Wear Behavior of Mechanically Alloyed Ti-Based Bulk Metallic Glass Composites Containing Carbon Nanotubes

The present paper reports the preparation and wear behavior of mechanically alloyed Ti-based bulk metallic glass composites containing carbon nanotube (CNT) particles. The differential scanning calorimeter results show that the thermal stability of the amorphous matrix is affected by the presence of CNT particles. Changes in glass transition temperature (Tg) and crystallization temperature (Tx) suggest that deviations in the chemical composition of the amorphous matrix occurred because of a partial dissolution of the CNT species into the amorphous phase. Although the hardness of CNT/Ti50Cu28Ni15Sn7 bulk metallic glass composites is increased with the addition of CNT particles, the wear resistance of such composites is not directly proportional to their hardness, and does not follow the standard wear law. A worn surface under a high applied load shows that the 12 vol. % CNT/Ti50Cu28Ni15Sn7 bulk metallic glass composite suffers severe wear compared with monolithic Ti50Cu28Ni15Sn7 bulk metallic glass.


Introduction
Ever since the first bulk metallic glasses (BMGs) were synthesized in 1989 [1], considerable progress has been made in the preparation of BMGs.A great number of multicomponent metallic glass alloys with strong glass-forming ability and high thermal stability have been discovered, such as Ti, Zr, Fe, Pd, Mg, Cu, Ni, and other metallic glass alloy systems [2][3][4].BMGs have attracted much attention because of their unique properties, such as their remarkable levels of strength, hardness, corrosion resistance, and wear resistance.However, BMGs usually exhibit little overall room temperature plasticity because of the formation of highly localized shear banding, which may result in catastrophic failures [5].Attempts have been made to enhance the ductility of BMGs by reducing their dimensions, thereby mitigating the catastrophic effects of instabilities on the material behavior [6].A decrease in the aspect ratio of the sample (i.e., samples with small dimensions in the directions perpendicular to the applied load) generates an interaction in the form of deflected shear bands with a wavy shape [7].Jiang et al. [8] conducted compression tests on specimens with different height (h) to diameter (d) (h/d) ratios for Zr 52:5 Cu 17:9 Ni 14.6 Al 10 Ti 5 .Plasticity reached almost 80% for the specimens with an h/d ratio of 0.5, which was much higher than the 2% plasticity of specimens with an h/d ratio of 2. The researchers attributed these differences to the effect of specimen geometry, in which the platen can hinder the excessive propagation of shear bands from the premature fracture.Metallic glass matrix composites have been developed to improve the mechanical properties of such single-phase metallic glass alloys.Such glasses are ex situ, reinforced by second-phase particulates, fibers, or in situ formed ductile phase precipitates [9].It was discovered that the existence of ceramic or insoluble metallic particles inside the glass matrix can suppress the propagation of shear bands, thereby increasing the toughness and ductility of the metallic glass matrix composite.
Ti-based BMGs were first reported in a Ti-Zr-Ni-Be system in 1994 [10].Many techniques have been applied successfully to prepare Ti-based BMGs, but most research efforts and industrial inquiries have focused on different methods of implementation for rapid solidification [11].An alternative method is using solid-state amorphization processing such as mechanical alloying (MA) to prepare amorphous powders that are suitable for further compaction and densification.This method facilitates the introduction of reinforcing particles into the glassy matrix.As previous investigations have demonstrated, MA has been successfully applied for the preparation of amorphous Cu-and Zr-based composite powders [12,13].Carbon nanotubes (CNTs) are endowed with exceptionally high material properties very close to their theoretical limits, such as electrical and thermal conductivity, strength, stiffness, and toughness.Therefore, CNTs are the latest reinforcement materials employed in composites, owing to their possessing structural properties stronger than those of any metal [14].Kim et al. [15] reported that carbon fiber-reinforced BMG composites can be successfully prepared by infiltrating liquid Zr-Ti-Cu-Ni-Be into carbon fiber bundles.Shortly after this, Bian et al. [16][17][18][19] synthesized CNT-reinforced BMG composites, which exhibit excellent wave absorption properties.
The advantages of BMGs render them most suitable for application in small structural parts.For industrial applications, detailed information about wear performance is required, but only a few previous studies have reported on this.The first reported wear test was a pin-on-disk study of the abrasive wear of Pd-Si-Cu metallic glasses [20].Gloriant classified Zr-based bulk glassy alloy in the hardened alloy group with regard to the relationship between its hardness and wear resistance [21]; however, despite increasing demand, few studies have been published on the mechanism of wear in bulk glassy alloys, and therefore numerous unique characteristics are yet to be defined.
Even though BMGs and CNTs possess many excellent properties, a literature survey indicated that no report exists on the formation of a Ti-based BMG composite powder containing CNT particles using the MA process.For the present study, amorphous Ti 50 Cu 28 Ni 15 Sn 7 powders both with and without the addition of CNT particles were prepared using MA.Subsequent consolidation of as-milled powders was performed, and the mechanical properties of the compacts were evaluated by Vickers microhardness and wear property testing.

Experimental
Elemental powders of Ti (99.7%, <325 mesh), Cu (99.9%, <325 mesh), Ni (99.9%, <300 mesh), Sn (99.999%, 100 mesh), and CNTs (99.999%, <325 mesh) were weighed to yield the desired Ti 50 Cu 28 Ni 15 Sn 7 or CNT/Ti 50 Cu 28 Ni 15 Sn 7 composition.The mixture of elemental metallic powders with the nominal composition of Ti 50 Cu 28 Ni 15 Sn 7 (in at.%) was mechanically alloyed with or without the addition of 4, 8, and 12 vol.% CNT powders.Milling was performed in a SPEX 8016 shaker ball mill (SPEX SamplePrep, Metuchen, NJ, USA) under an Ar atmosphere.Specific details of the MA process are described elsewhere [12].The as-milled composite powders were consolidated in a vacuum hot pressing machine to prepare bulk samples, each of which had a 10 mm diameter and a 2 mm thickness.Hot pressing was performed at 723 K under a pressure of −1.20 GPa and a high-vacuum (approximately 10 −3 Torr) environment.
The structures of both the as-milled powders and bulk samples were analyzed using a Siemens D 5000 X-ray diffractometer (Siemens AG, Karlsruhe, Germany) and a Hitachi-4100 scanning electron microscope (SEM) (Hitachi High-Technologies Corporation, Tokyo, Japan).Thermal analysis was investigated using a Dupont 2000 differential scanning calorimeter (DSC) (TA instruments, New Castle, PA, USA), in which the sample was heated from room temperature to 900 K under a purified Ar atmosphere at a constant heating rate of 40 K/min.In addition, the Vickers microhardness tests of consolidated BMG specimens were measured with a Matsuzawa MXT50-UL machine (Matsuzawa Co., Ltd., Akita-shi, Japan) using a static load of 0.5 N. The friction and wear behaviors of Ti 50 Cu 28 Ni 15 Sn 7 composite specimens both with and without CNT particles were tested under unlubricated conditions on a block-on-roller TE53 wear tester from Plint Co., Kingsclere, UK.The roller material was hardened and tempered AISI O1 tool steel in 65 HRC (832 HV in microhardness).The wear test was performed under selected contact pressures of 5, 10, 15, and 20 N at a revolution speed of 200 rpm.The weight loss of the specimen was recorded using an OHAUS AP250D-0 electronic balance (OHAUS Co., Parsippany, NJ, USA).Samples of JIS SKH-2 high-speed tool steel with a hardness of 24 and Ti 50 Cu 28 Ni 15 Sn 7 crystalline alloy with a hardness of 29 were also tested for comparative purposes.

Results and Discussion
Figure 1 shows the X-ray diffraction (XRD) patterns of Ti 50 Cu 28 Ni 15 Sn after the addition of 4-12 vol.% CNT particles after 8 h of milling.For the plain Ti 50 Cu 28 Ni 15 Sn alloy and the matrix with 4 vol.% CNT particles, only a broad diffraction peak appears around 2θ = 42 • , indicating the formation of fully amorphous powders.This may be attributed to the small volume fraction of CNT particles and their small crystalline size.The findings of the present study are similar to the observations regarding the preparation of the Al 2 O 3 /NiAl intermetallic-matrix composite reported by Lin et al. [22], in that, for small Al 2 O 3 additions in mechanically alloyed Ni-Al alloys, no Al 2 O 3 phase was detected by XRD after 10 h of milling.In the case of 12 vol.% CNT powders, crystalline TiC diffraction peaks were observed superimposed on the amorphous broad diffraction peaks.Similar behavior was reported by Sun et al. [23], who examined Cu 47 Ti 34 Zr 11 Ni 8 with graphite flakes.It is well known that metalloid carbon can form strong covalent bonds with metallic constituents, causing the subsequent formation of easily nucleated crystalline compounds [24,25].The activation energy of TiC formation was calculated as 731.6 kJ/mol [26].
Metals 2016, 6, 289 3 of 10 Ti50Cu28Ni15Sn7 composite specimens both with and without CNT particles were tested under unlubricated conditions on a block-on-roller TE53 wear tester from Plint Co., Kingsclere, UK.The roller material was hardened and tempered AISI O1 tool steel in 65 HRC (832 HV in microhardness).
The wear test was performed under selected contact pressures of 5, 10, 15, and 20 N at a revolution speed of 200 rpm.The weight loss of the specimen was recorded using an OHAUS AP250D-0 electronic balance (OHAUS Co., Parsippany, NJ, USA).Samples of JIS SKH-2 high-speed tool steel with a hardness of 24 and Ti50Cu28Ni15Sn7 crystalline alloy with a hardness of 29 were also tested for comparative purposes.

Results and Discussion
Figure 1 shows the X-ray diffraction (XRD) patterns of Ti50Cu28Ni15Sn after the addition of 4-12 vol.% CNT particles after 8 h of milling.For the plain Ti50Cu28Ni15Sn alloy and the matrix with 4 vol.% CNT particles, only a broad diffraction peak appears around 2θ = 42°, indicating the formation of fully amorphous powders.This may be attributed to the small volume fraction of CNT particles and their small crystalline size.The findings of the present study are similar to the observations regarding the preparation of the Al2O3/NiAl intermetallic-matrix composite reported by Lin et al. [22], in that, for small Al2O3 additions in mechanically alloyed Ni-Al alloys, no Al2O3 phase was detected by XRD after 10 h of milling.In the case of 12 vol.% CNT powders, crystalline TiC diffraction peaks were observed superimposed on the amorphous broad diffraction peaks.Similar behavior was reported by Sun et al. [23], who examined Cu47Ti34Zr11Ni8 with graphite flakes.It is well known that metalloid carbon can form strong covalent bonds with metallic constituents, causing the subsequent formation of easily nucleated crystalline compounds [24,25].The activation energy of TiC formation was calculated as 731.6 kJ/mol [26].Differential scanning calorimetry was applied to investigate the glass transition and crystallization behavior of the as-milled powders.The DSC scans of the 8 h as-milled Ti50Cu28Ni15Sn7 monolithic glass and the composites with CNT particles are presented in Figure 2. The glass transition (Tg) and crystallization (Tx) temperatures were defined as the onset temperature of the endothermic and exothermic DSC events, respectively.The Tg and Tx were 719 and 765 K, respectively, for the Ti50Cu28Ni15Sn7 without CNT particles.The glass transition temperature (Tg) and the crystallization temperature (Tx) of the composite powders were higher than those of single-phase amorphous Ti50Cu28Ni15Sn7 alloys.The addition of CNT particles into the glassy matrix induced a shift to higher temperatures for both Tg and Tx.This suggests that a small amount of CNT particles might be alloyed into the matrix upon milling, thereby changing the overall composition of the glassy phase.Similar behavior was reported for small ZrC additions by MA in a Zr-Al-Cu-Ni alloy [27]; hence, slight deviations must be taken into account when explaining changes in the thermal stability of the Differential scanning calorimetry was applied to investigate the glass transition and crystallization behavior of the as-milled powders.The DSC scans of the 8 h as-milled Ti 50 Cu 28 Ni 15 Sn 7 monolithic glass and the composites with CNT particles are presented in Figure 2. The glass transition (T g ) and crystallization (T x ) temperatures were defined as the onset temperature of the endothermic and exothermic DSC events, respectively.The T g and T x were 719 and 765 K, respectively, for the Ti 50 Cu 28 Ni 15 Sn 7 without CNT particles.The glass transition temperature (T g ) and the crystallization temperature (T x ) of the composite powders were higher than those of single-phase amorphous Ti 50 Cu 28 Ni 15 Sn 7 alloys.The addition of CNT particles into the glassy matrix induced a shift to higher temperatures for both T g and T x .This suggests that a small amount of CNT particles might be alloyed into the matrix upon milling, thereby changing the overall composition of the glassy phase.
Similar behavior was reported for small ZrC additions by MA in a Zr-Al-Cu-Ni alloy [27]; hence, slight deviations must be taken into account when explaining changes in the thermal stability of the composite powders.According to the heat of thermodynamic mixing (∆H) [28], the heat of mixing in Ti and C is −151 kJ/mol, which is the highest of all the elements in the present study (Ti, Cu, Ni, Sn, and CNT).It was observed that Ti and a metalloid carbon with strong covalent bonds could be synthesized as TiC, thereby changing the matrix component in the composite materials.
Metals 2016, 6, 289 4 of 10 composite powders.According to the heat of thermodynamic mixing (ΔH) [28], the heat of mixing in Ti and C is −151 kJ/mol, which is the highest of all the elements in the present study (Ti, Cu, Ni, Sn, and CNT).It was observed that Ti and a metalloid carbon with strong covalent bonds could be synthesized as TiC, thereby changing the matrix component in the composite materials.Based on the DSC results, the 8 h as-milled composite powders were consolidated into a disk through vacuum hot pressing with a 10 mm diameter and a 2 mm thickness.The powders were hot pressed at 723 K under a pressure of 1.2 GPa for 30 min.Figure 4 shows a consolidated sample of a bulk metallic glass composite; it exhibited a smooth outer surface and a metallic luster.XRD and DSC analyses were performed to confirm the amorphization status of the consolidated samples.Though  Based on the DSC results, the 8 h as-milled composite powders were consolidated into a disk through vacuum hot pressing with a 10 mm diameter and a 2 mm thickness.The powders were hot pressed at 723 K under a pressure of 1.2 GPa for 30 min.Figure 4 shows a consolidated sample of a bulk metallic glass composite; it exhibited a smooth outer surface and a metallic luster.XRD and DSC analyses were performed to confirm the amorphization status of the consolidated samples.Though Based on the DSC results, the 8 h as-milled composite powders were consolidated into a disk through vacuum hot pressing with a 10 mm diameter and a 2 mm thickness.The powders were hot pressed at 723 K under a pressure of 1.2 GPa for 30 min.Figure 4 shows a consolidated sample of a bulk metallic glass composite; it exhibited a smooth outer surface and a metallic luster.XRD and DSC analyses were performed to confirm the amorphization status of the consolidated samples.Though not shown here, consolidated BMG Ti 50 Cu 28 Ni 15 Sn 7 and the composite samples exhibited similar XRD and DSC results as those of the corresponding powders shown in Figure 1 (XRD) and Figure 2 (DSC).The cross-sectional view of the consolidated Ti50Cu28Ni15Sn7 BMG and composite samples examined by SEM is shown in Figure 5.No significant pores were observed in the cross-sectional view taken at magnification of 20,000 times, which reveals that relatively dense Ti50Cu28Ni15Sn7 bulk composite samples can be successfully fabricated by using vacuum hot pressing.Meanwhile, the CNT particle is hard to observe in the SEM image.The main reason for this is probably due to its nanometer size.The density value of the monolithic Ti50Cu28Ni15Sn7 bulk sample as measured by the Archimedes method was 6.51 g/cm 3 , while those for the composite samples containing 4%, 8%, and 12% vol.CNT particles were 6.44, 6.43 and 6.37 g/cm 3 , respectively, leading to a relative density value of 97.8%, 94.7%, 94.4%, and 94.0%, correspondingly.The cross-sectional view of the consolidated Ti 50 Cu 28 Ni 15 Sn 7 BMG and composite samples examined by SEM is shown in Figure 5.No significant pores were observed in the cross-sectional view taken at magnification of 20,000 times, which reveals that relatively dense Ti 50 Cu 28 Ni 15 Sn 7 bulk composite samples can be successfully fabricated by using vacuum hot pressing.Meanwhile, the CNT particle is hard to observe in the SEM image.The main reason for this is probably due to its nanometer size.The density value of the monolithic Ti 50 Cu 28 Ni 15 Sn 7 bulk sample as measured by the Archimedes method was 6.51 g/cm 3 , while those for the composite samples containing 4%, 8%, and 12% vol.CNT particles were 6.44, 6.43 and 6.37 g/cm 3 , respectively, leading to a relative density value of 97.8%, 94.7%, 94.4%, and 94.0%, correspondingly.
Figure 6 shows the mechanical properties of BMG composite samples evaluated using Vickers microhardness testing.Because a relatively high density was achieved, the influence of porosity on the Ti 50 Cu 28 Ni 15 Sn 7 BMG and its composites in the present study can be neglected.The Vickers microhardness for the Ti 50 Cu 28 Ni 15 Sn 7 BMG disc was 6.85 GPa, which was higher than the values of the Ti-based amorphous ribbons (which ranged from 6.02 to 6.56 GPa) prepared by melt spinning [2], mechanically alloyed Zr 55 Al 10 Cu 30 Ni 5 BMG composites (5.6 to 6.1 GPa), and Zr 65 Al 7.5 Cu 17.5 Ni 10 composites (4.9 to 5.7 GPa) [29].The microhardness of the Ti 50 Cu 28 Ni 15 Sn 7 BMG composites with 12% vol.CNT particles was 9.34 GPa.No cracks were observed around the indentation either from the corners or sides, which implies that the BMG composites are sufficiently tough to prevent fractures.The microhardness readings of the specimens of SKH-2 high-speed steel and crystalline Ti 50 Cu 28 Ni 15 Sn 7 alloy were 5.41 and 6.69 GPa, respectively.The difference in microhardness may be attributed to the in situ formed TiC particles and residual stress generated as a result of the differences in the coefficients of thermal expansion between the amorphous matrices and the TiC particulates, resulting in a high density of dislocations [30,31].composite samples can be successfully fabricated by using vacuum hot pressing.Meanwhile, the CNT particle is hard to observe in the SEM image.The main reason for this is probably due to its nanometer size.The density value of the monolithic Ti50Cu28Ni15Sn7 bulk sample as measured by the Archimedes method was 6.51 g/cm 3 , while those for the composite samples containing 4%, 8%, and 12% vol.CNT particles were 6.44, 6.43 and 6.37 g/cm 3 , respectively, leading to a relative density value of 97.8%, 94.7%, 94.4%, and 94.0%, correspondingly.Figure 6 shows the mechanical properties of BMG composite samples evaluated using Vickers microhardness testing.Because a relatively high density was achieved, the influence of porosity on the Ti50Cu28Ni15Sn7 BMG and its composites in the present study can be neglected.The Vickers microhardness for the Ti50Cu28Ni15Sn7 BMG disc was 6.85 GPa, which was higher than the values of the Ti-based amorphous ribbons (which ranged from 6.02 to 6.56 GPa) prepared by melt spinning [2], mechanically alloyed Zr55Al10Cu30Ni5 BMG composites (5.6 to 6.1 GPa), and Zr65Al7.5Cu17.5Ni10composites (4.9 to 5.7 GPa) [29].The microhardness of the Ti50Cu28Ni15Sn7 BMG composites with 12% vol.CNT particles was 9.34 GPa.No cracks were observed around the indentation either from the corners or sides, which implies that the BMG composites are sufficiently tough to prevent fractures.The microhardness readings of the specimens of SKH-2 high-speed steel and crystalline Ti50Cu28Ni15Sn7 alloy were 5.41 and 6.69 GPa, respectively.The difference in microhardness may be attributed to the in situ formed TiC particles and residual stress generated as a result of the differences in the coefficients of thermal expansion between the amorphous matrices and the TiC particulates, resulting in a high density of dislocations [30,31].Samples of the Ti-based BMG composite, crystalline Ti-based alloy and SKH-2 specimen were tested for wear resistance and friction on a block-on-roller wear tester under unlubricated conditions.Quenched and tempered JIS SKS3 (AISI O1) tool steel with 65 HRC (832 HV) was selected as the roll material.Contact pressures of 5, 10, 15, and 20 N were exerted on the tops of the specimens for the wear test, and the roller rotation speed was set to 200 rpm.The steady-state friction coefficients for the six different specimens tested on block-on-roller devices are illustrated in Figure 7.Under these wear configurations, the coefficients of friction (COFs) of the Ti-based BMG composites and conventional materials were measured within the range 0.45-0.85;their values within the range depended on the normal load and the test environment.In a comparison of the steady-state COFs of various specimens, the COFs of SKH-2 and crystalline Ti-based alloy were higher than those of the Ti-based BMG composites; furthermore, the COFs decreased as the applied load increased [21,32].It is notable that COFs of the Ti50Cu28Ni15Sn7 BMG matrix can be reduced with the addition of CNT particles.In contrast to frictional behavior, the results of the present study indicate noteworthy differences in performance; under a high load (20 N), the BMG composites containing 12% vol.CNT particles exhibited the highest COFs, whereas the wear performance levels of the monolithic Ti-based BMG and the matrices with 4-8 vol.% CNT particles were higher than those of the matrix with 12 vol.% CNT particles.The soft SKH-2 was found to possess lower wear resistance, whereas the hard Samples of the Ti-based BMG composite, crystalline Ti-based alloy and SKH-2 specimen were tested for wear resistance and friction on a block-on-roller wear tester under unlubricated conditions.Quenched and tempered JIS SKS3 (AISI O1) tool steel with 65 HRC (832 HV) was selected as the roll material.Contact pressures of 5, 10, 15, and 20 N were exerted on the tops of the specimens for the wear test, and the roller rotation speed was set to 200 rpm.The steady-state friction coefficients for the six different specimens tested on block-on-roller devices are illustrated in Figure 7.Under these wear configurations, the coefficients of friction (COFs) of the Ti-based BMG composites and conventional materials were measured within the range 0.45-0.85;their values within the range depended on the normal load and the test environment.In a comparison of the steady-state COFs of various specimens, the COFs of SKH-2 and crystalline Ti-based alloy were higher than those of the Ti-based Metals 2016, 6, 289 7 of 10 BMG composites; furthermore, the COFs decreased as the applied load increased [21,32].It is notable that COFs of the Ti 50 Cu 28 Ni 15 Sn 7 BMG matrix can be reduced with the addition of CNT particles.In contrast to frictional behavior, the results of the present study indicate noteworthy differences in performance; under a high load (20 N), the BMG composites containing 12% vol.CNT particles exhibited the highest COFs, whereas the wear performance levels of the monolithic Ti-based BMG and the matrices with 4-8 vol.% CNT particles were higher than those of the matrix with 12 vol.% CNT particles.The soft SKH-2 was found to possess lower wear resistance, whereas the hard Ti-based BMG composites containing 12 vol.% CNT particles did not exhibit the highest wear performance levels.Similar results were also found in various BMG systems of different microhardness levels; Fleury et al. [32] reported that Ni-based BMGs exhibit higher microhardness levels than Ti-or Zr-based BMGs do, but also that wear resistance levels of Ni-based BMGs are not as high as those of Ti-or Zr-based BMGs.  Figure 8 illustrates the cumulative wear mass losses of various specimens tested under the contact pressure of different applied loads on the block-on-roller wear tester.The wear mass losses of the Ti-based BMGs with both 4 and 8 vol.% CNT particles were lower than those of SKH-2 and crystalline Ti-based alloy specimens; however, the wear mass loss of hardness in the matrix with 12 vol.% particles was higher than that of the other Ti-based BMG composites.We can conjecture that the hardness level of the matrix with 12 vol.% CNT particles is higher than that of the roller (8.23 GPa), and that the friction wear force of the roller repeatedly wears on the matrix with 12 vol.% CNT particles, thereby altering the wear configuration.The wear behavior of the matrix with 12 vol.% CNT particles was transferred to the roller for reinforcement.The reinforced matrix with 12 vol.% CNT particles peeled easily as a result of wear from the roller, subsequently causing debris, and the peeled debris became attached to the roller.When the roller motion was repeated, the wear behavior was transferred to the debris of the matrix with 12 vol.% CNT particles, thereby inducing an increase in both COFs and the wear behavior.Figure 8 illustrates the cumulative wear mass losses of various specimens tested under the contact pressure of different applied loads on the block-on-roller wear tester.The wear mass losses of the Ti-based BMGs with both 4 and 8 vol.% CNT particles were lower than those of SKH-2 and crystalline Ti-based alloy specimens; however, the wear mass loss of hardness in the matrix with 12 vol.% particles was higher than that of the other Ti-based BMG composites.We can conjecture that the hardness level of the matrix with 12 vol.% CNT particles is higher than that of the roller (8.23 GPa), and that the friction wear force of the roller repeatedly wears on the matrix with 12 vol.% CNT particles, thereby altering the wear configuration.The wear behavior of the matrix with 12 vol.% CNT particles was transferred to the roller for reinforcement.The reinforced matrix with 12 vol.% CNT particles peeled easily as a result of wear from the roller, subsequently causing debris, and the peeled debris became attached to the roller.When the roller motion was repeated, the wear behavior was transferred to the debris of the matrix with 12 vol.% CNT particles, thereby inducing an increase in both COFs and the wear behavior.
Concerning the mechanisms of abrasive wear in Ti-based BMG composites, Figure 9 shows the surface tracks for the monolithic Ti-based BMG, the matrix with 12 vol.% CNT composites, the crystalline Ti-based alloy, and the SKH-2 specimen after tests were performed under different applied loads.It can be observed from the micrographs of SKH-2 that a ditch-like microstructure formed on the surface along the sliding direction.This wear mechanism possesses the characteristics of typical adhesive wear.Adhesive wear occurs when the contact pressure between rolling surfaces is high enough to cause local plastic deformation and welding between the contacting asperities [33], whereas the crystalline Ti-based alloy exhibited less adhesive wear on the surface.The Ti-based BMG and the matrix with 12 vol.% CNTs exhibited an abrasive wear track.Abrasive wear occurs when hard debris particles are indented and form a groove on the rolling surface of the material [33].Energy-dispersive X-ray spectroscopy (EDS) analysis of the wear tracks indicated that the dark areas of the JIS SKS3 tool steel were rich in Fe, suggesting that material transfer occurred from the soft counterpart rolling to the hard Ti-based BMG composites.Similar behavior has been reported by Fleury et al. and Liu et al. [32,34], regarding examinations of different BMG systems and Si 3 N 4 -based composites.Kalin et al. [35] reported that the model for wear behavior should include chemical reactions, spot temperatures, thermal conductivity, and phase transformations.Figure 8 illustrates the cumulative wear mass losses of various specimens tested under the contact pressure of different applied loads on the block-on-roller wear tester.The wear mass losses of the Ti-based BMGs with both 4 and 8 vol.% CNT particles were lower than those of SKH-2 and crystalline Ti-based alloy specimens; however, the wear mass loss of hardness in the matrix with 12 vol.% particles was higher than that of the other Ti-based BMG composites.We can conjecture that the hardness level of the matrix with 12 vol.% CNT particles is higher than that of the roller (8.23 GPa), and that the friction wear force of the roller repeatedly wears on the matrix with 12 vol.% CNT particles, thereby altering the wear configuration.The wear behavior of the matrix with 12 vol.% CNT particles was transferred to the roller for reinforcement.The reinforced matrix with 12 vol.% CNT particles peeled easily as a result of wear from the roller, subsequently causing debris, and the peeled debris became attached to the roller.When the roller motion was repeated, the wear behavior was transferred to the debris of the matrix with 12 vol.% CNT particles, thereby inducing an increase in both COFs and the wear behavior.Concerning the mechanisms of abrasive wear in Ti-based BMG composites, Figure 9 shows the surface tracks for the monolithic Ti-based BMG, the matrix with 12 vol.% CNT composites, the crystalline Ti-based alloy, and the SKH-2 specimen after tests were performed under different applied loads.It can be observed from the micrographs of SKH-2 that a ditch-like microstructure formed on the surface along the sliding direction.This wear mechanism possesses the characteristics of typical adhesive wear.Adhesive wear occurs when the contact pressure between rolling surfaces is high enough to cause local plastic deformation and welding between the contacting asperities [33], whereas the crystalline Ti-based alloy exhibited less adhesive wear on the surface.The Ti-based BMG and the matrix with 12 vol.% CNTs exhibited an abrasive wear track.Abrasive wear occurs when hard debris particles are indented and form a groove on the rolling surface of the material [33].Energy-dispersive X-ray spectroscopy (EDS) analysis of the wear tracks indicated that the dark areas of the JIS SKS3 tool steel were rich in Fe, suggesting that material transfer occurred from the soft counterpart rolling to the hard Ti-based BMG composites.Similar behavior has been reported by Fleury et al. and Liu et al. [32,34], regarding examinations of different BMG systems and Si3N4-based composites.Kalin et al. [35] reported that the model for wear behavior should include chemical reactions, spot temperatures, thermal conductivity, and phase transformations.

Conclusions
In the present study, amorphous Ti50Cu28Ni15Sn7 and its composite powders were successfully synthesized through an MA process performed on powder mixtures of pure Ti, Cu, Ni, Sn, and CNT after 8 h milling.The metallic glass composite powders exhibited a large supercooled liquid region

Conclusions
In the present study, amorphous Ti 50 Cu 28 Ni 15 Sn 7 and its composite powders were successfully synthesized through an MA process performed on powder mixtures of pure Ti, Cu, Ni, Sn, and CNT after 8 h milling.The metallic glass composite powders exhibited a large supercooled liquid region before crystallization.The thermal stability of the amorphous matrix was affected by the presence of CNT particles.Changes in T g and T x suggest that deviations in the chemical composition of the amorphous matrix occurred because of a partial dissolution of the CNT species in the amorphous phase.BMG composite discs were obtained by consolidating the 8 h as-milled composite powders through a vacuum hot pressing process.The Vickers microhardness test results were 6.85 GPa for the Ti 50 Cu 28 Ni 15 Sn 7 BMG, and 7.96, 8.42, and 9.34 GPa for the composites containing 4%, 8%, and 12% vol.CNT particles, respectively.Notable levels of hardness (approximately 30%) were achieved for the Ti 50 Cu 28 Ni 15 Sn 7 BMG composites comprising 12 vol.% CNT particles; however, the wear mass loss of the Ti-based BMG composite with 12 vol.% CNT particles (the composite with the highest hardness) was higher than that of the other Ti-based composites containing 4 and 8 vol.% CNT particles when tested under different applied loads.The results demonstrate that the wear resistance levels of Ti-based BMG composites do not correlate with their hardness levels because they do not follow the wear law.Wear surfaces under high applied loads demonstrate that the matrix with 12 vol.% CNT particles suffered severe wear compared with the monolithic Ti-based BMG.
here, consolidated BMG Ti50Cu28Ni15Sn7 and the composite samples exhibited similar XRD and DSC results as those of the corresponding powders shown in Figures 1 (XRD) and 2 (DSC).

Figure 4 .
Figure 4. Outer morphology of Ti50Cu28Ni15Sn7 with 4% vol.CNT particles after vacuum hot pressing at 723 K under a pressure of 1.2 GPa.

Figure 5 .
Figure 5. Cross-sectional scanning electron microscope (SEM) image of the Ti50Cu28Ni15Sn7 bulk metallic glasses (BMG) and composite samples after vacuum hot pressing at 723 K under a pressure of 1.2 GPa.

Figure 4 .
Figure 4. Outer morphology of Ti 50 Cu 28 Ni 15 Sn 7 with 4% vol.CNT particles after vacuum hot pressing at 723 K under a pressure of 1.2 GPa.

Figure 5 .
Figure 5. Cross-sectional scanning electron microscope (SEM) image of the Ti50Cu28Ni15Sn7 bulk metallic glasses (BMG) and composite samples after vacuum hot pressing at 723 K under a pressure of 1.2 GPa.

Figure 5 .
Figure 5. Cross-sectional scanning electron microscope (SEM) image of the Ti 50 Cu 28 Ni 15 Sn 7 bulk metallic glasses (BMG) and composite samples after vacuum hot pressing at 723 K under a pressure of 1.2 GPa.
-or Zr-based BMGs do, but also that wear resistance levels of Ni-based BMGs are not as high as those of Ti-or Zr-based BMGs.

Figure 7 .
Figure 7. Effects of CNT volume fraction on the friction coefficients of Ti50Cu28Ni15Sn7 BMG composites under different loads.

Figure 7 .
Figure 7. Effects of CNT volume fraction on the friction coefficients of Ti 50 Cu 28 Ni 15 Sn 7 BMG composites under different loads.

Figure 7 .
Figure 7. Effects of CNT volume fraction on the friction coefficients of Ti50Cu28Ni15Sn7 BMG composites under different loads.

Figure 8 .
Figure 8. Mass loss of CNT/Ti50Cu28Ni15Sn7 bulk metallic glass composites under different loads.Figure 8. Mass loss of CNT/Ti 50 Cu 28 Ni 15 Sn 7 bulk metallic glass composites under different loads.

Figure 8 .
Figure 8. Mass loss of CNT/Ti50Cu28Ni15Sn7 bulk metallic glass composites under different loads.Figure 8. Mass loss of CNT/Ti 50 Cu 28 Ni 15 Sn 7 bulk metallic glass composites under different loads.