Materials and Design 32 (2011) 2350–2354

Contents lists available at ScienceDirect

Materials and Design

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

Short Communication

Influence of surface roughness on the corrosion behaviour of magnesium alloy

R. Walter, M. Bobby Kannan ⇑
Discipline of Chemical Engineering, School of Engineering and Physical Sciences, James Cook University, Townsville, Queensland 4811, Australia

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

Article history: In this study, the influence of surface roughness on the passivation and pitting corrosion behaviour of
Received 3 September 2010 AZ91 magnesium alloy in chloride-containing environment was examined using electrochemical tech-
Accepted 6 December 2010 niques. Potentiodynamic polarisation and electrochemical impedance spectroscopy tests suggested that
Available online 13 December 2010
the passivation behaviour of the alloy was affected by increasing the surface roughness. Consequently,
the corrosion current and the pitting tendency of the alloy also increased with increase in the surface
roughness. Scanning electron micrographs of 24 h immersion test samples clearly revealed pitting corro-
sion in the highest surface roughness (Sa 430) alloy, whereas in the lowest surface roughness (Sa 80) alloy
no evidence of pitting corrosion was observed. Interestingly, when the passivity of the alloy was dis-
turbed by galvanostatically holding the sample at anodic current for 1 h, the alloy underwent high pitting
corrosion irrespective of their surface roughness. Thus the study suggests that the surface roughness
plays a critical role in the passivation behaviour of the alloy and hence the pitting tendency.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction the passivation behaviour of magnesium alloys with different

surface finish to correlate the surface roughness to their general
Magnesium alloys have gained significant interest for automo- corrosion and pitting tendency.
tive and aerospace applications for their high-strength-to-weight In this study the corrosion behaviour of AZ91 magnesium alloy
ratio. However, their wide-spread use is limited by their high cor- was studied using electrochemical techniques, such as potentiody-
rosion susceptibility, particularly in chloride-containing environ- namic polarisation and electrochemical impedance spectroscopy,
ment. In recent years, a substantial amount of work has been especially to understand the passivation behaviour of the alloy
carried out to improve the corrosion resistance of magnesium with different surface roughness.
and its alloys through alloying and surface coatings for various
applications [1–5]. A few researchers have also studied the influ- 2. Experimental procedure
ence of microstructure on the corrosion behaviour of magnesium
alloys [6,7]. Sand-cast AZ91 magnesium alloy was used as the test material
Generally, the surface roughness plays a role on the corrosion in this study. The chemical composition of the alloy is given in
behaviour of metallic materials. It has been reported that an Table 1. For obtaining different surface roughness, the samples
increase in the surface roughness of stainless steels increases the were grinded/polished with different grits of silicon carbide (SiC)
pitting susceptibility [8–10] and general corrosion rate [11]. A sim- (i.e., 320, 600 and 1200) and 3 lm diamond-paste. The surface
ilar trend has been reported for other metals, such as copper [12], roughness was analysed using atomic force microscopy (AFM).
and titanium-based alloys [13]. However, the literature on the ef- The corrosion behaviour of the alloy was studied using electro-
fect of surface roughness on the corrosion behaviour of magnesium chemical techniques, such as potentiodynamic polarisation and
and its alloys is limited. Interestingly, the only work by Alvarez electrochemical impedance spectroscopy (EIS). A potentiostat and
et al. [14] on AE44 magnesium alloy is in contrast to the trend frequency response analyser (Model VersaSTAT 3) driven by
reported for other metallic materials. Based on immersion test VersaStudio software were used for potentiodynamic polarisation
results, the authors [14] reported that the general corrosion and EIS experiments, respectively. A typical three electrode system
decreased as the alloy’s surface roughness increased. Further they consisting of graphite as a counter electrode, Ag/AgCl electrode as a
reported that the polished alloy allowed greater initial pitting and reference electrode and sample (0.75 cm2 exposed area) as a
higher pitting volume than the semi-polished alloy. Typically, the working electrode were used. The test electrolyte was 0.5 wt.%
general and localized corrosion behaviour of alloys would depend NaCl solution. Prior to testing, the samples were allowed for 2 h
on their passivation behaviour. Hence, it is important to know to reach a relatively stable open circuit potential. EIS experiments
were performed over the frequency range of 105–102 Hz at AC
⇑ Corresponding author. Tel.: +61 7 4781 5080; fax: +61 7 4781 6788. amplitude of 5 mV. Potentiodynamic polarisation experiments
E-mail address: [email protected] (M.B. Kannan). were done at a scan rate of 0.5 mV/s. In addition, immersion tests

0261-3069/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.matdes.2010.12.016
 R. Walter, M.B. Kannan / Materials and Design 32 (2011) 2350–2354 2351

Table 1 Table 2
Chemical composition of the AZ91 alloy. Surface roughness of AZ91 magnesium alloy with different surface finish.

Al Zn Mn Si Fe Mg Sample finish Sa (nm) Standard deviation (nm)

AZ91 sand-cast 9.18 0.78 0.20 0.01 0.002 Bal. 320 SiC 430 183
600 SiC 248 86
1200 SiC 145 42
3 lm diamond-paste 80 21
were carried out for 24 h, and galvanostatic tests were carried out
at 1.25 mA/cm2 anodic current for 1 h, and the post-corrosion sam-
ples were examined using scanning electron microscope (SEM) to
identify the mode of corrosion attack.

3. Results and discussion

The AFM images for the different surface finishes are shown in
Fig. 1. It was clearly evident that the surface roughness of the alloy
decreased as the grit size of the grinding paper used was increased.
Further, the diamond-paste polished alloy showed a relatively
smooth surface as compared to the SiC grinded alloy. The mea-
sured surface roughness and standard deviation are listed in Table
2. As expected the standard deviation in the surface roughness
values decreased as the surface of the alloy became smoother.
Fig. 2 shows the Nyquist plots of the alloy with different surface
roughness tested in chloride-containing environment. In the SiC
grinded samples, a similar behaviour was observed i.e., a high Fig. 2. Nyquist plots of AZ91 magnesium alloy, with different surface roughness,
frequency capacitive loop and low frequency inductive loop. It is tested in 0.5 wt.% NaCl.
reported in literature that the high frequency capacitive loop
corresponds to the charge transfer and film effect [15,16] and the second capacitive loop in the mid frequency is related to the
low frequency inductive loop indicates pitting of the alloy [17]. relaxation of mass transport through the corrosion product layer,
Interestingly, the diamond-paste polished alloy showed one or in other words, suggests the presence of a protective film [15].
capacitive loop at high frequency and another depressed capacitive It also means that the absence or lack of evidence of the second
loop at medium frequency. Moreover, the alloy showed no evidence capacitive loop in the mid frequency is an indication of scarcely/
of an inductive loop, which suggested that the sample with smooth no protective film. Although the formation of magnesium hydroxide
surface has not undergone pitting. It is reported in the literature that via the reaction: 2Mg þ 2H2 O ! MgðOHÞ2 þ H2 , is plausible under
for rare-earth containing magnesium alloys, an observation of a different surface roughness conditions, the formation of a continu-

Fig. 1. Surface topography of AZ91 magnesium alloy grinded/polished up to (a) 320 grit SiC; (b) 600 grit SiC; (c) 1200 grit SiC; and (d) 3 lm diamond-paste.
2352 R. Walter, M.B. Kannan / Materials and Design 32 (2011) 2350–2354

exhibits higher passivation tendency and pitting resistance than the

high surface roughness alloy.
The polarisation curves of AZ91 alloy, with different surface
roughness, tested in chloride-containing solution are shown in
Fig. 3. The electrochemical corrosion parameters from the polarisa-
tion curves are listed in Table 3. The corrosion potential (Ecorr) of
the alloy shifted towards the noble direction as the surface rough-
ness decreased. Notably, the corrosion current (icorr) decreased as
the surface roughness decreased. However, there was no signifi-
cant difference in the cathodic current within the samples having
different surface roughness, which suggested that the shift in the
Ecorr and the difference in icorr were solely due to the anodic behav-
iour of the alloy. The alloy with the highest surface roughness
showed a sharp increase in the anodic current just above the Ecorr.
This phenomenon (i.e., sharp increase in the anodic current) was
an indication of pitting, and in this case the pitting potential was
Fig. 3. Potentiodynamic polarisation curves of AZ91 alloy, with different surface pinned with the Ecorr. Interestingly, as the surface roughness de-
roughness, tested in 0.5 wt.% NaCl.
creased the alloy revealed a passive-like behaviour before a sharp
break-down. It was noticed that the passive-potential region (Epass,
ous protective film would be higher on a smooth surface alloy than the difference between the corrosion potential and the break-down
on an irregular surface. Hence, the alloy with relatively high surface potential, Ebd) increased when the surface roughness of the alloy
roughness showed no evidence of the second capacitive loop, but decreased. Hence, the alloy with the lowest surface roughness
instead showed an inductive loop in the low frequency indicative showed a passive-potential region of about 79 mV, whereas the
of pitting. In the case of the low surface roughness alloy, the highest surface roughness alloy showed no evidence of passive
evidence of depressed second capacitive loop and the non-existence region. Although the passivity could be mainly due to the forma-
of low frequency inductive loop suggested that smooth surface alloy tion of magnesium hydroxide, the presence of aluminium (forms
aluminium oxide in aqueous environment) in the alloy may also
have some influence on the passivation behaviour.
Table 3 The SEM micrographs of the samples immersed in chloride-
Electrochemical corrosion parameters of AZ91 alloy (with different surface rough-
ness) obtained from potentiodynamic polarisation curves.
containing solution at the open circuit potential for 24 h (Fig. 4)
clearly revealed that the alloy with the highest surface roughness
Surface finish icorr (lA/cm2) Ecorr (V) Ebd (V) Epass (mV) underwent high pitting corrosion, whereas the alloy with the low-
320 grit SiC 6.92 1.447 1.447 0 est surface roughness showed no evidence of localized attack. The
600 grit SiC 4.79 1.427 1.392 35 alloy having mid-range surface roughness did show some evidence
1200 grit SiC 3.73 1.418 1.362 56
of pitting corrosion, however substantially lower than in the alloy
3 lm diamond-paste 2.19 1.378 1.299 79
with highest surface roughness.

Fig. 4. SEM micrographs of AZ91 magnesium alloy grinded/polished (for different surface roughness) up to (a) 320 grit SiC; (b) 600 grit SiC; (c) 1200 grit SiC; and (d) 3 lm
diamond-paste, and immersed in 0.5 wt.% NaCl for 24 h.
 R. Walter, M.B. Kannan / Materials and Design 32 (2011) 2350–2354 2353

Fig. 5. SEM micrographs of AZ91 magnesium alloy grinded/polished up to (a) 320 grit SiC; (b) 600 grit SiC; (c) 1200 grit SiC; and (d) 3 lm diamond-paste, after galvanostatic
testing.

Interestingly, the SEM micrographs of galvanostatically-held in chloride-containing environment. The electrochemical experi-
alloy revealed a large number of pits, irrespective of their surface ments showed that an increase in the surface roughness of the
roughness (Fig. 5). Alvarez et al. [14] also found pitting in both pol- alloy affects the passivation tendency and consequently increases
ished and semi-polished AE44 magnesium alloy. Interestingly, they the pitting susceptibility of the alloy. However, when the passivity
reported that the density of pitting was relatively higher in pol- of the alloy is disturbed then the influence of surface roughness on
ished alloy as compared to semi-polished alloy. In fact, a closer the pitting corrosion susceptibly becomes less significant.
look at the SEM micrographs of galvanostatically-held alloy, sug-
gested that the alloy with lowest surface roughness exhibits a References
slightly higher number of pits as compared to the alloy with high-
est surface roughness. However, Alvarez et al. [14] observed larger [1] Song G, Atrens A. Understanding magnesium corrosion – a framework for
improved alloy performance. Adv Eng Mater 2003;5:837–58.
pits in semi-polished alloy as compared to polished alloy. [2] Bobby Kannan M, Dietzel W, Blawert C, Atrens A, Lyon P. Stress corrosion
In order to understand the differences in the pitting behaviour cracking of rare-earth containing magnesium alloys ZE41, QE22 and electron
of magnesium alloy with different testing methods, the fundamen- 21 (EV31A) compared with AZ80. Mater Sci Eng A 2007;480:529–39.
[3] Bobby Kannan M, Raman RK. In vitro degradation and mechanical integrity of
tal corrosion mechanism of magnesium has been reviewed. It is
calcium-containing magnesium alloys in modified-simulated body fluid.
well documented in the literature that magnesium dissolution Biomaterials 2008;29:2306–14.
increases the local pH at cathodic sites of the sample, which tends [4] Gray JE, Luan B. Protective coatings on magnesium and its alloys – a critical
review. J Alloys Compd 2002;336:88–113.
to facilitate corrosion-product film, or in other words passivates
[5] Bobby Kannan M, Gomes D, Dietzel W, Abetz V. Polyoxadiazole-based coating
the alloy [1]. However, in the presence of chloride ions the passive for corrosion protection of magnesium alloy. Surf Coat Technol 2008;202:
film on magnesium breaks down, causing pitting corrosion. Alvarez 4598–601.
et al. [14] conducted the testing in 3.5% NaCl solution, which is not [6] Song G, Atrens A, Dargusch M. Influence of microstructure on the corrosion of
diecast AZ91D. Corros Sci 1998;41:249–73.
different to the test solution in this study; however it should be [7] Bobby Kannan M. Influence of microstructure on the in-vitro degradation
noted that they have aerated the solution throughout the experi- behaviour of magnesium alloys. Mater Lett 2010;62:739–42.
ment. Although oxygen (in air) has no significant influence on [8] Hong T, Nagumo M. Effect of surface roughness on early stages of pitting
corrosion of type 301 stainless steel. Corros Sci 1997;39:1665–72.
the corrosion behaviour of the magnesium [18], the stirring effect [9] Zuo Y, Wang H, Xiong J. The aspect ratio of surface grooves and metastable
caused by aeration could reduce the local pH change and conse- pitting of stainless steel. Corros Sci 2002;44:25–35.
quently affect the passivation tendency of the alloy. Hence, they [10] Sasaki K, Burstein GT. The generation of surface roughness during slurry
erosion–corrosion and its effect on the pitting potential. Corros Sci 1996;38:
observed pitting corrosion even in the polished alloy under immer- 2111–20.
sion testing. However, in the case of the galvanostatically-held [11] Shahryari A, Kamal W, Omanovic S. The effect of surface roughness on the
alloy, the anodic current was high enough to break the passive film efficiency of the cyclic potentiodynamic passivation (CPP) method in the
improvement of general and pitting corrosion resistance of 316LVM stainless
of the alloy under all surface roughness (refer Fig. 3), and hence pit-
steel. Mater Lett 2008;62:3906–9.
ting corrosion was observed in all the samples irrespective of their [12] Li W, Li DY. Influence of surface morphology on corrosion and electronic
surface roughness. behaviour. Acta Mater 2006;54:445–52.
[13] Cabrini M, Cigada A, Rondelli G, Vicentini B. Effect of different surface
finishing and of hydroxyapatite coatings on passive and corrosion current of
4. Conclusions Ti6Al4V alloy in simulated physiological solution. Biomaterials 1997;18:
783–7.
[14] Alvarez RB, Martin HJ, Horstemeyer MF, Chandler MQ, Williams N, Wang PT,
The study clearly suggests that the surface roughness plays a et al. Corrosion relationships as a function of time and surface roughness on a
critical role in the corrosion behaviour of AZ91 magnesium alloy structural AE44 magnesium alloy. Corros Sci 2010;52:1635–48.
2354 R. Walter, M.B. Kannan / Materials and Design 32 (2011) 2350–2354

[15] Zucchi F, Grassi V, Frignani A, Monticelli C, Trabanelli G. Electrochemical [17] Jin S, Amira S, Ghali E. Electrochemical impedance spectroscopy evaluation of
behaviour of a magnesium alloy containing rare earth elements. J Appl the corrosion behavior of diecast and thixocast AXJ530 magnesium alloy in
Electrochem 2006;36:195–204. chloride solution. Adv Eng Mater 2007;9:75–83.
[16] Guo LF, Yue TM, Man HC. Excimer laser surface treatment of magnesium alloy [18] Makar GL, Kruger J. Corrosion of magnesium. Int Mater Rev 1993;38:
WE43 for corrosion resistance improvement. J Mater Sci Lett 2005;40:3531–3. 138–53.