation of the bond properties between
concrete and reinforcement as a function of the degree of reinforcement
corrosion. Cem Concr Res 2002;32(8):13138.
[9] Fang C, Lundgren K, Chen L, Zhu C. Corrosion inuence on bond in reinforced
concrete. Cem Concr Res 2004;34(11):215967.
[10] Abosrra L, Ashour AF, Yousef M. Corrosion of steel reinforcement in concrete
of different compressive strengths. Constr Build Mater 2011;25(10):391525.
[11] Pfeifer DW, Landgren RJ, Zoob A. Protective systems for new prestressed and
substructure concrete (FHWA-RD-86-193). McLean, VA: Federal Highway
Administration; 1987.
[12] ASTM C1543. Standard test method for determining the penetration of
chloride ion into concrete by ponding; 2002.
[13] ASTM G109. Standard test method for determining the effects of chemical
admixtures on the corrosion of embedded steel reinforcement in concrete
exposed to chloride environments; 2002.

189

[14] ASTM C 234. Standards test method for comparing concretes on the basis of
the bond developed with reinforcing steel; 1991.
[15] Chapman RA, Shah SP. Early-age bond strength in reinforced concrete. ACI
Mater J 1987;84(6):50110.
[16] Danish Standards Organization, Pullout Test. DS 2082, Copenhagen; 1980. p. 2.
[17] ASTM C876. Standards test method for half-cell potentials of uncoated
reinforcing steel in concrete; 1999.
[18] ASTM G59. Standard test method for conducting potentiodynamic polarization
resistance measurements; 1997.
[19] Fu X, Chung L. Effect of watercement ratio, curing age, silica fume, polymer
admixtures, steel surface treatments, and corrosion on bond between concrete
and steel reinforcing bars. ACI Mater J 1998;95(6):72534.
[20] CEB Bulletin, No.243. Strategies for testing and assessment of, concrete; 1998.