A. Microstructure

The micrograph of the AZ91 alloy under various conditions is presented in Fig. 2. As presented in Fig. 2(a), the microstructure of the as-received samples consisted of the β-intermetallic phase, Mg₁₇Al₁₂, precipitated in the α-Mg matrix. To obtain a better view acquisition, the SEM micrograph is presented in Fig. 2(b). It could be observed that the β-phase formed the reticulate structure with the eutectic α-Mg in the grid. A lamellar α + β eutectic mixture is attached to the precipitates. Similar observations were reported in as-cast AZ91 magnesium alloys by other researchers. ^{Reference Chye, Zamzuri, Norbahiyah, Ismail, Derman and Illias8–Reference Luong, Shunmugasamy, Cox, Gupta and Rohatgi10} Feng et al. ^{Reference Feng and Ma11} and Lü et al. ^{Reference Lü, Wang, Ding, Zeng and Zhu12} reported that the interface between the coarse β-Mg₁₇Al₁₂ and α-Mg was easy to produce stress concentration and tended to generate cracks in the process of deformation, which had an adverse effect on the mechanical properties. Figure 2(c) presents the optical micrograph of the surfacing alloys following the T4 heat treatment. Subsequently to the solid solution, the Mg₁₇Al₁₂ phase dissolved into α-Mg and produced a microstructure consisted of the supersaturated α-Mg phase, which led to the poor electrical conductivity and low hardness in Table II.

FIG. 2.

Micrograph of the surfacing alloys: (a) as-received; (b) the SEM micrograph of (a); (c) surfacing alloy-T4; (d) surfacing alloy-T6; (e) the FESEM micrograph of (d); (f) the corresponding enlarged view of (e).

TABLE II.

Electrical conductivity and microhardness of the tested surfacing alloys.

The optical micrograph of the AZ91 surfacing alloys treated by T6 is presented in Fig. 2(d). It could be observed that high amounts of precipitates existed following aging. Figure 2(e) demonstrates the FESEM micrographs of the samples. The pictures display that the β-phase was precipitated from the supersaturated α-Mg solid solution in two forms: the discontinuous precipitation and the continuous precipitation. ^{Reference Zhou, Shen and Aung13–Reference Duly, Simon and Brechet15} The discontinuous precipitation is the cellular growth of the β-phase and α-Mg alternating layers at the grain boundaries [Fig. 2(e)]. ^{Reference Celotto5} It could be observed that the β-phase varied in size, ranging from 0.5 to 6 μm. Also, the continuous precipitation led to the formation of dot-shaped, fine needle like β-precipitates within the grains [Fig. 2(f)], which were believed to improve the strength through the dislocation–particle interaction by the Orowan process. Consequently, the surfacing alloy of AZ91 following the T6 treatment had the highest macrohardness (Table II).

The measured and calculated X-ray diffractogram of the surfacing alloys is presented in Fig. 3. As presented in the picture, the calculated intensities (I _cal.) were compared with those observed (I _obs.) and the factors R _wp and results of quantitative analysis were also analyzed. As can be seen, the solution heat treatment (T4) was responsible for the diffraction peak intensities disappearance of the β-phase [Fig. 3(b)]. The aluminum atoms dissolution was effective in the low electric conductivity of the T4 treated surfacing layer (Table II). The β-phase peaks were detected again following the T6 treatment [Fig. 3(c)]. The quantitative analysis results demonstrated that the volume fraction of β-Mg₁₇Al₁₂ in the as-received alloy was 5.9% [Fig. 3(a)], whereas in the T6 treated sample it was 11.3% [Fig. 3(c)]. In consequence, the amount of β-Mg₁₇Al₁₂ was significantly increased due to the T6 heat treatment.

FIG. 3.

The measured and calculated X-ray diffractogram of the surfacing alloys. (a) As-received; (b) surfacing alloy treated by T4; (c) surfacing alloy treated by T6.

B. Wear properties

The friction coefficients for all three alloys are plotted against the applied loads in Fig. 4(a). As presented in the chart, the friction coefficients of the alloys with various heat treatment conditions exhibited a similar trend as the load increased, such as the former rapidly decreased at a low load, gradually decreased at a medium load, and finally decreased to the lowest level at a high load. The load affected the wear behavior by contact area and deformation. During sliding, the metal surface was in an elastic–plastic state. The real contact area consisted of quite low contacts at the asperity tips and was not linearly related to the load. With increasing applied load, these asperities participated in sustaining the normal pressure by generating a higher plastic deformation, leading to the decrease of the friction coefficient. ^{Reference Ramezani and Ripin16} On the other hand, as the load increased, the thermal softening and melting of the contact surface caused by the high amount of frictional heat allowed an easier sliding at the interface. ^{Reference Zafari, Ghasemi and Mahmudi17}

FIG. 4.

(a) Friction coefficients and (b) wear rates of the tested alloys under different applied loads.

Fig. 4(b) presents the variation of the volumetric wear rates versus the applied load for the surfacing alloys. The volumetric wear rate was estimated in mm³/m by the mass loss division by the density of the alloys. It was evident that at the low applied loads of 5, 25, and 50 N, the wear rates increased linearly. Also, no drastic increase existed until the highest load was reached, whereas the wear rates of the sample at 100 N were approximately 1.5 times higher than the load of 50 N. This indicated that a transition from the mild wear to the severe wear had occurred. These results coincide with the reported results in other investigations. ^{Reference Selvan and Ramanathan18,Reference An, Li, Lu, Chen, Xu, Chen and Wang19}

At all applied loads, the AZ91 surfacing alloy without heat treatment had a higher wear rate than the specimens subjected to T4 at constant loads, whereas the alloy subjected to T4 had the lowest hardness. It did not follow the Archard’s law ^{Reference Archard20} and the results were obtained by Xu Zhen et al. ^{Reference Taltavull, Torres, López and Rams21,Reference Zhen, Zhihao, Dongyue, Qingqiang and Zhanzhi22} The massive β-Mg₁₇Al₁₂ phase existing in the as-received surfacing alloy might have been the main reason for this difference. During sliding, the massive phase might detach from the matrix and become particles that adhere between the asperities, possibly hindering sliding and increasing the frictional force. ^{Reference Ramezani and Ripin16} Furthermore, the interface between the crude β-phase and α-Mg was easy to produce stress concentration and generate a mass of twins, as presented in Figs. 5(a) and 5(b). As the wear continued, the cracks tended to grow and spread on the surface, which aggravated the delamination wear. This was the main factor for the poor wear resistance during the severe wear stage (a load of 100 N). The T4 treatment resulted in the dissolution of the massive β-Mg₁₇Al₁₂ phase, reducing the number of crack sources and leading to the wear resistance improvement. Besides, the deformed layers of the surfacing sample subjected to T4 were shallower than the as-received sample layers [Fig. 5(c)].

FIG. 5.

Micrographs of the subsurface of surfacing alloys after sliding under an applied load of 100 N. (a) As-received; (b) the corresponding enlarged view of (a); (c) surfacing alloy-T4; (d) surfacing alloy-T6.

In contrast, the wear rate of the surfacing alloy of AZ91 treated by T6 was the least in value with respect to the various alloys under constant sliding conditions. This phenomenon might be related to the β-phases precipitation during aging. The β-phases distributed at the grain boundaries could pin the boundaries and hinder sliding and migration. Moreover, the fine f-compounds in the form of continuous precipitates were strongly bonded into the matrix, which resisted the hard particles wear effect. As a result, no twins were observed in the micrographs [Fig. 5(d)] and the depth of the deformed layers of the surfacing alloy treated by T6 was the shallowest in Fig. 5.

C. Wear mechanisms

The wear properties’ change of the surfacing samples was correlated with various wear mechanisms under each wear condition. For the wear mechanisms that occurred in each wear condition test to be identified, the worn surface and debris of the worn specimens were analyzed. The results disclosed four different wear operating mechanisms: abrasion, oxidation, delamination, and plastic deformation.

Figure 6 presents the SEM micrograph of the worn surface following sliding under the 100 N applied load. It could be observed that severe delamination occurred during the as-received AZ91 surfacing alloy wear [Fig. 6(a)]. A mass of lumps stripped from the contact surface of the sample and formed irregular and high amount of wear debris, as presented in Fig. 6(d). Furthermore, both the area and depth of the delaminated craters were significantly reduced in the surfacing metal treated by T4 [Fig. 6(e)]. As aforementioned in a previous paper, the solid solution treatment resulted in the dissolution of the massive β-phase and retarded the speed of crack initiation and growth. As it could be observed in Fig. 6(c), the worn surface of the T6 treated surfacing alloy under the 100 N applied load was quite smoother than the untreated alloy surface. This occurred because the presence of fine and dispersed β-phases precipitated in the grains, which were strongly bonded into the matrix, hindered the dislocation motion, and resisted the hard particles wear effect. As a result, the sequential surface destruction was difficult to occur. The debris morphologies of the alloys treated by T6 are presented in Fig. 6(f). The band-like and fine debris were suggested to be the result from the abrasive wear, whereas the sheets or the irregular lumps resulted from delamination. It appeared that the debris average size generated by the alloys treated by T6 was lower than the other alloys, which was indicated as a relatively moderate wear.

FIG. 6.

SEM micrograph of the worn surface and debris after sliding under an applied load of 100 N. (a, d) As-received; (b, e) surfacing alloy-T4; (c, f) surfacing alloy-T6.

The SEM images of the worn surface of the surfacing AZ91 magnesium alloy treated by T6 under various applied loads ranging from 5 to 100 N are presented in Fig. 7. It could be observed that numerous grooves were parallel to the sliding direction on the surfaces of most pins, which are presented dominant in Figs. 7(a), 7(c), and 7(d). These scratches were the characteristics of abrasive wear, which could be attributed to the hard asperities on the steel counter-face or the detached particles that were removed from the pin and placed at surface contact. ^{Reference Archard20,Reference Moussa, Waly and El-Sheikh23} This phenomenon is generally recognized as the common wear mechanism at low loads for the sliding wear of magnesium alloys. ^{Reference Yan and Wang6,Reference An, Li, Lu, Chen, Xu, Chen and Wang19,Reference Jiang, Bi, Zzhao, Li, Lian and Jiang24} The increased applied load resulted in the stronger cutting and plowing on the pin, which could be responsible for the deeper and wider grooves.

FIG. 7.

SEM micrograph and the results of EDS of the worn surface of the AZ91 surfacing alloys with T6 after sliding under different applied loads: (a) 5 N; (b) EDS analysis of region A depicted in (a); (c) 25 N; (d) 50 N; (e) EDS analysis of region B depicted in (d); (f) 100 N.

As presented in Figs. 7(b) and 7(e), the energy disperse spectroscopy (EDS) analysis of regions A and B marked in Figs. 7(a) and 7(d) identified the strong O peaks, which indicated the oxide films existence on the sample pin surfaces. These results were indicative of oxidative wear. It was believed that the oxidation of the surface could be attributed to frictional heating during sliding and the easily oxidized nature of magnesium. ^{Reference Yan and Wang6} An et al. ^{Reference An, Li, Lu, Chen, Xu, Chen and Wang19} believed that the thick oxide layer could effectively protect the sliding surface and result in a mild wear condition. Selvan et al. ^{Reference Selvan and Ramanathan18,Reference Stott and Wood25} reported that the sinking on the wear surface could be filled with oxide fragments, which could become compacted into a protective layer that would prevent the metal-to-metal contact resulting in a low wear rate.

With increasing load to 50 N, a series of short and shallow cracks, approximately perpendicular to the sliding direction, could be observed on the wear surface [Fig. 7(d)], indicating the appearance of the wear delamination mechanism. In this mode, the normal and shear stress were transferred through the contacting asperities to elicit a plastic strain on the surface, which promoted the nucleation and propagation of subsurface cracks in the matrix–reinforcement interfacial area. ^{Reference Yan and Wang6} As the wear continued, the cracks tended to grow and finally lead to the surface layer detachment. As the applied load was enhanced to 100 N [Fig. 7(f)], the deep and wide delaminated craters could be evidently observed and the surface became smooth due to plastic deformation, which was generally associated to the delamination severity. Corresponding with the wear rate variation, the transition from mild to severe wear was also suggested.

The SEM images of the protruded layers on the worn surface corresponding to the 100 N load for the surfacing sample treated by T6 are presented in Fig. 8. It could be observed that the high applied load resulted in a severe plastic deformation. The deformed surface layers were extruded along the sliding direction and consequently spread on the sample. As a result, the metallic layers were observed to be protruded slightly on the contact surface (Fig. 8) and form irregular lumps [Fig. 6(f)].

FIG. 8.

SEM images of protruded layers on the worn surface corresponding to 100 N load for the surfacing alloy with T6.