Corrosion Science 46 (2004) 169181

www.elsevier.com/locate/corsci

Electrochemical impedance study on galvanized

steel corrosion under cyclic wetdry
conditionsinuence of time of wetness
A.P. Yadav *, A. Nishikata, T. Tsuru
Department of Metallurgical Engineering, Graduate School of Science and Engineering,
Tokyo Institute of Technology, 2-12-1, O-okayama, Meguro-ku, Tokyo 152-8552, Japan
Received 17 February 2003; accepted 11 April 2003

Abstract

Electrochemical impedance technique has been applied to study the corrosion behavior of
galvanized steel under wetdry cyclic conditions with various drying periods. The wetdry
cycles were carried out for the period of 336 h by exposure to alternate conditions of 1 h
immersion in a 0.5 M NaCl solution and drying for various time periods (11, 7 and 3 h) at 298
K and 60% RH. During the wetdry cycles, the polarization resistance, Rp , and solution re-
sistance, Rs , were continuously monitored. The instantaneous corrosion rate of the coating
was estimated from the obtained R
1 p and time of wetness was determined from the Rs values.
The corrosion potential, Ecorr , was also measured only during the immersion period of each
wetdry cycle. In all cases, the corrosion was accelerated by the wetdry cycles in the early
stage, and started to decrease at a certain cycle and nally became similar to that at the initial
cycle. The underlying steel corrosion commenced after the corrosion rate started to decrease.
The shorter drying period in each cycle led to higher amount of corrosion of the coating
because the surface was under wet conditions for longer periods. On the other hand, time to
red rust appearance due to occurrence of the underlying steel corrosion became shorter as the
drying period increased, although the total amount of corrosion was smaller. The corrosion
mechanism of substrate steel under various drying conditions has been discussed, the galvanic
coupling eect being taken into account.

2003 Elsevier Ltd. All rights reserved.

Keywords: Wetdry cyclic test; EIS; Galvanized steel; Atmospheric corrosion; Corrosion monitoring

*
Corresponding author. Fax: +81-3-5734-2835.
E-mail address: [email protected] (A.P. Yadav).

0010-938X/$ - see front matter
2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0010-938X(03)00130-6
170 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181

1. Introduction

The wide spread use of galvanized steels in a variety of structural applications such
as buildings and bridges are well known. The voluminous use of zinc for coating is
attributed to their excellent performance in corrosion protection of steel structure,
particularly in marine environment [1]. The demands for more corrosion resistant
coating are on rise, while more knowledge regarding the behavior of galvanized steel
corrosion in atmospheric environment is in need. In atmospheric environments,
where corrosion proceeds under a thin electrolyte layer during wetdry cycles, the
corrosion process is very dierent from that in the bulk solutions because the diusion
of oxygen is greatly enhanced under a thin electrolyte layer [26]. Also the solvation
capacity of a thin electrolyte layer for the dissolved species becomes very limited,
which aects the process of formation and transformation of corrosion products [7].
The data from eld exposure test in dierent geographical locations have shown
that the time to red rust formation and coating life are largely a linear function of
coating thickness [8]. In the real atmospheric environment, cycles of dierent wet and
dry periods are encountered. However, there are limited data on how the dierent
nature of wet and dry cycles aects the time to red rust formation and the coating life
of galvanized steel.
Electrochemical impedance spectroscopy (EIS) is proved to be quite useful
technique for the monitoring of atmospheric corrosion [911]. Short time exposure
of galvanized steel in chloride solution with dierent Al contents has shown that the
addition of Al more than 5 mass% is eective in increasing the corrosion resistance
[12]. The corrosion rate has been reported to be almost constant throughout the
short exposure time. Similar study was made on electroplated zinc coating for short
time exposure and the eect of parameters like pH, temperature and sulfate and
chloride ions were discussed [13]. Since these measurements were made for short time
exposure, the long-term corrosion behavior cannot be explained. Therefore, in the
present study, a wetdry cyclic test, which simulates marine environments, has been
used to evaluate the corrosion behavior of galvanized steel. Special attention has
been paid to inuence of time of wetness on the time to red rust appearance due to
occurrence of the underlying steel corrosion.

2. Experimental

2.1. Electrode system

A two-electrode cell conguration with the exposed area of 1 cm2 each, as shown
in Fig. 1, was used as a probe for the corrosion monitoring. A commercial hot dip
galvanized steel sheet with the coating thickness of 1820 lm was used as the probe
electrode. A pair of identical electrodes was cut into the dimension of 14 mm 14
mm 2.5 mm from the galvanized steel sheet and mounted in an ambient temper-
ature solidifying epoxy resin with the separation gap of 0.1 mm. In order to avoid
crevice corrosion on the exposed area and on the edges separating the two metal
 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181 171

0.5mm
10mm
10mm Metal Bank
plate

10mm
10mm
10mm
Epoxy

Gap 0.1mm
Gap 0.1mm

Lead wires
40mm
(a) (b)

Fig. 1. Schematic diagram of two-electrode cell used for monitoring of impedance (a) top view (b)
transverse cross-sectional view.

sheets, these were covered with a thin layer of epoxy resign. A three-electrode cell
conguration was employed for polarization measurement with an Ag/AgCl, KCl
(saturated) reference electrode (SSE) and Pt sheet as a counter electrode.

2.2. Wetdry cycle

The wetdry cycle was conducted by exposing the sample to an alternate condi-
tion of 1 h immersion in a 0.5 M NaCl solution and various hours of drying at 298 K
and 60% RH. The solution was prepared from analytical grade chemical of NaCl
and distilled water. The drying times were set at 11, 7 and 3 h. Duplicate samples
were exposed in all sets of experiments to check the reproducibility. To ensure the
constancy of electrolyte layer thickness at the onset of each dry cycle, a peripheral
bank of 0.5 mm thick was xed on the electrode surface. This gave a layer thickness
of about 0.5 mm on the initiation of the dry period. The detail of cyclic wetdry set-
up was described elsewhere [11].

2.3. Corrosion monitoring

Continuous monitoring of corrosion rate of the zinc coating was performed by

using an AC corrosion monitor (RIKEN DENSHI CT-3) with a multiplexer con-
trolled by a personal computer. A special software was used for the data acquisition
at the time interval of 12 min. This system operates at two frequencies, high fre-
quency, 10 kHz (ZH ) and low frequency, 10 mHz (ZL ). The impedances were si-
multaneously measured at these frequencies, and the polarization resistance, Rp was
given by subtracting the high frequency impedance from the low frequency one as
follows
Rp ZL
ZH 1
This is based on the assumption that the high frequency impedance gives a solution
resistance and the low frequency provides a sum of solution resistance and polari-
zation resistance.
172 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181

The corrosion current density, icorr can be calculated from this Rp value by using
SternGeary equation [13] given by
icorr kR
1
p 2
where the value of k is a proportional constant and is given by the expression
b a bc
k 3
2:303ba bc
where ba and bc are the anodic and cathodic Tafel slopes for zinc dissolution and
oxygen reduction, respectively.
The average corrosion mass loss, DM (g cm
2 ) of zinc coating after the exposure
periods of ts was calculated by using the relation as follows

Z t
MZn
DM icorr dt 4
2F 0

where MZn is atomic weight of zinc ( 65.4 g mol
1 ), and F is Faraday constant
( 96485 C mol
1 ). Average coating thickness loss, DT was calculated by dividing
DM value by the density of zinc (7.14 g cm
3 ).
After the wetdry cyclic exposure test, the whole frequency impedance charac-
teristics of corroded samples were measured in the range of 1 mHz to 20 kHz using
a Schlumberger SI 1280 frequency response analyzer.

3. Results and discussion

3.1. Corrosion monitoring under wetdry cycles

The wetdry was carried out by exposure to alternate conditions of 1 h immersion

in a 0.5 M NaCl solution and drying at 25 C and 60% RH. Three dierent drying
periods, 11, 7 and 3 h, were employed to investigate the eect of drying on time to
red rust appearance. During the wetdry cycles, the R
1
1
p and Rs values were mon-

1
itored every 12 min. The Rp was considered an index of the corrosion rate and the
R
1
s was used to calculate time of wetness (TOW). The monitoring results of the Rp

1

values are shown in Fig. 2. Fig. 2(a)(c) are the results for 11, 7 and 3 h dryings,
respectively. These were measured using the samples placed horizontally with their
surface facing upward. Fig. 2(d) is a result obtained under the same wetdry con-
dition as the (c), but using the probe set at 30 to the horizontal.
In all cases, the corrosion was accelerated in the early stage and decreased in the
later stage. The exposure of the samples to longer drying periods led to early decrease
of the R
1
p . As was reported in the previous study [13], red rusts which were formed
due to underlying steel corrosion, were rst observed after the R
1 p showed a maxi-
mum, except for the specimen of Fig. 2(c). Arrows in Fig. 2 indicate the cycles where
red rusts were rst observed with the naked eyes. It was found from the monitoring
results that red rusts were not observed only for the specimen exposed horizontally
to the shortest drying period (3 h drying) during 336 h exposure, and that the time to
 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181 173

8 8
(a) 11h drying (b) 7h drying

Rp-1/ x 10--3 -1 cm-2

Rp-1 / x 10-3 -1 cm-2

6 6

4 4

2 2

0 0
0 48 96 144 192 240 288 336 0 48 96 144 192 240 288 336
Time, t/h Time, t/h
8 8
(c) 3h drying (d) 3h (Inclined)
Rp-1 / x 10-3 -1 cm-2

Rp-1 / x 10-3 -1 cm-2

6 6

4 4

2
2

0
0 48 96 144 192 240 288 336 0
0 48 96 144 192 240 288 336
Time, t/h
Time, t/h
t
Fig. 2. Corrosion monitoring results of galvanized steel during wet and dry cycle of immersion in 0.5 M
NaCl at 298 K for 1 h and drying at 298 K 60% RH for (a) 11 h, (b) 7 h and (c) 3 h on a horizontal
specimen surface and (d) 3 h on an inclined specimen surface.

red rust appearance was signicantly shortened by exposure to wetdry cycles with
longer drying period. This means that when the drying period in each cycle is suf-
ciently long, the zinc coating does not protect the substrate steel by galvanic action.

3.2. Corrosion products

After the wetdry cyclic test, the corrosion products were analyzed using X-ray
diraction (XRD). Many peaks were obtained, however only ZnO and ZnCl2
4Zn(OH)2 , were identied. There were no peaks for iron compounds probably they
existed in amorphous form. The XRD and SEM photographs showed that there is
no dierence in the composition and morphology of the corrosion products among
the samples exposed to the wetdry conditions with dierent drying periods.

3.3. Polarization behavior

The polarization behavior of zinc in 0.5 M NaCl solution is shown in Fig. 3. The
currentpotential curve was measured at a scan rate of 1 mV/s from the open circuit
potential towards either anodic or the cathodic direction at 298 K, using as polished
samples. Before start of the measurement, sample was kept in 0.5 M NaCl solution
for 30 min. Accordingly, its surface was considered to be covered with a very thin
174 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181

cm--2
10 2

Current density, i /mA

10 1 40mV/decade

/
10 0

10 -1

10 -2

10 -3
--1.6 --1.4 --1.2 -1.0
- -0.8
- --0.6 -0.4
-

E/V vs. Ag/AgCl

Fig. 3. Polarization curve of galvanized steel measured in 0.5 M NaCl solution at scan rate of 1 mV/s after
immersion for 30 min.

oxide layer of zinc. The presence of oxide layer has been conrmed by measuring
deferential capacitance [14]. A diusion limiting current of oxygen reduction can be
seen in the early stage of polarization. However, at lower potentials (less than )1.5
V), the reduction of H2 O becomes more dominant, leading to a progressive increase
in the current density. The Tafel slope of the anodic polarization curve, ba was found
to be about 0.040 V/decade, which is in general agreement with the reported value
[15]. On the other hand, the Tafel slope of the cathodic side, bc was determined to be
innite (1), assuming that the oxygen reduction rate is controlled by diusion
process. These values of the Tafel slopes will be used for estimating the corrosion
rate of the coating in a latter section.

3.4. Time of wetness

The corrosion rate averaged over whole exposure time (336 h) is often used to
evaluate the coating performance. However, as the corrosion proceeds at negligibly
small rates when the surface is completely dried up, it is better to consider an average
corrosion rate during the sample surface is covered with a conductive electrolyte
layer. This will give better understanding of the inuence of drying or wet periods on
the corrosion process. For this calculation, the time period over which the sample
surface remained covered with an electrolyte layer, i.e. TOW was determined from
the high frequency part of the impedance data, which gives the solution resistance, Rs
Fig. 4 shows a typical monitoring results of solution conductance, R
1 s for the sample
subjected to 7 h drying. It is apparent that during the 1 h immersion, the R
1 s value
ranges from 5.56.0 10
2 X
1 , and it dropped to 0.81.0 10
2 X
1 on the onset of
drying period and slowly decreased to zero when the surface dried up completely.
Therefore, TOW can be taken as the time periods when the solution conductance
is greater than 1 10
3 X
1 . The TOW is aected by the formation of corrosion
products, which may change the average wetting time during a drying period.
Therefore, to calculate the TOW, total number of data having R
1 s value higher than
1 10
3 X
1 was obtained which was multiplied by the time interval in which one
 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181 175

10
7h drying

Rs -1 / x 10 -2 -1 cm -2
8

0
0 48 96 144 192 240 288 336
Time, t/h
Fig. 4. Monitoring results of reciprocal of solution resistance obtained from the high frequency impedance
at 10 kHz under 7 h drying condition.

Table 1
Summary of corrosion rate of galvanized steel calculated by integrating R
1
p vs. time plot in Fig. 2

Drying Total Time of wetness Iavr 10
6 A cm
2 Iavr 10
6 A cm
2
periods (h) cycle no. TOW/h
11 28 96 8.0 29.0
7 42 180 16.0 36.0
3 84 271 34.0 42.0
3 (inclined) 84 111 15.0 46.0

data was obtained. The TOW values thus obtained are shown in Table 1. For the
sample of 3 h drying and horizontal set, the TOW was calculated to be 271 h, which
corresponds to about 81% wet of the total exposure period (336 h). The sample
exposed to the 11 h drying had the shortest wet period (96 h, 29%) among them.

3.5. Average corrosion rate

To estimate the average corrosion rate, Iavr over whole exposure time, the R
1
p vs.
time curves in Fig. 2 were integrated and the integrated values were divided by the
total exposure time (336 h), i.e.
Z 336
k
Iavr R
1
p dt 5
336 0

On the other hand, the average corrosion rate Iavr over time of wetness (TOW) was
estimated by dividing the integrated values by TOW, i.e.
Z 336
 k
Iavr R
1
p dt 6
TOW 0
The summary of these two types of average corrosion rates is given in Table 1. The k
value of 0.017 V was used for the calculation, which was determined by Eq. (3), using
the anodic Tafel slope (ba 0:040 V) and cathodic Tafel slope (bc 1) of the
176 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181

polarization curve of Fig. 3, assuming that the k value does not change during the
whole cycles. It can be seen in Table 1 that the average corrosion rate, Iavr during the
total exposure period (336 h) becomes larger as the TOW value is higher because
corrosion progresses only when the material surface is covered with an electrolyte
layer. It is apparent that the Iavr of horizontal sample (TOW: 271 h) is two times

higher than that of the inclined sample (TOW: 111 h). On the other hand, the Iavr
which is average corrosion rate during time period when the surface is wet, becomes
higher as the drying period of each cycle is shorter, i.e. as the total cycle number
increases (Table 1). This indicates that the corrosion is accelerated by the cyclic wet
dry.
It is shown in Table 1 that all the samples, except under the inclined condition,
remained wet over 3 h during the drying period in each cycle. As mentioned before,
all the samples, except the horizontal sample under 3 h drying condition, showed the
occurrence of the underlying steel corrosion (Fig. 2). The average thickness loss
DT total of the coating layer was calculated from Eq. (4) and is presented in Table 2
along with the average mass loss, DMtotal . It is interesting to note that though the
sample exposed under the 3 h drying underwent a highest loss of the coating
thickness (12 lm), it did not show red rusts during the entire exposure time. In
contrast, the sample subjected to 11 h drying condition had least average thickness
loss (2.9 lm), but the red rusts were observed earlier than the other samples. This
indicates the signicance of drying time available in each cycle.
In addition, it should be noted that the underlying steel corrosion commences,
although the zinc coating still remains because average values of the corrosion
thickness loss are much lower than the initial thickness (1820 lm) of the coating.
Average values of the thickness loss DT max until the cycle where the R
1 p shows a
maximum value, R
1 p max were calculated and presented in Table 2. We have claried
that the corrosion front reaches the coating/substrate boundary on around the cycle
of R
1
1
1
p max [16]. In other words, the appearance of Rp max on the Rp vs. time curve
is an indication that the underlying steel is partially exposed to chloride solutions. It
can be seen in Fig. 2 that the corrosion front reaches the boundary earlier as the
longer drying period is employed, though the DT max value is smaller. This indicates
that corrosion of the coating progresses heterogeneously at the initial stage and the
corrosion becomes more localized by exposing to wetdry conditions with longer
drying periods, because, in case of the longer drying, the coating is exposed to more
concentrated chloride environments for the longer periods.

Table 2
Thickness loss and mass loss of galvanized steel under various wetdry cyclic conditions
Drying periods (h) Mass loss Mtotal Thickness loss Thickness loss at
10
3 g cm
2 T total =lm R
1
p max =lm

11 2.1 2.9 1.3

7 4.7 6.6 5.0
3 8.4 11.8 9.3
3 (inclined) 3.8 5.3 4.0
 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181 177

-0.80
11h dry
7h dry

Ecorr / V vs. Ag/AgCl

-0.85
3h dry
-0.90 3h dry (300)

-
-0.95

-1.00

-1.05
0 0.5 1
Normalized cycle Number

Fig. 5. Time variation of corrosion potential measured during corrosion monitoring once a day in im-
mersed conditions.

3.6. Corrosion potential

The time variation of corrosion potential, Ecorr of all the samples under dierent
cyclic environments are shown in Fig. 5. These were measured once a day only
during immersed conditions while monitoring the Rp . In this gure, the Ecorr values
are plotted against the normalized cycle number divided by the total cycle number.
The arrows indicate the cycle in which red rusts were rst observed with the naked
eyes. In all cases, for rst few cycles the Ecorr shifts to less noble values due to en-
hancement of anodic dissolution of the coatings [13]. As the cycle proceeds, the
corrosion products are deposited on the surface and the potential gradually shifts in
noble direction. However, the shift of Ecorr to noble values depends on the type of
wetdry cyclic conditions employed. The appearance of red rusts was observed when
the Ecorr was in the range of )0.950 V. This range of potential is suciently negative
than potentials used for cathodic protection of ordinary steels, )0.77 V, where it is
expected that the zinc coating layer around a defect would act as a sacricial anode
[17]. The monitoring results of the Ecorr imply that the red rusts appeared during the
drying periods.

3.7. Impedance characteristics

The impedance characteristics of the samples measured at the corrosion potential

after the wetdry cyclic test are shown in Fig. 6. The points in the plots indicate the
experimental data and the lines are the results of curve tting. The results revealed
that the interface response contains only one time constant and that a characteristic
frequency, where the phase shift shows a minimum, shifted in the low frequency side
as the corrosion progressed. This shift indicates an increase in the capacitance value.
Due to frequency dispersion of impedance data, the constant phase element ap-
proach was used to t the data and the equivalent circuit is shown in Fig. 7. It
consists of a solution resistance, Rs in series with a parallel circuit of a charge transfer
178 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181

104
(a) 11h
7h

IZI / cm2
103 3h
3h (300)

102

101

20
(b)
Phase shift, /degree

--20

--40

-60

--80
10 -3 10 -2 10 -1 10 0 10 1 10 2 10 3 10 4 10 5
Frequency, f/Hz

Fig. 6. Impedance characteristics obtained on corroded galvanized steel samples under dierent wetdry
cyclic conditions in 0.5 M NaCl solution.

Fig. 7. Equivalent circuit used for tting the impedance results of Fig. 6.

resistance, Rct and a constant phase element, CPE. Though the experimental data
and the results of curve tting are in fairly good agreement, it needs to evaluate the
physical signicance of parameters used for tting.
The parameters used for the tting are shown in Table 3. The parameters T and P
are related to the CPE by the equation
ZCPE T jxP 
1 7
2
where j is the imaginary number j
1, T is a constant representing capacitance,
x is the angular frequency, and P a=p=2, and a is the phase angle of the CPE
[18]. The factor P is adjustable parameter that usually lies between 0 and 1. The CPE
describe an ideal capacitor when P 1, a depressed semicircle appears when P < 1.
In most corrosion system, the capacitive semicircle exhibits some deviations from an
ideal semicircle. This has been referred to as frequency dispersion attributed to
surface inhomogeneities. The values of P lie in between 0.71 to 0.87, indicating that
 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181 179

Table 3
Parameters used for tting the impedance data in Fig. 6
Drying period (h) Parameters
Rs (X cm2 ) Rct (X cm2 ) T (mF/cm2 ) P
11 13 2120 15 0.87
7 15 1800 14 0.87
3 13 1450 7.0 0.82
3 (inclined) 18 1600 3.6 0.71

the sample surface was slightly inhomogeneous, because it may be covered with the
corrosion products of dierent thickness. The capacitive element T has an abnor-
mally high value in the range of millifaraday/cm2 . It cannot be considered as the
double layer capacitance.
There are several research papers discussing the electrochemical impedance
characteristics of metals under thick corrosion products layer. Bonnel et al. [19] have
reported an abnormal high value of capacitance and have attributed to the porous
and conductive nature of underlying deposits. They have shown that the increase of
roughness could not give such a high value of capacitance. Macdonald [20] has in-
terpreted the impedance behavior of metals covered with porous lms using a
transmission line (TML) model. In this model, current distribution on pore wall of
porous lms is considered and the impedance is characterized by the fact that if the
current distributes unevenly on the wall surface, the phase shift does not exceed
)45. The uneven current distribution should remarkably appear at high frequency
regions in this model. In case of the impedance behavior of zinc obtained in this
study, the surface was covered with thick and porous corrosion products, but a
typical behavior of TML type was not observed. Accordingly, the abnormally large
capacitance can not be explained by this model.
In this study, the abnormally large capacitance could not be interpreted, but from
the whole frequency impedance, it was conrmed that the impedance at 10 mHz
employed for monitoring is close to the polarization resistance, which is dened as
impedance at a low frequency limit.

3.8. Corrosion mechanism of underlying steel

The corrosion monitoring as well as corrosion potential measurements have

shown that the drying periods available in each cycle play an important role in
determining the coating performance of galvanized steel. As can be seen from the
monitoring results of Fig. 2, the corrosion rate increases in the rst half and then
decreases in the second half. It is to be noted that the surface of galvanized steel was
covered with air formed oxide layer as it was exposed to the corrosion medium
without polishing. In addition, the surface of commercial galvanized specimen is
heterogeneous. Accordingly, the acceleration of corrosion at the initial stage is at-
tributed to breakdown of the air-formed oxide lm and an increase in active
dissolution sites on the coating surface due to an increase in concentration of
chloride ions during the drying period. As mentioned before, the corrosion at the
180 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181

initial stage progressed heterogeneously and was more localized by subjecting the
sample to longer drying conditions.
It has been reported that the corrosion fronts reach the boundary of the coating/
substrate when a rapid reduction of the R
1p of the coating layer occurs [16], followed
by commencement of the underlying steel corrosion. However, as shown in Fig. 2,
the underlying steel corrosion did not occur for the sample under the condition of 3 h
drying although decrease of R
1 p was observed (Table 2). This means that the inhi-
bition is a necessary condition, but not sucient condition for the commencement of
underlying steel corrosion. The sucient drying time available in each dry period
seems to be necessary for the early commencement of underlying steel corrosion.
Another important feature under the 3 h drying is that when the R
1 p decreases, the
corrosion potential measured in the immersed condition became even more noble
than under the 11 h drying, but red rusts was not observed.
This dierence in the behavior of samples towards red-rusting under dierent
drying periods can be explained by considering the galvanic eect during the drying
periods. When the surface of the samples becomes covered with thick zinc corrosion
products, the coating surface loses its ability as a sacricial anode. However, if the
surface is covered with relatively thick electrolyte throughout the drying periods or
for longer periods in each cycle, a very large area around a defect could act as the
sacricial anode. The defect means a site where the corrosion front reaches the
substrate steel and the steel is directly exposed to the environment. As a result,
sucient galvanic current will be supplied to the defect for the cathodic protection
and red rusts will not appear. For this reason, despite a highest thickness loss of
coating layer, the sample under 3 h drying did not show red rusts. In contrast, if the
surface remains dry for suciently long time in each dry period, the coating would
be covered with invisible layers of electrolyte and anode area to a defect (cathode)
could become extremely limited. In such a condition, the sucient galvanic current
will not be supplied to the defect for the cathodic protection [21]. Thereafter, the
underlying steel corrosion starts, as found under the 11, 7 h and inclined 3 h drying
conditions despite a low coating thickness loss. This also explains why the red rust
appearance commenced even though the corrosion potential under immersed con-
ditions was in the range of )0.950 V, which is considered to be protective to steel.

4. Conclusions

The corrosion monitoring of galvanized steel was carried out under wetdry cyclic
conditions containing chloride ions. The following conclusions are drawn:

1. At early stage, the corrosion of zinc coating was accelerated by the wetdry cycles.
The localized type of corrosion took place because the coating is exposed to the
concentrated chloride solution during drying.
2. The underlying steel corrosion commenced when the reciprocal of polarization re-
sistance, which was considered an index of corrosion rate of the zinc coating in
this study, started to decrease.
 A.P. Yadav et al. / Corrosion Science 46 (2004) 169181 181

3. The high amount of corrosion of the coating layer was obtained when the drying
periods in each cycle was short because this kept the surface wet for longer peri-
ods. On the other hand, presence of long drying periods in each cycle shortened
the time to red rust appearance, although the total amount of corrosion was smal-
ler. This indicates that the galvanic protection ability of zinc coating layer is not
sucient under a very thin electrolyte layer.

References

[1] X.G. Zhang, Corrosion and Electrochemistry of Zinc, Plenum Press, New York, 1996.
[2] I.L. Rosenfeld, Atmospheric Corrosion of metals, Some Questions of theory, in: the 1st International
Congress on Metallic Corrosion, London, April 1991, p. 243.
[3] M. Stratmann, H. Streckel, Corros. Sci. 30 (1990) 697.
[4] F. Mansfeld, J.V. Kenkel, Corros. Sci. 16 (1976) 111.
[5] N.D. Tomashov, Corrosion 20 (1964) 76.
[6] A. Nishikata, Y. Ichihara, Y. Hayashi, T. Tsuru, J. Electrochem. Soc. 114 (1997) 1244.
[7] D.C. Stiles, E.O. Edney, Corrosion 23 (1967) 15.
[8] K. Videm, R. Myrdal, in: The Electrochemical behavior of Steel in Concrete and How to Evaluate the
Corrosion Rate, National Association of Corrosion Engineers, Corrosion
96, paper no. 348, 1996.
[9] A. Nishikata, Y. Ichihara, T. Tsuru, Corros. Sci. 37 (1995) 897.
[10] A. Nishikata, Y. Ichihara, T. Tsuru, Electrochim. Acta 41 (1996) 1057.
[11] A. Nishikata, Y. Yamashita, H. Katayama, T. Tsuru, A. Usami, K. Tanabe, H. Mabuchi, Corros.
Sci. 37 (1995) 2059.
[12] H. Katayama, Y.C. Tay, A.S. Viloria, A. Nishikata, T. Tsuru, Mater. Trans. JIM 38 (1998) 1089.
[13] G.A. El-Mahdy, A. Nishikata, T. Tsuru, Corros. Sci. 42 (2000) 183.
[14] L.M. Baugh, Electrochim. Acta 24 (1979) 657.
[15] W. Walter, Corros. Sci. 32 (1991) 1331.
[16] A.P. Yadav, A. Nishikata, T. Tsuru, Corros. Sci., Submitted for publication.
[17] R.A. Humble, Corrosion 4 (1948) 358.
[18] B.A. Boukamamp, Solid State Ionics 20 (1980) 31.
[19] A. Bonnel, F. Dabosi, C. Deslouis, M. Dupart, M. Keddam, B. Tribollet, J. Electrochem. Soc. 130
(1983) 753.
[20] D.D. Macdonald, in: Techniques for Characterization of Electrodes and Electrochemical Processes,
John Wiley and Sons, New York, 1991, p. 515.
[21] X.G. Zhang, Corrosion 56 (2000) 134.