Corrosion Science 74 (2013) 379386

Contents lists available at SciVerse ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

Effect of carbon dioxide on the atmospheric corrosion of ZnMgAl

coated steel
Nathalie LeBozec a,, Dominique Thierry a, Michael Rohwerder b, Dan Persson c, Gerald Luckeneder d,
Linda Luxem e
a
Institut de la Corrosion, 220 Rue Pierre Rivoalon, F-29200 Brest, France
b
Max-Planck-Institut fr Eisenforschung GmbH, Max-Planck-Str. 1, 40237 Dsseldorf, Germany
c
Swerea KIMAB, Isafjordsgatan 28A, 114 28 Stockholm, Sweden
d
voestalpine, voestalpine Strasse 3, 4021 Linz, Austria
e
Thyssen Krupp Steel Europe, Eberhardstrasse 12, 44145 Dortmund, Germany

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

Article history: The atmospheric corrosion of line hot dip ZnMgAl coating was investigated at low and ambient concen-
Received 29 March 2013 tration of CO2 as a function surface chloride concentration and temperature and compared to conven-
Accepted 15 May 2013 tional hot dip galvanised (GI) and Galfan coatings. The corrosion of zinc coatings was enhanced in low
Available online 24 May 2013
CO2 conditions and ZnMgAl material was more affected than GI, and in the range of the Galfan coating.
An obvious pH effect was underlined in low CO2 conditions. Layered double hydroxide (LDH) and simo-
Keywords: nkolleite were mainly formed on ZnMgAl coating in the absence of CO2 while hydroxycarbonate and
A. Metal coating
simonkolleite were dominating in ambient air.
A. Zinc
C. Atmospheric corrosion
2013 Elsevier Ltd. All rights reserved.
B. Weight loss
B. XRD

1. Introduction aluminium in the enhanced corrosion properties of ZnMg and

Hot dip galvanised steel (HDG) is widely used for applications in
the automotive, building and house hold appliance industry. This is
linked to the corrosion properties of HDG which provides a barrier
and a galvanic protection of the steel substrate. In the automotive
industry, there is a need to move to new coatings with improved
corrosion resistance in order to reduce the thickness of the coating
and to eliminate the need of additional corrosion protection in con-
ned areas such as hem anges.
In the last 23 decades, a large body of literature has been pub-
lished on further enhancements of the corrosion properties of HDG
by additions of magnesium, aluminium or a combination of both
[110]. New coating alloys containing aluminium and magnesium
have been developed for continuous galvanised steel sheet. These
coatings normally contain 0.211% of magnesium and 0.13% of
aluminium which are added to zinc. The addition of magnesium
and aluminium results in a complex microstructure containing pri-
mary zinc, MgZn2, binary eutectic MgZn2Zn, binary eutectic ZnAl
and a ternary eutectic depending upon the composition of the
coating. Despite the large body of literature on the subject, there
is still no clear explanation of the role of magnesium and
Corresponding author. Tel.: +33 298 051552; fax: +33 298 050894.
E-mail address: nathalie.lebozec@institut-corrosion.fr (N. LeBozec).
ZnMgAl coated steel compared to HDG. Several explanations
are available in the literature. The proposed mechanisms includes:
(1) buffering effect of dissolved Mg which buffers the pH at catho-
dic sites [4], (2) stabilization/formation of zinc hydroxy chloride on
the surface providing a more compact morphology [10], (3) forma-
tion of a layered double hydroxide of type Zn2Al(OH)6(CO3)1/2xH2O
or/and Mg6Al2(OH)16CO3 [7], and (4) barrier effect of a magnesium
rich oxide limiting the overall corrosion reaction [5]. It should also
be pointed out that several of these mechanisms can act together
and thus cannot be regarded as isolated. Besides, the exact mech-
anism or the combination of these mechanisms is probably largely
dependent upon exposure conditions as it was discussed else-
where [11].
Most of the conclusions obtained on the performance and the
mechanisms of corrosion of these new coatings have been obtained
after rather complex accelerated corrosion tests or/and natural
exposures. Under these conditions, a 220 improvement in corro-
sion resistance has been observed for ZnMgAl coated steel in
open bare conguration compared to hot-dip galvanised steel
[11]. However, in a recent paper, Le Bozec et al. have shown that
in conned areas simulating hem anges conguration in cars,
the improvement of corrosion resistance for ZnMgAl coated
steel was less obvious and largely dependent upon testing condi-
tions. These results were explained in terms of a drastic increase

0010-938X/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.corsci.2013.05.011
380 N. LeBozec et al. / Corrosion Science 74 (2013) 379386
of the local pH in the connement due to a possible limited supply
of CO2 in the conned zone [12].
The effect of CO2, in presence of NaCl, on the atmospheric cor-
rosion of zinc, aluminium and magnesium alloys have been previ-
ously studied [1316]. It has been shown that the absence of CO2
increased the corrosion rate of these materials by a factor of 34
for zinc [13,14] and 730 for aluminium and aluminium alloys
[15,16]. In the case of zinc, the decrease in the corrosion rate in
CO2 containing atmosphere was explained by the enhanced forma-
tion of simonkolleite resulting from a decrease conductivity as
chloride was removed from the electrolyte. On the other hand,
for aluminium, the explanation was linked to the neutralization
of the cathode in presence of CO2.
The aim of the present study was to study the effect of CO2 on
the corrosion performance of ZnMgAl coated steel and to com-
pare the results with those obtained for hot-dip galvanised and
Galfan (e.g. Zn5Al) coated steel. In addition, this study shall help
to provide better understanding of the atmospheric corrosion of
these coatings.
2. Materials and experimental methods
2.1. Materials
Line hot dip ZnMg2%Al2% (ZMA2) coated steel as well as con-
ventional hot-dip galvanised (GI) and Zn5%Al (Galfan) coated
steel were used in the present study. The composition of the differ-
ent coatings is given in Table 1. The expected coating thickness was
7.5 lm. All coatings were hot-dipped on steel (e.g. DX54D at a
thickness of 0.8 mm). Pure magnesium was also used as reference.
The samples had a geometry of 50  50 mm with a test surface of
approximately 16 cm2 (40  40 mm) due to the edge protection.
Before exposure, the samples were ultrasonically cleaned in hep-
tane, dried in air and stored in a desiccator over silica gel for
24 h before being weighed. Then, the edges were protected in order
to avoid any corrosion of steel at the cut-edge.
The microstructure of ZMA2 coating was identied on cross-
section using a Scanning Electronic Microscope (SEM) Zeiss Supra
35 (back-scattered electron mode) combined with a EDAX LEO
1530 allowing the identication of the different phases. In addi-
tion, Glow Discharge Optical Emission Spectroscopy (GDOES) was
carried out using a GDA750 equipment from Spectruma Analytik
(750 mm focal width, around 40 element channels and one optical
diffraction grating (2400 streaks/mm)) in order to assess the com-
position of the coating as a function of the depth.
2.2. Exposure conditions
NaCl was added by spraying a given volume of solution of meth-
anol containing NaCl to get three different surface concentrations
e.g. 100, 50, 20 lg/cm2 of chloride (Cl). These concentrations were
selected to able a comparison with the previous literature [13,14].
It should be pointed out that even the highest value of chloride
deposition does not correspond to extreme marine environments
as chloride deposition in the range or higher than 50 lg/cm2 per
day may be reported [17]. The surface concentration of NaCl was
Table 1
Material composition of line hot dip galvanised coatings.
Material Composition wt%
Zn Al Mg
ZnAl2Mg2 (ZMA2) 96 2 2
Hot dip galvanised steel (GI) 99.8 <0.2
Galfan (Galf) 95 5 7.5 lm
veried by weighing the samples after 24 h storage in a desiccator.
The distribution of salt on the samples was quite even. In each
experiment, ve parallel samples were exposed among which 3
were used for the metal loss calculation and 2 for the analyses of
corrosion products. In addition, two samples without chloride
were used as references.
The experimental set-up is schematically presented in Fig. 1. It
consists of a transparent cell of 39 L immersed in a water tank held
at a selected temperature e.g. 20 or 30 C. The samples were ex-
posed in horizontal position. The gas ow was 5 L/min. The relative
humidity was regulated by mixing dry air and humidied air to get
90 2% R.H. In the experiments without CO2, the air was puried
by letting it pass through lters before entering the system. A con-
centration of less than 3 ppm of CO2 was achieved. In the experi-
ments with CO2, a concentration of 350400 ppm of CO2 was
measured. In all cases, the exposure duration was limited to
14 days in order to avoid red rust formation as this generally intro-
duces large errors in the mass loss determination. Preliminary
experiments in CO2 free atmosphere with 100 lg/cm2 of chloride
have indeed shown that red rust was formed after 1516 days of
exposure on ZMA2.
2.3. Methods of evaluation
After completion of the exposure, the sample were dried into
the exposure cell for about 4 days by introducing dry air or CO2 free
air (R.H. <10%) before being weighed (dry mass gain). Water solu-
ble corrosion products and unreacted NaCl were rst removed by
leaching in pure water for 3 min. Then, the samples were pickled
in saturated glycine for up to 20 min in order to remove the corro-
sion products and the weight loss of the metallic coating was gravi-
metrically calculated. The amount of Cl in water or glycine was
determined by ion chromatography (detection limit below
12 mg/m2). On one sample, the surface pH was visually mapped
by placing a thin layer of agar containing a wide range of pH indi-
cator (WRI). Details on the preparation of the WRI agar may be
found elsewhere [18]. This allowed a visualisation of the anodic
and cathodic areas upon the material and exposure conditions.
2.4. Corrosion products analysis
In order to identify the corrosion products, Infrared Reection
Absorption Spectroscopy (IRRAS) measurements were performed
with a Varian FTS 7000 spectrometer. Additional measurements
were performed using X-ray diffraction with a Theta-2-Thetha dif-
fractometer from GE Seiffert, equipped with a Co-X-ray tube and a
Fe lter. The measurements were carried out in the grazing inci-
dence mode, where the incidence beam is xed to an incident an-
gle of 2 to ensure good layer to substrate sensitivity. The
TwoTHETA step size was 0.05 and count time was 20 s each step.
SEM and EDX were performed with a CamScan series 4 SEM.
3. Results
3.1. Initial microstructure of ZnMg2Al2
Fig. 2 shows a cross section of line hot-dip ZnMg2%Al2%
coated steel obtained by SEM. The microstructure consists in a zinc
phase containing small amount of aluminium, MgZn2, binary eu-
Thickness
tectic and ternary eutectic aluminium, MgZn2. This is in agreement
with the literature on the subject [6,7]. Fig. 3 shows the composi-
tional depth proles obtained by GDOES. The coating thickness is
7.5 lm in the range of 8 lm (e.g. close to the expected value). As further
7.5 lm
inferred from Fig. 3, Mg and to lesser extent Al are enriched on
the surface compared to the bulk values.
 N. LeBozec et al. / Corrosion Science 74 (2013) 379386 381

Dry purified air

Used for
CO2 adsorber CO2 free
experiments

Gas flow generator

CO2
Analyzer

Exposure chamber Mixer Humidifier

(heated)

Water tank at constant temperature

Fig. 1. Schematic drawing of the experimental setup.

Binary eutectic
Primary Zn
Elem Wt % At %
Elem Wt % At %
-----------------------------
---------------------------
MgK 11.55 25.70
MgK 0.00 0.00
AlK 0.96 1.92
AlK 0.83 1.99
ZnK 87.49 72.39
ZnK 99.17 98.01

Ternary eutectic
Elem Wt % At %
----------------------------
MgK 1.35 2.66 2 m
AlK 23.67 42.18
ZnK 74.99 55.16

Fig. 2. Microstructure of ZnMg2Al2.

100 10 different CO2 content in the air. Thus, in air with natural concentra-
90 Zn Fe 9 tion of CO2, hot-dip galvanised presented as expected a higher me-
tal loss than Galfan and ZMA2, where ZMA2 had the lowest mass
wt.-% Mg, Al and O

80 8
loss, in good agreement with previous studies conducted in similar
wt.-% Fe and Zn

70
60
50
40
Mg
30
20
Al
10
O inuence of CO2, the ratio of the metal loss for low CO2
0 0
0 2 4 6 8 10 12
Depth (m)
Fig. 3. Compositional depth proles of ZnMg2Al2 prior to exposure.
3.2. Kinetics of corrosion in ambient and CO2 free atmospheres
Fig. 4 presents the metal loss of the different metallic coatings
e.g. ZMA2, GI and Galfan after 2 weeks of exposure at 90% R.H. with
or without CO2 for different surface concentrations of chloride. The
results shown in Fig. 4 clearly indicate that carbon dioxide has an
important impact on the corrosion of zinc coatings. An enhance-
ment of the corrosion rate was observed in low CO2 conditions
independently on the coating composition. This is in good agree-
ment with the work of Lindstrm and co-workers who studied
the inuence of CO2 on pure zinc [14]. The metal loss increased
with the concentration of surface chloride. It is also interesting to
note that no major differences were observed for the different
metallic coatings irrespective of CO2 content and temperature,
when the samples were contaminated with a low amount of chlo-
ride e.g. 20 lg/cm2. However, for higher chloride concentrations,
the metal loss varied considerably for the metallic coatings with
7
6 conditions [6,7,11,12]. The situation was totally reversed when the
5 experiments were conducted in CO2 free atmosphere, ZMA2 coat-
4
ing showing a higher metal loss than conventional zinc coating.
These results were observed at both temperatures 20 and 30 C
3
with an obvious tendency to induce more metal loss when the
2
temperature was increased, as expected. In order to highlight the
1
14 16 (<3 ppm)/Air is given in Fig. 5 for the different metallic coatings
versus chloride concentration after 2 weeks of exposure at 90%
R.H. Thus, at low CO2 concentration, the metal loss was approxi-
mately 2 times higher on hot-dip galvanised steel whatever the
chloride concentration compared to ambient CO2 concentration.
It should be underlined that this ratio is higher when considering
ZMA2 and Galfan and is increasing with chloride concentration,
particularly for the ZMA2 coating where a metal loss 12 times
higher was measured at the highest concentration of chloride in
CO2 free atmosphere compared to air. On pure magnesium, a cor-
rosion rate of 8.5 and 20 lm were measured at 30 C and 90%
R.H. in presence of 100 lg/cm2 of chloride in ambient and CO2 free
environments, respectively. This indicates that on pure magne-
sium, the inuence of CO2 is in the same range as that measured
on pure zinc by Lindstrm et al. [14], but much less compared to
that reported for pure aluminium or on aluminium alloys [15].
As shown in Fig. 6, only a small fraction of water soluble chlo-
ride was found on hot-dip galvanised steel in CO2 containing air
indicating that most of the chloride was incorporated in the corro-
sion products as insoluble form, while in CO2 free atmosphere,
most of chloride was water soluble. Regarding ZMA2 and Galfan
coatings, it is interesting to note that the fraction of water soluble
chloride is signicantly more important when exposed to CO2
382 N. LeBozec et al. / Corrosion Science 74 (2013) 379386

1,5
4,5
GI GI
4,0 ZMA2
ZMA2
3,5 Galfan Galfan

Metal loss, m
Metal loss, m
3,0 1,0

2,5
2,0
1,5 0,5
1,0
0,5
0,0 0,0
0 20 40 60 80 100 120 0 20 40 60 80 10 0 120
[Cl-], g/cm [Cl-], g/cm

(a) (b)
3,0
GI
2,5 ZMA2
Galfan
Metal loss, m

2,0

1,5

1,0

0,5

0,0
0 20 40 60 80 10 0 120
[Cl-], g/cm
(c)
Fig. 4. Inuence of CO2 and temperature on the weight loss of metallic coated steel in open conguration exposed 14 days at 90% R.H. (a): Low CO2 30 C, (b): atmosphere
30 C, (c): low CO2 20 C.

14 totally removed (e.g. in the range of 3 ppm). Note that hydrozincite

GI 30C -90% RH was not detected in XRD (but detected by FTIR) presumably due to
Ratio metal loss low CO2 /Air

12 ZMA2 its amorphous structure. As further inferred from Table 2, ZMA2

10
Galfan coating revealed the presence of layered double hydroxides
(LDH) and simonkolleite as the main compounds at low CO2 con-
8 centration while some hydroxy carbonate and simonkolleite were
dominating in air. It should be noticed that in this case, neither
6
XRD nor FTIR techniques were able to detect LDH on ZMA2 after
4 14 days of exposure in air. The distribution of LDH forming at
low CO2 concentrations can be directly visualised by SEM/EDX
2
(see Fig. 7). As can be seen, under these conditions, large areas pre-
0 sumably cathodes are formed showing high aluminium and oxy-
0 20 40 60 80 100 120 gen contents, as expected for LDH compounds.
[Cl-], g/cm
3.4. Surface pH after exposure
Fig. 5. Ratio of the metal loss low CO2/Air as a function of chloride concentration.
30 C, 90% R.H.
Fig. 8 shows the surface pH measured using an agar containing
pH indicator that was applied on the different samples after the
containing air than in low CO2 conditions, while it was opposite on experiments. From Fig. 8, it is clear that in the absence of CO2,
hot-dip galvanised steel. This is in line with the low metal loss the pH became very alkaline (see the dark blue area on Fig. 8) for
measured on ZMA2 in air. all investigated metallic coatings. From this gure, it is also clear
that most of the exposed surface is at alkaline pH. This is in con-
trast to the pH measured in air with CO2, which remained in the
3.3. Nature of the corrosion products formed in ambient and CO2 free neutral region apart from some localised areas indicated by a blue
atmospheres colour in the picture.

Table 2 shows the corrosion products that were identied using 4. Discussion
FTIR and XRD technique. As observed in Table 2, a predominance of
zincite (ZnO) was noticed on hot-dip galvanised steel in the ab- 4.1. Atmospheric corrosion of ZMA2 in the presence of CO2
sence of CO2 while simonkolleite (Zn5Cl2(OH)8, H2O) and hydro-
zincite were the main compounds formed in CO2 containing The mechanisms of atmospheric corrosion of zinc and hot-dip
conditions. Some traces of carbonate containing compound were galvanised steel in chloride containing conditions have been stud-
also found in CO2 free atmosphere probably because CO2 was not ied in details by different authors [1314,1922]. The corrosion of
 N. LeBozec et al. / Corrosion Science 74 (2013) 379386 383

100 100
Chloride contamination: 50 g/cm Chloride contamination: 100 g/cm
90 90
Low CO2 Low CO2
80 80
Atm Atm

Water soluble chloride, wt%

Water soluble chloride, wt%

70 70

60 60

50 50

40 40

30 30

20 20

10 10

0 0
ZMA2 GI GALF ZMA2 GI GALF

Fig. 6. Water soluble chloride in corrosion products formed on zinc coated steel after 2 weeks if exposure at 30 C and 90% R.H. Inuence of CO2 and chloride concentration.

Table 2 CO2 g 2OH $ CO3 H2 O 4

Summary of major compounds identied by XRD and FTIR spectroscopy in the
corrosion products formed on GI and ZMA2 upon CO2 level and chloride concentra-
tion. The main corrosion products are in bold.
CO2 aq H2 O ! HCO3 H 5

Material Exposure Chloride Detected phases

conditions concentration
HCO3 $ CO2
3 H

6
(lg/cm2)
The formation of zinc hydroxycarbonates at the cathodic sites is
ZMA2 Low CO2 20 Zn0.61Al0.39(OH)2(CO3)0.195H2O possible according to reaction (7):
(LDH)
Zn5(OH)8Cl2H2O 5Zn2 2CO2 
3 6OH $ Zn5 OH6 CO3 2 7
ZnO
100 Zn0.61Al0.39(OH)2(CO3)0.195H2O It should also be noticed that even simonkolleite may undergo
(LDH) some transformations into hydrozincite during the carbonatization
Zn5(OH)8Cl2H2O
ZnO
of the surface according to reaction (8):
Ambient 20 Zn5(OH)8Cl2H2O 
Zn5 OH8 Cl2 H2 O 2HCO3 $ Zn5 OH6 CO3 2 3H2 O 2Cl 8
Zn5(OH)6(CO3)2
100 Zn5(OH)8Cl2H2O The results obtained in this work for hot-dip galvanised steel
Zn5(OH)6(CO3)2
are in good line with the literature on the subject [13,14]. Indeed,
GI Low CO2 20 Zn5(OH)8Cl2H2O
ZnO simonkolleite and hydrozincite were found as the main corrosion
100 Zn5(OH)8Cl2H2O products, independently on the applied chloride concentration.
ZnO As shown in Fig. 8, the surface pH remained rather neutral promot-
Ambient 20 Zn5(OH)8Cl2H2O
ing the formation of these corrosion products. In the case of ZMA2,
Zn5(OH)6(CO3)2
100 Zn5(OH)8Cl2H2O similar corrosion products were found as on GI. It should be no-
Zn5(OH)6(CO3)2 ticed that under these exposure conditions, the relative perfor-
mance of ZMA2 compared to GI steel was largely dependent
upon the surface chloride concentration (e.g. better performance
at higher chloride concentrations). This is in good agreement with
zinc starts with the oxidation of zinc at anodic sites. The hydrolysis the literature on the subject showing that the relative performance
of the Zn cations may locally lead to a slight decrease in pH locally of zinc magnesium aluminium coated steel is generally superior to
according to Eq. (1) which explains the formation of ZnO. Simo- HDG in harsh environments [11,12]. It is generally accepted that
nkolleite is thus formed at anodic sites due to migration of chloride Mg and Zn are dissolved selectively from the coating whereas the
to anodic sites according to Eqs. (2) and (3): dissolution of aluminium is linked to an increase of the local pH
at cathodic sites. In a recent paper, Persson et al. have shown that
Zn2 H2 O ! ZnO 2H 1 LDH are formed on ZMA2 coating within a few hours of exposure
under rather similar exposure conditions, indicating under these

5ZnO 2Cl 6H2 O $ Zn5 OH8 Cl2 H2 O 2OH 2 exposure conditions that the local pH increases sufciently to dis-
solve Aluminium [23]. This is in contrast to the results reported by
 Volovitch et al. who indicate two steps mechanisms of corrosion of
5Zn2 2Cl 9H2 O ! Zn5 OH8 Cl2 H2 O 8H 3
Zn24%Mg24%Al coating with the formation of LDH at the later
Due to the presence of CO2 in air (typically 350400 ppm), car- stage [10]. It should be noticed that a direct comparison of the
bonates are often reported as corrosion products of zinc and hot- work of Persson et al. [23], Volovitch et al. [10] and the present
dip galvanised steel. The formation of carbonates depends on the work is rather difcult as exposure conditions and duration are dif-
adsorption of carbon dioxide into the surface electrolyte and the ferent as well the composition of ZnMgAl coating in [10]. As seen
formation of bicarbonate and carbonate according to reactions in Fig. 2, Al was enriched in the surface layer of ZMA2. It is thus
(4)(6) resulting in a decrease of the surface pH. possible that at the very initial stage as reported by Persson
384 N. LeBozec et al. / Corrosion Science 74 (2013) 379386

Fig. 7. SEM and EDX map of the surface of a ZMA2 sample contaminated with 100 lg/cm2 Cl, exposed to low CO2 atmosphere.

GI ZMA2 Galfan
100 g/cm Cl-
Air - 30C- 90%
Low CO2 20C - 90% 100 g/cm Cl-

Fig. 8. Photograph of surface pH from WRI containing agar after exposure on corroded GI, ZMA2 and Galfan metallic coating in top: Air at 30 C/90% R.H. and bottom: CO2 free
at 20 C/90% R.H. 100 lg/cm2 Cl. Dark blue = pH 10 or more; green: pH 7; red: pH 4.

et al., more aluminium is dissolved favouring the formation of LDH.
At later stage, the corrosion rate seems to be more governed by the
dissolution of zinc with the formation of zinc corrosion products as
observed in this work.
On the other hand in very harsh conditions such as the VDA cor-
rosion test or the neutral salt spray, LDH seems to be formed prob-
ably due to an increase of the local pH leading to a higher
dissolution of the surrounding aluminium phase [6,11]. Thus, from
this work as well as other work on the subject, it seems that the
nature of the corrosion products formed on ZMA2 depends largely
upon exposure conditions and are most likely only the reect of
the local pH generated at the cathodic sites. As presented in
Fig. 4, the corrosion rate of ZMA2 in CO2 containing atmosphere
was low independently on the surface chloride concentration. On
the contrary, a strong dependency of chloride surface concentra-
tion was seen for GI coated steel. Moreover, as inferred from
Fig. 6, most of the chloride is bonded in the corrosion products
for GI steel whereas it is unreacted in the case of ZMA2 coated
 N. LeBozec et al. / Corrosion Science 74 (2013) 379386
steel. It is likely that this reects an increasing surface pH on HDG
steel compared to ZMA2 when increasing the surface chloride
concentration.
4.2. Atmospheric corrosion of ZMA2 in the absence of CO2
In the absence of CO2, the corrosion rate of GI steel increased by
a factor of 23 depending upon the surface chloride concentration.
This is in good agreement with the data of Lindstrm et al. ob-
tained on pure zinc [14]. It should be noticed that a similar ratio
(e.g. 23) was also found for pure magnesium in the present study.
The presence of up to 4045% water soluble chloride without CO2
compared to 520% with CO2 seems to indicate a decrease in the
amount of simonkolleite formed and that the surface conductivity
was high in low CO2 environments. It should be noted that in CO2
free environments, the corrosion became more localised (see Fig. 7)
with large cathodic areas. As summarized in Table 2, zincite was
the dominating corrosion products for GI steel in CO2 free environ-
ments indicating a high pH on the surface due to an increase catho-
dic activity (as also observed in Fig. 8). The results obtained with
ZMA2 coating in CO2 free environments clearly indicate that the
performance of this coating was inferior to GI steel under these
conditions. It should also be noted that the corrosion rate of
ZMA2 was close to that of Galfan at both 20 and 30 C and indepen-
dent on the surface chloride concentration. This seems to indicate
that the poor performance of ZMA2 in this environment is linked to
its aluminium content. Indeed, as observed by Bengtsson et al., alu-
minium and aluminium alloys show a rapid corrosion in CO2 free
environments due to an increased pH at the cathodic sites
[15,16]. The increased rate of dissolution of aluminium in the case
of ZMA2 coating in CO2 free atmospheres was conrmed by the
presence of large amounts of LDH in the corrosion products. This
together with the presence of zincite seems to indicate that the
pH was very alkali under these conditions. This has been conrmed
by the surface pH measurements as shown in Fig. 8. Hence, it is
possible that under these conditions, aluminium is dissolved due
to an increase pH at the cathodic areas. In a recent paper by Pers-
son et al., a mechanism of corrosion of ZMA2 was proposed for the
rst stages in chloride containing atmospheres [23]. In that situa-
tion, the corrosion product formation is a result of the anodic dis-
solution of ZnMgZn2 phases and cathodic oxygen reduction
taking place on the Zn phase. It is proposed here that the high
pH at the cathodes also leads to signicantly higher pH at the
boundaries where Al-rich phases are located which induces a rapid
dissolution of the aluminium oxide. It cannot be excluded that the
cathodic reaction of oxygen reduction also takes place on the
aluminium rich phase. This has led to a poor performance of
ZMA2 under these conditions. A similar mechanism is probably
valid for Zn5%Al coating as this material also shows very poor
performance in CO2 free atmospheres.
The present results are important for the application of zinc
magnesium aluminium coated steel as situations where CO2 can
be depleted, may be encountered in practical applications such
as in hem-anges. This is the subject for further investigations in
our laboratory.
5. Conclusions
The effect of CO2 on the atmospheric corrosion of line hot dip
ZnMgAl coatings was investigated under carefully controlled con-
ditions at ambient concentration of CO2 and low CO2 content as a
function of chloride concentration. Two temperatures 20 and
30 C were tested and two conventional zinc coatings e.g. GI and
Galfan were also included in the study as referenced materials.
385
From this investigation, the absence of CO2 clearly enhanced the
corrosion of all zinc coatings and the metal loss increased with the
concentration of chloride and the temperature. Interestingly, ZMA
coatings were noticeably more affected than conventional GI coat-
ing. While GI was approximately twice more corroded in the ab-
sence of CO2 compared to ambient concentration, ZMA was 12
times more affected when exposed to low CO2 conditions, particu-
larly at high concentration of chloride. An obvious pH effect in the
absence of CO2 was underlined. The analysis of corrosion products
by FTIR and XRD indicated the presence of layered double hydrox-
ide (LDH) and simonkolleite on ZMA2 coating in the absence of CO2
while hydroxycarbonate and simonkolleite were dominating in
ambient air, just as for GI. These ndings indicated that the good
corrosion performance of zinc alloyed coating cannot be explained
strictly by the formation of certain corrosion products. Rather, the
detected compounds are reecting the actual corrosion conditions
on the sample.
Acknowledgements
The nancial support of European Commission through Re-
search Fund for Coal and Steel (Contract RFSR-CT-2009-00014) is
acknowledged. Ari Peltola (Ruukki Metals) and Giancarlo Marchi-
aro (Centro Ricerche Fiat) are also acknowledged for valuable
discussion.
References
[1] S. Tanaka, K. Honda, A. Takahashi, Y. Morimoto, M. Kurosaki, H. Shindo, K.
Nishimura, M. Sugiyama, The performance of ZnMgAlSi hot-dip galvanised
steel sheet, in: Proceedings of the Galvatech 01, International Conference on
Zinc and Zinc Alloy Coated Steel, June 2628, Brussels, Belgium, 2001, pp. 153
160.
[2] T. Tsujimura, A. Komatsu, A. Andoh, Inuence of Mg content in coating layer
and coating structure on corrosion resistance of hot-dip ZnMgAl alloy
coated steel sheet, in: Proceedings of the Galvatech 01, International
Conference on Zinc and Zinc Alloy Coated Steel, June 2628, Brussels,
Belgium, 2001, pp. 145152.
[3] T. Koll, K. Ullrich, J. Faderl, J. Hagler, A. Spalek, Properties and potential
applications of ZnMg-alloy-coatings on steel sheet by PVD, in: Proceedings of
Galvatech 04, Association for Iron & Steel Technology, Chicago, IL, United
States, 2004, p. 803.
[4] C. Hosking, M.A. Strm, P.H. Shipway, C.D. Rudd, Corrosion resistance of zinc
magnesium coated steel, Corros. Sci. 49 (2007) 36693695.
[5] T. Prosek, A. Nazarov, U. Bexell, D. Thierry, J. Serak, Corrosion mechanism of
model zincmagnesium alloys in atmospheric conditions, Corros. Sci. 1 (2008)
22162231.
[6] S. Schuerz, G.H. Luckeneder, M. Fleischanderl, P. Mack, H. Gsaller, A.C. Kneissl,
G. Mori, Chemistry of corrosion products on ZnMgAl alloy coated steel,
Corros. Sci. 52 (2010) 32713279.
[7] S. Schuerz, M. Fleischanderl, G.H. Luckeneder, K. Preis, T. Haunschmied, G.
Mori, A.C. Kneissl, Corrosion behaviour of ZnMgAl coated steel sheet in
sodium chloride-containing environment, Corros. Sci. 51 (2009) 23552363.
[8] P. Volovitch, C. Allely, K. Ogle, Understanding corrosion via corrosion product
characterization: I. Case study of the role of Mg alloying in ZnMg coating on
steel, Corros. Sci. 51 (2009) 12511262.
[9] T. Prosek, N. Larche, M. Vlot, F. Goodwin, D. Thierry, Corrosion performance of
ZnMgAl coatings in open and conned zones in conditions simulating
automotive applications, Mater. Corros. 61 (2010) 412421.
[10] P. Volovitch, T.N. Vu, C. Allely, A. Abdel, K. Ogle, Understanding corrosion via
corrosion product characterization: II. Role of alloying elements in improving
the corrosion resistance of ZnMgAl coatings on steel, Corros. Sci. 53 (2011)
24372445.
[11] D. Thierry, T. Prosek, N. Le Bozec, E. Diler, Corrosion protection and corrosion
mechanisms of continuous galvanised steel sheet with focus on new coating
alloys, in: Proceedings of Galvatech11, International Conference on Zinc and
Zinc Alloy Coated Steel, June 2124, Genova, Italy, 2011.
[12] N. Le Bozec, D. Thierry, A. Peltola, L. Luxem, G. Luckeneder, G. Marchiaro, M.
Rohwerder, Corrosion performance of ZnMgAl coated steel in accelerated
corrosion tests used in the automotive industry and eld exposures, Mater.
Corros. 64 (2013).
[13] T. Falk, J.E. Svensson, L.G. Johansson, The inuence of CO2 and NaCl on the
atmospheric corrosion of zinc: a laboratory study, J. Electrochem. Soc. 145
(1998) 29932999.
[14] R. Lindstrm, J.E. Svensson, L.G. Johansson, The atmospheric corrosion of zinc
in the presence of NaCl: the inuence of carbon dioxide and temperature, J.
Electrochem. Soc. 147 (2000) 17511757.
386 N. LeBozec et al. / Corrosion Science 74 (2013) 379386

[15] D. Bengtsson Blcher, R. Lindstrm, J.E. Svensson, L.G. Johansson, The effect of [20] I. Odenwall, M. Westdahl, Zinc chlorohydroxosulfates: newly-discovered
CO2 on the NaCl-induced atmospheric corrosion of aluminium, J. Electrochem.; corrosion products on zinc. Structure determination of
Soc. 148 (2001) B127B131. NaZn4Cl(OH)6SO46H2O and X-ray study of Zn4Cl2(OH)4SO45H2O, Corros. Sci.
[16] D. Bengtsson Blcher, J.E. Svensson, L.G. Johansson, Carbon dioxide: the 34 (1993) 12311242.
unknown factor in the atmospheric corrosion of aluminum?, in: Proceedings of [21] I. Odnevall, C. Leygraph, Formation of NaZn4Cl(OH)6SO46H2O in a marine
the 15th ICC Conference, September 2227, Granada, Spain, 2002, pp. 523 atmosphere, Corros. Sci. 34 (1993) 12131229.
530. [22] D. Thierry, N. Le Bozec, Corrosion products formed on conned hot dip
[17] N. LeBozec, J.P. Pautasso, P. Le Calv, Assessment of the aggressiveness and galvanised steel in accelerated cyclic corrosion test, Corrosion 65 (2009) 718
environmental conditions on ships, in: Proceedings of Eurocorr 2009, 725.
September 610, Nice, France, 2009, paper 7745. [23] D. Persson, T. Prosek, N. Le Bozec, D. Thierry, In situ infrared reection
[18] H.S. Isaacs, G. Adzic, C.S. Jeffcoate, WR Whitney Award Lecture: Visualizing spectroscopy studies of the initial atmospheric corrosion of ZnMgAl coated
Corrosion, Corrosion 56 (2000) 971978. steel, Corros. Sci. 72 (2013) 5463.
[19] T.E. Graedel, Corrosion mechanisms for zinc exposed to the atmosphere, J.
Electrochem. Soc. 136 (1989) 193C203C.