Reciprocating wear mechanisms in a Zr-based bulk metallic glass
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Reciprocating wear mechanisms in a Zr-based bulk
metallic glass
H.W. Jin, R. Ayer, and J.Y. Koo
Advanced Structural Materials Section, ExxonMobil Research & Engineering Company,
Annandale, New Jersey 08801
R. Raghavan and U. Ramamurtya)
Department of Materials Engineering, Indian Institute of Science, Bangalore-560012, India
(Received 20 July 2006; accepted 12 October 2006)
The dry sliding friction coefficient and the wear volume loss W, in a
zirconium-based bulk metallic glass (BMG) under high-frequency (50 Hz)
reciprocating conditions, were investigated with the objective of assessing the influence
of free volume and crystallization on the wear behavior of amorphous metals. The
BMG samples were annealed either below the glass transition temperature Tg to induce
structural relaxation and hence reduce the free volume that controls plasticity through
shear-band formation or above Tg to crystallize the amorphous BMG prior to wear
testing. Results show that the wear behavior of both the as-cast and relaxed glasses
was dominated by the oxidation of the surface layers. A sharp transition in the contact
electrical resistance complemented by a marked increase in was noted. This was
attributed to the formation of a thick tribo film with high oxygen concentration and its
subsequent delamination. The values, before as well as after the transition, in the
relaxed glasses were similar to those for the as-cast alloy. However, a gradual decrease
in W with annealing temperature was observed. A good correlation between W and
nanohardness was noted, implying that the intrinsic hardness in the BMGs controlled
the wear rate.
I. INTRODUCTION Two distinct deformation modes have been reported in
Amorphous metals or metallic glasses were first syn- amorphous alloys.5 At higher temperatures and low
thesized by rapidly quenching Au75Si25 alloy melts in stresses, the plastic deformation occurs through homoge-
1960.1 As a result of the extreme cooling rates (∼106 K/s) neous (viscous) flow. At lower temperatures and higher
required to achieve a high degree of undercooling in the stresses, the deformation is quite inhomogeneous, being
melt, metallic glasses could only be produced in thin concentrated in shear bands that form in orientations
ribbons (∼100 m) or in powder form. Hence, despite close to the maximum resolved shear stress plane. Be-
their unique combinations of mechanical properties such cause of these fundamental differences, the wear/friction
as high elastic limits, nearly theoretical strength com- properties of amorphous alloys are also distinct from
bined with good toughness, the widespread application of those of crystalline metals.
amorphous alloys has been significantly restricted. The Metallic glasses are generally expected to exhibit out-
development of bulk metallic glasses (BMGs) based on standing wear resistance due to their high hardness.
multicomponent alloy systems that remain amorphous However, a review of the literature shows conflicting
even when the cooling rates are low (∼1 K/s) rejuvenated trends regarding the role of amorphous structure on the
interest in these alloys for practical applications.2–4 Due wear performance. It has been reported that the Fe-based
to a lack of long-range order, these alloys exhibit quite metallic glasses exhibit higher wear resistance than their
distinct deformation mechanisms and the absence of dis- respective crystallized states.6,7 On the contrary, there
location-mediated slip, which is the major plastic defor- are reports indicating that the wear performance of zir-
mation mechanism in conventional crystalline alloys. conium (Zr)-based BMGs are significantly inferior to
that of the crystallized alloy8 and traditional crystalline
alloys.9 Fu et al.10 have examined the unlubricated slid-
a)
Address all correspondence to this author. ing characteristics of Zr-based BMGs and have observed
e-mail: ramu@materials.iisc.ernet.in that extensive material flow took place under sliding
DOI: 10.1557/JMR.2007.0048 conditions, indicating that plasticity at the contact plays
264 J. Mater. Res., Vol. 22, No. 2, Feb 2007 © 2007 Materials Research SocietyH.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass
an important role in the wear/friction characteristics of Oliver-Pharr method,15 to extract the elastic modulus and
the BMGs. the hardness. Microhardness measurements were per-
It is well known that plasticity in amorphous metals is formed using a Vickers diamond indenter and a 500-g
controlled by the free volume. For example, it has been load with a dwell time of 15 s.
shown that free-volume reduction [through structural re-
laxation anneals below the glass transition temperature
(Tg)] leads to severe embrittlement due to the reduced B. Wear testing
propensity for shear-band formation.11,12 Also, the crys- A high-frequency reciprocating rig (HFRR) instru-
tallization induces embrittlement in a tough, as-cast, ment (PCS Instruments, London, UK) with “ball-on-
amorphous alloy.13,14 Thus, both structural relaxation plate/disc” geometry was used for measuring the friction
and crystallization may yield different deformation and coefficient and wear volume loss (W). The 6-mm-
flow behavior. However, a systematic study that explores diameter WC ball (Salem Specialty Balls, Salem, MA)
the relationship between the plasticity and its wear char- and the 1.5-mm-thick BMG specimens were fixed tightly
acteristics has not yet been reported in the literature. in specimen holders that have appropriate recesses ma-
Such understanding will facilitate the design of BMGs chined into them. The upper specimen holder has a load-
from a wear resistance point-of-view. Therefore, the ob- ing pin to which a dead weight load was applied. An
jective of the present study was to conduct a detailed electromagnetic vibrator drives the upper specimen
study on the friction and wear characteristics of a Zr- holder back and forth via a pushrod. A linear variable
based BMG in which the propensity for shear banding differential transducer (LVDT) attachment to this allows
was controlled systematically by annealing. Nanoinden- for monitoring of the motion of the pushrod, and adjusts
tation was used to characterize the changes in intrinsic the frequency and amplitude of the vibrator. A piezoelec-
plasticity in the BMG due to annealing and was corre- tric force transducer connected to the lower block mea-
lated with wear rates. sures the friction force generated between the ball and the
specimen as they slide relative to each other. Prior to the
wear tests, all of the samples were cleaned with toluene,
II. MATERIAL AND EXPERIMENTS acetone, and hexane successively to have a clean surface.
The initial arithmetic average roughness of the samples,
A. Heat treatment and characterization after these treatments and estimated using the profilome-
A 14-mm-diameter rod of the Zr-based BMG, with a ter, was ∼0.8 m, whereas it is approximately 艋2 m
commercial designation LM-1 was obtained from Liq- after wear testing. Note that the latter would be signifi-
uidmetal Technologies Ltd. (Lake Forest, CA). The cantly affected by the presence of wear debris and oxide
nominal chemical composition of the alloy (in at.%) was flakes that form during the wear process.
Zr42Ti15.5Cu14.5 Ni3.5Be24.5. Discs of ∼1.5 mm thickness Unless otherwise mentioned, all of the wear tests were
were electrodischarge-machined from the rod. The pri- conducted with an applied load of 100g for 2 h at a
mary objective of this work is to examine how structural frequency of 50 Hz. The sliding distance for each pass
relaxation changes the wear characteristics of the amor- was kept constant at 2 mm. A resistance heater attached
phous alloys. For this purpose, the BMG discs were an- to the lower specimen holder allowed for control of the
nealed for 1 h in a flowing argon environment at 400, temperature through a thermocouple that was located
469, 531, and 687 K, which correspond to 0.62, 0.73, close to the specimen. All of the tests reported in this
0.83, and 1.07 times the Tg (643 K), respectively. The article were conducted at 303 K.
last temperature, being above the Tg, induces crystalliza- The HFRR instrument is capable of measuring the
tion of the BMG, whereas the first three temperatures of friction coefficient, temperature, and the electrical con-
annealing induce structural relaxation. The as-cast and tact resistance ⍀ between the specimen and the ball. The
annealed samples, polished to a 0.5-m finish, were electronics unit applies a bias voltage through the ball,
characterized by optical microscopy, x-ray diffraction measures ⍀, and scales it between 0% and 100%. When
(XRD) and differential scanning calorimetry (DSC). the ⍀ was zero, metal-to-metal contact exists. When
The variation in mechanical properties, specifically there was no electrical contact between the specimen and
elastic modulus and hardness, of the BMG with the heat the ball, ⍀ ⳱ 100%. Thus, ⍀ gives a qualitative indica-
treatment temperature were assessed by nanoindentation tion of the presence or absence of an oxide film.
experiments using a Hysitron Tribometer (Hysitron Inc, W was measured by scanning the wear scars using a
Minneapolis, MN) with a Berkovich tip. All of the tests laser profilometer with a depth resolution of 1 m and a
were conducted with a maximum load of 9 mN, with a spatial resolution of 30 m. For each case, at least five
prescribed loading rate of 0.36 mN/s with a 10-s hold repeat experiments were conducted, and the average val-
time at the peak load. The loads, P, versus the depth-of- ues are reported herein. Scanning electron microscopy
penetration, h, curves were analyzed by employing the (SEM) of the wear scars was conducted to understand the
J. Mater. Res., Vol. 22, No. 2, Feb 2007 265H.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass wear mechanisms. Energy-dispersive spectroscopy was performed, where necessary, to identify the nature of the scales formed on the surface. III. RESULTS A. Structural relaxation Inspection of the polished cross section of the as-cast alloy revealed that the BMG was not fully amorphous, but contained crystalline regions with needle-shaped and dendritic particles in the amorphous matrix. The volume fraction of these crystalline particles was very small (
H.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass
TABLE I. Summary of the DSC experiments.
Enthalpy at the Glass transition Enthalpy at the Peak crystallization
Annealing glass transition, temperature, crystallization, temperature,
temperature (°C) ⌬Hg (J/g) Tg (°C)a ⌬Hx (J/g) Tx (°C)a
As-cast 6.79 ± 0.09 370.3 ± 1.8 80.7 ± 2.2 464.4 ± 1.6
127 (0.62 × Tg) 6.71 ± 0.25 370.6 ± 1.7 84.3 ± 5.4 464.4 ± 1.9
196 (0.73 × Tg) 6.42 ± 0.12 369.6 ± 1.2 83.9 ± 0.8 463.5 ± 0.3
258 (0.83 × Tg) 5.93 ± 0.03 372.4 ± 1.6 85.0 ± 1.6 463.3 ± 1.0
414 (1.07 × Tg) … … … …
Values given as mean ± SD.
a
Inflection temperature.
(profuse shear-band activity) mode. However, Greer
et al.18 have suggested that the smoothening of the P-h
curves was an instrumental artifact rather than a transi-
tion in flow behavior.
C. Friction and wear properties
Figure 3 shows a typical plot of the ⍀ and the friction
coefficient as a function of time of the reciprocation of
the WC ball on the as-cast BMG disc. Note that ⍀ re-
flects the electrical resistance of the contact, therefore
zero or low ⍀ means metal-to-metal contact, whereas a
high ⍀ value can be inferred to be due to the formation
of a thick, nonconducting tribolayer. The details of this
transition are discussed in Sec. III. D. As seen from Fig.
3, the value initially was low (∼0.62), but after about
10 min, a sharp transition was seen, with increasing to
about 1. Further, the noise in the variation with time
increased markedly. A concomitant transition in ⍀, from
FIG. 2. Typical load, P-h curves obtained on the as-cast and annealed
a near-zero value to 100% was also noted. This behavior
alloys. The inset highlights the reduced susceptibility for serrated flow was observed in all of the different annealing conditions
(a result of the shear-band formation and propagation) with increasing tested. Average values of before and after the transition
annealing temperature. as well as the average transition times tT are listed in
Table II. Here, tT is defined as the time at which the ⍀
from 6.91 GPa in the as-cast alloy to 8.49 GPa in the crosses 50% for the first time.
alloy annealed at 531 K, a marked increase (to 10.8 GPa) The data presented in Table II show that the friction
occurred in the crystallized alloy. coefficient after the transition a was, within the experi-
The P-h curves obtained during the nanoindentation of mental scatter, independent of the heat treatment with a
amorphous alloys generally tended to be serrated, as seen value of ∼0.95. This was also true for the friction coef-
from the response of the as-cast alloy shown in Fig. 2. It ficient before the transition b for the structurally relaxed
has been well established that the serrations are a result glasses (i.e., those annealed at temperatures below the
of the nucleation and propagation of shear bands, the Tg), with a value of ∼0.62. Likewise, the tT values were
carriers of plasticity in amorphous metals. Since defor- also similar. In the case of the crystallized BMG, a
mation through shear banding is inhomogeneous in na- slightly higher value for b (∼0.70) was noted. However,
ture, each shear-band activity is associated with a dis- the tT in this case was much smaller (20 s) when com-
crete displacement burst, which resulted in a serration in pared to the tT of ∼600 s measured in the as-cast and
the loading part of the P-h curve.16 With increasing an- relaxed glasses. This implies that the transition in the
nealing temperature, we noted that the P-h curves got crystallized glass was essentially instantaneous.
smoother, with the fully crystallized alloy showing no A typical profile, measured using a laser-scanning pro-
serrations at all, which is highlighted through the inset in filometer, of the wear scar produced after 2 h of testing
Fig. 2. Schuh et al.17 have argued that P-h curves of is shown in Fig. 4(a). As seen, the wear scar is symmetri-
amorphous alloys are smoother at higher loading rates cal and uniform. A line across the middle of the scar [Fig.
because of a transition in flow behavior from a hetero- 4(b)] indicates a cylindrical groove. More importantly, it
geneous (discrete shear-band activity) to homogenous shows the edges to be flat, implying that there is no
J. Mater. Res., Vol. 22, No. 2, Feb 2007 267H.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass
TABLE II. Summary of the indentation, friction, and wear experiments.
Reduced Maximum Micro-Vickers Friction Friction
Annealing modulus, depth, Nanohardness hardness Transition coefficient coefficient
temperature (°C) Er (GPa) hmax (nm)a (GPa) (VHN) time, tT (s) before tT, b after tT, a W (×106 m3)
As-cast 106.2 ± 0.6 254.5 ± 2.5 6.9 ± 0.2 753.4 ± 7.3 550 ± 131 0.62 ± 0.02 0.93 ± 0.05 60.6 ± 50.1
127 110.3 ± 3.2 250.6 ± 5.9 7.1 ± 0.4 752.8 ± 8.9 650 ± 151 0.63 ± 0.01 0.95 ± 0.01 80.2 ± 25.6
196 104.5 ± 4.6 250.1 ± 6.7 7.4 ± 0.4 740.6 ± 4.4 605 ± 105 0.62 ± 0.02 0.95 ± 0.01 58.9 ± 15.3
258 104.2 ± 0.6 242.7 ± 0.8 8.5 ± 0.16 740.8 ± 4.7 517 ± 221 0.61 ± 0.01 0.96 ± 0.05 40.0 ± 19.9
414 119.2 ± 3 220.0 ± 3.4 10.8 ± 0.7 902.2 ± 18.8 20 0.7 ± 0.03 0.95 ± 0.02 5.4 ± 1.0
Values given as mean ± SD.
a
Depth at the maximum load.
confirmed that the scales observed indeed contained
large concentrations of oxygen [Fig. 6(d)] with the pres-
ence of a strong oxygen peak, whereas the spectrum
obtained from other areas did not show an oxygen peak
[Fig. 6(c)]. Among the alloying elements present in the
BMG, Zr, titanium (Ti), and beryllium (Be) are prone to
easy oxidation. Fu et al.10 reported similar layers con-
taining large amounts of oxygen due to sliding wear in
air, whereas they were absent when tested in a vacuum.
The tribolayer produced in air may indeed be some
mixed oxide; however, further studies on its chemical
composition or structure (amorphous or crystalline) were
not performed in this study. Hence, it is referred to sim-
ply as a tribolayer.
The tribolayers appear to crack extensively, with the
periodically spaced cracks perpendicular to the recipro-
cating ball direction, as shown in Fig. 7(a). These
cracked films eventually peel off, due to the action of the
FIG. 3. Typical variation of ⍀ and as a function of time. Note that
⍀ ⳱ 0% implies no tribolayer, whereas ⍀ ⳱ 100% implies a com- tangential stress applied by the sliding WC ball, exposing
plete loss of contact. the unoxidized BMG. Such a peeled-off region is shown
in Fig. 7(b), with EDAX values obtained from this region
showing the absence of oxygen. Figure 7(b) also indi-
pileup of material (due to plastic flow of the material
cates that some bulk material being chipped off along
underneath the WC ball) around the edges. The W during
with the scale (i.e., the tribolayer when it peels off) does
the wear process is estimated by a three-dimensional in-
not debond cleanly along the oxide/bulk interface.
tegration process and is listed in Table II. Dry sliding
To investigate the microscopic origins of the step-like
wear in most metallic materials was observed to obey
transition in ⍀ and variations with time (Fig. 4), wear
Archard’s law, W ⳱ K(SN/H), where S is the total sliding
tests were interrupted after 300 s (recall that the transition
distance, N is the normal load, and K is the wear coeffi-
time is around 600 s), and the resulting scars were ex-
cient. The least-square fit of the above equation is used
amined. The low- and high-magnification images ob-
for the data plotted in Fig. 5. While considerable scatter
tained from these scars are shown in Figs. 8(a) and 8(b),
exists in the data (especially in the as-cast alloy), a gen-
respectively. Clearly, no scales could be seen, although
eral trend of a decrease in W with an increase of nano-
indications of the early stages of tribolayer formation
hardness (annealing temperature) can be seen.
could be seen. Also, the surface morphology was much
smoother, indicating that the mechanism of material re-
D. Wear mechanisms moval was different in this regime. Sliding wear tests
The worn surfaces were examined by SEM to under- conducted in vacuum by Fu et al.10 show that a soft layer
stand the microscopic reasons for the wear of the BMG. develops during sliding, with a low friction coefficient,
Figure 6(a) is a lower magnification image of the wear and that the wear rate is lower than when sliding in air.
scar in the as-cast BMG, showing abrasive wear marks. These observations, in conjunction with those made in
A higher magnification image [Fig. 6(b)] shows that the the present study, indicate that the micromechanism of
worn surfaces were partially covered with tribofilms. En- wear in the BMG is abrasive wear, with the high-oxygen-
ergy dispersive x-ray spectroscopy (EDAX) analyses containing tribolayers acting as the abrasive medium
268 J. Mater. Res., Vol. 22, No. 2, Feb 2007H.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass
FIG. 4. (a) Topography of the wear scar and (b) a line scan across the scar showing the profile of the scar.
any features and appeared to be very smooth. Most im-
portantly, no tribolayers could be seen. EDAX spectra
obtained from regions marked 1, 2, and 3 in Fig. 9(b)
were identical. An example is shown in Fig. 9(c).
IV. DISCUSSION
A gradual reduction in the wear rate with increasing
annealing temperature can be seen in Fig. 5. This reduc-
tion in wear rate was due to the gradual increase in the
hardness with the annealing temperature. Van den Beu-
kel and Sietsma19 developed a model that described the
functional form of the DSC curves based on the free-
volume theory. According to their model, for every su-
percooled liquid, there exists an equilibrium free-
volume, vfe, at a given temperature. In the as-cast alloy,
an excess amount of free volume, vf > vfe, is frozen-in
due to the nonequilibrium processing conditions, which
FIG. 5. Variation of the W measured during high-frequency- during continuous heating in a DSC experiment annihi-
reciprocating wear tests, as a function of H of the BMG. lates and approaches vfe. The reduction in vf gives rise to
heat release ⌬H, when the glass sample is heated in a
(because of their high hardness and low toughness, the lat- DSC. It was experimentally verified by Slipenyuk and
ter facilitating comminution of the films once they peel off). Eckert20 that ⌬Hg ⬀ ⌬vf. Since the free volume gets
The SEM images obtained from the BMG samples that reduced during the structural relaxation anneals, ⌬Hg
were annealed at temperatures below the Tg were very gradually decreases with increasing annealing tempera-
similar, and were consistent with the ⍀, , and W data ture (Table I), implying a systematic reduction in the
given in Table II, indicating that the wear mechanisms do ability of the BMG to accommodate plastic deformation.
not change significantly with relaxation heat treatments The reduction in the serration step sizes with increasing
below the Tg. Wear scars in the crystallized BMGs (i.e., annealing temperature indicates that, at the nanoscale,
the samples annealed at 687 K) showed a different wear increasing the structural relaxation annealing tempera-
morphology altogether. The low- and high-magnification tures leads to a reduction of free volume. This reduction
SEM micrographs obtained from the wear scars in the of free volume diminishes the ability of the BMG to
crystallized alloy are shown in Fig. 9. In contrast to those accommodate plastic deformation, which is consistent
shown in Figs. 7(a) and 7(b), respectively, which were with earlier reports on annealing induced embrittlement
obtained at the same magnifications but in the as-cast in BMGs.11,12
alloy, wear scars in the crystallized alloy are devoid of Interestingly, the microhardness of the BMG appeared
J. Mater. Res., Vol. 22, No. 2, Feb 2007 269H.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass FIG. 6. (a) A lower magnification SEM image of the wear scar in the as-cast BMG, showing abrasive wear marks. (b) A higher magnification image showing a tribolayer formed on the wear surface. EDAX spectrums obtained from (c) the region marked 1 and (d) the region marked 2 in (b). Note the presence of a strong oxygen peak in the latter, confirming that the scale indeed contains a large amount of oxide. to be relatively insensitive to the relaxation anneals (with pear to alter the wear performance/friction coefficient Vickers hardness number data ranging between 740 and significantly. One possibility that needs consideration is 753), whereas a marked rise was seen in the crystallized the role of localized frictional heating and flow soften- alloy, similar to that seen in nanohardness. The origin for ing24 of the BMG. If the temperature rise at the surface the differences in the nanohardness and microhardness of layer due to vigorous scratching of the BMG against the the alloys that were annealed below the Tg is discussed as hard surface is significant, it could lead to a transition follows. The deformation underneath an indenter is in- from a heterogeneous to a homogeneous mode of trinsically stable (i.e., the contact area always increases deformation.25 This in turn can drastically alter the unless it is interrupted by cracking in the case of brittle wear performance. However, the average sliding speed materials being indented with sharp indenters) with an (∼0.2 mm/s) is not high enough to give large temper- increasing depth of penetration of the indenter. The stress ature increases from friction and hence to alter defor- singularity at the tip of the Vickers indenter induces plas- mation mechanisms. tic deformation through shear banding, even in the an- The experimental results and observations indicate nealed material, instead of cracking.21 It is possible that that the predominant wear mechanism in the as-cast and the stress singularity at the indenter tip is mitigated structurally relaxed Zr-based BMG investigated in this through the homogeneous deformation of the glass, work was oxidation of the alloy. This is not surprising which is induced by the viscoplasticity that is generally considering the high chemical affinity of Zr, Ti, and Be dominant at low strain rates, followed by shear-band nu- with oxygen to form respective oxides. The formation of cleation. The high stresses that prevail under the Vicker’s high-oxygen-containing tribolayers and their subsequent indenter (estimated to be equal to 35.8%) could promote debonding under the action of the reciprocating WC ball shear banding.22 While the nanoindentation is highly constitutes the main mechanism of wear in the as-cast sensitive to relatively small changes (reflected in the and structurally relaxed alloys. Then, the tT is considered P-h curves as well as in the nanohardness), microin- to be the time required for sufficiently thick tribolayers to dentation measures the response of the BMG averaged form. The residual stress in the tribofilm (that arises due over a significantly larger volume and hence is insensi- to the differences in the molar volumes of the oxide layer tive to the underlying subtle changes in micromecha- and that of the substrate that is oxidized) increases with nisms of deformation. the increasing tribolayer thickness. This, coupled with It has been proposed by previous studies10,23 that the the applied frictional force, appears to determine the wear/friction properties could be affected by plastic de- process of delamination. formation (i.e., shear-banding propensity), since the wear An interesting observation in the present study is that process essentially involves deformation. Based on the the crystallized alloy exhibited minimal scale formation, present investigation, structural relaxation does not ap- while the as-cast and structurally relaxed alloys showed 270 J. Mater. Res., Vol. 22, No. 2, Feb 2007
H.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass
FIG. 8. (a) Low-magnification and (b) high-magnification images of
the wear scars obtained after 300 s (before the transition) from the
as-cast BMG.
has been suggested that the formation of crystalline
Zr2Cu and Zr2Al was responsible for improving the
oxidation resistance due to their higher oxidation
resistance and promotion of the development of Al2O3
and oxides of copper. These results indicate that the
crystalline phase of Zr2Cu is much more resistant to
oxidation than the amorphous state. Paljevia et al.28 re-
ported that the oxidation reaction onset temperature of
ZrAl2 is ∼700 K.
FIG. 7. (a) SEM micrograph showing extensive cracking of the tri-
bolayer formed on the wear surface. (b) Micrograph showing a region Following annealing above the Tg, the oxidation resis-
where the layer has peeled off. (c) EDAX spectrum from the circled tance remarkably improved with crystallization so that
region of (b) showing the absence of oxygen. film formation was rarely observed, possibly due to its
insignificant amount. As the result of their superior
extensive scale formation. The oxidation behaviors of oxidation resistance, the crystallized alloys were worn
Zr-based amorphous alloys have been reported by sev- without the intervention of the oxidation under testing
eral researchers.26–29 It has been reported by Wu et al.27 conditions in the present study. However, there are contra-
that the oxidation rate of Zr–Al–Cu metallic glass speci- dicting reports on the effect of crystallization on oxida-
mens was higher than that of annealed samples above the tion. Kai et al.29 reported that the oxidation rate of Zr–
Tg and that the oxidation resistance of the specimen de- Cu–Al–Ni BMG is lower than that of the crystalline alloy
pended on the amount of crystalline precipitates. It composed of Zr2Cu, Zr2Ni, and ZrAlCu2 at temperatures
J. Mater. Res., Vol. 22, No. 2, Feb 2007 271H.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass
above 698 K. It is noteworthy that the oxidation resis-
tance of Zr-based amorphous alloys strongly depends on
the chemistry as well as the degree of crystallinity.
Therefore, the effect of crystallization/chemistry on the
oxidation behavior of the Zr-based BMG should be con-
sidered when these alloys are designed for use under
wear/friction circumstances where optimized stability
against oxidization is required. Although it is possible for
the surface layers of the crystalline BMG to reamorphize
during sliding wear (which in turn oxidize in a manner
similar to that of amorphous alloys), it does not appear to
be the case in the present context. The combination of
relatively low sliding velocities and loads may be the
reason for this.
V. SUMMARY
An experimental investigation into the high-frequency,
dry-sliding wear behavior of a Zr-based BMG was con-
ducted. The BMG was systematically annealed at differ-
ent temperatures (both below and above the Tg), and the
role of structural relaxation on the wear resistance of the
BMG was investigated. On the basis of the experimental
results the following conclusions can be drawn.
(1) The wear behavior of both the as-cast and relaxed
glasses was dominated by the oxidation of the surface
layers. Formation of the oxide layers and their subse-
quent peel-off constituted the main mechanism of wear
for the experimental conditions of this study. A sharp
transition in the oxide film thickness complemented by a
sharp increase in after ∼600 s were observed.
(2) With increasing annealing temperature (below the
Tg), the nanohardness of the BMG increased. This hard-
ness increase was proposed to be due to the reduced
propensity for shear-band formation, caused by the an-
nihilation of the free volume present in the BMG. A
reasonable correlation between W and nanohardness was
identified.
(3) The crystallized alloy has considerably higher
wear resistance and also higher hardness. In contrast to
the as-cast and structurally relaxed alloys, the wear
mechanism was predominantly abrasive wear in the crys-
tallized alloy. The formation of oxide scales was insig-
nificant, indicating superior oxidation resistance in crystal-
lized alloy to the as-cast and structurally relaxed glasses.
ACKNOWLEDGMENTS
The authors acknowledge S. Bellamare, H. Jun,
J. Krylowski, C. Marzinsky, A. Schwartzman, T. Smith,
and M.J.N.V. Prasad for their assistance with different
FIG. 9. SEM micrographs of the wear scars in the crystallized alloy experiments reported in this article. We are grateful to
showing (a) lower magnification and (b) higher magnification
Professor S. Suresh of Massachusetts Institute of Tech-
views. EDAX spectrums obtained from regions marked 1, 2, and 3
were identical, as shown in (c). Compare these images with those nology and Dr. M. Webster of ExxonMobil Research and
obtained at the same magnification and shown in Figs. 7(a) and 7(b), Engineering Company (EMRE) for their advice and sup-
respectively. port. U. Ramamurty acknowledges the support he has
272 J. Mater. Res., Vol. 22, No. 2, Feb 2007H.W. Jin et al.: Reciprocating wear mechanisms in a Zr-based bulk metallic glass
received from ExxonMobil Research and Engineering 15. W.C. Oliver and G.M. Pharr: An improved technique for deter-
Company for a summer visit to their Corporate Strategic mining hardness and elastic modulus using load and displacement
sensing indentation experiments. J. Mater. Res. 7, 1564 (1992).
Research Labs, which facilitated the research reported in
16. C.A. Schuh, T.G. Neih, and Y. Kawamura: Rate dependence of
this article. Constructive comments and suggestions for serrated flow during nanoindentation of a bulk metallic glass.
the overall improvement of the article by Dr. S. J. Mater. Res. 17, 1651 (2002).
Karthikeyan as well as anonymous reviewers are grate- 17. C.A. Schuh, A.C. Lund, and T.G. Nieh: New regime of homoge-
fully acknowledged. neous flow in the deformation map of metallic glasses: Elevated
temperature nanoindentation experiments and mechanistic mod-
eling. Acta Mater. 52, 5879 (2004).
REFERENCES 18. A.L. Greer, A. Castellero, S.V. Madge, I.T. Walker, and
J.R. Wilde: Nanoindentation studies of shear banding in fully
1. W. Klement, R.H. Willens, and P. Duwez: Noncrystalline struc- amorphous and partially devitrified metallic alloys. Mater. Sci.
ture in solidified gold-silicon alloys. Nature 187, 869 (1960). Eng., A 375–377, 1182 (2004).
2. A. Inoue: Bulk amorphous and nanocrystalline alloys with high 19. A. Van Den Beukel and J. Sietsma: The glass transition as a free
functional properties. Mater. Sci. Eng., A 304–306, 1 (2001). volume related kinetic phenomenon. Acta Metall. Mater. 38, 383
3. S. Schneider: Bulk metallic glasses. J. Phys.: Condens. Matter 13, (1990).
7723 (2001). 20. A. Slipenyuk and J. Eckert: Correlation between enthalpy change
4. W.H. Wang, C. Dong, and C.H. Shek: Bulk metallic glasses. and free volume reduction during structural relaxation of
Mater. Sci. Eng., R 44, 45 (2004). Zr55Cu30Al10Ni5 metallic glass. Scripta Mater. 50, 39 (2004).
5. F. Spaepen: A microscopic mechanism for steady state in homo- 21. M.N.M. Patnaik, R. Narasimhan, and U. Ramamurty: Spherical
geneous flow in metallic glasses. Acta Metall. 25, 407 (1977). indentation response of metallic glasses. Acta Mater. 52, 3335
6. M. Anis, W.M. Rainforth, and H.A. Davies: Wear behavior of (2004).
rapidly solidified Fe68Cr18Mo2B12. Wear 172, 135 (1994). 22. U. Ramamurty, S. Jana, Y. Kawamura, and K. Chattopadhyay:
7. T. Gloriant: Microhardness and abrasive wear resistance of me- Hardness and plastic deformation in a bulk metallic glass. Acta
tallic glasses and nanostructured composite materials. J. Non- Mater. 53, 705 (2005).
Cryst. Solids 316, 96 (2003). 23. R. Bhowmick, R. Raghavan, K. Chattopadhyay, and U. Ramamurty:
8. G. Li, Y.Q. Wang, L.M. Wang, Y.P. Gao, R.J. Zhang, Z.J. Zhan, Plastic flow softening in a bulk metallic glass. Acta Mater. 54,
L.L. Sun, J. Zhang, and W.K. Wang: Wear behavior of bulk 4221 (2006).
Zr41Ti14Cu12.5Ni10Be22.5 metallic glass. J. Mater. Res. 17, 1877
24. X. Fu, M.L. Falk, and D.A. Rigney: Sliding behavior of metallic
(2002).
glass: Part II. Computer simulations. Wear 250, 420 (2001).
9. P.J. Blau: Friction and wear of a Zr-based amorphous alloy under
dry and lubricated conditions. Wear 250, 431 (2001). 25. C.A. Schuh and T.G. Neih: A survey of instrumented indentation
10. X. Fu, T. Kasai, M.L. Falk, and D.A. Rigney: Sliding behavior of studies on metallic glasses. J. Mater. Res. 19, 46 (2004).
metallic glass: Part I. Experimental investigations. Wear 250, 409 26. S.K. Sharmal, T. Strunskus, H. Ladebusch, and F. Faupel:
(2001). Surface oxidation of amorphous Zr 65 Cu 17.5 Ni 10 A l7.5 and
11. P. Murali and U. Ramamurty: Embrittlement of a bulk metallic Zr46.75Ti8.25Cu7.5Ni10Be27.5. Mater. Sci. Eng., A 304–306, 747
glass due to sub-Tg annealing. Acta Mater. 53, 1467 (2005). (2001).
12. U. Ramamurty, I.M.L. Lee, J. Basu, and Y. Li: Embrittlement of 27. Y. Wu, T. Nagase, and Y. Umakoshi: Effect of crystallization
a bulk metallic glass due to low temperature annealing. Scripta behavior on the oxidation resistance of a Zr–Al–Cu metallic glass
Mater. 47, 107 (2002). below the crystallization temperature. J. Non-Cryst. Solids 352,
13. N. Nagendra, U. Ramamurty, T.T. Goh, and Y. Li: Effect of 3015 (2006).
crystallinity on the impact toughness of a La-based bulk metallic 28. M. Paljevia: High-temperature oxidation behaviour in the Zr-Al
glass. Acta Mater. 48, 2603 (2000). system. J. Alloys Compd. 204, 119 (1994).
14. J. Basu, N. Nagendra, Y. Li, and U. Ramamurty: Microstructure 29. W. Kai, H.H. Hsieh, T.G. Nieh, and Y. Kawamura: Oxidation
and mechanical properties of partially-crystallized la-based bulk behavior of a Zr–Cu–Al–Ni amorphous alloy in air at 300–425 °C.
metallic glass. Philos. Mag. 83, 1747 (2003). Intermetallics 10, 1260 (2002).
J. Mater. Res., Vol. 22, No. 2, Feb 2007 273You can also read
