Journal of Applied Electrochemistry (2006) 36:375–383  Springer 2005

DOI 10.1007/s10800-005-9083-x

Evaluation of the corrosion behavior of galvannealed steel in chloride aqueous

solution and in tropical marine environment

PAULO S. G. DA SILVA1, ALBERTO N. C. COSTA2, OSCAR R. MATTOS3, ADRIANA N. CORREIA1 and

P. DE LIMA-NETO1,*
1
Departamento de Quı´mica Analı´tica e Fı´sico-Quı´mica, Universidade Federal do Ceará, C.P. 6035, 60455-970,
Fortaleza, CE, Brazil
2
Gereˆncia de Pesquisa e Desenvolvimento, CSN, Volta Redonda, RJ, Brazil
3
Laboratório de Corrosão Prof. Manoel de Castro, PEMM/COOPE/UFRJ, Rio de Janeiro, RJ, Brazil
(*author for correspondence: e-mail: pln@ufc.br)

Received 12 January 2005; accepted in revised form 28 September 2005

Key words: corrosion, electrochemistry, galvanized steel, galvannealed steel, tropical marine environment

Abstract

An investigation was made into the corrosion behavior of commercial galvannealed steel in 10)2 mol dm)3 NaCl
aqueous solutions and in a tropical marine environment, using scanning electron microscopy (SEM), galvanostatic
electrochemical stripping (GES), potentiodynamic linear polarization (PLP), and electrochemical impedance
spectroscopy (EIS) techniques and open circuit measurements (Eoc). For purposes of comparison, a commercial
galvanized steel was also subjected to similar corrosion tests. GES and SEM techniques allowed for the identifi-
cation of f, d and G intermetallic phases and revealed cracks in the galvannealed steel. The PLP, EIS and Eoc results
indicated that the galvannealed coating was more corrosion resistant than galvanized coating in an aqueous
medium, but that their corrosion behaviors were similar in the marine environment. The corrosion behavior of the
galvannealed steel was affected by the evolution of the cracking process in the Zn–Fe layer due to the dissolution of
zinc-rich phases, while the galvanized steel displayed generalized corrosion in the aqueous medium and localized
corrosion in the marine environment.

1. Introduction [12], who studied the corrosion of galvannealed steel

exposed to an industrial environment for a year,
It is well known that galvannealed (Zn–Fe) coatings concluded that the corrosion process was associated
possess higher corrosion resistance and better painting with the presence of depressions in the external coating
and welding properties than galvanized (Zn) coatings. surface related to the presence of n and d Zn–Fe
Galvannealed coatings, which are produced by rapid intermetallic phases. However, the literature offers little
annealing of galvanized steel at temperatures ranging data about the corrosion behavior of these coatings in
from 490
C to 540
C, consist of several zinc–iron natural environments, particularly in tropical marine
intermetallic phases: the f, the d and the G phases [1–3]. environments.
Due to its superior properties, galvannealed steel has been Therefore, this study investigated the corrosion
increasingly used in the automotive industry and interest behavior of unpainted galvannealed coatings undergo-
in understanding its corrosion behavior has grown [4–12]. ing long-term exposure to both aerated NaCl aqueous
Although galvannealed steel is applied on an indus- solutions and to Brazil’s northeastern tropical marine
trial scale, reports in the literature regarding its corro- environment, hoping to shed light on the corrosion
sion behavior are mainly based on laboratory ageing mechanism of Zn–Fe coating. For purposes of compar-
tests. Samples of galvanized and galvannealed coatings ison a galvanized coating was also studied under the
in alkaline mediums have revealed that the products of same conditions. Environmental corrosion tests were
corrosion are modified by the chloride content in conducted in Fortaleza, a tourist and industrial city
solution. Moreover, the passivation process for both located on Brazil’s northeastern coast. Fortaleza’s
zinc-rich n and d phases is diffusion-controlled over the environmental conditions are extremely aggressive to
range of pH 9–14 [6]. The electrochemical stripping metals and alloys, with an average annual ambient
technique has been used to characterize the Zn–Fe temperature of about 30
C, a summer-like climate
intermetallic phases present in galvannealed coatings throughout the year, and constant winds blowing inland
[7–11] and in synthesized alloys [9]. Fabri-Miranda et al. off the ocean.
376
2. Experimental details sample holder exposed about 55 cm2 in the EIS mea-
surements. Pt foil and saturated calomel (SCE) were
2.1. Material used, respectively, as auxiliary and reference electrodes.
The GES experiments were galvanostatically con-
Samples of galvannealed and galvanized steel sheets trolled at 2 mA cm)2 using 3.4 mol dm)3 NaCl and
were supplied by Companhia Siderúrgica Nacional 0.6 mol dm)3 ZnSO4 as electrolyte. The evolution of the
(CSN), Volta Redonda, RJ, Brazil. The iron content EIS diagrams with immersion time (long-term laboratory
in the galvannealed coating, which was determined by tests) for the samples immersed in the 10)2 mol dm)3
localized EDAX analyses at seven different points of the NaCl aqueous solution was assessed at Eoc. The EIS
Zn–Fe surface, presented an average value of 12.5 at% experiments were conducted in a frequency range of
(±1.1 at%). 5 mHz to 10 kHz and an amplitude perturbation of
5 mV. All the working solutions were prepared with
2.2. Physical characterization Milli-Q water and analytical grade chemical reagents. A
potentiostat/galvanostat AUTOLAB PGSTAT 30
The microstructure of the as-received and tested samples linked to a PC microcomputer and controlled by GPES
was analyzed by scanning electron microscopy (SEM) and FRA software was used to acquire and analyze the
using a Philips XL-30 microscope. Samples for SEM electrochemical data.
cross-section analysis were previously embedded in
epoxy resin, polished with SiC sandpaper up to 600
grit, then with alumina (/=0.3 lm) or diamond paste
(/=0.3 lm), and finally subjected to chemical etching in
3. Results and discussion
a solution of 5% HNO3 in ethanol. Chemical analyses
of the samples before and after the corrosion tests were
3.1. Electrochemical and morphological characterizations
conducted using a Link Analytical QX-2000 energy
dispersive X-ray analyzer (EDAX) attached to the SEM
Figure 1 shows the potential-time curves obtained from
apparatus.
galvanostatic stripping of the galvannealed and galva-
nized coatings. The galvannealed coating consistently
2.3. Corrosion tests
showed more positive potential values than did the
galvanized coating due to its content of iron, an element
The marine environment corrosion tests involved the
with a nobler thermodynamic standard electrochemical
3-year exposure of galvannealed and galvanized steel
potential. While the curve corresponding to the galva-
samples, according to the ASTM G-50 Standard [13].
nized coating presented a single and well defined plateau
The deterioration of exposed samples was assessed using
at 1.0 V, the curve for the galvannealed coating pre-
a method established in our group [12, 14]. This method
sented three potential plateaus, two of them well defined
consists of periodically removing samples from the
at )0.91 V and )0.75 V and a short broad one at
corrosive environment to which they are exposed and
)0.82 V. These results are in close agreement with those
subjecting them to laboratory analyses by electrochem-
previously reported [5–11]. Based on these studies, the
ical impedance spectroscopy (EIS) at open circuit
plateau observed at )1.0 V can be attributed to the g
potential (Eoc), after which they are returned to the
phase, corresponding to the Zn layers, while those
corrosive environment. The electrolyte was a
observed at )0.91 V, )0.82 V and )0.75 V can be
10)2 mol dm)3 NaCl aqueous solution. The samples
ascribed, respectively, to the n, d, and G Zn–Fe
were analyzed in the conditions of exposure, with the
intermetallic phases. The short potential plateau
products of corrosion on their surfaces. This is a non-
destructive test that does not influence the subsequent
corrosion behavior. All the corrosion tests performed in
the laboratory were conducted at room temperature
(25±1
C).
Mass loss tests were performed by immersing the
samples in a NaCl solution for 20 days, followed by
chemical attack in a 10% glycine solution for 10 min,
according to the method proposed by Beccaria [15]. The
electrochemical laboratory tests involved the following
electrochemical techniques: galvanostatic electrochemi-
cal stripping (GES), potentiodynamic linear polarization
(PLP), EIS and Eoc measurements. All the electrochem-
ical experiments were conducted on duplicate samples.
Fig. 1. Potential-time curves obtained from galvanostatic stripping
The electrochemical cell consisted of a PVC tube glued to of the galvannealed (a) and galvanized (b) coatings in 3.4 mol dm)3
the surface, exposing an area of approximately 11 cm2 NaCl and 0.6 mol dm)3 ZnSO4 at 2 mA cm)2 and room tempera-
during the GES and PLP tests, while a cell with a Teflon ture.
 377

Fig. 2. SEM micrographs showing the surface morphology of the as-received galvanized (a, b) and galvannealed (c, d) coatings.

observed at )0.52 V corresponds to the dissolution of The measurements taken at different points of the
Fe from the steel substrate. coating indicated that the average thickness of the Zn
Figure 2 shows a SEM image of the surface morphol- layer was 15.36±0.5 lm. SEM analyses of the tested
ogy of as-received galvanized and galvannealed steels. galvanized samples revealed that the Zn coating had
The galvanized steel displayed a homogeneous surface dissolved completely by the end of the immersion tests in
devoid of surface cracks (Figure 2a). A detailed SEM 10)2 mol dm)3 NaCl, indicating a generalized corrosion
analysis (Figure 2b) revealed a few superficial defects process. The SEM micrograph of the samples exposed
which probably resulted from the industrial hot dip to the tropical marine environment for 14 months
galvanizing process [1]. On the other hand, the as- (Figure 3b) showed the occurrence of localized dissolu-
received galvannealed steel surface was granular, with tion of the layer, which was also observed in samples
hairline cracks (Figure 2c). A detailed SEM analysis exposed for 36 months. The localized corrosion of the
(Figure 2d) enabled us to identify hexagonal and mono- exposed galvanized steel tested here was congruent with
clinic crystals associated, respectively, with d (FeZn7) the behavior observed by ref. [15] in Zn electrocoatings
and n (FeZn13) intermetallic phases [7, 12]. The presence also exposed to a marine environment.
of hairline cracks is typical of other Zn–alloy electro- Figure 4a shows the as-received galvannealed layer,
coatings, such as Zn–Ni electrodeposits [14], and may be which displayed cracks from the surface down to the
related to the internal stress of the coating that develops substrate. A comparison of the cross-section in this
as the coating is formed. Cracking in Zn–Fe coatings is micrograph with those presented in the literature [12, 16]
caused by heat-treating galvanized steel [12]. The SEM confirmed the presence of the d and G phases. The
micrograph of the as-received galvannealed steel does thinner G phase had a higher Fe content (FeZn4) and
not show depressions on the surface coating, such as was located at the substrate/coating interface. This
those observed by Fabri-Miranda et al. in galvannealed phase was considered responsible for the lack of
coating samples produced on an industrial continuous adherence between the Zn–Fe coating and the steel
galvanizing line [12]. during the stamping process. After the G phase, the
Figures 3 and 4 show the metallographic cross-section thicker d phase (FeZn7) with a lower Fe content was
of the as-received and tested galvanized and galvan- found to prevail in the bulk of the coating. This phase
nealed samples. The micrograph in Figure 3a reveals a is considered the optimum phase for the stamping
homogeneous and compact as-received galvanized layer. process [17]. Additionally, our SEM analyses of the
378

Fig. 3. SEM micrographs showing the cross section of the as-received galvanized steel (a) and as-exposed in a marine atmosphere during
14 months (b).

cross-section indicated that this layer had a homoge- marine environment, as indicated in Figure 4c. The Fe
neous thickness of 10.5±0.2 lm. content in the grain remained about 12.5 at%, while this
At the end of the immersion tests in aerated content in the cracked region was 94.2 at%.
10)2 mol dm)3 NaCl aqueous solution and after expo- Our GPES, SEM and EDAX analyses showed that
sure to the natural tropical marine environment, the the morphological evolution of the cracks was associ-
cross-sections of the galvannealed samples showed a ated with the dissolution of the phases’ zinc-rich
similar morphological evolution toward a more cracked contents. The literature contains reports of similar
layer (Figure 4b–d). Our SEM analysis revealed cracks results on the evolution of cracking of the Zn–Ni layer
distributed throughout the coating. Localized corrosion during the process of corrosion [14].
was also observed in samples exposed for 36 months
(Figure 4e). These analyses indicate that the corrosion 3.2. Corrosion tests
process in the galvannealed coating was similar in both
aqueous solution and marine environment and that it After 20 days of immersion in 10)2 mol dm)3 NaCl
led to and aggravated the occurrence of cracking. The aqueous solution, the mass loss of the galvannealed and
appearance of cracks at the d/G interface, which were galvanized samples was 124 mg cm)2 d)1 and 184 mg
absent from the as-received samples, were attributed to cm)2 d)1, respectively, confirming that, in an aqueous
galvanic coupling between these phases. These SEM medium, galvannealed steel is more corrosion resistant
analyses are in close agreement with those of Fabri- than galvanized steel.
Miranda et al. [12], who studied the corrosion of The potentiodynamic polarization curves of both
galvannealed steel in an industrial environment over a materials in aerated 10)2 mol dm)3 NaCl aqueous
1-year period. solution were plotted in order to assess their cathodic
Localized EDAX analyses were carried out on the and anodic behaviors, which are shown in Figure 5. The
grain and in the cracked region of the tested galvan- corrosion potentials derived from these curves were
nealed samples after 14 months of exposure to the )1.071 V for galvanized and )0.969 V for galvannealed
 379

Fig. 4. SEM micrographs showing the cross section of the as-received galvannealed steel (a), as-tested galvannealed steel after the end of
immersion tests in 10)2 mol dm)3 NaCl aqueous solution (b), as-exposed in a marine atmosphere during 14 months (c) and 36 months (d, e).

Fig. 5. Potentiodynamic polarization curves obtained for galvan- Fig. 6. Evolution of the open circuit potential of the galvannealed
nealed steel and galvanized steel in aerated 10)2 mol dm)3 NaCl (n) and galvanized (m) samples with time for samples immersed in
aqueous solution at 1 mV s)1 and at room temperature. 10)2 mol dm)3 NaCl aqueous solution at room temperature.
380

Fig. 7. Evolution of the impedance diagrams with immersion time for galvanized steel samples immersed in 10)2 mol dm)3 NaCl.

steel, while the polarization resistance indicated by these temperature. The end of the process was determined
curves was 0.73±0.08 kW cm)2 for galvannealed steel when the surfaces were covered with white corrosion
and 0.62±0.15 kW cm)2 for galvanized steel. These products and isolated red spots.
results indicate that Zn–Fe coating is nobler and more This figure confirms that the Eoc values obtained from
resistant to generalized corrosion than Zn coating in an the long-term immersion tests are congruous with those
aqueous medium. obtained in the GES experiments. The Eoc correspond-
On the cathodic side, the Tafel slope for the oxygen ing to the galvanized coating remained approximately
reduction reaction (¶E/¶logi) was about 0.1 V decade)1 constant at 1.0 V for 3 days, followed by a gradual
for both materials, which is highly congruent with that increase up the end of the immersion test. The galvan-
obtained for bare iron [18]. This finding suggests that the nealed coating displayed an initial decrease in the early
cathodic reaction mechanism is the same for both layers. hours of immersion, followed by a gradual increase until
The presence of iron in the galvannealed coating inhib- the end of the test. The galvannealed sample presented
ited the cathodic reaction, since the cathodic current higher Eoc values than the galvanized layer, but lower
densities of galvannealed steel are lower. Additionally, than the Eoc of the steel substrate, indicating that the
the anodic branch of the polarization curves indicated Zn–Fe coating acted as a sacrificial anode protecting the
that these materials showed no passivation in the steel substrate, but that its tendency to corrode was
chloride medium, with the galvannealed steel showing lower than that of the Zn coating. The initial decrease in
lower anodic current densities than the galvanized steel. the Eoc of the galvannealed steel is related to the
The monitoring Eoc is a sensitive measurement to dissolution of the residual Zn in the coating. The
detect the various stages of dissolution of a coating in increasing Eoc with immersion time is explained by the
long-term immersion tests. Figure 6 shows the evolution fact that the phases with zinc-rich contents dissolve
of the Eoc with immersion time for the two materials preferentially, which gradually leads to alloy ennoble-
immersed in aerated 10)2 mol dm)3 NaCl at room ment. The dependence of Eoc on immersion time
 381

Fig. 8. Evolution of the impedance diagrams with immersion time for galvannealed steel samples immersed in 10)2 mol dm)3 NaCl.

observed in the Zn–Fe coating is in close agreement with times the impedance diagrams evolve to a single
that reported by Baldwin et al. for Zn–Ni electrocoa- capacitive loop.
tings [20]. Moreover, Besseyrias et al. [8, 9] pointed out Figure 8 shows the evolution of the galvannealed steel
that the similarity between the Eoc and GES results for impedance diagrams as a function of immersion time.
Zn–Fe coating confirms the existence of galvanic cou- After 1 h of immersion, the impedance diagram of the
pling between different phases of the coating and that galvannealed sample presents a single capacitive arc
galvanic corrosion is important in this electrolyte, a relating to the electron transfer reaction at the metal/
statement confirmed by our SEM micrographs in solution interface, with the measured values of capac-
Figure 4. itance ranging from 120 to 150 lF cm)2. At immersion
Figure 7 shows the evolution of the impedance times of 4–24 h a diffusion component was observed in
diagrams for the galvanized layer as a function of the the lowest frequency range, probably related to the
immersion time in 10)2 mol dm)3 NaCl aqueous solu- dissolution of the residual Zn present in the coating.
tion; the characteristic frequencies are indicated in each Only a capacitive loop is observed at immersion times of
Nyquist plot. The impedance diagrams of the galvanized more than 24 h.
sample were characterized by three loops in the first 8 h Figures 9 and 10 depict the impedance results
of immersion. The loop in the high frequency range was obtained, for the samples of galvanized and galvan-
attributed to the charge transfer resistance (Rct) in nealed steels respectively, tested in the natural marine
parallel with the double layer capacitance (Cdl), with environment for 36 months. The impedance diagrams
measured values ranging from 75 to 100 lF cm)2, which for the Zn layer (Figure 9) presented two capacitive
are characteristic of zinc layers [14, 21]. The presence of loops up to 21 months of exposure time, while the
inductive loops for Zn in an almost neutral solution has diagram for the sample exposed for 36 months showed
been reported previously [22]. With longer immersion three capacitive loops. The impedance diagrams of the
382

Fig. 9. Evolution of the impedance diagrams with exposing time for galvanized steel samples exposed in tropical marine environment.

Zn–Fe coatings (Figure 10) showed only one capacitive immersion tests the galvannealed steel displayed higher
loop except for the one obtained after 21 months. resistance values than the galvanized steel. Additionally,
As Magalhães et al. [23] pointed out, the value of the the resistance value of the galvanized steel rose rapidly
resistive component of impedance measured at a fixed on the first day of immersion, declining thereafter until
frequency and in the range of the lower frequencies is the sixth day, when it again showed an upward trend at
very suitable for industrial routines and relates to the longer immersion times. On the other hand, the resis-
corrosion of the coating. Figure 11 shows the values of tance of the galvannealed steel remained approximately
the real part of the impedance measurement obtained at constant in the first 8 h of immersion, increasing with
10 mHz as a function of immersion time in NaCl longer immersion times. Figure 11b shows that the
solution (Figure 11a) and as a function of exposure time behavior of the two layers was similar in the natural
in the natural marine environment (Figure 11b). In the marine environment, increasing initially up to a

Fig. 10. Evolution of the impedance diagrams with exposing time for galvannealed steel samples exposed in tropical marine environment.
 383
behavior of the galvannealed steel was affected by
cracking in the Zn–Fe layer pursuant to the dissolution
of zinc-rich phases, while the galvanized steel displayed
generalized corrosion in an aqueous medium and
localized corrosion in the marine environment. The
galvannealed steel showed greater corrosion resistance
than the galvanized steel in the aqueous medium.
However, the two coatings displayed similar corrosion
behavior in the tropical marine environment.

Acknowledgements

The authors thank CNPq-CTPETRO (Proc. 460033/

01-8), CAPES and FINEP (Proc. 22.01.0762.00),
Brazil, for financial assistance.

References

1. D.T. Llewellyn and R.C. Hudd, Steels: Metallurgy and Applica-

tions, 3rd ed., (Butterworth Heinemann, Oxford, England, 1998).
2. A.R. Marder, in G. Krauss and D.K. Matlock (Eds.), Zinc-Based
Coatings System: Metallurgy and Performance, (TMS, Warren-
dale, 1990), p. 55.
3. A.R. Marder, Prog. Mater. Sci. 45 (2000) 191.
4. C.R. Shastry and H.E. Towsend, Corrosion 45 (1989) 103.
5. H.H. Lee and D. Hiam, Corrosion 45 (1989) 852.
6. G. Sergi, N.R. Short and C.L. Page, Corrosion 41 (1985) 618.
Fig. 11. Zreal at 10 mHz vs. immersion time in 10)2 mol dm)3 NaCl 7. X.G. Chang and I.C. Bravo, Corrosion 50 (1994) 308.
solution (a) and vs. exposure time in tropical marine environment 8. A. Besseyrias, F. Dalard, J.J. Rameau and H. Baudin, Corros. Sci.
(b). 37 (1995) 587.
9. A. Besseyrias, P. Giraud, F. Dalard, J.J. Rameau and H. Baudin,
Revue de Métallurgie 95 (1998) 641.
maximum and then declining gradually until the end of 10. C. Xoffer, H. Dillen and B.C. De Cooman, J. Appl. Electrochem.
the test. 29 (1999) 209.
The increase of the impedance is attributable to the 11. A. Besseyrias, F. Dalard, J.J. Rameau and H. Baudin, Corros. Sci.
presence of corrosion products on the surface, which 39 (1997) 1883.
12. F.J. Fabri Miranda, J.D. da Silva Filho, I.C.P. Margarit and O.R.
improves the barrier properties. However, this improve- Mattos, in Proceedings of the 7th International Symposium on
ment in barrier properties does not prevent the localized Electrochemical Methods in Corrosion Research, paper 95.
corrosion process from continuing in both materials (see 13. ASTM G-50 Standard, 03–02, West Conshohocken, PA: ASTM,
Figures 3 and 4). In the case of the galvannealed 183 (1995).
samples, another possible explanation for the increase 14. F.J. Fabri Miranda, I.C.P. Margarit, O.R. Mattos, O.E. Barcia
and R. Wiart, Corrosion 55 (1999) 732.
in total impedance as a function of exposure time is the 15. A.M. Beccaria, Corrosion 46(11) 906.
preferential dissolution of the zinc-rich phases. This 16. C.E. Jordan, K.M. Goggins, A.O. Benscoter and A.R. Marder,
assumption is in close agreement with the findings of Mater. Characterization 31 (1993) 107.
Lee and Hiam [5], who showed that the dissolution of 17. Y. Lin, Galvatech’98 (ISIJ, Chiba, Japan, 1998), pp. 236.
zinc-rich phases caused the values of polarization 18. C. Cachet, B. Saidani and R. Wiart, J. Electrochem. Soc. 139
(1992) 644.
resistance to increase. 19. J. O’M Bockris and S.U.H. Khan, Surface Electrochemistry, A
Molecular Leval Approach (Plenum Press, New York, 1993).
20. K.R. Baldwing, M.J. Robinson and C.J.E. Smith, Corros. Sci. 36
4. Conclusions (1994) 1115.
21. G. Barceló, M. Sarret, C. Müller and J. Pregonas, Corros. Sci. 43
(1998) 13.
Electrochemical measurements and SEM and EDAX 22. S. Magaino, M. Soga, K. Sobue, A. Kawaguchi, N. Ishida and
analyses showed the differences in the corrosion behav- H. Imai, Electrochim. Acta 44 (1999) 4307 and references cited
ior of galvanized and galvannealed steel. The corrosion therein.