Corrosion Science 49 (2007) 211–219

www.elsevier.com/locate/corsci

Corrosion and corrosion test methods of zinc

coated steel sheets on automobiles
Sakae Fujita *, Daisuke Mizuno
JFE Steel Corporation, 1 Kawasaki-cho, Chuo-ku, Chiba 260-0835, Japan

Available online 25 July 2006

Abstract

The effect of zinc and zinc alloy coated steel sheets on perforation corrosion in actual automobiles
and the relevant accelerated corrosion test methods were studied. The main factor affecting corrosion
in the crevice of lapped panels was the coating weights of zinc and zinc alloys rather than the type of
coating. Perforation corrosion process of galvanized steel in the crevice of lapped portion in auto-
mobiles was divided in four stages. Based on the analysis of corrosion in actual automobiles, a Per-
foration Corrosion Index, PCI for lapped steel panels was proposed. Assuming PCI for various
accelerated corrosion test methods, the corrosion resistance of various types of coated steel sheets
in actual environments was evaluated.
Ó 2006 Published by Elsevier Ltd.

Keywords: A. Zinc; A. Zinc alloy; A. Coating; A. Steel; B. Corrosion test; C. Corrosion; C. Perforation;
C. Automobile

1. Introduction

Automotive corrosion is classified into three types, perforation corrosion, cosmetic

corrosion, and edge corrosion. Perforation corrosion is the most important problem rather
than cosmetic corrosion and edge corrosion. Corrosion-resistance quality targets for
the automobile have been established in North America and northern Europe to reduce

*
Corresponding author. Tel.: +81 43 262 2451; fax: +81 43 262 2909.
E-mail address: [email protected] (S. Fujita).

0010-938X/$ - see front matter Ó 2006 Published by Elsevier Ltd.

doi:10.1016/j.corsci.2006.05.034
212 S. Fujita, D. Mizuno / Corrosion Science 49 (2007) 211–219

automotive corrosion, including the Canada Code I (1978), Canada Code II (1981), Nor-
dic Code (1983), and the voluntary targets of US Big 3 (GM, Ford, Chrysler; 1989). Sub-
sequently, requirements for corrosion resistance have become progressively higher, as seen
in the voluntary 12-year corrosion guarantee proposed by the major European automak-
ers in 1998.
Since the 1990s, studies of the actual condition of automotive corrosion in areas where
deicing salt was used have been conducted in Europe, North America, and Japan [1–9].
Rendahl [5] collected a large number of auto parts from vehicles used in northern Europe
and investigated corrosion in these specimens. The results revealed that the main factor
affecting perforation corrosion in automobiles was not the type of coating, but the coating
weight. The authors [6–9] also reported the main factor affecting perforation corrosion was
coating weight regardless of type of coating, based on corrosion analysis of actual auto-
mobiles collected in North America. It is difficult to achieve the perfect agreement between
the corrosion phenomena in accelerated corrosion tests and those in actual vehicles.
Efforts have been made to establish the correlation between the test results and the corro-
sion resistance in actual automobiles.
In the present study, the perforation corrosion resistance in zinc- and zinc alloy-
coated steel sheets were investigated quantitatively. Based on the analysis of actual
automobiles and the evaluation of accelerated corrosion test methods, the corrosion resis-
tance for perforation in actual environment using accelerated corrosion tests will be
discussed.

2. Experimental procedure

2.1. Corrosion analysis method of actual automobiles

Automobiles driven for 5–11 years in North America were collected. The recovery
area was Detroit, Michigan (USA), where deicing salt was used. The lapped steel panels
in the recovered automobiles were disassembled, the paint film was stripped, and the
corrosion products were removed. The surface corroded in the crevice was then divided
into unit sections. Using a micrometer with a protruding tip, the maximum depth of
corrosion in the respective sections was measured as the difference with the sheet thickness
in normal parts. An extreme value (double exponential probability) analysis of the
maximum corrosion depth occurring in each part was performed using the Gumbel distri-
bution [10] and the mode of maximum corrosion depth in each part of automobiles was
obtained.

Table 1
Accelerated corrosion test methods
Test no. Test method
T1 Corrosion test of automaker A
T2 Corrosion test of automaker B
T3 JASOM609-91
T4 SAE J2234 – 5days type
T5 SAE J2234 – 7days type
 S. Fujita, D. Mizuno / Corrosion Science 49 (2007) 211–219 213

2.2. Accelerated corrosion test methods

Table 1 shows accelerated corrosion test methods used in evaluations of perforation

corrosion. Corrosion test methods T1 and T2 are used by two automakers, respectively,
and T3, T4 and T5 have been standardized as the cosmetic corrosion test methods by
the Society of Automotive Engineers of Japan and the Society of Automotive Engineers
(SAE), respectively. The corrosion tests were performed by wet and dry cycle, including
the wet period of immersion, salt spray or humid environment. Table 2 shows the types
of coated steel sheets used in the corrosion tests. The structure of the specimen used in
the corrosion tests is shown in Fig. 1. The specimen was assembled by stacking and spot
welding coated steel sheets of the same type. After spot welding, zinc phosphate-type
chemical conversion treatment was performed (manufactured by Nihon Parkerizing; coat-
ing weight on outer side; 3 g/m2), and electrodeposited paint was applied (manufactured
by Kansai Paint; film thickness on outer side: 20 lm).

Table 2
Materials used for accelerated corrosion tests
Sign Materials Coating weight (g/m2)
CRS Cold rolled steel sheet 0
EG20 Electro galvanized steel sheet 20
EG50 50
EG70 70
ZnNi30 Zn–Ni Electro galvanized steel sheet 30
GA40 Galvannealed steel sheet 40
GA45 45
GI50 Galvanized steel sheet 50
GI60 60
GI90 90
GI100 100

180

150
70

40

45
Spot welding
Fig. 1. Lapped type of specimen used for accelerated corrosion tests.
214 S. Fujita, D. Mizuno / Corrosion Science 49 (2007) 211–219

3. Results

3.1. Perforation corrosion in actual automobiles

Fig. 2 shows the relationship between the mode of maximum corrosion depth in the
lapped part of the side-sill and the used period. Wax was not applied to these parts.
Assuming the thickness of steel substrate in the outer panel is 0.7 mm, it can be under-
stood that perforation corrosion will not occur in panels for at least 14 years with GI
(coating weight: 120 g/m2), whereas, the perforation life is estimated at approximately 8
years with GA (coating weight: 35 g/m2).

1.0
Mode of maximum corrosion depth / mm

SIDE SILL
without wax
0.8

0.6
GA(35g/m 2)
GA(35g/m
GI (120g/m22)
0.4

0.2

0.0
0 2 4 6 8 10 12 14 16
Used Period in field / year
Fig. 2. Relationship between the mode of maximum corrosion depth in the lapped part of the side-sill and the
used period.

1.0
Mode of maximum corrosion depth / mm

0.9 CRS
EG50 CRS
0.8
GA45
0.7 ZnNi30
0.6
ZnNi30
0.5
0.4
0.3
GA45
0.2
EG50
0.1
0.0
0 50 100 150
Test cycles

Fig. 3. Relationship between the mode of maximum corrosion depth and the number of test cycles in corrosion
test T1.
 S. Fujita, D. Mizuno / Corrosion Science 49 (2007) 211–219 215

3.2. Perforation corrosion evaluation by accelerated corrosion tests

Fig. 3 shows the relationship between the mode of the maximum corrosion depth and
the number of test cycles in corrosion test method T1 (automaker A). It can be understood
that lapped specimens began to be corroded immediately after the start of the test cycle
and then corrosion progressed in a basically linear manner with cold rolled steel (CRS).
In comparison, the onset of corrosion was delayed and corrosion of the steel sheets began
after a certain incubation period of time with Zn and zinc alloy-coated steel sheets.

4. Discussion

4.1. Corrosion process of perforation

The authors [9] proposed the corrosion process affecting Zn coated steel sheets in
lapped panels of automobiles dividing by periods s1–s4. Stage 1: Stage in which the Zn
coating film covers the entire surface of the substrate steel and the coating film itself is cor-
roded (s1), Stage 2: Stage in which the Zn coating is partially lost and the Zn film is sac-
rificially corroded protecting the substrate steel (s2), Stage 3: Stage in which the substrate
steel sheet is corroded but corrosion is restrained by zinc corrosion products (s3), and
Stage 4: Stage in which the substrate steel is corroded at the same rate as uncoated steel
(s4). This corrosion process of GI (120 g/m2) with the corrosion-resistant sides facing each
other in North America is shown schematically in Fig. 4. The corrosion-resistance period
s1–s3 during which Zn coating displayed the protective effect was approximately 9 years.
Within this period, the corrosion-resistance period during which corrosion of the substrate
steel was restrained by Zn corrosion products (s3) was 4 years, and thus accounted for
approximately 45% of the total corrosion-resistance period attributable to the Zn coating
(s1  s3).

Fig. 4. Schematic corrosion process of zinc coated steel in the crevice of lapped panels in automobiles.
216 S. Fujita, D. Mizuno / Corrosion Science 49 (2007) 211–219

4.2. Life prediction method using accelerated corrosion test

Figs. 5 and 6 show the relationship between the mode of maximum corrosion depth and
the coating weight in the corrosion tests T3 and T1 in this study. As shown in Fig. 5 cor-
rosion test T3 gives the results in which the coating weight is not the controlling factor, and
the use of Zn–Ni or GA is a more important factor in the corrosion resistance than the
coating weight. As shown by arrows in this figure, the corrosion resistance of GA40
and ZnNi30 can be estimated to be equal to pure zinc with coating weights of 142 g/m2
and 68 g/m2, respectively. Thus, it can be understood that corrosion test T3 is not repro-
ducible the phenomena observed in actual automobiles. On the other hand, as shown in
Fig. 6, corrosion test T1 identified heavier coating weights as the primary factor in sup-

0.8
Mode of maximum corrosion depth / mm

0.7

0.6
C RS
CRS
0.5

0.4
G I6 0
GI60
0.3
EG20
E G 20 EG
E G 50
50 G I1 00
GI100
0.2
Z nN 3
ZnNi30
i0
0.1
G A 40
GA40
0
0 20 40 60 80 100 120 140 160
Coating weight / g/m2

Fig. 5. Relationship between the mode of maximum corrosion depth and the coating weight of specimen in 150
cycles of corrosion test T3.

1.0
Mode of maximum corrosion depth / mm

0.9
C RS
CRS
0.8
0.7
0.6
E G 20ZZnNi30
EG20 nN 3
i0
0.5
0.4
E G 40
EG40
0.3
0.2 G A 45
GA45
E G 50 EG70
EG50 E G 70
0.1
EG90
E G 90
0.0
0 10 20 30 40 50 60 70 80 90
Coating weight / g/m2

Fig. 6. Relationship between the mode of maximum corrosion depth and the coating weight of specimen in 120
cycles of corrosion test T1.
 S. Fujita, D. Mizuno / Corrosion Science 49 (2007) 211–219 217

pressing perforation corrosion. GA45 and ZnNi30 show similar corrosion resistance to
that of zinc coating (EG, GI) with the same weights. Thus, it can be said that corrosion
test T1 reproduces the phenomena observed in actual automobiles with comparatively
good accuracy.
From the viewpoint of atmospheric corrosion, Zhang [11] has reported on the relation-
ship between corrosion test methods and corrosion in actual environments, focusing on
the ratio of the corrosion rates of zinc and steel. In order to clarify the relationship
between these corrosion test methods and corrosion in actual vehicles, the authors studied
the corrosion-resistance period (s1 + s2 + s3) attributable to protection by the pure zinc
coating such as EG or GI and the corrosion period of steel substrate (s4). Actual corrosion
conditions affect not only the period s1 + s2 + s3 but also the period s4. Thus for the eval-
uation of perforation corrosion the authors proposed a Perforation Corrosion Index (PCI)
for lapped steel panels including all periods of corrosion s1, s2, s3 and s4 as follows:
Perforation Corrosion Index ðPCIÞ ¼ V Steel =V Coating ð1Þ
where, VSteel and VCoating (VZn) are defined as follows:
V Zn ¼ tcoating =ðs1 þ s2 þ s3 Þ ð2Þ
V Steel ¼ tsteel =s4 ð3Þ
tcoating and tsteel are thickness of coating and substrate steel. The corrosion-resistance
effect P of the coating layer relative to the substrate steel sheet is defined as follows:
P ¼ ðs1 þ s2 þ s3 Þ=s4 ð4Þ
The value of PCI for the side-sill in an actual environment (Detroit, North America)
was calculated from Fig. 4 as follows:
V Zn ¼ 120 ðg=m2 Þ=9 ðyearÞ ¼ 13:3 ðg=m2 =yearÞ ¼ 1:8 lm=year ð5Þ
V Steel ¼ 750 ðlmÞ=5 ðyearÞ ¼ 150 lm=year ð6Þ
PCI ¼ 82 ð7Þ
Thus the accelerated corrosion test conditions resulting in the PCI value close to 82
would be similar to the actual corrosion conditions for automobiles in Detroit.
Table 3 shows VZn, VSteel and PCI evaluated in each corrosion test. VZn was evaluated
using the data of pure zinc coating steel sheets (EG or GI) with heavy coating weight. PCI
of each corrosion test method is different to each other.
Fig. 7 shows the protective effect of coatings, P for GI100, GA45 and ZnNi30 as a func-
tion of PCI in addition to the data of corrosion in actual automobile environment for
automobile.

Table 3
VZn, Vsteel and PCI evaluated in accelerated corrosion tests and real environment
Test no VZn (lm/cycle) Vsteel (lm/cycle) PCI
T1 0.073 7.3 98
T2 0.47 7.7 16
T3 0.18 3.4 19
T4 0.23 6.6 29
T5 0.13 7.2 56
Actual automobile 1.8 (lm/year) 150 (lm/year) 82
218 S. Fujita, D. Mizuno / Corrosion Science 49 (2007) 211–219

Side sill of
T2 T3 T4 T5 Automobile
Automaker B JASO M609 SAE-5days SAE-7days in Detroit T1
α = 16 α = 19 α = 29 α = 56 α = 82 Automaker A
α = 98
2.0
1.8 Steel Thickness=0.7mm
GI100g/m 2
1.6
τ4

1.4
60g/m 2
1.2
P, (τ 1 + τ 2 +τ 3 )

50g/m 2
1.0
GA45g/m 2 40g/m 2
0.8
0.6 30g/m 2

0.4 ZnNi30g/m 2
0.2
0.0
0 10 20 30 40 50 60 70 80 90 100
PCI

Fig. 7. The protective effect of zinc and zinc alloy coatings, P for each test and an actual environment of
automobiles as a function of Perforation Corrosion Index, PCI.

P of zinc coating for GI100 was directly calculated from the relation P = PCI
(tcoating/tsteel) for tsteel = 0.7 mm and tcoating = 0.014 mm. For P of zinc alloy-coated steel
sheets of GA45 and ZiNi30 the same figures as Figs. 5 and 6 were obtained for tests
T2, T3 and T4. P values for GA45 and ZnNi30 were estimated using Eqs. (2)–(4) from
VZn and Vsteel given in Table 3 and a steel sheet thickness of 0.7 mm in addition to tcoating
of zinc corresponding to the weight of pure zinc estimated from the corrosion depth of
GA45 and ZnNi30 in figures including Figs. 5 and 6. Accordingly, the effect of zinc alloy
coatings is evaluated in terms of the weight of zinc coating. It is clear from the figure that
under inadequate experimental conditions such as low PCI, both zinc alloy coatings
behaved as quite a thick zinc coatings, whereas under the actual corrosion conditions
for automobiles, that is, PCI 82, the effective thickness of zinc for these alloy coatings
was far lower than thick zinc coatings such as GI100.
Consequently, the data for the individual coating under the given accelerated corrosion
test conditions should be plotted in the P–PCI relation, for estimation of the performance
of the coating in the actual corrosion conditions and for selection of the accelerated cor-
rosion test conditions close to the actual corrosion conditions of automobiles.
As it can be expected that the PCI will also differ depending on the position in the auto
body, this method can be expanded to the corrosion life prediction for individual parts of
the automobile. Moreover, the effect of new materials in the actual environment can be
evaluated by extrapolation/interpolation of PCI to actual automobiles, combining various
corrosion test methods.

5. Conclusions

(1) The main controlling factor affecting perforation corrosion on actual automobiles
was coating weight regardless of type of coating.
(2) The perforation corrosion process in Zn coated steel sheets was divided into four
stages.
 S. Fujita, D. Mizuno / Corrosion Science 49 (2007) 211–219 219

(3) Based on the analysis of corrosion in actual automobiles, Perforation Corrosion

Index, PCI for lapped steel panels was proposed.
(4) PCI for lapped portion in the side-sills of vehicles driven in North America was cal-
culated to be approximately 82.
(5) PCI is considered as the effective index for evaluation of the corrosion test conditions
in various accelerated corrosion test methods in relation to the actual environment
for automobiles.

References

[1] Peter Elliott, in: Corrosion/’91, NACE, Cincinnati, 1991, Paper No. 378.
[2] Staffan Hedlund, Automotive Corrosion and Prevention (ACAP), SAE, Dearborn, 1993, pp. 255–261.
[3] T.C.Simpson, A.W. Bryabt, G. Hook, R.J. Swinko, R.W. Miller, in: Corrosion/95, NACE, Orlando, 1995,
Paper No. 394.
[4] M. Nakazawa, Y. Miyoshi, D. Davidson, in: Corrosion/’95, NACE, Orlando, 1995, Paper No. 375.
[5] Bo. Rendahl, in: Corrosion/98, NACE, Paper No. 741, 1998.
[6] Sakae Fujita, Hiroshi Kajiyama, in: Corrosion/95, NACE, Paper No. 378, 1995.
[7] Sakae Fujita, Hiroshi Kajiyama, in: Corrosion/98, NACE, Paper No. 541, 1998.
[8] Sakae Fujita, in: Proc. of 4th International Conference on Zinc and Zinc Alloy Coated Steel Sheet
(Galvatech98), ISIJ, Chiba (1998) pp. 686–691.
[9] Sakae Fujita, Hiroshi Kajiyama, Zairyo to Kakyo 50 (2001) 115–123.
[10] E.J. Gumbel, Statistics of Extreames, Columbia Univ. Press, 1958;
M. Hirabayashi, H. Iwasaki, transl., Fundamentals of X-ray Crystallography (Translation), Tokyo,
Maruzen, 1978, pp. 529.
[11] X.G. Zhang, J. Hwang, W.K. Wu, in: Proc. 4th Int. Conf. on Zinc and Zinc Alloy Coated Steel Sheet
(Galvatech’ 98), ISIJ, Chiba, 1998, pp. 410–415.