J. An, Y.X. Tian, C.Q. Feng

Key Laboratory of Automobile Materials, Ministry of Education, Department of Materials Science and Engineering, Jilin University,Changchun 130025, PR China

Received 25 November 2019; received in revised form 8 January 2020; accepted 13 February 2020

Available online 29 September 2020

Abstract

Keywords: Mg-alloys; Elevated temperatures; Wear rate; Wear mechanism; Wear transition.

1. Introduction

Among Mg-based alloys,Mg97Zn1Y2 (in atomic ratio)alloy is possibly the most attractive one owing to its outstanding mechanical properties at both room and elevated temperatures and a special type of structure phase, i.e. long-period stacking ordered (LPSO) phase [1-5]. Kawamura et al. [6] produced a high-strength Mg97Zn1Y2 alloy using rapidly solidifie powder metallurgy (RS/PM) technology in 2001. The alloy presented high yield strength of 600MPa at room temperature and 510MPa at 150°C. The LPSO structure phase X-Mg12ZnY was thought to contribute a lot to the strengthening of RS/PM Mg97Zn1Y2 alloy besides the strengthening of fine-graine Mg matrix of 100-200nm in diameter.Thereafter, numerous efforts have been devoted to understanding of every aspect of LPSO structure phases. In recent years other unique characteristics of LPSO structure phases have also been revealed, such as high thermal stability up to 500°C, kink-band deformation mode and preventing effect on the growth of deformation twin in Mg matrix [7-9].Therefore, Mg97Zn1Y2 alloy is thought to be more suitable for making structural components working in both room and elevated-temperature environments than conventional Mg-Al-Zn (AZ) alloys, and it is even suitable for making wear components such as piston, sliding bearing and low-load bearing gear. Mg97Zn1Y2 alloy also exhibited a much better dry sliding wear performance than the widely used AZ91 alloy at 0.785m/s under room temperature condition [10]. For these reasons, it is necessary to conduct research in wear properties and wear transition of Mg97Zn1Y2 alloy at elevated temperatures for future practical applications. Mg-based alloys have been found to display two different types of wear behavior,namely mild wear and severe wear. The former is acceptable in engineering applications because of its steady wear state,low wear rate and slight surface damage, while the latter is opposite [11-17].

The dry sliding wear performance of Mg97Zn1Y2 alloy at room temperature was comprehensively studied within wide sliding velocity and load ranges [18]. The critical load conditions for M-S wear transition at sliding velocities of 0.5-4.0m/s were also determined, which depicted a mild wear regime where Mg97Zn1Y2 alloy component could work safely at room temperature [19]. However, up to now, only few investigations have been conducted on friction and wear characteristics of Mg-based alloys at elevated temperatures,for the most part, they focused on presenting the influence of load and test temperature on wear rate and identifying wear mechanisms [20-26]. The more complicated problems associated with M-S wear transition have not been fully touched,including M-S wear transition mechanism, criterion for MS wear transition and evaluation of the critical load or test temperature for M-S wear transition. The solutions of these problems are conductive to better understanding of M-S wear transition mechanism under elevated temperature conditions,also beneficia for engineering applications of Mg97Zn1Y2 alloy.

The objective of the present study is to investigate wear performance of Mg97Zn1Y2 alloy under room and elevated temperatures.The influence of experimental parameters(load and temperature) on wear rate and wear mechanisms were demonstrated under different wear conditions. These results were then summarized in a wear mechanism transition map.The microstructure and hardness in subsurfaces of worn samples were further examined to clarify the real reason for MS wear transition. The correlation between test temperature,load and wear transition was finall discussed according to the surface DRX temperature (SDT) criterion for M-S wear transition presented by Liang et al. [12].

2. Experimental details

2.1. Material

Mg97Zn1Y2 alloy was prepared by gravity casting method. The raw materials including high purity Mg(＞99.95wt.%)),Zn(＞99.95wt.%))and Mg-20Y(wt.%)master alloy were firstl melted at 760°C,and then hold at 720°C for a period of about 20 min under protection of a mixture gas(CO2and SF6).The melt was finall poured in a steel die,which yielded a cast ingot with a dimension of 95mm diameter and 200mm height. The optical microstructure of as-cast Mg97Zn1Y2 alloy consisted of the coarse primary α-Mg dendrites surrounded by network of X-Mg12ZnY eutectic phase,as shown in Fig. 1. The compressive properties of the alloy were measured on a MTS810 material testing system as follows: 120MPa yield strength, 207MPa compressive strength and 17.9% strain limit. The hardness was measured to be 79HV on a HVS-1000 Vickers hardness tester. The phase analysis and microstructure examination of Mg97Zn1Y2 alloy were performed by means of Rigaku D/MAX 2500PC X-ray diffractometer (XRD) and LEXT-OLS3000 confocal scanning laser microscope (CSLM), respectively.

Fig. 1. The optical microstructure of Mg97Zn1Y2 alloy.

Fig. 2. MG-2000 high-speed and high-temperature pin-on-disk type wear testing machine.

2.2. Elevated-temperature wear tests

Wear tests were carried out on a MG-2000 high-speed and high-temperature pin-on-disc type wear machine at temperatures of 20-200°C under unlubricated sliding condition.Fig. 2 shows the wear machine equipped with an electric furnace. The test temperatures were controlled accurately with error less than ±5°C by thermocouple. Other testing parameters were selected as follows: sliding velocity 0.5m/s,wear distance 565m, wear track of discs 60mm diameter,load rang of 20-360N. Wear samples were cylindrical pins of diameter 6mm and 13mm height. They were wire-electrode machined from the cast ingot.The material used for preparing discs was a middle carbon chromium steel 50Cr. The discs were heat treated to reach Vickers hardness of 570 HV by quenching and low-temperature tempering. The surfaces of pins and discs were prepared by grinding with SiC paper and then polishing to roughness Ra about 0.4μm. The wear rate of pin was expressed by the volume loss per unit sliding distance. A digital micrometer with an accuracy of 0.001mm was used to measure the height reduction of pins, and from which the volume loss was obtained.

Identificatio of wear mechanisms and analysis of compositions of the worn surfaces were performed by means of SEM equipped with EDS.The friction-induced microstructure change in subsurfaces of several samples was examined under CSLM. These samples were selected among those tested in mild wear and severe wear, and then prepared by grinding and polishing the cross sections along the sliding direction.The microhardness was measured across subsurfaces of the selected samples,and then plotted as a function of depth from surface.

3. Results and discussion

3.1. Effect of applied load and test temperature on wear rates

Fig. 3 shows variations of wear rates of Mg97Zn1Y2 alloy with load at various test temperatures. In the case of sliding wear at temperatures of 20-100°C, it is observed from Fig. 3(a) that on the one hand, wear rate increases on the whole with increasing load at each test temperature; on the other hand, wear rate essentially increases with increasing temperature from 20 to 50°C or from 20 to 100°C under a given constant load. However, the wear rate was even lower at 100°C than at 50°C in the load range of 20-100N. When applied load was higher than 100N, the wear rate at 100°C started surpassing that measured at 50°C, and maintained a plateau within 100-160N. The lower wear rate that occurred at 100°C in the load range of 20-100N is apparently contrary to popular expectation that higher test temperature usually lead to higher wear rate under a constant load. These phenomena will be discussed later in Section 3.2 on the basis of morphological and chemical analyses of worn surfaces.Besides these details, a common point is also noted that each wear rate-load curve can be roughly divided into two distinct stages in terms of the curve slope. Wear rate moderately increased in a low slope with increasing load in the firs stage,but rapidly went up in a high slope in the second stage. The turning points between the firs and second stages can be clearly identifie from the wear rate curves. They are 220N at 20°C, 200N at 50°C and 160N at 100°C, respectively. In the case of sliding wear at temperatures of 150 and 200°C,as seen from Fig. 3(b), the wear rate was generally higher at 200°C than at 150°C within 20-170N, and both wear rate curves also varied with applied load in a similar way to those at temperatures of 20-100°C. That is, they can also be divided into two stages, i.e. the firs stages are from 20 to 120N at 150°C and from 20 to 80N at 200°C, while the second stages begin from 120 to 220N at 150°C and from 80 to 200N at 200°C, respectively. The turning points between the firs and second stages are 120N at 150°C and 80N at 200°C, respectively.

Fig. 3. Variations in wear rate with load at temperatures of 20-100°C (a)and 150-200°C (b).

These turning points of wear rate curves can be preliminarily postulated to be M-S wear transition points according to the notation of M-S wear transition proposed by Chen and Alpas [11]. In addition, it is found that at lower temperatures of 20 and 50°C, the wear rates at the turning points are 39.8×10-12mm3m-1and 45.3×10-12mm3m-1respectively. They are close to 39.9×10-12mm3m-1, the wear rate at M-S transition point for AZ91D alloy under sliding at 0.5m/s and room temperature [11]. Of course, the finaconfirmatio of the M-S wear transition load at each test temperature ought to be further considered together with SEM observation of worn surfaces, since it is commonly accepted that the beginning of severe plastic deformation (SPD) wear mechanism for Mg-based alloys typically corresponds to M-S wear transition at room and elevated temperatures [11-23].

Table 1 Contents of major elements on worn surfaces at various temperatures (wt%).

3.2. Wear mechanisms and wear mechanism transition map

Fig. 4. SEM images of worn surfaces under different sliding conditions: (a)20N and 100°C, (b) 80N and 100°C, (c) 180N and 100°C, (d) 200N at 100°C, (e), 20N and 200°C, (f) 60N and 200°C, (g) 120N and 200°C, (h)170N and 200°C .

The morphological features and compositions of worn surfaces were examined to identify wear mechanisms associated with different testing parameters. The contents of main elements on worn surfaces were detected by EDS and given in Table 1. SEM images of worn surfaces at 100 and 200°C are shown in Fig. 4. In the case of sliding wear at 100°C, as the load was increased from 20 to 160N, i.e. in the firs stage of wear rate-load curve, worn surfaces demonstrated a typical feature of delamination wear mechanism. For example, at 20N, delamination scars were observed together with surface cracks, meanwhile several oxide patches were found at localized areas (Fig. 4(a)). The oxygen content reached 21.42% on the worn surface. At 80N, a few cracks became apparently vertical to the sliding direction (Fig. 4(b)), while the oxygen content still maintained a high level of 15.6%. As seen from the image of pin edge at top right corner of Fig. 4(b), the surface was not severely plastic deformed since no extruded edge was formed. Delamination and surface oxidation were the dominant wear mechanisms in the firs stage. When load was applied between 180 and 240N i.e.in the second stage of wear rate-load curve, worn surfaces exhibited features of severe wear. For example, at 180N, surface material underwent a severe plastic deformation (SPD), resulting in a flatte surface without surface crack (Fig. 4(c)), while the oxygen content was 10.78%. At 200N, the surface material was plasticfl wed more apparently along the sliding direction, producing an evident extruded edge (Fig. 4(d)). Meanwhile the oxygen content reached 14.29%, indicating a heavily oxidized surface. In addition, the oxidized surface material cracked and broke off from edge due to its fragile nature under such a high applied load. SPD and spalling of oxide layer were the predominant wear mechanisms in the second stage. In the case of sliding wear at 200°C, as the load was increased from 20 to 60N in the firs stage of wear rate-load curve,worn surface was apparently flattened and the amount of cracks relating with delamination wear significantl decreased (Fig. 4(e) and(f)).Meanwhile,oxygen content still maintained a medium-tohigh level of 7.25%-16.54%. Therefore, delamination, mild plastic deformation and surface oxidation were the dominated wear mechanisms in the firs stage. When wear entered into the second stage, at 120N, for instance, the worn surface was severely deformed along the sliding direction, which resulted in an extruded edge, but no cracks were produced on the surface(Fig.4(g)).The oxygen content reduced to only 3.3%.At high load of 170N, the surface material suffered from melting since there were signs of melt waves and a sleek edge(Fig. 4(h)), while the oxygen content was 6.0%. Therefore,SPD and surface melting were the main wear mechanisms in the second stage.SEM observation results prove that the wear mechanism transition from the firs stage to the second stage actually corresponds to M-S wear transition.

The reasons for lower wear rate at 100°C than at 50°C within 20-120N and plateau occurrence within 120-160N could be qualitatively explained by the change of surface composition. It is found from elemental mappings that the surfaces were oxidized much heavier at 100°C than at 50°C within 20-40N. Therefore, under loads of 20 and 40N, a heavily oxidized layer might be responsible for a lower wear rate occurring at 100°C since oxide layer is usually harder than Mg matrix [23]. However, when the load was 60N, oxygen content reached as high as 14.97% at 50°C, much higher than 7.46% at 100°C, suggesting that surface oxidation was not the only reason for lower wear rate at 100°C.It notes that the contents of Y and Zn elements are 6.32% and 2.29% at 100°C,higher than 5.59% and 2.02% at 50°C. A comparison was made between elemental mappings of worn surfaces at 50 and 100°C under 60N, as illustrated in Fig. 5. The fin Yrich and Zn-rich particles were found on the worn surface at 100°C. These particles are possibly Mg12ZnY phase without being worn off the surface due to fair deformation ability of Mg12ZnY phase, and their existence can strengthen the surface material or oxide layer through fibe reinforcement effect[7].Such a reinforcement effect was also found in maintaining a wear rate plateau within 120-160N at 100°C, for example,the contents of Y and Zn elements were 6.29% and 2.20% at 140N, while the oxygen content was only 7.84%.

Fig. 5. SEM images and EDS mappings of worn surfaces at 50°C and 100°C under 60 N: SEM images at 50°C (a) and 100°C (b), Mg at 50°C(c) and 100°C (d), O at 50°C (e) and 100 °C (f), Y at 50°C (g) and 100°C(h), Zn at 50°C (i) and 100°C (j).

Fig. 6. Wear rate map (a) and wear mechanism transition map (b), AAbrasion;O-Oxidative wear; D-Delamination;S.O-Surface oxidation;M.P.DMild plastic deformation; S.P.D-Severe plastic deformation; S.O.L-Spall of oxide layer; Ad-Adhesion; S.M-Surface melting.

A wear rate map and a wear mechanism transition map were plotted on load-temperature coordinate system for Mg97Zn1Y2 alloy. Fig. 6(a) is the wear rate map that lists the wear rate data at various loads and temperatures. The data were expressed in m3m-1×10-12. At M-S wear transition loads, the wear rate generally decreased with temperature along the solid line AA′,decreasing from 45.3×10-12m3m-1at 50°C to 32.0×10-12m3m-1at 200°C.It is also found that the difference between the wear rates at M-S wear transition loads and those just above line AA′appears to be increasingly small with temperature increasing. However, the surface damage extent is much larger than what is reflecte from wear rate. Fig. 6(b) is the wear mechanism transition map. It consists of two wear regimes i.e. mild and severe wear regimes.These two regimes are separated from each other by line AA′.Bellow line AA′is the mild wear regime. This regime is also thought as a safe area for wear parts made of the studied material working steadily. The boundary line AA′was drawn based on several attentive issues including the turning point in wear rate curve and beginning of SPD mechanism. The mild wear regime includes three sub-regimes, which are separated each other by dashed lines CC′and DD′′. Each sub-regime represents two or three wear mechanisms such as abrasion,oxidative wear,dalamination,surface oxidation and mild plastic deformation.The boundary lines CC′and DD′were mainly confirme by disappearance of grooves(abrasion)and appearance of irregular scars and/or surface cracks (delamination),and occurrence of rather flatte surface with few cracks but almost no scars (delamination+mild deformation). The severe wear regime also includes three sub-regimes, which are separated each other by dashed lines BB′and DD′. The surface melting sub-regime is delimited by the dashed line CC′,which was confirme by observations of surface melting evidence such as melt waves on surface, multilayered structure or sleek edge by SEM, and low coefficien of friction due to the melt-induced lubricating effect. The boundary line DD′was confirme by absence of spalling of oxide layer and EDS detection of a rather low content of oxygen element (less than 3.4%) on severely deformed surface with extruded edge (SPD mechanism).

3.3. Microstructures evolution in subsurfaces

SEM observation results revealed an important wear mechanism transition occurring on the worn surfaces during M-S wear transition process. It was that once SPD mechanism of surface material controlled the wear process, there was no formation of surface cracks even when subjected to very high applied loads. Owing to their inherent nature of hexagonal close-packed structure, Mg alloys typically have a moderate ductility, and that is why surface cracks are often found in delamination mechanism. However, when severe wear commences, the surface material demonstrates an incredibly high ductility. Therefore, it is reasonable to postulate that a certain type of microstructure transformation occurs and hence changes the mechanical properties of surface material, especially improves the ductility.In order to testify this deduction,the microstructure and hardness in subsurfaces in mild and severe wear regimes were analyzed.

In the case of sliding wear at 100°C,the worn samples that were subjected to loads of 40,80,140 and 160N in mild wear and 230N in severe wear were chosen for observation of microstructures evolution in the subsurfaces, as shown in Fig. 7.At 40N, the material in subsurface suffered from plastic deformation,consequently leading to formation of a deformation zone beneath the surface (Fig. 7(a)). The deformation zone reached a depth of about 85μm.In addition,it was found that part of worn surface was covered by mechanically mixed layer(MML).At 80N,the deformation zone extended to a depth of about 110μm. It was observed that MML almost covered the whole worn surface from the high magnificatio photograph(Fig.7(b)),and the α-Mg dendrites together with LOSP structure phase were elongated towards the sliding direction beneath MML. It is known that the LOSP structure phase has a certain extent of deformability through slip and kinking band mechanisms [8]. At higher loads of 140 and 160N, deformation zones were further extended to larger depth of about 177 and 195μm respectively, and the material presented a number of fl w lines towards the sliding direction in the top part of the deformation zone (Fig. 7(c)). When subjected to 200N in severe wear, a friction-affected zone (FAZ) of a thickness of about 244μm was produced beneath the surface (Fig. 7(d)).As seen from the magnificatio microstructure (Fig. 7(e)), a different sub-zone was produced within the depth range of 0-48μm, where most of the plastically deformed dendrites changed and transformed into newly formed fin microstructure. The grains were so fin that they could not be recognized well under the resolution-limited microscope. This sub-zone could be composed of DRX microstructure since large plastic deformation and frictional heating could promote DRX transformation near the surface. There was a deformed microstructure sub-zone beneath DRX sub-zone, where the fl w lines and deformation twins could be observed. At 230N in surface melting mechanism, FAZ further increased to a depth of about 274μm, and it consisted of four different microstructure sub-zones. They were solidification DRX,deformation+DRX and deformation sub-zones, respectively.The thickness values of these sub-zones were about 20, 39,117 and 74μm, respectively. The details of microstructure in the solidification DRX and deformation+DRX mixed microstructure sub-zones are shown in the magnificatio microphotograph (Fig. 7(f)). The features in the solidificatio sub-zone were that LOSP structure phase transformed from original strips to a large amount of fin particles embedded in the α-Mg phase, these particles presented smooth edge and they were arranged in necklace along sliding direction. The change of LOSP structure phase indicates that the material is partly molten during wear testing. Large stripes of Mg12ZnY present the sign of erosion by melt,while those thin Mg12ZnY phase strips in the eutectics (Mg12ZnY phase+α-Mg phase)locally melt, and are broken into fin particles in the fl wing deformation. A quite amount of fin grains in micrometer scale can be roughly recognized in DRX sub-zone. The microstructure in deformation+DRX mixed microstructure subzone essentially maintains the deformed morphology and only a little amount of DRX grains can be recognized in the α-Mg dendrites.

Fig. 7. Microstructures in the subsurfaces of pins worn at 100°C and different loads: (a) 40N, (b) 80N, (c) 140N, (d) 200N, (e) 200N, (f) 230N.

In the case of sliding wear at 200°C, in mild wear, at 40N, for example, FAZ reached a depth of about 102μm(Fig. 8(a)). The material adjacent to the surface suffered from a large plastic deformation, resulting in fl w lines extending towards the surface in an angle of about 20° (Fig. 8(b)). At 60N, FAZ was deepened to about 180μm (Fig. 8(c)), and the material near the surface was subjected to much more intensifie plastic deformation (Fig. 8(d)). However, at 120N in severe wear, FAZ extended to a depth of about 242μm(Fig. 8(e)). The most part of FAZ was essentially composed of DRX microstructure, but the DRX grain size was found to vary with location. Fine-grain DRX sub-zone was located within the depth range of 0-42μm, where SPD brought about fl w lines along the sliding direction. Coarse-grain DRX subzone was formed within the depth range of 42-186μm,where DRX grains in micrometer scale could be clearly recognized(Fig. 8(f)).

Fig. 8. Microstructures in the subsurfaces of pins worn at 200°C and different loads: (a) and (b) 40N, (c) and (d) 60N, (e) and (f) 120N.

It is well known that the requirements for DRX microstructure transformation are a large enough plastic deformation and a high enough temperature. The friction-induced plastic deformation in subsurface can be described by an approach of equivalent plastic strain 27,[28]. The equivalent strain can be expressed by Eq. (1).

Fig. 9. Subsurface equivalent plastic strain vs. depth from surface.

Fig. 10. Variations in hardness with depth from surface at 100 and 200°C under different loads: (a) 40, 80, 140 and 160N at 100°C, (b)160, 200 and 230N at 100°C, (c) 40 and 60N at 200°C, (d) 60 and 120N 200°C.

where ε is equivalent strain, z is depth, θ is the shear angle of fl w line at depth z. The subsurface plastic deformation strains of two samples worn at 100 and 200°C in mild wear regime are illustrated in Fig. 9. The equivalent plastic strain is rather high near the surface, and then reduces with depth.The equivalent plastic strain can maintain above 1.0 within a certain depth range, for example, 0-130μm at 140N and 100°C, and 0-60μm at 60N and 200°C. Apparently, such a big plastic deformation strain satisfie the requirement of plastic deformation for DRX transformation.In addition,when the applied load surpasses a critical value, the surface temperature can reach the critical temperature for DRX transformation with the help of friction and furnace heating. Therefore, DRX microstructures were observed in subsurfaces of pins worn in severe wear regime.

3.4. Hardness change in subsurfaces

The microhardness change that was accompanied by the microstructure evolution in the subsurfaces was also measured, as shown in Fig. 10. Under the condition of test temperature 100°C, the hardness-depth curves at various loads of 40, 80, 140 and 160N in mild wear are shown in Fig. 10(a).At each load, the hardness decreased monotonously with increasing depth until the hardness of about 77 HV. This hardness represents the unaffected substrate. In the depth ranges:0-90μm at 40N, 0-120μm at 80N, 0-180μm at 140N and 0-200μm at 160N, the hardness is found to be higher than that of the substrate. These depth ranges are essentially equal to those for plastic deformation zones observed in Fig. 7.Apparently, it is the strain-hardening effect in theses depth ranges that gives rise to a higher hardness, and the hardness enhancement is increasingly significan with increasing load,indicating strain-hardening effect is intensifie with load rising. The hardness-depth curves at 200 and 230N in severe wear are shown in Fig. 10(b). For the purpose of comparison,the hardness variation at 160N is also included in Fig. 10(b).At 200N, the variation of hardness with depth consisted of two stages. The firs stage was from 0 to 60μm, where hardness decreased to a relatively low value of 80 HV. The second stage began from 60 to 260μm, where the hardness rose again at depth of 60-110μm and finall went down until the substrate hardness. In the firs stage, the hardness was much lower than it was at 160N, suggesting that a significant softening effect occurred. In addition, the depth range for the firs stage is essentially equal to the depth range for DRX sub-zone. These phenomena prove that the softening effect is mainly caused by DRX microstructure transformation. The hardness rising and subsequent decreasing at the second stage apparently indicates that strain hardening gradually overwhelms the softening within the depth range of 60-110μm, and then it totally governs the hardness variation within the depth range of 110-260μm. However, at 230N,the variation of hardness with depth could be divided into four stages. The firs stage was from 0 to 20μm (forming a platform), the second stage range was from 20 to 60μm(showing a low valley), the third stage was from 60 to 180(presenting a plateau), the fourth stage began from 180 to 290μm (exhibiting a decreasing slope). It was also found that in the firs stage the hardness was almost equal to the hardness of the substrate, but in the second and third stages the hardness was lower than the hardness of the substrate, especially much lower in the second stage. In the fourth stage,the hardness gradually decreased to the substrate hardness.Referred to the microstructure features in the subsurface as shown in Fig. 7(f), it is prostituted that the solidifie microstructure transformation brings about a hardness platform comparable to the substrate hardness in the firs stage, an almost complete DRX microstructure transformation results in the lowest hardness valley in the second stage, the mixed microstructure of the deformed and partly DRXed grains causes a hardness plateau in the third stage, and plastic deformation produces a decreasing hardness slope in the fourth stage.Under the condition of test temperature 200°C, the variations of hardness with depth at different loads of 40, 60 and 120N were also measured. At applied loads of 40 and 60N in mild wear, these two hardness curves presented the typical strain-hardening effect in their respective FAZs (Fig. 10(c)).However, at 120N in severe wear, at depth of 0-220μm, the hardness was even lower than it was at 60N (Fig. 10(d)),suggesting that softening effect controls the hardness variation. The depth range of softening could be divided into two subranges. The firs one was from 0 to 40μm, and the second one began from 40 to 220μm. It was found that the hardness was much higher in the firs one than in the second one.This phenomenon in fact verifie the results of microstructure observation of Fig. 8(f). The high hardness in the firs subrange is due to formation of fin DRX grains induced by high plastic strain, while low hardness in the second subrange is attributed to the coarse DRX grains induced by low plastic strain. The difference in DRX grain size between the firs and second subranges apparently agrees with DRX theory, i.e. the larger the plastic deformation, the fine the DRX grains grow.At depth of 220-250, the hardness decreased to the substrate hardness, indicating that plastic deformation sub-zone takes account only a small part of FAZ.

Fig. 11. Schematic diagram showing the microstructure change: (a) before M-S wear transition, (b) after M-S wear transition.

The above experimental results shows that strain hardening dominates the mechanical properties of material in subsurface in mild wear, while DRX softening takes over in severe wear.The microstructure transformation in subsurfaces before and after M-S wear transition was schematically shown in Fig.11.According to Archard’s wear law [29], wear resistance relies on the hardness of metallic materials, and increasing hardness favors improving resistance to wear. Therefore, in mild wear,increasing load deepens the plastic deformation zone, and intensifie the strain hardening of surface material (Fig. 11(a)).Consequently, the gradual increasing of wear rate with load in mild wear regime is ascribed to strain hardening effect.However, a combination of the frictional and environmental heating promotes the DRX microstructure transformation in subsurface in severe wear regime (Fig. 11(b)), which consequently results in DRX softening.DRX softening brings about a rapid rising of wear rate. That is, DRX softening activates the M-S wear transition at elevated temperatures. Therefore,it is reasonable to deduce that the elevated-temperature M-S wear transition should still obey the contact surface DRX temperature criterion proposed by Liang et al. [12]. The criterion tells that M-S wear transition starts when surface temperature rises up to the critical DRX temperature of material, i.e.TS≥TDRX.

3.5. M-S wear transition loads

M-S wear transition load is plotted against test temperature in Fig. 12. Obviously, the transition load FTexhibits an approximately linear relation with test temperature T. The transition load-temperature curve is thus linearly fitted The fitte line can also be written as Eq. (2) using the line slope,k=-0.7842, and the T-axis intercept, TC=302.91°C.

Fig. 12. M-S wear transition loads at different test temperatures.

In accordance with SDT criterion for M-S wear transition,i.e.Eq.(3),Liang et al.[12]proposed a formula for evaluating transition load, as expressed by Eq. (4).

where Ts is the contact surface temperature, TDRXthe critical temperature for DRX transformation of surface material at sliding speed v, T the test environmental temperature, μ the coefficien of friction. KDRXin Eq. (4) is a constant that depends on wear apparatus and physical properties of pin and disc, and it is expressed by Eq. (5).

where Anis the nominal contact area, Kmpand lbare the thermal conductivity mean diffusion distance of pin respectively,α is the fraction of heat conducted into pin.

In the present study, it was found that μ varied a little at each transition load within test temperature range of 20-200°C, i.e. it ranged from 0.30 to 0.32. On the other hand,temperature has a little influenc on Kmpof Mg97Zn1Y2 alloy,as temperature increased from 50 to 200°C,Kmpincreases a little, from 57 to 64Wm-1K-1[30]. Other parameters α, lband Anare also almost constant under critical conditions of M-S wear transition. Therefore, if the criterion of M-S wear transition and the formula for evaluating transition load is applicable to elevated-temperature severe wear transition, and KDRXis still considered as a constant, FTshould present a linear relation with T. The critical surface temperature TDRXis reported to be 311.6°C for M-S wear transition at the sliding speed of 0.5m/s[17].KDRXis therefore calculated to be 8.816 using the sliding data at 100°C: FT160N, v 0.5m/s, μ 0.30,TDRX311.6°C. The transition loads at other test temperatures are estimated using Eq. (4). The calculated transition loads are also included in Fig. 12. The calculated transition loads have a good consistence with those measured ones. The error ranges from 0.2% to 1.83% within the temperature range of 20-150°C.The big deviation occurred at 200°C,4.4N,about 5.5% of the measured transition load. Moreover, if the M-S wear transition follows the SDT criterion,the TCin Eq.(2)actually corresponds to TDRXin Eq. (4), and –k corresponds to KDRXμv. The calculated TDRXby DRX theory is 311.6°C,KDRXμv is 0.7561, while the linearly fitte TCis 302.9°C, -k is 0.7842. The difference between the two groups of data is small. Therefore, under given elevated temperature wear conditions, M-S wear transition of Mg97Zn1Y2 alloy still obeys the SDT criterion.

4. Conclusions

1. At test temperatures of 20-200°C, wear rates increased with increasing load. Each wear rate vs. load curve exhibited two different developing stages that corresponded to mild wear and severe wear respectively.

2. Wear mechanisms operating in mild wear and severe wear were indicated in the wear mechanism transition map.

3. After M-S wear transition, the primary changes of microstructure and mechanical property in subsurfaces were DRX microstructure transformation and the resulting softening. DRX softening in subsurface is the major cause of M-S wear transition.

4. M-S wear transition load decreased linearly with temperature rising.

5. M-S wear transition at temperatures of 20-200°C obeys the SDT criterion.

Acknowledgements

The authors would like to thank the support from National Natural Science Foundation of China (Grant No.51775226).