Construction and Building Materials xxx (xxxx) xxx

Contents lists available at ScienceDirect

Construction and Building Materials

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

Corrosion rate measurement by using polarization resistance method for

microcell and macrocell corrosion: Theoretical analysis and
experimental work with simulated concrete pore solution
Lijie Chen, Ray Kai Leung Su ⇑
Department of Civil Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong, China

h i g h l i g h t s

Polarization resistance method (PRM) is applied to microcell and macrocell corrosion (MMC)

Stern-Geary equation is modified to account for the effects of macrocell current.

Dispersion of the imposed polarization current affects the corrosion rate measured by PRM for MMC.

Macrocell current is correlated with the measured corrosion rate of passive rebars during MMC.

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

Article history: The applicability of the polarization resistance method to microcell and macrocell corrosion (MMC) has
Received 12 May 2020 been often ignored. To address this research gap, this study investigates corrosion rate measurement by
Received in revised form 9 September 2020 using the polarization resistance method on MMC of reinforced concrete. A modified Stern-Geary equa-
Accepted 15 September 2020
tion for MMC is proposed to consider the effects of the macrocell current. The Tafel kinetic parameters
Available online xxxx
and polarization resistance are obtained by conducting potentiodynamic polarization tests in a simulated
concrete pore solution. The proposed equation is validated for both active and passive steel rebars. A sig-
Keywords:
nificant overestimation is found for passive steel rebars while a substantial measurement error is found
Reinforcement corrosion
Corrosion rate
for active steel rebars due to the macrocell current.
Macrocell corrosion Ó 2020 Elsevier Ltd. All rights reserved.
Polarization resistance method
Stern Geary equation
Tafel slope

1. Introduction on-site measurements. Compared with other corrosion monitoring

methods, the polarization resistance method is direct, quick and
The corrosion of reinforcement steel is one of the major reasons accurate. A promising testing method in [5] has been developed
for the deterioration of reinforced concrete (RC) structures. In based on the polarization resistance method without an electrical
China, the costs related to corrosion in the infrastructure sector connection to the corroding rebar.
was 74.91 billion RMB (~10.5 billion USD) in 2014 [1] but an esti- Based on whether a macrocell current is found on the reinforce-
mated 25% to 30% of such corrosion related costs could be elimi- ment, the corrosion can be categorized as [6]:
nated with the use of appropriate anti-corrosion strategies [2].
An effective monitoring method of the rate of corrosion therefore (a) Microcell corrosion: there is no macrocell current on the
has great significance for preventative purposes by identifying reinforcement. Anodes and cathodes are immediately adja-
the initiation of the corrosion and determining the efficiency of cent, and
corrosion prevention methods and repair work. (b) MMC: there is a macrocell current on the reinforcement.
As the only method that can be used to directly determine the Anodes and cathodes are spatially isolated but electrically
corrosion current density, the polarization resistance method connected.
[3,4] has been a common practice for both research purposes and
In particular, note that the polarization resistance method
⇑ Corresponding author. assumes that the metal subjected to testing shows the presence
E-mail address: klsu@hku.hk (R.K.L. Su). of microcell corrosion. However, this assumption is open to debate

https://doi.org/10.1016/j.conbuildmat.2020.121003
0950-0618/Ó 2020 Elsevier Ltd. All rights reserved.

Please cite this article as: L. Chen and Ray Kai Leung Su, Corrosion rate measurement by using polarization resistance method for microcell and macrocell
corrosion: Theoretical analysis and experimental work with simulated concrete pore solution, Construction and Building Materials, https://doi.org/
10.1016/j.conbuildmat.2020.121003
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

with RC corrosion. In general, chloride induced corrosion is associ- in previous studies and the pore solution of concrete is simulated
ated with MMC [7–9]. The anode under a chloride attack in RC is by mixing tap water and demineralized water. Therefore, the use
usually localized as shown in Fig. 1, and surrounded by a larger of the polarization resistance method on cathodes and anodes
cathode area within the same rebar. Thus, a macrocell current is under different C/A ratios and electrolyte environments needs to
generated between the anode and cathode. The macrocell current be further investigated.
has been shown to dominate in chloride induced corrosion. The In this paper, the feasibility of using the polarization resistance
contribution of the macrocell current to the total corrosion current method for MMC is studied by means of a theoretical analysis and
has been found to be over 0.5 in different cathode to anode (C/A) experiments with a simulated concrete pore solution. The polariza-
ratios [10]. The MMC in cracked concrete can be further acceler- tion resistance is obtained by using the linear polarization resis-
ated due to the increased penetration of oxygen, chloride and tance technique. Both the anodes and cathodes are examined
moisture [11–13]. The ratio of the macrocell to microcell current with different C/A ratios and in different electrolyte environments.
density is 31 after an electrical connection is created between The results are discussed in consideration of increasing the accu-
the localized anode and surrounding cathode in the pitting corro- racy of measurements conducted by using the polarization resis-
sion of cracked concrete [14]. However, a large macrocell current tance method for MMC with respect to both anodes and cathodes.
makes it difficult to measure the maximum depth of the pitting
corrosion, which has been found to be 8 times that of the measured 2. Theoretical background
average corrosion penetration in general [15]. Carbonation-
induced corrosion has been conventionally regarded as microcell The alkaline environment of the pore solution of RC facilitates
corrosion. However, concrete carbonation is a slow chemical pro- the formation of a passive film on reinforcements. This passive
cess and the concrete around the rebars is partially carbonated film, which can protect the reinforcement steel from corrosion,
most of the time as shown in Fig. 2, which means that a macrocell ruptures due to concrete carbonation or chloride attacks. Thus,
current can flow from the active to the passive rebars given enough the corrosion of the reinforcement commences and propagates.
oxygen and moisture. More than 50% of the mass loss could be Redox reactions are used to describe the process of RC corrosion
caused by the macrocell current of partially carbonated RC [16]. as follows:
With an increase in the C/A ratio [17], the macrocell current can
accelerate the corrosion rate in carbonation-induced corroded (a) Iron oxidation (anodic reaction):
structures by 10% to 200%.
The polarization resistance method has been extensively used Fe ! Fe2 þ þ 2e -
ð1Þ
in both fieldwork and laboratory research. Nevertheless, the
assumption that there is microcell corrosion might be neglected
by engineers and researchers, so that measurements carried out
by using the polarization resistance method for MMC might be (b) Oxygen reduction (cathodic reaction):
not appropriately interpreted. Although the applicability of the - -
polarization resistance method for MMC has been debated [8,18–
O2 þ 2H2 O þ 4e ! 4OH ð2Þ
20], there are few studies that have examined the feasibility of Anodic reactions only occur on active reinforcements with
using this method for MMC. Angst and Büchler [21] stated that a unstable passive film [23], while oxygen could be reduced at the
well-defined mixed potential does not exist and the use of the surface of both active and passive reinforcements.
polarization resistance method is fundamentally incorrect for The volume expansion of rust results in cracking and even spal-
MMC. They found errors that may be caused by the use of an ling of the concrete cover, which significantly accelerate the prop-
apparent electrode potential and polarization current; however, agation of corrosion. As a result of the corrosion, the cross sectional
under some circumstances, acceptable results can be obtained area of the embedded rebars is reduced, which leads to the degra-
because the errors cancel each other out. In considering the disper- dation of the structural performance.
sion of an applied polarization current, they further developed a
new approach to reduce the errors in the galvanostatic pulse tech- (1) Tafel kinetics for iron oxidation and oxygen reduction
nique (which is a polarization technique) [22]. It should be noted Iron oxidation and oxygen reduction at the interface
that only anodes under a high C/A ratio have been investigated between steel and the concrete pore solution can be

O2 O2 O2 Microcell Localized O2 O2 O2
current anode

Macrocell current

Fig. 1. Chloride-induced pitting corrosion.

2
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

Cathode Anode
Carbonated concrete
Electrical connection
by chair or bent-up bar

CO2 O2
Fig. 2. Partially carbonated RC slab.

described by using Eqs. (1) and (2). The electrode kinetics @E

ba ¼ ð5Þ
can be described by using the relationship between the cur- @logia
rent density and electrode potential. Assuming that there is
a large overpotential from the equilibrium state, the anodic @E
bc ¼ ð6Þ
and cathodic reactions of corrosion can be simplified as Tafel @logic
behaviour [24]:
The macrocell current density is defined as:
EEoa
ia ¼ ioa e2:3 ba ð3Þ imac ¼ ia  ic ð7Þ

EEoc
ic ¼ ioc e2:3 bc ð4Þ
where ia is the anodic current density, ic is the cathodic current den- (2) Microcell corrosion
sity, ioa is the anodic exchange current density, ioc is the cathodic In microcell corrosion, the corrosion is uniformly distributed
exchange current density, ba is the anodic Tafel slope, bc is the over the surface of the steel. The steel surface is electrically
cathodic Tafel slope, Eoa is the anodic equilibrium potential, and isolated from the steel at different states of corrosion. Micro-
Eoc is the cathodic equilibrium potential. cell corrosion is in a dynamic equilibrium state, in which the
The Tafel slope is defined as the slope of the E-logi curve (Fig. 3): anodic current density is equal to the cathodic current den-
sity. Thus, no macrocell current flows through the steel–con-
crete interface:

imac ¼ ia  ic ¼ 0 ð8Þ
Eqs. (3) and (4) are substituted into Eq. (8):
Ecor Eoa Ecor Eoc
icor ¼ ioa e2:3 ba ¼ ioc e2:3 bc ð9Þ

where icor is the corrosion current density, and Ecor is the corrosion
potential. The corrosion takes place at a rate of icor, and the product
of iron oxidation, ferrous ion, is locally consumed by the hydroxide
ions produced with oxygen reduction.
In RC corrosion, different levels of depassivation are generally
found regardless whether it is pitting corrosion or corrosion of par-
tially carbonated concrete as mentioned above. If it is assumed that
the active and passive steel surfaces are electrically isolated, then
their microcell corrosion systems should be different. Fig. 4 shows
the microcell corrosion systems for active and passive steel sur-
faces. They are calculated as follows:
For active steel surfaces:
a a a
imac ¼ ia  ic ¼ 0
Ea Ea
2:3 cor a oa 2:3
Ea a
cor Eoc
ð10Þ
a a a
icor ¼ ioa e b
a ¼ ioc e ba
c

Fig. 3. Tafel slopes of anodic and cathodic reactions. For passive steel surfaces:
3
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

0.5 0.5 i aa
iaa
a
i a i mac
mac

a
Eco p
r < Ecor
iap
iap iaa > ica

i (uA/cm2)
i (uA/cm2)

0 0
iaa = ica iap = icp iap < icp
p p
i mac i mac
p
i icp
ica c
ica

-0.5 -0.5

-800 -600 -400 -200 0 -800 -600 -400 -200 0

E (mV vs. SCE) E (mV vs. SCE)

Fig. 4. Microcell corrosion systems on active and passive steel surfaces without Fig. 5. Electrical connection in MMC system.
electrical connection.

a a
ia > ic
p p p
imac ¼ ia  ic ¼0 a
imac > 0 ð15Þ
p
E
p p
E
p
ð11Þ a a a
ia ¼ ic þ imac
E E
2:3 cor oa 2:3 cor oc
p p p p p
icor ¼ ioa e b
a ¼ ioc e b
c

For passive steel surfaces, the macrocell current density is

where superscripts a and p denote the active and passive steel sur-
cathodic:
faces, respectively. The parameters in Eqs. (10) and (11) have the
p p
same meaning as their counterparts in Eqs. (3), (4), (8) and (9). ia < ic
p
The cathodic reaction of steel might be different due to the different imac < 0 ð16Þ
local oxygen diffusion conditions, which determine the values of p p
ia ¼ ic
p
þ imac
the cathodic electrochemical parameters. The anodic reaction of
steel is directly related to the level of depassivation. An increase The intensity of the macrocell current of the active steel surface
in the level of depassivation of the steel would increase the anodic is equal to that of the passive steel surface:
exchange current density, and reduce the anodic Tafel slope and Iamac ¼ Ipmac ð17Þ
anodic equilibrium potential [6,25]. Thus:
a
a p where imac
is the macrocell current density of the active steel sur-
ioa > ioa p
face, imac is the macrocell current density of the passive steel sur-
Eaoa < Epoa ð12Þ
face, Iamac is the intensity of the macrocell current of the active
baa < bpa steel surface, and Ipmac is the intensity of the macrocell current of
Compared to the active steel surface, the microcell corrosion on the passive steel surface.
the passive steel surface is based on a higher corrosion potential Eqs. (15) and (16) show that unlike microcell corrosion, the
because of the higher anodic equilibrium potential according to anodic and cathodic current densities are different in MMC. The
Eq. (12): corrosion rate of MMC should be represented by the rate of iron
a p
oxidation ia or ia , which is related to the cathodic and macrocell
Eacor < Epcor ð13Þ
current densities. Active steel is polarized by the anodic macrocell
This potential difference between the active and passive steel current, which results in a higher anodic current density. The ano-
surfaces is shown in Fig. 4. dic current density of the passive steel surface is reduced due to
the cathodic polarization. Therefore, the behaviour of MMC greatly
(3) Microcell and macrocell corrosion depends on the macrocell current.
After the active and passive steel surfaces are electrically
connected, an MMC system is established as shown in 3. Theoretical analysis
Fig. 5. The potential difference between the active and pas-
sive steel surfaces allows the macrocell current to flow Stern and Geary [3] first proposed the Stern-Geary equation or
through the concrete pore solution from the active steel sur- the polarization resistance equation based on the assumption that
face to the passive steel surface. Thus: there is microcell corrosion:
B
imac ¼ ia  ic –0 ð14Þ icor ¼ ð18Þ
Rp
The rates of iron dissolution are affected by the macrocell cur-
rent flow with respect to both the active and passive steel surfaces. with
For active steel surfaces, the macrocell current density is B ¼ 2:3ðb
ba bc
a þbc Þ
anodic: ð19Þ
Rp ¼ @i@Enet

4
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

where icor is the corrosion rate, Rp is the polarization resistance, E is The notional corrosion rate is proportional to the macrocell cur-
the electrode potential, inet is the net current density of the elec- rent density on passive steel surfaces.
trode, and B is a Stern-Geary coefficient. It should be noted that accurate measurements of the notional
The well-known Stern-Geary equation; that is, Eq. (18), is often corrosion rate icorr by using the polarization resistance method are
interpreted by using the corrosion rate and microcell current den- difficult with MMC. Errors in measuring the true polarization resis-
sity. However, this interpretation is based on the assumption that tance Rp in MMC might be due to the ohmic drop in the electrolyte,
there is microcell corrosion. and failure to evaluate the polarization area and current flow path
In this section, the theoretical errors in the polarization resis- [8,15,21,26]. Thus, the following relationship is proposed:
tance method for MMC are examined.
Without an external polarization current or stray current, the Rp
k¼ ð32Þ
net current density can be represented by using the macrocell cur- Rapp
p
rent density:
where k is the ratio of the true polarization resistance Rp to the
inet ¼ imac ¼ ia  ic ð20Þ apparent polarization resistance Rapp
p .
According to Eq. (20), the partial derivative of the net current Eq. (32) is substituted into Eqs. (26) and (27), which provides:
density inet with respect to the electrode potential E is:
1 app
@inet @ðia  ic Þ @ia @ic 2:3ia @logia 2:3ic @logic ia ¼ i þ a1 imac ð33Þ
¼ ¼  ¼  ð21Þ k corr
@E @E @E @E @E @E
with
Eqs. (5) and (6) are then substituted into Eq. (21):
1 app
@inet 2:3ia 2:3ic 2:3ia bc þ 2:3ic ba i ¼ icorr ð34Þ
¼ þ ¼ ð22Þ k corr
@E ba bc ba bc
The cathodic current density is substituted with the relation- app B
icorr ¼ ð35Þ
ship in Eq. (20): Rapp
p

@inet 2:3ia bc þ 2:3ðia  inet Þba 2:3ia ðbc þ ba Þ  2:3inet ba

¼ ¼ ð23Þ app
where icorr is the apparent corrosion rate by using the polarization
@E ba bc ba bc
resistance method for MMC. Eq. (33) is used to correlate the actual
According to the definition of polarization resistance: corrosion rate with the apparent corrosion rate of steel.
1 @inet In considering Eq. (33), the relative error with the use of the
¼ ð24Þ apparent corrosion rate can be derived as follows:
Rp @E
app app
Eq. (24) is substituted into Eq. (23), and rearranged to give: icorr  ia i
drel
corr ¼ ¼ ð corr  1Þ ð36Þ
ia ia
ba bc 1 ba
ia ¼ þ inet ð25Þ
2:3ðba þ bc Þ Rp ba þ bc app
icorr kðia  a1 imac Þ imac
¼ ¼ k  ð1  a1 Þ ð37Þ
In considering that the net corrosion current density is equal to ia ia ia
the macrocell current density before external polarization occurs
as shown in Eq. (20), Eq. (25) can be written as: imac
drel
corr ¼ k  ð1  a1 Þ1 ð38Þ
ia ¼ icorr þ a1 imac ð26Þ ia
with where drel
corr is the relative error of using the apparent corrosion rate
B as the total corrosion rate for MMC.
icorr ¼ ð27Þ It can be observed in Eqs. (33)–(38) that these errors directly
Rp
depend on the measurement of the polarization resistance, Tafel
ba bc slopes and macrocell current of the steel. In order to obtain the
B¼ ð28Þ polarization resistance of steel, the galvanostatic pulse technique
2:3ðba þ bc Þ
[27–29], potentiodynamic polarization technique [20,30,31], and
ba electrochemical impedance technique [32–34] have been pro-
a1 ¼ 2 ð0; 1Þ ð29Þ posed. The formula of the ratio of the true polarization resistance
ba þ bc
to the apparent polarization resistance, k, varies depending on
where ia is the actual corrosion rate in the MMC, icorr is the notional the measurement techniques used and whether the circuits are
corrosion rate by using the polarization resistance method for MMC, assumed to be equivalent.
B is a Stern-Geary coefficient, and a1 is a modification factor and
depends on the Tafel slopes. Eq. (25) can be simplified as Eq. (18)
when the macrocell current between the electrode-electrolyte 4. Experiments and results
interface is negligible.
Special attention should be given to the case of passive steel In this section, a series of experiments on galvanic cells are dis-
surfaces, where the actual corrosion rate can be regarded as zero: cussed. Both the active and passive steel rebars were examined by
varying the rebar diameter, C/A ratio, macrocell current intensity
ia ¼ 0 ð30Þ and electrolyte composition. The Tafel slopes and polarization
Eq. (30) is substituted into Eq. (26) which yields: resistance were obtained through potentiodynamic polarization.
Eqs. (21) and (31) are validated for the active and passive steel
icorr ¼ a1 imac ð31Þ
rebars respectively.

5
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

A total of six sets of galvanic cells were designed, which are

listed in Table 1. Different concentrations of chloride ions in the
anolyte solution were used to produce different levels of depassi-
vation. The catholyte solutions consisted of 0.1 M sodium hydrox-
ide (NaOH). In some cases, 0.01 M, 0.05 M and 0.1 M of chloride
ions were mixed into the catholyte solutions to simulate the chlo-
ride contamination around passive steel rebars. As for the solvent,
all of the solutions used deionized water. NaOH pellets were added
into the catholyte solution of Group A-20-0.6-C-20-0.1-0.05 if sig-
nificant carbonation of the catholyte occurred, so that the catholyte
could acclimatize to the highly alkaline environment. These solu-
tions were exposed to air for 236 to 853 h.
Fig. 6 illustrates the configuration of the electrochemical mea-
surements. First, a Tafel potentiodynamic test was conducted
from 250 mV to +250 mV relative to the open circuit potential
(OCP) on the anodes and cathodes respectively to obtain the Tafel
(a) Galvanic cell simulating MMC in RC pore solution
slopes. The working electrode (WE), namely the anode or cathode
in each part, corresponding reference electrode (RE) and counter
electrode (CE) were connected to an electrochemical workstation
RE WE CE (PARSTAT 3000A-DX bipotentiostat, Princeton Applied Research,
USA).
After the Tafel tests were conducted on the anodes and cath-
Ipd Electrochemical odes, the switch of the zero resistance ammeter (ZRA) was turned
Ip
ZRA
workstation off to form a circuit between the active and passive steel rebars.
During the MMC process, the amount of macrocell current that flo-
Ipe Tape wed between the anodes and the cathodes was routinely recorded
Anolyte Catholyte by using the ZRA. Aside from measurements of the amount of
macrocell current, periodic linear polarization resistance (LPR)
tests [4] on the anodes and cathodes were also conducted respec-
Cathode tively, in which the potential of the WE was scanned from 30 mV
to +30 mV relative to the previous potential (PP). The numbers of
Anode
periodic LPR tests on anodes range from 7 (A-10-0.1-C-20-0.1-
0.01) to 92 (A-10-0.6-C-10-0.1-0.1). As shown in Fig. 6, the WE
could be polarized by a fraction of the imposed polarization cur-
Mortar disc
rent during the LPR test [21,22]:
(b) Electrochemical configuration. ZRA: Zero Resistance Ammeter;
Ip ¼ Ipe þ Ipd ð39Þ
RE: Reference Electrode;
WE: Working Electrode; CE: Counter Electrode where Ip is the imposed polarization current by the CE, Ipe is the
polarization current that flows through the WE, and Ipd is the dis-
Fig. 6. Experimental setup. persed polarization current that flows through the electrode cou-
pled with the WE in this macrocell.
The ZRA was read as follows:
(1) Experimental procedures
One of the galvanic cells used in the experiment is shown in IZRA ¼ Imac  Ipd ð40Þ
Fig. 6(a). This cell is further illustrated in Fig. 6(b). A mortar The net current that flowed through the WE is:
disc separates the galvanic cell into two parts. The catholyte
solution on the right side of the cell was used to simulate the Inet ¼ Ipe þ Imac ¼ Ip þ IZRA ð41Þ
alkaline environment around a passive steel rebar, while the The shift of the electrode potential and imposed polarization
anolyte solution on the left side of the cell was used to current from 10 mV to +10 mV [4] relative to the PP was used
depassivate the reinforcements. All of the solutions were to calculate the apparent polarization resistance as follows:
exposed to air and kept unstirred in a laboratory environ-
ment. During the experiments, the pH value and solution DE
Rapp ¼ A ð42Þ
temperature were monitored by using a pH meter. The
p
DIp
reduction in the level of the solutions due to evaporation where A is the exposed area of the WE. The changes in the reading
was measured by using a stainless ruler. The area and cur- of the ZRA during the LPR test were also recorded at PP10 mV and
rent density of the electrodes were calculated accordingly. PP + 10 mV to calculate the true polarization resistance:
The electrodes are Grade 500B ribbed steel rebars with diame-
ters of 10 mm and 20 mm. The upper and lower part of the rein- @E DE DE
Rp ¼  A¼ A ð43Þ
forcements were covered with tape to tailor the electrode areas @inet DInet DIp þ DIZRA
as shown in Fig. 6. Prior to the immersion of the anodes, they were
grinded with sand paper in deionized water, soaked in a rust The k can be obtained by taking Eqs. (42) and (43) into
remover solution and then weighed. The cathodes were also consideration:
grinded with sand paper in deionized water, cleaned with ethanol Rp DE=ðDIp þ DIZRA Þ DI p
in an ultrasonic bath, and then rinsed with deionized water as out- k¼  ¼ ð44Þ
Rapp
p DE=DIp DIp þ DIZRA
lined in [35] prior to immersion.

6
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

Table 1
Electrode diameters and solution compositions of experimental groups.

Group* Anode diameter(mm) Anolyte composition Cathode diameter(mm) Catholyte composition Immersion time(h)
A-10–0.1-C-20–0.1–0.01 10 0.1 M NaCl 20 0.1 M NaOH + 0.01 M NaCl 409
A-10–0.3-C-20–0.1 10 0.3 M NaCl 20 0.1 M NaOH 596
A-10–0.6-C-10–0.1–0.1 10 0.6 M NaCl 10 0.1 M NaOH + 0.1 M NaCl 841
A-20–0.1-C-20–0.1 20 0.1 M NaCl 20 0.1 M NaOH 236
A-20–0.3-C-20–0.1 20 0.3 M NaCl 20 0.1 M NaOH 359
A-20–0.6-C-20–0.1–0.05 20 0.6 M NaCl 20 0.1 M NaOH + 0.05 M NaCl 853

* Each experimental group is labeled as follows: A-Anode diameter-Concentration of NaCl in anolyte-C-Cathode diameter-Concentration of NaOH in catholyte-Concentration
of NaCl in catholyte.

After the periodic LPR testing, the switch of the ZRA was turned of depassivation and anodic Tafel slope are affected by the chloride
on and the Tafel potentiodynamic tests were repeated on the content and pH value of the anolyte solution respectively due to
anodes and cathodes. To eliminate the transient response led by the change in the [Cl-]/[OH–] ratio as shown in Table 2. The effects
the capacitor behaviour of the WE, a thermodynamic equilibrium of the macrocell current on the Tafel slopes are also important for
state was confirmed prior to each step during the entire measure- the active steel rebars. There are two possible effects. First, the cor-
ment scheme. The WE was considered to be steady if a stable elec- rosion of the active steel rebars is exponentially accelerated by the
trode potential was obtained. The scan rate of the Tafel and LPR macrocell current, and a thick layer of corrosion product may inhi-
tests was 0.16 mV/s based on the recommendation in [30]. The bit both the anodic and cathodic reactions to some degree. Sec-
gravimetric weight loss of the active steel rebar was measured at ondly, the local [Cl-]/[OH–] ratio may increase due to the electric
the end of the experiment, based on the instructions in [36]. forces, thus increasing the level of depassivation and reducing
the anodic Tafel slope. After MMC occurs, a significant decline in
(2) Results of Tafel tests the anodic Tafel slopes of the active steel rebars in Groups A-10-
Typical potentiodynamic polarization curves in this study 0.3-C-20-0.1 and A-20-0.1-C-20-0.1 may result due to the decrease
are presented in Fig. 7. Most of the polarization curves show in the pH value of the anolyte solution. It appears that there is a
a clear Tafel region in the anodic and cathodic branches (e.g. threshold [Cl-]/[OH–] ratio, above which the anodic Tafel slope
Fig. 7(b)), in which the applied potential is linearly propor- remains linear; for instance, the anodic Tafel slope of the active
tional to the logarithmic polarization curves. The Tafel steel rebars in Groups A-20-0.3-C-20-0.1 and A-20-0.6-C-20-0.1-
slopes are determined by the linear fit of the Tafel region. 0.05. The increase in the cathodic Tafel slope of the active steel
For the anodic polarization curves of the anodes, some of rebars suggests that the deposition of the corrosion products dur-
them seem to slope over a certain potential range, e.g. the ing MMC increases the thickness of the diffusion layer for the oxy-
anode of Group A-10-0.3-C-20-0.1 in Fig. 7(a). This may indi- gen reduction reaction.
cate the formation of a non-passive film due to corrosion Particular attention should be given to the effects of the pH
product deposition or impurities in the iron [37]. This defor- value on the cathodic Tafel slopes. For the passive steel rebars, a
mation can be remedied by connecting the start and end of reduced pH value causes a reduction in the amount of hydroxyl
the slope by using a straight line. A good Tafel region and ions and thickness of the passive film [23,39]. The reduction of oxy-
anodic Tafel slope can be obtained based on the modified gen on the passive steel rebars tends to takes place more rapidly
slope. In Fig. 7(c), there is no Tafel region in the anodic due to the reduced amount of hydroxyl ions and thinner diffusion
branch of the polarization curve after MMC takes place, layer. Thus, the cathodic Tafel slope tends to decrease. As shown in
and therefore the corresponding anodic Tafel slopes are Table 3, the decrease in the cathodic Tafel slope of the passive steel
regarded as infinite. rebars after MMC occurs might be caused by a lower pH value and
aging of the passive film [35].
The Tafel slopes and related data are summarized in Tables 2
The values of B (Stern-Geary coefficient) and a1 (modification
and 3. The corrosion rate is calculated by using Eq. (18), in which
factor) can be calculated given the anodic and cathodic Tafel
a polarization resistance that ranges from OCP-10 mV to
slopes. Therefore, the level of depassivation also determines the
OCP + 10 mV is used. Tables 2 and 3 show that there is a consider-
value of B and a1 of the active steel rebars as shown in Table 2.
able increase in the corrosion rate of the anodes after MMC occurs
The values of B and a1 of the active steel rebars decrease with
except for Groups A-10-0.1-C-20-0.1-0.01 and A-20-0.1-C-20-0.1,
increases in the [Cl-]/[OH–] ratio until the lower limit is reached,
while the corrosion rate of the cathodes is reduced after MMC
which is 18 mV and 0.22 respectively in this study. As for the pas-
occurs. The corrosion rate of the active steel rebars is significantly
sive steel rebars in the alkaline environment, decreases in B after
higher compared with that of the passive steel rebars within the
MMC occurs are mainly led by the reduction of the cathodic Tafel
same cell, which indicates a higher buffering capacity of the polar-
slope, which might be related to the decreasing pH value and
ization current. This may explain for the smaller shift in the OCP
longer exposure time as mentioned previously.
after MMC occurs in the active steel rebar as opposed to the pas-
The potentiodynamic polarization test with high overpotential
sive steel rebar. Except for the catholyte in Group A-20-0.6-C-20-
is considered as a destructive test [40,41]. The potential response
0.1-0.05 (the group with supplementary NaOH pellets), there is a
after polarization is found to be unstable in [40]. The OCP after con-
downward trend in the pH value of the solutions in this study
ducting Tafel test is shown in Table 2 in order to evaluate the effect
due to their carbonation. The average temperature of all of the
of high polarization on the corrosion system. It can be concluded
groups ranges from 19.9 °C to 21.9 °C, with the highest standard
that the shift in OCP after conducting Tafel test is slight for those
deviation of 1.0 °C. Thus, the solution temperatures are not given
groups in which the values of B and a1 change significantly after
in Tables 2 and 3.
MMC occurs. Although a large potential shift is found in Group
The anodic Tafel slope is dependent on the level of depassiva-
A-10-0.6-C-10-0.1-0.1, the values of B and a1 remain similar after
tion of the reinforcement [25,38], and the cathodic Tafel slope
MMC occurs. As mentioned before, the Tafel kinetic parameters
depends on the oxygen supply conditions [38], temperature and
are controlled by the depassivation level of anodes. Thus, the
aging of the passive film [35]. For the active steel rebars, the level

7
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

Before MMC occurs Before MMC occurs

-300 After MMC occurs -400 After MMC occurs

-400 -500
Potential (mV vs. SCE)

Potential (mV vs. SCE)

Corrosion product
-500
deposition -600

-600
-700

-700
-800
-800
-900
-900
-1000
10-1 100 101 102 103 104 105 106 100 101 102 103 104 105 106 107
Current (uA) Current (uA)

(a) A-10-0.3-C-20-0.1 anode (b) A-20-0.3-C-20-0.1 anode

200
Before MMC occurs
100 After MMC occurs

0
Potential (mV vs. SCE)

-100

-200

-300

-400

-500

-600

10-1 100 101 102 103 104

Current (uA)
(c) A-20-0.3-C-20-0.1 cathode

Fig. 7. Typical polarization curves.

Table 2
Results of Tafel tests on active steel rebars.

Group OCP/OCP after Tafel test (mV ba (mV/ bc (mV/ B a1 pH Corrosion rate (lA/ lg([Cl-]/[OH–]) in
vs. SCE) dec) dec) (mV) cm2) anolyte
A-10–0.1-C-20– Before MMC occurs 542/-541 318 341 72 0.48 10.68 27.39 2.3
0.1–0.01 After MMC occurs 580 322 201 54 0.62 9.79 14.38 3.2
A-10–0.3-C-20–0.1 Before MMC occurs 646/-649 348 105 35 0.77 11.58 0.91 1.9
After MMC 641 67 109 18 0.38 9.30 6.13 4.2
A-10–0.6-C-10– occursBefore 656/-674 63 112 18 0.36 11.14 3.29 2.6
0.1–0.1 MMCOccurs
After MMCoccurs 633 58 147 18 0.28 8.94 10.12 4.8
A-20–0.1-C-20–0.1 Before MMCoccurs 676/-681 198 149 37 0.57 10.33 24.83 2.7
After MMC occurs 734 72 162 22 0.31 8.97 12.11 4.0
A-20–0.3-C-20–0.1 Before MMC occurs 699/-705 59 145 18 0.29 9.70 11.78 3.8
After MMC occurs 710 66 193 21 0.25 9.17 26.92 4.3
A-20–0.6-C-20– Before MMC occurs 692/N.A. 61 218 21 0.22 9.66 18.27 4.1
0.1–0.05 After MMC occurs 673 61 222 21 0.22 7.98 102.34 5.8

Note: MMC: microcell and macrocell corrosion

8
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

Table 3
Results of Tafel tests on passive steel rebars.

Group OCP (mV vs. SCE) ba (mV/dec) bc (mV/dec) B (mV) a1 pH Corrosion rate (lA/ cm2)
A-10–0.1-C-20–0.1–0.01 Before MMC occurs 387 1320 107 43 0.93 12.39 0.38
After MMC occurs 308 1 37 16 1.00 10.21 0.00
A-10–0.3-C-20–0.1 Before MMC occurs 363 702 139 50 0.83 13.66 0.30
After MMC 277 786 63 25 0.93 10.45 0.06
A-10–0.6-C-10–0.1–0.1 Occurs Before MMC Occurs 453 850 227 78 0.79 13.35 3.23
After MMC occurs 388 433 195 58 0.69 10.06 1.19
A-20–0.1-C-20–0.1 Before MMC occurs 310 1030 125 48 0.89 13.23 0.19
After MMC occurs 261 1 34 15 1.00 10.99 0.00
A-20–0.3-C-20–0.1 Before MMC occurs 343 3220 118 50 0.96 13.39 0.34
After MMC occurs 125 1 43 19 1.00 10.47 0.00
A-20–0.6-C-20–0.1–0.05 Before MMC occurs 522 1073 135 52 0.89 13.31 1.32
After MMC occurs 347 307 128 39 0.71 13.85 0.30

Z te X
n1
effects of potentiodynamic polarization test on Tafel kinetic 1
parameter are assumed to be negligible compared with those of Q¼ Idt  ðt iþ1  t i Þ ½Iiþ1 þ Ii  ð47Þ
ts i¼1
2
MMC in this study.
where m is the weight loss, Msteel is the molar mass of steel
(3) Results of active steel rebars (55.85 g/mol), z is the transferred electron number (2), F is the Fara-
Fig. 8 shows the validation of Eq. (26) for the anodes during day’s constant (96485.3C/mol), Q is the electrical charge, n is the
MMC. For each group, the values of B and a1 in Eqs. (28) and number of periodical LPR tests, Ii,ti are the current intensity and cor-
(29) are obtained by: responding elapsed time in i-th LPR test. To calculate the weight
t s
losses caused by the macrocell current induced error (a1 imac ),
B ¼ Bs þ tLPR
t e ts
ðBe  Bs Þ macrocell current (imac ) and notional corrosion rate (icorr ), the values
t s
ð45Þ
a ¼ as þ tLPR
t e ts
ðae  as Þ of I and t can be substituted accordingly.
The weight losses before and after MMC occur, which are calcu-
where t LPR is the elapsed time when conducting LPR test, t s is the lated by the corrosion rates measured in the Tafel tests, are
elapsed time when conducting potentiodynamic polarization test deducted from the overall gravimetric weight loss to obtain the
before MMC occurs, t e is the elapsed time when conducting poten- gravimetric weight loss during MMC, namely the left part of Eq.
tiodynamic polarization test after MMC occurs, Bs ; Be ; as ; ae are the B (46).
values and a1 values before and after MMC occurs, which can be As shown in Fig. 8, Eq. (26) is proven to be valid. The gravimetric
found in Table 2. weight loss and the sum of the weight losses caused by icorr and
Faraday’s law [4] is used to calculate the electrochemical weight a1 imac show a good agreement. The average deviation from the pro-
loss: posed Eq. (26) is 3.1% with the highest deviation of 16%. The con-
tribution of the macrocell current is mainly associated with the
M steel Q C/A ratio: a higher C/A ratio means a greater contribution from
m¼ ð46Þ
zF the macrocell current. It appears that a higher chloride content also

Fig. 8. Comparison of weight losses caused by a1 imac , imac and icorr with gravimetric.

9
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

increases the contribution of the macrocell current when groups

with 0.3 M of sodium chloride (NaCl) are compared to those with
0.1 M of NaCl. The sum of the weight losses caused by icorr and
a1 imac in groups with a high C/A ratio (A-20-0.1-C-20-0.1 and A-
20-0.3-C-20-0.1) is lower than the gravimetric weight loss. This
shows that there may be a notable macrocell current induced error
of 38% when the C/A ratio exceeds 10.

(4) Results of passive steel rebars

The corrosion rate of the passive steel rebars in the experi-
ment can be neglected because cathodic protection occurred
during the macrocell coupling [42]. Moreover, a visual
inspection showed that the passive steel rebars did not pro- Fig. 10. Equivalent circuit of MMC under external polarization current.
duce any corrosion rust in the experiment after MMC
occurred. As proposed in Eq. (31), the notional corrosion rate
coupled with the WE. The electrolyte resistance can be
icorr is compared with the a1 Imac in Fig. 9. For each group, ignored due to the addition of NaCl and NaOH. The capacitor
the values of B and a1 in Eqs. (28) and (29) are obtained by is used to describe the transient behaviour of the electrode.
Eq. (45). The corresponding values of Bs, Be, as, ae are listed In this study, the electrode potential of the WE was scanned
in Table 3. In considering that the passive film of the passive from 30 mV to +30 mV relative to the PP by 0.16 mV/s, dur-
steel rebar might be unstable and the surface morphology of ing which the polarization resistance was obtained from
the passive steel rebar might change under a pH value of 10 mV to +10 mV. Therefore, the capacitor can be treated
11.5 [43], the results that correspond to the pH values that as an open circuit after it is charged. The apparent polariza-
are higher than 11.5 are shown in Fig. 9. tion resistance is determined by using:
There is a strong correlation between the notional corrosion Rp ðRc þ Rpd Þ
rate icorr and a1 Imac . The average deviation from the proposed Rapp ¼ ð48Þ
p
Rp þ Rc þ Rpd
Eq. (31) is 3.6%. This correlation is observed for different C/A
ratios, macrocell current densities, cathode diameters, pH values The ratio of the true polarization resistance to apparent polar-
and level of chloride contamination in the catholyte. Note that ization resistance is:
for the passive steel rebars, the lowest value of a1 is 0.69, with
Rp Rp þ Rc þ Rpd
an average value of 0.89, as shown in Table 3. Based on the rela- k¼ ¼ >1 ð49Þ
tionship shown in Eq. (31), icorr is directly related to the macrocell Rapp
p Rc þ Rpd
current when a high value of a1 is considered.
1 Rc þ Rpd
¼ <1 ð50Þ
5. Discussion k Rp þ Rc þ Rpd

(1) Relative error of active steel rebars k depends on the true polarization resistance of the anodes and
The galvanic cells in this study can be simplified into the cir- cathodes in the galvanic cell, C/A ratio and the ohmic resistance
cuit shown in Fig. 10. Rc is the ohmic resistance of the mortar of the mortar disc. The level of imposed polarization current
disc, and Rpd is the polarization resistance of the electrode attracted by the WE is 1k, while k reflects the degree of the dispersion

15 15

10 10
)

5 5

0 0
0 5 10 15 0 5 10 15
) )

x=y A-10-0.1-C-20-0.1-0.01 x=y pH>13 pH 12.5~13

A-10-0.3-C-20-0.1 A-10-0.6-C-10-0.1-0.1
A-20-0.1-C-20-0.1 A-20-0.3-C-20-0.1
pH 12.0~12.5 pH 11.5~12.0
A-20-0.6-C-20-0.1-0.05

(a) by different experimental groups (b) by different pH values

Fig. 9. Correlation between icorr and a1 imac of cathodes during MMC (Average deviation from proposed relationship: 3.6%).

10
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

C/A :14.5-11.8
1- 1.65
2.0

C/A :12.9-10.5
C/A :3.5-2.4
1.37
1.5 1.30
C/A :3.0-2.6
C/A :1.0-0.8 C/A :9.5-2.7
1.18
1.10
1.04
1.0 0.95
0.89 0.92
0.80 0.81
0.75

0.5

0.0
A-10-0.1 A-10-0.3 A-10-0.6 A-20-0.1 A-20-0.3 A-20-0.6
-C-20-0.1-0.01 -C-20-0.1 -C-10-0.1-0.1 -C-20-0.1 -C-20-0.1 -C-20-0.1-0.05

Fig. 11. Relative errors of active steel rebars during periodic LPR tests.

of the imposed polarization current. A higher k value means a a1 and imac

ia
determine the macrocell current induced relative error.
greater dispersion of the polarization current. As shown in Table 2, a1 decreases with an increase in the level of
Fig. 11 shows the value of k and macrocell current induced rel- depassivation of the anode and the contribution of the macrocell
ative error 1  a1 imac
ia
during the periodic LPR tests of the active steel current to the total corrosion current imac mainly depends on the
ia
rebars. C/A ratio. The highest relative error triggered by the macrocell cur-
The value of k is calculated by using Eq. (44). As shown in Eq. rent would be expected for a slightly active anode with a large C/A
(49), k reflects the relationship between the polarization resistance ratio based on Eq. (51). As shown in Fig. 11, the highest average rel-
of the anode and the sum of the ohmic resistance of the mortar disc ative error triggered by the macrocell current of 25% is found in
and the polarization resistance of the cathodes. Theoretically, a lar- Group A-20-0.1-C-20-0.1, which may explain for the large differ-
ger C/A ratio would increase the dispersion of the polarization cur- ence of 38% between the nominal corrosion rate and gravimetric
rent by reducing the polarization resistance of the cathodes and weight loss as shown in Fig. 8. The considerably high relative error
highly depassivated anodes would attract more of the polarization triggered by the macrocell current might be due to the large C/A
current with a lower polarization resistance. As shown in Fig. 11, ratio (Group A-20-0.3-C-20-0.1) or low active state (Group A-10-
the decrease in the k of Group A-10-0.3-C-20-0.1 is caused by the 0.3-C-20-0.1).
reduced polarization resistance of the anode with higher corrosion Fig. 12 shows the total relative error of the active steel rebars,
activity compared to Group A-10-0.1-C-20-0.1-0.01. The low val- namely Eq. (38). It is interesting that the average total relative
ues of k in Groups A-10-0.6-C-10-0.1-0.1 and A-20-0.6-C-20-0.1- errors are negligible except for that of Group A-20-0.3-C-20-0.1.
0.05 suggest that more than 90% of the imposed polarization cur- As mentioned before, the increase in the C/A ratio and reduction
rent is attracted by the anodes. This could be the result of the C/ in the level of depassivation increase the relative error caused by
A ratios that fall within a low to moderate range and the highly the macrocell current and the dispersion of the imposed current.
active anodes. On the other hand, a substantial amount of the As a result, the decrease in the value of 1  a1 imac may counter
ia
imposed polarization current is dispersed into the cathodes in
the increase in the value of k. This may explain why acceptable
Groups A-20-0.1-C-20-0.1 (27%) and A-20-0.3-C-20-0.1 (39%),
results were obtained in previous studies [5,44] without consider-
where the C/A ratios are larger than 10. This is consistent with
ing the dispersion of the imposed polarization current and relative
the results found in [21]. The concrete resistivity is also important
error caused by the macrocell current during MMC. However, a sig-
for the distribution of the current flow. High concrete resistance
nificant instantaneous relative error of 116% is found in Group A-
would inhibit the lateral dispersion of the imposed polarization
20-0.3-C-20-0.1 with an average relative error of 35%. The error
current as shown in Fig. 10. To conclude, the level of the depassi-
caused by the current dispersion is also believed to be cancelled
vation of the anodes, electrolyte conductivity and C/A ratio all
out due to the failure to accurately measure the potential of the
determine the ability of the anodes to attract the imposed polariza-
corroding surface [21]. The mechanism under which these errors
tion current.
cancel each other out therefore needs to be further investigated.
Based on Eq. (26), the macrocell current induced error is app
obtained by using: Moreover, note that the use of the apparent corrosion rate icorr
to calculate the total weight loss of the anode for MMC is based
imac imac on accurately measuring B. A value of B = 26 mV is recommended
1  a1 ¼ 1  a1 <1 ð51Þ
ia icorr þ a1 imac with respect to active steel rebars for simplicity purposes [4,44].

11
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

130% 25%~75%
C/A :14.5-11.8
120% Mean
110% Data
100%
90%
Total relative error 80%
70%
C/A :12.9-10.5
60%
50%
C/A :3.5-2.4
40% 35%
30% C/A :3.0-2.6
20% 16% C/A :1.0-0.8 C/A :9.5-2.7
10% 3% 1%
1%
0%
-10% -6%
-20%
-30%

A-10-0.1 A-10-0.3 A-10-0.6 A-20-0.1 A-20-0.3 A-20-0.6

-C-20-0.1-0.01 -C-20-0.1 -C-10-0.1-0.1 -C-20-0.1 -C-20-0.1 -C-20-0.1-0.05

Fig. 12. Total relative errors of active steel rebars.

According to the B values listed in Table 2, the use of B = 26 mV B

Bp ¼ ¼ Rp  jimac j ð52Þ
provides acceptable results for most of the cases in this study. a1
However, Group A-10–0.1-C-20–0.1–0.01 is significantly underes-
timated, as this group has a measured value of B that ranges from Bp
¼ jimac j ð53Þ
54 to 72 mV. The Tafel slopes of the steel in RC are dependent on Rp
the active state of the steel rebars, oxygen supply conditions, and
Fig. 13 shows the values of Bp and k of the passive steel rebars at
ambient temperatures. Thus, the RC structures are categorized in
pH values that are higher than 11.5. Note that with respect to the
accordance with [38] and the corresponding ranges of the Tafel
passive steel rebars, a B value of 52 mV is widely used in the liter-
slopes, B and a1 are listed in Table 4. These values of the Tafel
ature [44,45,47,48]. The average Bp value in the groups without
slopes are selected from previous experimental and numerical
chloride contamination is close to 52 mV. The Bp values seem to
studies [38,45,46] and the experimental results in this study.
be affected by the chloride contamination of the catholyte. The
According to Table 4, a larger error may be caused by the recom-
Bp value for the slightly chloride contaminated group is on average
mended value of B = 26 mV.
32.6 mV, while the average Bp value increases with chloride con-
tamination. The average Bp value is found to be 83.5 mV in Group
(2) Macrocell current induced error in passive steel rebars
A-10-0.6-C-10-0.1-0.1, which is close to the upper limit of the B
The notional corrosion rate icorr has been proven to be
value for the passive steel rebars as shown in Table 4. The k value
directly related to the macrocell current that flows through
for passive steel rebars also shows the amount of dispersed polar-
the surface of the passive steel rebars in this study. Given
ization current. It can be observed in Fig. 13 that a large amount of
an electrode with a negative potential, the macrocell current
the imposed polarization current is dispersed in most of the pas-
can polarize passive steel rebars towards the cathodic direc-
sive steel rebars in this study. This could be attributed to the high
tion and reduce the polarization resistance of the passive
electrolyte conductivity and higher polarization resistance com-
steel rebars. Table 5 shows the polarization resistance of
pared to the coupled active steel rebars in this study.
the passive steel rebars at pH values that are higher than
The effect of the dispersion of the imposed polarization current
11.5. Both the true and apparent polarization resistances of
on the polarization resistance of the passive steel rebars has to be
the passive steel rebars are significantly reduced during
considered for practical measurement purposes. By multiplying
MMC due to the macrocell current for all of the groups in
the apparent polarization resistance Rapp p with the macrocell cur-
this study.
rent density imac , the value of Bapp is:
In order to associate the macrocell current with the polarization Bp
resistance, the value of Bp can be calculated by using: Bapp ¼ ¼ Rapp
p jimac j ð54Þ
k

Table 4
Range of electrochemical parameter values.

Case Concrete state ba (mV/ bc (mV/ B (mV) a1

dec) dec)
Active steel rebar and sufficient oxygen Carbonated, dry; chloride contaminated with sufficient oxygen 30 ~ 350 34 ~ 350 7 ~ 76 0.08 ~ 0.91
Active steel rebar and oxygen diffusion control Chloride contaminated, uncarbonated, water-saturated 30 ~ 350 1 13 ~ 152 0
Passive steel rebar and sufficient oxygen Uncarbonated, low chloride content, dry 1 34 ~ 242 15 ~ 105 1
Passive steel rebar and oxygen diffusion control Uncarbonated, low chloride content, wet 1 1 1 0~1

12
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

Table 5
Changes in polarization resistance of passive steel rebars due to macrocell current.

Group Rp before MMC occurs Average value (standard deviation) of Rp Average value (standard deviation) of Rapp Rp after MMC occurs
(103O∙cm2) during MMC (103O ∙cm2) during MMC (103O ∙cm2) (103O ∙cm2)
A-10–0.1-C-20– 112.0 13.6 (0.7) 10.7(0.4) 1
0.1–0.01
A-10–0.3-C-20– 165.4 6.8 (0.3) 3.7 (0.1) 460.8
0.1
A-10–0.6-C-10– 24.1 7.7 (2.0) – 49.3
0.1–0.1
A-20–0.1-C-20– 261.7 5.5 (0.7) 3.4 (0.3) 1
0.1
A-20–0.3-C-20– 145.5 4.1 (0.6) 2.4 (0.2) 1
0.1
A-20–0.6-C-20– 39.6 9.2 (3.0) 4.7 (0.6) 132.8
0.1–0.05

100
2.4 83.5
Bp 1.92
2.2

2.0 80
1.81 1.72
1.8 1.64
1.6
55.1 60
1.4 48.7

Bp (mV)
46.3
1.27 48.4
1.2

1.0 40
32.6
0.8

0.6
20
0.4

0.2
N.A.
0.0 0
A-10-0.1 A-10-0.3 A-10-0.6 A-20-0.1 A-20-0.3 A-20-0.6
-C-20-0.1-0.01 -C-20-0.1 -C-10-0.1-0.1 -C-20-0.1 -C-20-0.1 -C-20-0.1-0.05

Fig. 13. Values of k and Bp for passive steel rebars.

Bapp
¼ jimac j ð55Þ
Rapp
p 40 25%~75%
When Bapp is used for passive steel rebars in the polarization Mean
resistance method during MMC without considering the measure- Data
ment error of the polarization resistance, the result is the macrocell
current density.
Fig. 14 shows the Bapp values of the passive steel rebars within 30 29.4 29.2
different ranges of pH. The Bapp values in Group A-10-0.6-C-10-0.1- 27.6
Bapp (mV)

0.1 are not included because the net current density of the passive 25.8
steel rebar was directly measured for this group. Despite the slight
increase in the Bapp value at higher ranges of pH as shown in
Fig. 14, the variations in the values of Bapp are limited in this study. 20
Furthermore, the recommended value of 26 mV for active steel
rebars is acceptable for groups that have less than moderate chlo-
ride contamination in this study. However, it should be noted that
this narrow range of Bapp values could be caused by the specific
conditions of this study. According to Eq. (54), the Bapp value would
be significantly affected by the dispersion of the imposed polariza- 10
tion current, which depends on the actual conditions of RC 11.5~12.0 12.0~12.5 12.5~13 >13
structures.
pH range
In MMC, the measured value of the corrosion rate of the passive
steel rebars by using the polarization resistance method has to be Fig. 14. Values of Bapp for passive steel rebars under different ranges of pH.

13
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

regarded differently. Due to the significant decrease in polarization

7. Data availability
resistance, a high corrosion rate is obtained. This corrosion rate
value could be interpreted as the rate of iron dissolution and it is
The raw/processed data required to reproduce these findings
expected that corrosion would very likely occur. However, in the
cannot be shared at this time as the data also forms part of an
case of MMC, the measured corrosion rate reflects the amount of
ongoing study.
macrocell current as discussed previously. As mentioned in Sec-
tion 2.3, the amount of cathodic protection of the passive steel
rebars is increased with a higher macrocell current density. There- CRediT authorship contribution statement
fore, when there is MMC, the corrosion rate of the passive steel
rebars is reduced based on a high measured value with the use Lijie Chen: Conceptualization, Methodology, Investigation, Data
of the polarization resistance method, which is contrary to conven- curation, Visualization, Writing - original draft. Ray Kai Leung Su:
tional interpretations. On the other hand, the corrosion rate of the Validation, Supervision, Resources, Writing - review & editing, Pro-
active steel rebars coupled with passive steel rebars is increased ject administration, Funding acquisition.
due to a higher macrocell current [10,17,49]. In other words, a high
measured corrosion rate of the passive steel rebars by using the Declaration of Competing Interest
polarization resistance method when there is MMC may indicate
a low risk of corrosion of the passive steel rebars and high proba- The authors declare that they have no known competing finan-
bility of corrosion of the active steel rebars along with the passive cial interests or personal relationships that could have appeared
steel rebars measured for corrosion rate. to influence the work reported in this paper.

References
6. Conclusions [1] B. Hou, X. Li, X. Ma, C. Du, D. Zhang, M. Zheng, W. Xu, D. Lu, F. Ma, The cost of
corrosion in China, npj Mater. Degrad. 1 (2017), https://doi.org/10.1038/
This study theoretically analyses the effects of the macrocell s41529-017-0005-2.
[2] G.H. Koch, M.P. Brongers, N.G. Thompson, Y.P. Virmani, J.H. Payer, Corrosion
current on the measured corrosion rate by using the polarization Cost and Preventive Strategies in the United States, Federal Highway
resistance method in a general and concise manner. Based on the Administration, United States, 2002. https://rosap.ntl.bts.gov/view/dot/
findings of this study, the following conclusions can be drawn: 39217.
[3] M. Stern, A.L. Geary, Electrochemical polarization I. A theoretical analysis of the
shape of polarization curves, J. Electrochem. Soc. 104 (1957) 56–63, https://
1. There is the need to be mindful of the effects of the macrocell doi.org/10.1149/1.2428496.
current and the measurement of the polarization resistance [4] C. Andrade, C. Alonso, Test methods for on-site corrosion rate measurement of
on corrosion rate measurements for MMC. The polarization steel reinforcement in concrete by means of the polarization resistance
method, Mat. Struct. 37 (2004) 623–643, https://doi.org/10.1007/BF02483292.
resistance is reduced due to the flow of the macrocell current [5] A. Fahim, P. Ghods, R. Alizadeh, M. Salehi, S. Decarufel, A new test method for
with respect to both the anode and cathode. Furthermore, the rebar corrosion rate measurement, in: Advances in Electrochemical
macrocell current makes it difficult to accurately measure the Techniques for Corrosion Monitoring and Laboratory Corrosion
Measurements, ASTM International, 2019, https://doi.org/10.1520/
polarization resistance. Different testing methods, level of stp160920170227.
depassivation, concrete resistivity and C/A ratios greatly affect [6] C. Cao, M.M.S. Cheung, B.Y.B. Chan, Modelling of interaction between
the effect of the measurement process of polarization resistance corrosion-induced concrete cover crack and steel corrosion rate, Corros. Sci.
(2013) 69–97, https://doi.org/10.1016/j.corsci.2012.11.028.
on the corrosion rate with MMC. [7] M. Raupach, Chloride-induced macrocell corrosion of steel in concrete—
2. The Tafel slopes and relevant electrochemical parameters deter- theoretical background and practical consequences, Constr. Build. Mater. 10
mine the accuracy of the polarization resistance method. Differ- (5) (1996) 329–338, https://doi.org/10.1016/0950-0618(95)00018-6.
[8] B. Elsener, Macrocell corrosion of steel in concrete – implications for corrosion
ent electrochemical parameters might be obtained under monitoring, Cem. Concr. Compos. 24 (2002) 65–72, https://doi.org/10.1016/
different pH values and chloride contents in electrolyte solu- s0958-9465(01)00027-0.
tions. Despite the deposition of corrosion products, the B and [9] M.M.S. Cheung, C. Cao, Application of cathodic protection for controlling
macrocell corrosion in chloride contaminated RC structures, Constr. Build.
a1 values of active steel rebars show a trend of decline with Mater. 45 (2013) 199–207, https://doi.org/10.1016/j.conbuildmat.2013.04.010.
an increase in the amount of depassivation until a lower limit [10] C. Andrade, P. Garcés, I. Martínez, Galvanic currents and corrosion rates of
is reached. reinforcements measured in cells simulating different pitting areas caused by
3. Substantial errors caused by the macrocell current with the chloride attack in sodium hydroxide, Corros. Sci. 50 (2008) 2959–2964,
https://doi.org/10.1016/j.corsci.2008.07.013.
polarization resistance method may occur with respect to active [11] A. Djerbi, S. Bonnet, A. Khelidj, V. Baroghel-bouny, Influence of traversing crack
steel rebars when there is MMC. Although the macrocell current on chloride diffusion into concrete, Cem. Concr. Res. 38 (6) (2008) 877–883,
induced error is compensated by the error from the dispersed https://doi.org/10.1016/j.cemconres.2007.10.007.
[12] V. Picandet, A. Khelidj, H. Bellegou, Crack effects on gas and water permeability
polarization current of active steel rebars during the LPR test of concretes, Cem. Concr. Res. 39 (2009) 537–547, https://doi.org/10.1016/j.
in most cases of this study, further efforts are still necessary cemconres.2009.03.009.
to determine all of the possible errors. [13] K. Wang, D.C. Jansen, S.P. Shah, A.F. Karr, Permeability study of cracked
concrete, Cem. Concr. Res. 27 (3) (1997) 381, https://doi.org/10.1016/s0008-
4. For passive steel rebars in the presence of MMC, the measured 8846(97)00031-8.
corrosion rate by using the polarization resistance method is [14] K.V. Subramaniam, M. Bi, Investigation of steel corrosion in cracked concrete:
directly related to the macrocell current density instead of the Evaluation of macrocell and microcell rates using Tafel polarization response,
Corros. Sci. 52 (2010) 2725–2735, https://doi.org/10.1016/
actual corrosion rate. Passive steel rebars might be considered j.corsci.2010.04.030.
to be active steel rebars because the macrocell current reduces [15] J.A. González, C. Andrade, C. Alonso, S. Feliu, Comparison of rates of general
the polarization resistance. However, this evaluation of the risk corrosion and maximum pitting penetration on concrete embedded steel
reinforcement, Cem. Concr. Res. 25 (1995) 257–264, https://doi.org/10.1016/
of a high corrosion rate is not reasonable for passive steel
0008-8846(95)00006-2.
rebars. A low probability of corrosion is expected for passive [16] A. Nasser, A. Clément, S. Laurens, A. Castel, Influence of steel–concrete
steel rebars provided that there is a large measured value with interface condition on galvanic corrosion currents in carbonated concrete,
the polarization resistance method. Further research is clearly Corros. Sci. 52 (2010) 2878–2890, https://doi.org/10.1016/
j.corsci.2010.04.037.
needed to evaluate the measured value by using the polariza- [17] A.B. Revert, K. Hornbostel, K. De Weerdt, M.R. Geiker, Macrocell corrosion in
tion resistance method for RC structures. carbonated Portland and Portland-fly ash concrete - contribution and

14
L. Chen and Ray Kai Leung Su Construction and Building Materials xxx (xxxx) xxx

mechanism, Cem. Concr. Res. 116 (2019) 273–283, https://doi.org/10.1016/j. Sci. Technol. 28 (2012) 905–912, https://doi.org/10.1016/s1005-0302(12)
cemconres.2018.12.005. 60149-2.
[18] M.G. Sohail, S. Laurens, F. Deby, J.P. Balayssac, Significance of macrocell [34] D.V. Ribeiro, J.C.C. Abrantes, Application of electrochemical impedance
corrosion of reinforcing steel in partially carbonated concrete: numerical and spectroscopy (EIS) to monitor the corrosion of reinforced concrete: a new
experimental investigation, Mater. Struct. 48 (2015) 217–233, https://doi.org/ approach, Constr. Build. Mater. 111 (2016) 98–104, https://doi.org/10.1016/
10.1617/s11527-013-0178-2. j.conbuildmat.2016.02.047.
[19] S. Laurens, P. Hénocq, N. Rouleau, F. Deby, E. Samson, J. Marchand, B. [35] S. Jaggi, B. Elsener, H. Bohni, Oxygen Reduction on Mild Steel and Stainless
Bissonnette, Steady-state polarization response of chloride-induced macrocell Steel in Alkaline Solutions, Eur. Fed. Corros. Publ., 2000, pp. 3–12. https://
corrosion systems in steel reinforced concrete — numerical and experimental www.researchgate.net/publication/284659710_Oxygen_reduction_on_mild_
investigations, Cem. Concr. Res. 79 (2016) 272–290, https://doi.org/10.1016/j. steel_and_stainless_steel_in_alkaline_solutions.
cemconres.2015.09.021. [36] ASTM G1–03, Standard practice for preparing, cleaning, and evaluating
[20] R. Francois, S. Laurens, F. Deby, in: Corrosion and Its Consequences for corrosion test specimens American Society for Testing and Materials 2011.
Reinforced Concrete Structures, Elsevier, 2018, https://doi.org/10.1016/c2016- DOI:10.1520/g0001-03r11.
0-01228-7. [37] H.J. Flitt, D.P. Schweinsberg, Evaluation of corrosion rate from polarisation
[21] U. Angst, M. Büchler, On the applicability of the Stern-Geary relationship to curves not exhibiting a Tafel region, Corros. Sci. 47 (2005) 3034–3052, https://
determine instantaneous corrosion rates in macro-cell corrosion: applicability doi.org/10.1016/j.corsci.2005.06.014.
of the Stern-Geary relation to galvanic corrosion, Mater. Corros. 66 (10) (2015) [38] G. Song, Theoretical analysis of the measurement of polarisation resistance in
1017–1028, https://doi.org/10.1002/maco.201407997. reinforced concrete, Cem. Concr. Compos. 22 (2000) 407–415, https://doi.org/
[22] U. Angst, M. Büchler, A new perspective on measuring the corrosion rate of 10.1016/S0958-9465(00)00040-8.
localized corrosion, Mater. Corros. (2020) 808–823, https://doi.org/ [39] A. Rossi, G. Puddu, B. Elsener, in: The Surface of Iron and Fe10Cr Alloys in
10.1002/maco.201911467. Alkaline Media, Corrosion of Reinforcement in Concrete, EFC Publication, 2007,
[23] N. Sato, T. Noda, K. Kudo, Thickness and structure of passive films on iron in pp. 44–61, https://doi.org/10.1533/9781845692285.44.
acidic and basic solution, Electrochim. Acta 19 (1974) 471–475, https://doi. [40] J.A. González, J.M. Miranda, N. Birbilis, S. Feliu, Electrochemical techniques for
org/10.1016/0013-4686(74)87026-x. studying corrosion of reinforcing steel: limitations and advantages, Corrosion
[24] A.J. Bard, L.R. Faulkner, J. Leddy, C.G. Zoski, Electrochemical Methods: 61 (2005) 37–50.
Fundamentals and Applications, Wiley New York, 1980. https://www. [41] W. Nguyen, J.F. Duncan, T.M. Devine, C.P. Ostertag, Electrochemical
wiley.com/en-us/Electrochemical+Methods%3A+Fundamentals+and polarization and impedance of reinforced concrete and hybrid fiber-
+Applications%2C+2nd+Edition-p-9780471043720. reinforced concrete under cracked matrix conditions, Electrochim. Acta 271
[25] L. Bertolini, E. Redaelli, Depassivation of steel reinforcement in case of pitting (2018) 319–336, https://doi.org/10.1016/j.electacta.2018.03.134.
corrosion: detection techniques for laboratory studies: detection of steel [42] C. Christodoulou, G. Glass, J. Webb, S. Austin, C. Goodier, Assessing the long
reinforcement depassivation, Mater. Corros. 60 (2009) 608–616, https://doi. term benefits of Impressed Current Cathodic Protection, Corros. Sci. 52 (2010)
org/10.1002/maco.200905276. 2671–2679, https://doi.org/10.1016/j.corsci.2010.04.018.
[26] P.V. Nygaard, M.R. Geiker, B. Elsener, Corrosion rate of steel in concrete: [43] C.-F. Chang, J.-W. Chen, The experimental investigation of concrete
evaluation of confinement techniques for on-site corrosion rate carbonation depth, Cem. Concr. Res. 36 (2006) 1760–1767, https://doi.org/
measurements, Mater. Struct. 42 (2009) 1059–1076, https://doi.org/10.1617/ 10.1016/j.cemconres.2004.07.025.
s11527-008-9443-1. [44] C. Andrade, J.A. González, Quantitative measurements of corrosion rate of
[27] C.J. Newton, J.M. Sykes, A galvanostatic pulse technique for investigation of reinforcing steels embedded in concrete using polarization resistance
steel corrosion in concrete, Corros. Sci. 28 (1988) 1051–1074, https://doi.org/ measurements, Mater. Corros. 29 (1978) 515–519, https://doi.org/
10.1016/0010-938x(88)90101-1. 10.1002/maco.19780290804.
[28] S. Sathiyanarayanan, P. Natarajan, K. Saravanan, S. Srinivasan, G. Venkatachari, [45] P. Garcés, M.C. Andrade, A. Saez, M.C. Alonso, Corrosion of reinforcing steel in
Corrosion monitoring of steel in concrete by galvanostatic pulse technique, neutral and acid solutions simulating the electrolytic environments in the
Cem. Concr. Compos. 28 (2006) 630–637, https://doi.org/10.1016/j. micropores of concrete in the propagation period, Corros. Sci. 47 (2005) 289–
cemconcomp.2006.03.005. 306, https://doi.org/10.1016/j.corsci.2004.06.004.
[29] B.Elsener,Corrosionrateofsteelinconcrete—MeasurementsbeyondtheTafellaw, [46] A.M. Hassanein, G.K. Glass, N.R. Buenfeld, Protection current distribution in
Corros. Sci.47 (2005) 3019–3033,https://doi.org/10.1016/j.corsci.2005.06.021. reinforced concrete cathodic protection systems, Cem. Concr. Compos. 24
[30] ASTM G59–97, Standard test method for conducting potentiodynamic (2002) 159–167, https://doi.org/10.1016/s0958-9465(01)00036-1.
polarization resistance measurements American Society for Testing and [47] P.V. Nygaard, M.R. Geiker, Measuring the corrosion rate of steel in concrete -
Materials 2014.DOI:10.1520/g0059. effect of measurement technique, polarisation time and current: measuring
[31] S.G. Millard, D. Law, J.H. Bungey, J. Cairns, Environmental influences on linear the corrosion rate of steel in concrete, Mater. Corros. 63 (2012) 200–214,
polarisation corrosion rate measurement in reinforced concrete, NDT E Int. 34 https://doi.org/10.1002/maco.201005792.
(2001) 409–417, https://doi.org/10.1016/s0963-8695(01)00008-1. [48] S. Ahmad, M.A.A. Jibran, A.K. Azad, M. Maslehuddin, A simple and reliable
[32] R.G. Duarte, A.S. Castela, R. Neves, L. Freire, M.F. Montemor, Corrosion behavior setup for monitoring corrosion rate of steel rebars in concrete, Sci. World J.
of stainless steel rebars embedded in concrete: an electrochemical impedance (2014) 1–10, https://doi.org/10.1155/2014/525678.
spectroscopy study, Electrochim. Acta 124 (2014) 218–224, https://doi.org/ [49] L. Bertolini, B. Elsener, P. Pedeferri, E. Redaelli, R. Polder, in: Corrosion of Steel
10.1016/j.electacta.2013.11.154. in Concrete, Wiley Online Library, 2013, https://doi.org/10.1002/
[33] J. Wei, X.X. Fu, J.H. Dong, W. Ke, Corrosion evolution of reinforcing steel in 9783527651696.
concrete under dry/wet cyclic conditions contaminated with chloride, J. Mater.

15