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 Society
H.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 265
H.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 267
H.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 2007
H.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 269
H.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 271
H.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 2007
H.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 273
You can also read