Dry sliding wear behaviour of aluminium–lithium alloys reinforced with SiC particles

Dry sliding wear behaviour of aluminium–lithium alloys reinforced with SiC particles

Wear 262 (2007) 292–300 Dry sliding wear behaviour of aluminium–lithium alloys reinforced with SiC particles J. Rodr´ıguez ∗ , P. Poza, M.A. Garrido,...

2MB Sizes 0 Downloads 41 Views

Wear 262 (2007) 292–300

Dry sliding wear behaviour of aluminium–lithium alloys reinforced with SiC particles J. Rodr´ıguez ∗ , P. Poza, M.A. Garrido, A. Rico Departamento de Ciencia e Ingenier´ıa de Materiales, Universidad Rey Juan Carlos, Tulip´an s/n, E28933 M´ostoles, Madrid, Spain Received 9 September 2005; received in revised form 7 April 2006; accepted 11 May 2006 Available online 27 June 2006

Abstract Several wear tests were carried out at different pressures and temperatures on Al-8090 and Al-8090 + 15 vol.% SiCp . Worn specimens and debris were also examined using SEM and EDX techniques to identify the dominant wear mechanisms. Wear rate increases about two orders of magnitude when temperature is above a critical one. The transition from mild to severe wear is dependent on nominal pressure. The composite transition temperature is higher than that of the unreinforced alloy. Within the mild wear regime, the wear rates for both materials exhibit a minimum over 100 ◦ C and are higher for the composite material than for the Al-8090 below the transition temperature. It has been also observed that the presence of mechanically mixed layers (MML) on the wear surface with varying morphology and thickness influenced the wear rate. The morphology and composition of the wear debris also change with the wear mechanism. © 2006 Elsevier B.V. All rights reserved. Keywords: Sliding wear; Metal–matrix composites (MMCs); Mechanically mixed layer (MML); Temperature effect

1. Introduction The dispersion of ceramic particles in aluminium alloys leads to considerable improvements in many mechanical properties. Benefits have been reported in stiffness, temperature performance and wear resistance, while other aspects mainly related to ductility and fatigue behaviour remain controversial [1,2]. The combination of cost and performance makes these discontinuously reinforced aluminium matrix composites (DRAMC) appealing for components in the automotive and aerospace industries. Under operating conditions where the contact between solids is expected, the tribological behaviour may become critical. Although in many circumstances lubrication is employed, dry sliding can be considered as the limit case. It has been pointed out [2] that wear resistance is not a material property and, consequently, the wear performance of DRAMC should be evaluated considering all the contributing factors (load, speed, temperature, materials involved, environment, geometry, etc.).

Corresponding author. Tel.: +34 914887159; fax: +34 914888150. E-mail address: [email protected] (J. Rodr´ıguez).

0043-1648/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2006.05.006

The tribological behaviour of aluminium alloys in dry sliding against steel counterfaces has been widely studied. The wear maps methodology developed by Lim and Ashby [3] has been successfully applied to aluminium alloys identifying several dominant mechanisms depending on load, speed and temperature [4]. At low loads and speeds, the mild wear regime is associated with an oxidational process. The formation of tribolayers seems to be crucial in this regime, although the classical wear models neglect it. As the load is increased, a transition to a severe wear regime is observed and the operating wear mechanisms changes to delamination with severe plastic deformation. DRAMC behave quite similar, but the presence of the reinforcement particles can alter the range where the different mechanisms become dominant. In spite of the great number of papers published treating the wear behaviour of DRAMC, general trends are difficult to identify. In fact, the matrix composition has a critical effect. Nevertheless, it is generally accepted that within the mild wear regime, the role of the reinforcement particles is to support the contact stresses preventing high plastic deformations. If the load is increased over a critical value, the particles will be fractured and comminuted, loosing their role as load supporters. In the severe wear regime, composites wear rate takes values similar or even worse than those of the unreinforced alloys.

J. Rodr´ıguez et al. / Wear 262 (2007) 292–300

Unlike load and speed, studies regarding the influence of the temperature are not common. Several studies showed that the mild to severe wear transition occurred at higher temperatures in the composites [5,6]. It is remarkable that, although it is not usually reported, no improvement is observed after adding reinforcement particles in several cases. For example, Alpas and Zhang studied the effect of particle size concluding that wear rates of composites exhibit similar wear resistance to that of the unreinforced alloys when the particle size is of the order of surface roughness [7]. Another negative effect may be the abrasive action of the hard reinforcement phase on the counterbody wear. In the last 15 years, most of the traditional aluminium alloys, such as 2000, 5000, 6000 and 7000 have been studied from the tribological point of view [1]. To the authors’ knowledge, no results are available about wear behaviour of aluminium–lithium alloys and DRAMC using these alloys as matrices. The presence of lithium increases the elastic modulus of aluminium alloys by approximately 6% for every 1 wt.% of lithium and the density is reduced by 3%, which improves significantly the specific modulus [8]. The addition of ceramic reinforcements provides further improvements in this direction and Al/Li–SiC composites exhibit elastic modulus over 100 GPa with relative densities around 2.6 leading to specific stiffness 50% above that of standard aluminium and titanium alloys [9,10]. In addition, the presence of lithium in the aluminium matrix helps to strengthen these materials through the precipitation of the ordered ␦ (Al3 Li) phase coherent with the matrix [11,12], and the Al–Li/SiC system has an excellent performance/cost ratio for applications where the stiffness is critical, as in aeronautical and aerospace applications. In this work, a systematic experimental study is performed to analyse the effect of load and temperature on the wear behaviour of an aluminium–lithium alloy Al-8090 and the same alloy reinforced with 15 vol.% of SiC particulates. The aim is to evaluate if the tendencies observed in more conventional aluminium alloys and their composites are also followed in this material. Microstructural and mechanical behaviour have been previously and extensively studied, which allows this work to be focused on the tribological performance. 2. Materials The 8090 Al alloy reinforced with 17 vol.% of SiC particles and the unreinforced alloy were supplied by Cospray Inc. (Banbury, United Kingdom) in the form of extruded rectangular bars of 25.4 mm × 62.5 mm cross-section. Composition of the alloy and the composite is given in Table 1. Both materials were produced by spray codeposition of the matrix and the particles onto a substrate [8]. They were extruded at 420 ◦ C into rectangular


bars, with an extrusion ratio of 25:1. The bars were solution heat treated at 530 ◦ C for 2 h, water quenched, and cold stretched up to 2% to relieve the residual stresses introduced during quenching. Afterward, they were artificially aged at 170 ◦ C, during 32 and 48 h for the composite and the unreinforced alloy, respectively, to reach the peak-aged condition (T651). The grain structure of the unreinforced alloy was highly anisotropic and grains larger than several hundreds of micrometers were found in the longitudinal direction. The average grain sizes in the long and short transversal directions, perpendiculars to the extrusion axis, were 21 and 8 ␮m, respectively. The ceramic particles homogenized the grain dimensions in the composite and the average grain size of the matrix was 12 ␮m in the longitudinal direction and 6 ␮m in the long and short transversal directions. The size of the reinforcements was 7.5 ± 2.4 ␮m with an aspect ratio of 2.4 ± 1.2. The particles were oriented with the longer axis in the extrusion direction. It is worth noting that very few (<4%) broken SiC particles were found and particle–matrix descohesion was never observed in the as-received material. The main precipitate found in the unreinforced alloy was the metastable and ordered ␦ (Al3 Li) phase, which is coherent with the aluminum matrix and nucleates homogeneously, with an average diameter around 25 nm. In order to refine the grain size, Zr is added to the 8090 alloy, leading to the development of ␤ (Al3 Zr) dispersoid. S (Al2 CuMg) phase was also observed. This precipitate, semicoherent with the matrix, has an orthorhombic crystal structure and is observed as needles oriented along the 1 0 0 directions in the matrix. The main difference between the composite and the unreinforced alloy was the development of a ␦ precipitate-free zone (PFZ), around 200 nm thick, along the matrix–reinforcement interface. Further details about the microstructure have been described elsewhere [10]. The stress–strain curves of these materials as a function of temperature have been recently reported [13] showing a degradation of the mechanical performance as the temperature is increased up to 190 ◦ C. The elastic modulus of the unreinforced alloy was found around 80 GPa at room temperature, while the ceramic reinforcements increased this value up to 101 GPa for the composite. The yield stress of both materials was quite similar being reported 537 MPa for the unreinforced alloy and 507 MPa for the composite at room temperature. The elastic modulus for the unreinforced alloy and the composite at 190 ◦ C were 71 and 91 GPa, respectively and the yield stress of both materials at this temperature was around 360 MPa. 3. Experimental techniques Tests were carried out in a wear testing machine with a pin on disc configuration under dry sliding conditions without eliminating the debris formed (see Fig. 1). Specimen and counterbody

Table 1 Chemical composition of Al-8090 and the composite Material









SiC (wt.%)

Al-8090 Al–Li/SiC

2.38 2.32

0.99 1.21

0.81 0.82

0.12 0.10

0.023 0.036

0.04 0.05

0.03 0.06

Balance Balance

– 17.3


J. Rodr´ıguez et al. / Wear 262 (2007) 292–300

Fig. 1. Schematic view of the pin-on-disc test configuration.

were cleaned using methanol to avoid the presence of humidity and other non-desirable films such as grease. Most requirements of the ASTM standard G99-04 were followed. Nevertheless, several modifications were introduced, mainly regarding the pin shape. Prismatic pins were made of the material under study (Al-8090 and its composite) with rectangular section of 2.5 mm × 6.3 mm. With this geometry, the nominal contact area was maintained constant during the tests in spite of the wear process. The disc, made of carbon steel (SAE 1045), rotates horizontally at sliding speed of 0.1 m/s. A dead weight loading system was used to perform the tests at nominal normal pressures of 6.3, 12.5, 16.5 and 50 MPa. The bulk test temperature was also modified from 20 to 300 ◦ C for the Al-8090 alloy and from 20 to 350 ◦ C for the composite material. Different temperatures were applied by means of a furnace designed to accommodate the main elements of the test: the pin and the disc. The temperature was measured with a thermocouple located at the centre of the pin support, providing, thus, a bulk temperature measurement. The furnace was programmed to ensure that a thermal steady state was reached before the beginning of the wear test (±5 ◦ C). The coefficient of friction was obtained by means of a torque transducer. The variation of the pin height was registered using a LVDT with ±1 ␮m of precision. The wear rate (␮m/m) was calculated as the slope of the sample height versus sliding distance. Both, friction coefficient and wear rate were continuously recorded during the test.

Finally, worn specimens were cross-sectioned using a diamond saw. To analyse the microstructure changes in the nearest zone at the pin contact surface, samples were prepared with conventional metallographic techniques using SiC papers up to 4000 grit finish, followed by polishing in a diamond slurry (up to 1 ␮m) and, finally, on SiO2 . These metallographic samples were observed in an Environmental Scanning Electron Microscope Philips XL 30 (ESEM) equipped with energy dispersive X-rays microanalysis (EDX). To contribute to the wear mechanisms identification, debris formed was also analysed by ESEM and EDX. Additionally, nanoindentation tests were carried out in some cross-sectioned specimens to evaluate the extent of plastic deformation in both materials. A MTS XP nanoindenter was used to apply a maximum load of 100 mN. 4. Results 4.1. Wear rate and friction coefficient The relative wear rates versus temperature are shown at different pressures in Fig. 2a and b using a logarithmic scale. The value of the wear rate is normalised by the wear rate measured at room temperature to better capture the tendency. As it can be appreciated in the figures, wear rate increased about two orders of magnitude when temperature is above a critical value, indicating a change in the mechanism from mild to severe wear. Although this effect is observed regardless of the pressure applied, the transition temperature is dependent on it. As it can also be seen in Fig. 2a and b, this temperature was higher for the composite material than for the unreinforced alloy, Al-8090. On the other hand, it has been observed that in both materials, the higher the normal pressure, the lower the transition temperature. Furthermore, for the lower nominal pressures (6.3 and 12.5 MPa), there was an initial decrease in wear rate with temperature, taking minimum values at 100 ◦ C. In fact, wear rate did not recover the room temperature values up to 200 ◦ C. Although the relative representation included in the last figures may be convenient to make the results easy to analyse, it

Fig. 2. Relative wear rates of the unreinforced alloy (a) and the composite material (b) vs. temperature.

J. Rodr´ıguez et al. / Wear 262 (2007) 292–300


Fig. 3. Wear rates and friction coefficients of the matrix and reinforced material vs. temperature.

does not allow a direct comparison between both materials. The absolute wear rate, determined from the height loss of the pin is shown in the upper part of Fig. 3. As it can be seen, the unreinforced alloy exhibited higher wear resistance than the composite material below the transition temperature. However, the beginning of the severe wear regime is shifted to higher temperatures in the composite. Additionally, in the lower part of Fig. 3, the average friction coefficients are included. The composite material presented higher values than those of the unreinforced alloy. An exception to this trend seems to appear when both materials are subjected to very demanding conditions, that is, within the severe wear regime. 4.2. Surface morphology Figs. 4 and 5 show cross-sections of the specimens observed by ESEM, corresponding to the aluminium alloy and the composite, respectively. The nearest zone of the surface presented a different microstructure than both the wearing material and the steel counterbody. It seems to be a mechanically mixed layer (MML) composed of debris particles, probably fractured and comminuted, coming from both sides of the contact. The presence of oxygen is also detected indicating oxidational processes.

The MML is also characterized by a high density of defects and microcracks. As an example, Figs. 4 and 5 show the different morphologies observed in the MML generated under conditions clearly corresponding to mild wear regime: 20, 100 and 200 ◦ C for a normal pressure of 6.3 MPa (the lowest used in this work). The MML exhibits significant differences as the temperature increases from room to higher values. At 20 ◦ C and 6.3 MPa (Fig. 4a) the layer was discontinuous, without covering the whole surface. The maximum thickness was over 30 ␮m. When temperature was increased to 100 ◦ C and the pressure was maintained at 6.3 MPa (Fig. 4b), the MML covered completely the wear surface, forming a continuous and protective layer. However, the thickness was not regular, ranging from 40 ␮m, as maximum value, to around 5 ␮m. Finally, subsequent increases in temperature (Fig. 4c) provided a MML with more regular thickness, but with the same maximum values of 40 ␮m. The observations performed in the composite material (Fig. 5) resulted in similar conclusions as the unreinforced alloy, but with thinner layers. MML measurements should be taken into account only as qualitative information because of the experimental difficulties associated with their obtaining. Small changes on the surfaces during the specimen preparation could modify the experimental data.


J. Rodr´ıguez et al. / Wear 262 (2007) 292–300

Fig. 4. SEM micrographs showing the MML formed on the contact surface of the unreinforced alloy: (a) 20 ◦ C, 6.3 MPa; (b) 100 ◦ C, 6.3 MPa; (c) 200 ◦ C, 6.3 MPa.

4.3. Wear debris SEM and EDX analyses of the wear debris were performed to identify its morphology and composition. As we have pointed out before, there are very marked differences in wear rate within the mild and severe wear regimes. Figs. 6 and 7 show some micrographs of the debris obtained during wear tests of the unreinforced alloy and the composite, respectively.

Fig. 5. SEM micrographs showing the MML formed on the contact surface of the composite: (a) 20 ◦ C, 6.3 MPa; (b) 100 ◦ C, 6.3 MPa; (c) 200 ◦ C, 6.3 MPa.

Fig. 6a (Al-8090 tested at 20 ◦ C and 6.3 MPa) shows a bimodal size distribution of fine particles and larger plate-like debris with sizes up to 200 ␮m. In tests carried out under severe conditions, 200 ◦ C and 16.5 MPa (Fig. 6c), large plate-like debris become dominant. These particles, with sizes beyond the mm, exhibited marked signals of plastic deformation as wear surface grooves and edge cracks. Some examples of the morphology of the composite debris particles are shown in Fig. 7a (20 ◦ C and 6.3 MPa) and Fig. 7c (350 ◦ C and 6.3 MPa). The situation is similar to that pointed out in Fig. 6 in both cases. The wear debris morphol-

J. Rodr´ıguez et al. / Wear 262 (2007) 292–300


Fig. 6. (a) Wear debris from Al-8090 tested under mild wear conditions (20 ◦ C, 6.3 MPa); (b) corresponding EDS analysis; (c) wear debris from Al-8090 tested under severe wear conditions (200 ◦ C, 16.5 MPa); (d) corresponding EDS analysis.

ogy was consistent with the wear rates measured during the tests. From EDX analysis of the particles (see Figs. 6b and d and 7b and d), it can be derived that the debris under mild wear conditions was composed of aluminium and iron in the case of the

unreinforced alloy. However, the large particles associated with the severe wear regime were composed mainly of aluminium. In the case of the composite, the results were very similar, except for the presence of broken silicon carbide particles coming from the sub-surface zone.

Fig. 7. (a) Wear debris from Al–Li/SiC tested under mild wear conditions (20 ◦ C, 6.3 MPa); (b) corresponding EDS analysis; (c) wear debris from Al–Li/SiC tested under severe wear conditions (350 ◦ C, 6.3 MPa); (d) corresponding EDS analysis.


J. Rodr´ıguez et al. / Wear 262 (2007) 292–300

5. Discussion According to the methodology developed by Lim and Ashby [3], the variables controlling wear mechanism are the normalised pressure, F¯ , and the normalised velocity, v¯ , defined as F¯ =

F An H


v¯ =

vr d

where F is the normal force, An , the nominal contact area, H the material hardness, v the sliding velocity, d, the thermal diffusivity and r is the radius of the circular nominal contact area. In the tests performed at room temperature, normalised pressures are lower than 3 × 10−2 and normalised velocities lower than 2.4. These values are associated with mild wear, but another factors like temperature softening or surface roughness may alter the conditions, placing the high temperature tests in the severe regime. The measured wear rates confirm this situation. As it has been pointed out in Section 4, variables that seem to control the transition from mild to severe wear are the load and the bulk temperature. A recent work on the mechanical properties of these aluminium lithium alloys reports deformation and failure mechanisms depending on temperature [13]. Actually, the ratio between the severity of the contact (pressure, friction coefficient and velocity) and the material resistance (hardness at the test temperature) controls the wear behaviour. This is confirmed by the experiments which show a clear dependency of the transition temperature with the normal pressure. Zhang and Alpas [14] showed that while the reinforcement particles support the loads, the wear rate is maintained at lower values. Mechanical predictions have been carried out in different metal matrix composites from different approaches: uniaxial models and finite element calculations [15]. In spite of the differences, the same conclusion can be established from the qualitative point of view: if the reinforcement particles fracture, the wear rate is considerably increased. This situation is also confirmed in this work for the Al-8090 + 15 vol.% SiC composite. The presence of the SiC particles provides a higher thermal stability and, consequently, the transition temperature from mild to severe wear increases more than 50 ◦ C in the composite. Within the mild wear regime, the reinforcement particles constraint the extent of plastic deformation. In Fig. 8a (Al-8090 tested at 200 ◦ C and 6.3 MPa) and Fig. 8b (composite tested at 200 ◦ C and 6.3 MPa), hardness measurements obtained from nanoindentation show hardening effect up to higher distance from the worn surface in the unreinforced alloy. Wear debris is another relevant aspect to be considered. At low pressures, it is mainly equiaxed with a composition mixed from the wearing material and the counterbody. When the load is increased, the dominant wear mechanism becomes to delamination and severe plastic deformation. With the morphology of the wear debris collected there is not doubt about this question, because the large wear particles show evident signal of gross plastic deformation (see Figs. 6c and 7c). On the other hand, the results obtained revealed that under conditions of mild wear, the wear rate is minimum at temperatures around 100 ◦ C. Initially, the wear rate decreases with temperature, but once the minimum is reached, subsequent

Fig. 8. (a) Al-8090 tested under mild wear conditions (200 ◦ C and 6.3 MPa). It is showed the three zones observed: mechanically mixed layer (MML), hardened zone (HZ) and not hardened zone (NHZ) (b) Al–Li/SiC tested under mild wear conditions (200 ◦ C and 6.3 MPa). It is showed the three zones observed: mechanically mixed layer (MML), hardened zone (HZ) and not hardened zone (NHZ).

increments are observed up to the transition temperature from mild to severe wear. This behaviour has been previously observed in other materials [14,16]. This type of variation should be due to a simultaneous effect of two competitive phenomena. The very well known material softening at elevated temperatures should be compensated by another effect addressed in the opposite direction. It is well established that within the mild wear regime, wear rate of aluminium and its composites is controlled by the formation of mechanically mixed layers rather than by the bulk strength of the material. Unfortunately, the actual role played by the MML is still under controversy. Some authors have indicated that the MML is a protective layer [17], while others have not observed any benefit. Ghazali et al. have recently published a systematic work focused on the effect of the aluminium alloy composition on the formation and resulting properties of the MML [18].

J. Rodr´ıguez et al. / Wear 262 (2007) 292–300


Fig. 9. Micrographs showing: (a) backscattered electron image of the Al-8090 MML (20 ◦ C, 6.3 MPa); (b) corresponding EDS analysis of Al-8090 MML.

Fig. 10. Micrographs showing: (a) backscattered electron image of the composite MML (100 ◦ C, 6.3 MPa); (b) corresponding EDS analysis of the composite MML.

Results are disappointing because no clear tendencies could be established and more fundamental research is needed. In the particular case of the aluminium lithium alloy studied in this work, the presence of this MML seems to be beneficial. No appreciable differences have been observed in the microstructure and composition of the MML of the specimens analysed. The microstructure of this layer is not uniform and consists of an agglomeration of smaller particles with a high level of porosity. These characteristics indicates that MML has a lubrication effect rather than to be a hard layer. As an example, Fig. 9 (Al-8090 tested at 20 ◦ C and 6.3 MPa) and Fig. 10 (Al-8090 + SiCp tested at 100 ◦ C and 6.3 MPa) show details of the MML formed in the unreinforced alloy and in the composite, respectively. Measured friction coefficients also show a minimum at temperatures higher than 20 ◦ C (see Fig. 3). Finally, there is a remaining question to be resolved: why is the wear rate of the composite higher than that of the unreinforced alloy in the mild wear regime? Although there is no experimental proof, a plausible explanation may be related to the abrasive role played by the reinforcement particles when coming to the contact surface by the wear process. As Fig. 3 shows, friction coefficient is always higher in the composite than in the unreinforced alloy.

6. Conclusions Tribological behaviour of Al–Li and Al–Li + 15% SiC composite have been experimentally analysed, leading to the following conclusions: • A temperature dependency transition from mild to severe wear has been observed for both materials, leading to changes of two orders of magnitude in wear rate. The temperature transition exhibits a clear dependency on nominal pressure. • The reinforcement benefit is limited to shift the transition temperature to higher values. Within the mild wear regime, composite wear rates are even higher than those of the reinforced alloy. • The formation of a mechanically mixed layer seems to be a key factor controlling the mild wear of these materials. Acknowledgement Authors are indebted to Comunidad de Madrid for the financial support of this work through grant 07N/0013/2002.


J. Rodr´ıguez et al. / Wear 262 (2007) 292–300

References [1] R.L. Deuis, C. Subramanian, J.M. Yellup, Comp. Sci. Tech. 57 (1997) 415–435. [2] I.M. Hutchings, Tribology, in: Friction and Wear of Engineering Materials, Edward Arnold, London, 1992. [3] S.C. Lim, M.F. Ashby, Acta Metall. 35 (1987) 1–24. [4] Y.B. Liu, R. Asthana, P. Rohatgi, J. Mater. Sci. 26 (1991) 99–102. [5] M.A. Mart´ınez, A. Mart´ın, J. Llorca, Scripta Metall. Mater. 28 (1993) 207. [6] L. Singh, A.T. Alpas, Scripta Metall. Mater. 32 (1995) 1099–1105. [7] A.T. Alpas, J. Zhang, Metall. Trans. A 25 A (1994) 1–14. [8] K.K. Sankaran, N.J. Grant, Mater. Sci. Eng. 44 (1980) 213–227. [9] T.C. Willis, Met. Mater. 4 (1988) 485–488.

[10] P. Poza, J. Llorca, Metal. Trans. A 30 (1999) 845–855. [11] W.S. Miller, J. White, D.J. LLoyd, The metallurgy of Al–Li based alloys, in: Aluminum Alloys: Their Physical and Mechanical Properties, EMAS, Warley, 1986, pp. 1799–1836. [12] T.H. Sanders, E.A. Starke Jr., in: T.H. Sanders, E.A. Starke Jr. (Eds.), Proceeding of the Fifth International Aluminium–Lithium Conference, MCEP, Birmingham, 1989, pp. 1–37. [13] C. Gonz´alez, A. Mart´ın, J. Llorca, Scripta Mater. 51 (2004) 1111– 1115. [14] J. Zhang, A.T. Alpas, Mater. Sci. Eng. A 161 (1993) 273–284. [15] J. Rodr´ıguez, A. Mart´ın, J. Llorca, Acta Mater. 48 (2000) 993–1003. [16] Y.-H. Lin, J. Du, S.-R. Yu, W. Wang, Wear 256 (2004) 275–285. [17] B. Venkataraman, G. Sundararajan, Wear 245 (2000) 22–28. [18] M.J. Ghazali, W.M. Rainforth, H. Jones, Wear 259 (2005) 490–500.