Intermetallics 42 (2013) 107e111

Contents lists available at SciVerse ScienceDirect

Intermetallics
journal homepage: www.elsevier.com/locate/intermet

Why does pitting preferentially occur on shear bands in bulk metallic

glasses?
Y.M. Wang a, C. Zhang a, Y. Liu a, K.C. Chan b, L. Liu a, *
a
School of Materials Science and Engineering, Huazhong University of Science and Technology, 430074 Wuhan, China
b
Department of Industrial and Systematic Engineering, The Hong Kong Polytechnic University, Hong Kong, China

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

Article history: The effect of shear bands on pitting corrosion of a pre-deformed Zr-based bulk metallic glass was studied
Received 3 April 2013 through immersion test in 1 M HCl solution. It was found that pitting always preferentially occurred at
Accepted 14 May 2013 the offsets of the shear bands, but this selective corrosion behavior disappeared when the shear offsets
Available online
were polished, although shear bands did exist at the same locations. This demonstrates that the shear
offsets, rather than the structural changes in the shear bands, is the dominant factor for the pitting
Keywords:
initiation.
B. Glasses, metallic
Ó 2013 Elsevier Ltd. All rights reserved.
B. Corrosion
D. Microstructure

1. Introduction Zr59Ti3Cu20Al10Ni8 BMG, they argued that the introduction of the

residual stress and the change in structure and compositions in
Bulk metallic glasses (BMGs) have attracted increasing the shear bands play an important role in pit initiation and
attention in recent years because of their unique mechanical and propagation. Nie et al. [11] also observed recently the same
chemical properties, including superior strength, high hardness, pitting behavior on shear bands that were formed by cold-rolling
relatively low Young’s modulus, large elastic limit, as well as in Zr64.13Cu15.75Ni10.12Al10 BMGs. It was suggested that the pitting
excellent wear and corrosion resistance [1e3]. Unlike crystalline was triggered mainly by structural changes (such as the free
materials, plastic deformation of BMGs at room temperature is volume intake and nanovoids formation) rather than the residual
always highly localized in some very narrow regions, named stress and surface offsets formed, based on the fact that the pits
shear bands [4], resulting in a limited plasticity of the super- were still distributed along the same shear band locations after
strong materials. The formation of shear bands is usually annealing or electropolishing. However, this evidence is not very
accompanied with structural changes, such as the accumulation convincing, as the distribution of pits, after electropolishing of
of free volume [5], the formation of nanovoids and nanocrystals the BMG, is actually stochastic rather than selective. In contrast,
[6,7], stress aggregation [8], as well as the formation of surface Hiromoto et al. [12] and Gebert et al. [13] revealed that the
offsets in their close vicinity, which are thought to significantly surface roughness plays a crucial role in pitting initiation in
influence the corrosion behavior of post-deformed BMGs in metallic materials, including BMGs. No consistent picture has
corrosive media [9,10]. The preferential pitting on shear bands been achieved so far on the factors that dominate preferential
was first reported by Pampillo [9], who suggested that the rela- pitting on shear bands. In this work, a comparative study on the
tively more disordered structure in the shear bands could cause a pitting behavior was carried out before and after polishing the
preferential etching. Recently, numerous studies have been con- shear offsets of the shear bands, which were intentionally
ducted to investigate the effect of shear bands on the corrosion designed by three-point-bending and uniaxial compression on
behavior of various Zr-based BMGs. For instance, Gebert et al. (Zr62Cu23Fe5Al10)97Ag3 BMGs. It was found that the shear offsets,
[10] found that pits were always initiated on the shear but not shear bands themselves, play a more important role in
bands formed by compression and three-point-bending in the initiation of pits of the post-deformed BMG in a corrosive
solution. The mechanism for pitting initiation on the shear bands
* Corresponding author. Tel.: þ86 27 87556894; fax: þ86 27 87554405. is discussed in terms of the galvanic effect between the peak and
E-mail address: [email protected] (L. Liu). valley of the shear offset.

0966-9795/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.intermet.2013.05.006
108 Y.M. Wang et al. / Intermetallics 42 (2013) 107e111

2. Experimental diameter 2 mm and length of 4 mm was compressed at a strain rate

of 104 s1 to form a single shear band (see Ref. [5] for the detailed
Alloy ingots with nominal composition (Zr62Cu23Fe5Al10)97Ag3 process). Then the rod with a pre-formed single shear band and after
[14] were prepared by arc-melting a mixture of pure Zr, Cu, Fe, Al and polishing was subsequently tested in 1 M HCl for 1 h of immersion.
Ag metals (purity > 99.5%) in a Ti-gettered purified argon atmo- The advantage of using a single shear band compared to multiple
sphere. The alloy ingots were melted at least five times to ensure shear bands is that the effect of structural change on corrosion
compositional homogeneity and then cast into a water-cooled behavior can be enlarged, as more severe structural changes
copper mold to produce BMG beams with the size of occurred in a single shear band [5].
50  3.2  3.2 mm. The amorphous structure of the as-cast alloys
was verified by X-ray diffraction (XRD, Philips c’ Pert Pro) and dif- 3. Results and discussion
ferential scanning calorimetry (DSC, PE-7). To create a high density
of shear bands with sufficiently large offsets, three-point-bending Fig. 1(a) shows typical pitting morphology on the surface of the
testing was carried out. Before testing, each surface of the rectan- tensile side of the BMG which was subjected to three-point-
gular beams was mechanically ground to remove about 0.1 mm on bending and subsequently to immersion in 1 M HCl solution for
each side to give a final size of 50  3  3 mm, aiming to avoid the 1 h. It was found that nearly all pits with sizes around 30e40 mm
effect of surface residual stress and superficial crystallization on were formed on the shear bands, indicating that the shear bands
pitting. The specimens were then finally polished to a mirror finish are indeed preferential sites for pitting initiation. Fig. 1(b) shows an
to offer the best condition for observing the shear bands. The three- amplified image of an individual pit, clearly showing that the pit is
point-bending testing was carried out on a Zwick/Roell 020 testing initiated at the band, and then grows radially into the surrounding
machine at a constant strain rate of 5  104 s1 until fracture. The amorphous matrix. Similar pitting behavior was also observed in
tensile sides of the bent samples were examined, as they had more the previous work by Gebert et al. [10] and Nie et al. [11]. Fig. 1(c)
shear bands with larger shear offset with respect to the compressive shows the 3D profile of a shear band containing a pit as an example,
sides. Pitting behavior of the bent samples was then studied by and indicates that a shear offset with a height of 5e10 mm was
immersion in 1 M HCl solution for 1 h at room temperature. The formed on the shear band. Fig. 1(d) shows the SEM image of the
morphology of the surfaces after corrosion test was characterized by polished sample (i.e., the shear offset and pits were ground away
scanning electron microscopy (SEM, FEI Quanta 200). Thereafter, the and then polished to a mirror finish) after immersion for 2 h. It can
same samples were carefully polished to remove all the shear offsets, be seen that, in this case, pits no longer appeared on the shear
and subsequently subjected to the same corrosion tests. In addition, bands (the yellow dash lines indicate the tracks where shear bands
the as-cast BMG beam, being intentionally scratched by a needle, appear, as originally shown in Fig. 1(a)), but are distributed sto-
was also immersed in the same corrosive solution for 1 h to illustrate chastically on the whole surface instead (see the insert in Fig. 1(d)).
the effect of surface roughness on the corrosion behavior of the BMG. In addition, the size of the pits (w10 mm after 2 h immersion) are
The 3D surface morphology of the shear offsets generated and the much smaller than those formed on the shear offsets (30e40 mm
induced scratch were examined by super three-dimensional depth after 1 h immersion), indicating that selective pitting did not occur
of field microscopy (VHX-1000C). For comparison, a BMG rod of when the shear offsets were polished out. Taking into account that

Fig. 1. (a) SEM image of the shear bands after three-point-bending test of (Zr62Cu23Fe5Al10)97Ag3 BMG and the distribution of pits after being immersed in 1 M HCl for 1 h, all
the corrosion pits were formed at the shear bands; (b) a magnified pit shown in (a); (c) a 3D profile showing a pit on a shear band; (d) SEM image of a polished sample
after being immersed in 1 M HCl for 2 h at the same position and the insert is the enlargement of a portion in (d), showing several pits stochastically distributed on the polished
surface.
 Y.M. Wang et al. / Intermetallics 42 (2013) 107e111 109

shear bands do exist at the same places, even though the offsets corrosion pits were found on the sample surface even prolonging
were removed, it is suggested that the shear offsets, but not shear the immersion time to 4 h (see Fig. 2(b)). Considering that a heavy
bands themselves, play a crucial role in the initiation of pits in structural change (i.e., significant intake of free volume in this case)
BMGs in corrosive solution. It is true that structural changes, e.g. the really occurred in the shear band, the disappearance of selective
accumulation of free volume, the formation of nanovoids or even pitting after elimination of the shear offset indicates that the
nanocrystals, can occur in the shear bands [5e7]. However, the structural changes in the shear bands have little effect on pit
present results imply that the change in structure may not be the initiation.
dominant factor for the initiation of pits on the shear bands. To To better understand the role of shear offsets in the preferential
further clarify this issue, a compressive test was conducted on a formation of pits, a mechanical scratch was intentionally made by a
BMG rod (f 2  4 mm) to form a single shear band under a needle on a mirror-finish BMG surface to form a sharp offset. The
controlled loading (see Ref. [5] for the detailed loading process), offset had a height of around 4.3 mm (see Fig. 3(a)), which is roughly
and the sample was subsequently immersed in 1 M HCl solution for in the same level as the shear offset formed by three-point-bending
1 h. The merit of using a single shear band is that severe plastic test (see Fig. 1(c)). Fig. 3(b) shows the SEM image of the scratched
deformation is constrained into a short region, leading to a drastic sample after immersion in 1 M HCl solution for 1 h. Many corrosion
structural change in the shear band, so that the effect of structural pits were formed on the sharp edges of the offset, while no pits
changes on corrosion, if any, could be enlarged. Our previous work were observed on the bed of the scratch where it is relatively
[5] revealed that a huge increase of 1.4% of free volume was smooth. It is noted that plastic deformation could take place not
introduced in the single shear band due to the constrained and only on the edges but also on the bed when a scratch was made.
severe plastic deformation. This caused a 35% reduction in hardness However the pits only existed at the edges but not on the bed,
in the band, which is almost 3 times lower than that in multiple which again confirms that the offset plays a dominant role in the
shear bands [15,16]. Fig. 2(a) shows that almost all the pits are initiation of pits.
formed on the sharp shear offset after being exposed in HCl solu- The question arises as to how and why the shear offsets of the
tion for 1 h. However, once the shear offset was removed, no shear bands promote pitting initiation on the deformed BMG. To

Fig. 2. (a) SEM image of pits on a single shear band after being immersed in 1 M HCl
for 1 h, the inset is the shear band morphology of the deformed sample after im- Fig. 3. (a) SEM image of the scratch induced by a needle, the inset is the line profile of
mersion; (b) SEM image of the sample after polishing the shear offset and immersion the scratch, showing a height of 4.3 mm of the offset; (b) SEM image of the scratched
in 1 M HCl for 4 h, and the inset is a cross-sectional view of the sample with a single sample after being immersed in 1 M HCl for 1 h, showing that all the pits were formed
shear band. at the edges of the scratch.
110 Y.M. Wang et al. / Intermetallics 42 (2013) 107e111

answer the questions, we have to understand where the passive the film there is usually weak and could be preferential broken, and
film on BMG surface is preferentially broken and how the galvanic become the possible sites for pitting. However, to make pitting
cells are formed. For an alloy containing metallic elements, passive proceed galvanic cells must be established, otherwise the broken
film could be spontaneously formed by the oxidation of the active film can be recovered by the consecutive oxidation reaction. As a
elements once it is immersed in corrosive solution. Actually, the matter of fact, the shear offsets on BMG surface with smaller cur-
formation of passive film is a dynamic process, i.e., dissolution and vature radius should have lower potential as compared with the
growth of the film simultaneously proceed [17]. The dynamic smooth region nearby. Therefore, the galvanic cell could be easily
equilibrium of the two processes could lead to the formation of formed between the shear offsets and nearby smooth surface.
stable passive film on the alloy surface, and protect the alloy from Combining the Kelvin’s equation [20], ln(P/P0) ¼ 2sM/rrRT, which
corrosion. However, the surface of the alloy is not completely correlates the surface vapor pressure (P) and the curvature radius
prefect and homogenous. The equilibrium of dissolving and (r) and the Nernst equation [21], E ¼ E4  ðRT=nFÞln ððP=P0 Þ=aMnþ Þ,
growing process may be dilapidated in the regions where defects or which correlates the electrode potential (E) and surface vapor
stress concentration exist. In these regions, the dissolution can pressure (P), we can drive the relationship between the electrode
prevail over the growth of the film, and eventually causes the potential and the surface curvature radius, as being simplified as
breakdown of the passive film [18,19]. In the present case, the shear
offsets on the BMG surface, originating from shear banding, are the
places where great tension stress and defects existed (see Fig. 4(a)),
1
Ef  (1)
r
The parameters in the above equations are defined as: P0 is the
vapor pressure of a metal surface with infinite curvature radius, s is
the surface tension, M and r are the molecular mass and density of
the metal, Ef is the standard electrode potential, n is the number of
electrons transferred in the reaction, F is the Faraday constant and
anþ
M is the activity of metal ion, which is a constant for a certain
metal in the same electrolyte, R and T are the gas constant and
temperature, respectively.
The equation (1) reflects clearly that the difference in curvature
radius r results in the difference in electrode potential E, which
means that the regions with smaller curvature radius exhibit lower
electrode potentials, vice versa. The theory can well explain the
preferential pitting at the shear offset in the present work. The peak
regions can have a lower electrode potential due to the smaller
radius r, and metals in these regions are more active, thus release
more easily electrons [22] with respect to the relatively smooth
regions nearby (see Fig. 4(b)). Therefore the peak regions (shear
offsets) can serve as anodes, while the valleys nearby serve as the
cathodes, leading to the formation of galvanic cells. Once the pas-
sive film ruptures at the peak regions, these galvanic cells start to
work via the exchange of ions and electrons, causing the initiation
of pitting [23], as show in Fig. 4(c). This phenomenon was actually
observed, especially on the specimen with a surface scratch, where
all the pits were only formed at the edges of the scratch (see
Fig. 3(b)).
Structural changes can occur in shear bands during plastic
deformation, including the increase of free volume due to shear
dilatation, and sometimes nanovoids formation and even crystal-
lization when the plastic deformation is very severe. Based on our
previous work, the intake of free volume is only the form of
structural change during the deformation and a maximum 1.4% of
free volume was introduced in the shear band. The increase in free
volume does not change the amorphous feature or the composi-
tion of the BMG, and thus cannot alter the chemical potential in
the shear bands. Therefore, selective corrosion could not take
place only by the increase of free volume. The formation of
nanocrystals or nanovoids may change the structure and compo-
sition of the BMG, causing alteration of the chemical potential in
the shear bands. However, a shear band usually has the width of
several tens nanometers (e.g. 10e20 nm [7,24]), meaning that the
structural change, if any, should be localized in such a narrow
Fig. 4. Schematic illustration of pitting corrosion in a BMG with shear offset: (a) failure region. This region with severe structural change, may alter locally
of the passive film caused by interfacial tension in the peak region; (b) electrons at the chemical potential of the shear band, but cannot induce the
sharp peak escape more easily than those in the flat area, lowering the potential of the
peak region; (c) the peak with a lower potential acts as an anode, while the region in
formation of a stable pit, as a stable pit usually has a size of larger
the flat area, with a higher potential, acts as a cathode, thus forming an electro- than ten micrometers, otherwise the pit can die by surface
chemical cell. repassivation [25].
 Y.M. Wang et al. / Intermetallics 42 (2013) 107e111 111

4. Conclusion References

The experimental results revealed that pitting corrosion always [1] Johnson WL. MRS Bull 1999;24:42.
[2] Inoue A. Acta Mater 2000;48:279.
preferentially occurred on the shear bands of plastically deformed [3] Wang WH, Dong C, Shek CH. Mater Sci Eng R 2004;44:45.
bulk metallic glasses, but pits tended to be formed stochastically [4] Argon AS. Acta Metall 1979;27:47.
after the shear offsets were removed. This indicates that the shear [5] Pan J, Chen Q, Liu L, Li L. Acta Mater 2011;59:5146.
[6] Bhowmick R, Raghavan R, Chattopadhyay K, Ramamurty U. Acta Mater
offsets, rather than shear bands themselves, play a dominant role in 2006;54:4221.
pitting initiation. The phenomenon can be accounted by the [7] Li J, Spaepens F, Hufnagel TC. Philos Mag A 2002;82:2623.
galvanic effect, in which the shear offsets, with higher chemical [8] Chen LY, Ge Q, Qu S, Jiang JZ. Scripta Mater 2008;59:1210.
[9] Pampillo CA. Scripta Mater 1972;6:915.
activity, serve as an anode, while the surrounding relatively flat [10] Gebert A, Gostin PF, Uhlemann M, Eckert J, Schultz L. Acta Mater 2012;60:
regions, with lower chemical activity, act as the cathodes. These 2300.
galvanic cells formed between shear offset and the surrounding [11] Nie XP, Cao QP, Wu ZF, Ma Y, Wang XD, Ding SQ, et al. Scripta Mater
2012;67:376.
regions could cause an occurrence of a fast dissolution at the shear
[12] Hiromoto S, Tsai AP, Sumita M, Hanawa T. Corros Sci 2000;42:2167.
offset, resulting in a preferential pitting corrosion. The structure [13] Gebert A, Gostin PF, Schultz L. Corros Sci 2010;52:1711.
changes, if any, in shear bands are constrained in a narrow region, [14] Liu Z, Chan KC, Liu L. J Alloys Compd 2009;487:152.
[15] Xie S, George EP. Acta Mater 2008;26:5202.
where stabilized pits are not able to be formed.
[16] Ramamurty U, Jana S, Kawamura Y, Chattopadhyay K. Acta Mater 2005;
53:705.
Acknowledgments [17] Heuer AH, Kahn H, Ernst F, Michal GM, Hovis DB, Rayne RJ, et al. Acta Mater
2012;60:716.
[18] Lin LF, Chao CY, Macdonald DD. J Electrochem Soc 1981;128:1194.
This work was financially supported by the National Nature [19] Sato N. J Electrochem Soc 1982;129:255.
Science Foundation of China (Grant Nos. 51071072 and 51271081). [20] Thomson (Kelvin) U. Phil Mag 1871;42:448.
The work was also partially supported by the National Fundamental [21] Akoi K. J Electroanal Chem 1991;1:310.
[22] Li W, Li DY. Acta Mater 2006;54:445.
Research Program of China (Grant No. 2011CB012806). The authors [23] Zhang C, Chan KC, Wu Y, Liu L. Acta Mater 2012;60:4152.
are grateful to the Analytical and Testing Center, Huazhong Uni- [24] Li J, Wang ZL, Hufnagel TC. Phys Lett B 2012;65:144201.
versity of Science and Technology for technical assistance. [25] Lucente AM, Scully JR. Electrochem Solid-State Lett 2007;10(5):C39.