Dry sliding wear behavior of Fe–28Al and Fe–28Al–10Ti alloys

Dry sliding wear behavior of Fe–28Al and Fe–28Al–10Ti alloys

Materials Science and Engineering A366 (2004) 127–134 Dry sliding wear behavior of Fe–28Al and Fe–28Al–10Ti alloys Xingsheng Guan1 , Kunihiko Iwasaki...

678KB Sizes 13 Downloads 57 Views

Materials Science and Engineering A366 (2004) 127–134

Dry sliding wear behavior of Fe–28Al and Fe–28Al–10Ti alloys Xingsheng Guan1 , Kunihiko Iwasaki∗ , Katsuhiro Kishi, Masafumi Yamamoto, Ryohei Tanaka Japan Ultra-high Temperature Materials Research Institute, 573-3 Okiube, Ube, Yamaguchi 755-0001, Japan Received 5 March 2003; received in revised form 29 August 2003

Abstract Dry or unlubricated sliding wear tests were carried on Fe–28Al and Fe–28Al–10Ti alloys produced by hot isostatic pressing (HIP) of powders. Since the Ti addition was found to lower both the wear rate and the coefficient of friction significantly, most of the detailed experiments were carried out on Fe–28Al–10Ti. The wear rate increased with the increase in applied load, speed and temperature, while the steady-state values of the coefficients of friction were affected little by them. Plastic deformation and the delamination or peeling-off of surface layers from the wear track were found to be the basic mechanisms in the mild sliding conditions (low load and speed) and in the severe ones (high load and speed), respectively. The wear rate was found to have a clear correlation with the strain-hardening rate but neither with the hardness nor with the yield strength. This implies that the wear properties may be related more keenly with dynamic mechanical properties than static ones. © 2003 Elsevier B.V. All rights reserved. Keywords: Fe3 Al-type intermetallic alloys; Wear properties; Wear mechanisms; Correlation between mechanical and tribological properties

1. Introduction Iron aluminides such as Fe3 Al have attracted great attention as potential structural materials in industrial applications because of their high strength/density ratio, remarkable corrosion resistance and low raw material cost [1]. Research efforts have been focused mainly on physical and mechanical properties and corrosion resistance. Their excellent properties, however, suggest that they will be also promising as tribo-materials to be used in aggressive conditions. Since loads are mostly compressive in tribological uses, poor tensile ductility generally found in intermetallic alloys is not considered to be so crucial as in other applications. By taking these advantages of the intermetallic alloys into account a series of works on their application to tribology have been performed by the present research group [2–4]. It has been found that Fe–Al alloys showed good wear properties and the wear resistance was significantly improved by the addition of high amount of C. Kim et al. [5,6] investigated room temperature dry sliding wear behavior of iron aluminides

of various composition ratios. They reported that the wear rate of the aluminides increased with the increase in applied load and sliding speed and the wear resistance decreased with the increase in aluminum contents. Hawk et al. [7] and Maupin et al. [8] reported that the addition of Ti to Fe3 Al was very effective in improving the anti-abrasive properties. Only these are the main works concerning the tribology of Fe–Al alloys as far as the authors know and more extensive works are highly desired for applying these alloys in the tribological fields. It is especially important from the practical point of view to clarify the effects of applied load, sliding speed and temperature on the tribological properties such as wear rate and the coefficient of friction. The present work was carried out along these requirements as a part of a national project to develop ecologically tailored tribo-materials. The correlation between mechanical and tribological properties was also examined in order to find a guideline for developing new tribo-materials. 2. Experimental procedure

∗ Corresponding author. Tel.: +81-836-51-7007; fax: +81-836-51-7011. E-mail address: [email protected] (K. Iwasaki). 1 Present address: Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Canada.

0921-5093/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2003.09.049

Specimens were prepared by a powder metallurgical method in the following ways. The powders were prepared by a gas-atomization process using commercial-purity Fe, Al and Ti bars. The raw materials with nominal compositions

128

X. Guan et al. / Materials Science and Engineering A366 (2004) 127–134

Fig. 1. SEM photographs illustrating the morphologies of the powders: (a) as gas-atomized, (b) after mechanical alloying (with 0.5% Y2 O3 addition).

of Fe–28Al and Fe–28Al–10Ti were induction-melted at about 1900 K under Ar atmosphere and kept at this temperature for 600 s for homogenization. The melted alloys were then gas-atomized into fine powders using Ar gas of 4.9 MPa in blowing pressure. The produced powders were classified by ultrasonic sieving and only those under 180 ␮m in size were used in the subsequent process. Although the gas-atomization process has been developed to produce high-quality powders, their sintering ability is rather low because of their spherical shape as seen in the SEM photo of Fig. 1(a). In order to improve the sintering ability, the gas-atomized powders were milled with steel balls at an agitation speed of 159 rpm for 180 ks under Ar atmosphere with a mechanical-alloying apparatus. The mass of the powders was 3 kg and the ball-to-powder weight ratio was 10:1. After the milling the powders became irregular in shape as shown in Fig. 1(b). This flaky shape and fine size of the powders improved the sintering ability significantly. During the milling 0.5 vol.% Y2 O3 powders were added and homogeneously dispersed because they had been known very effective for improving yield strength and room temperature ductility [9]. The powders were then sieved again to select only those under 100 ␮m, degassed at 773 K for 3.6 ks and sealed into mild steel cans in vacuum to be consolidated by hot isostatic pressing (HIP). The consolidation was performed at 1173 K under a pressure of 980 MPa for a holding time of 10.8 ks. Tribological properties were evaluated with a ball-on-disk type tribometer (CSEM Instrument) in air. The specimen was of rectangular shape (20 mm × 20 mm × 5 mm) and the surface to be measured was polished to a roughness of 0.05 ␮m. The specimen was put on a rotation deck and an alumina ball of 6 mm in diameter and 2400 HV in hardness was pushed against it as a counter material. The wear of alumina ball is negligible in comparison with the specimen. The tests were conducted in dry or unlubricated condition at sliding speeds between 0.05 and 0.2 m/s under loads between 1 and 10 N at temperatures between room temperature and 773 K with total disk revolutions of 4000. The frictional

force was measured with a load cell and converted into the coefficient of friction. The wear rate was evaluated by dividing the volume of the wear scar by the load and the sliding distance. Three kinds of sliding radii, 4, 6 and 8 mm, were used for each specimen, where the rotational speed of the disk was adjusted so as to maintain a constant sliding speed. The data shown below are the averages of those taken at three different radii. Mechanical properties were evaluated in compressive modes, that is, by Vickers hardness and compressive tests. The Vickers hardness was measured under a load of 98 N with the same specimens as were used for the ball-on-disk measurements. Rectangular specimens with dimensions of 3 mm × 3 mm × 6 mm were used for the compressive tests. The tests were carried out in air from room temperature up to 1073 K at an initial strain rate of 1.4 × 10−3 s−1 with Shimadzu high temperature material testing system. Microstructural observation of the wear scars and wear debris was carried out by scanning electron microscopy (SEM). The composition of the specimen was analyzed by X-ray diffraction (XRD) and electron probe microanalysis (EPMA).

3. Experimental results 3.1. Effect of Ti addition Effects of Ti addition on the wear behavior are shown in Fig. 2. Fig. 2(a) shows the variations of the coefficients of friction of Fe–28Al and Fe–28Al–10Ti alloys as a function of the number of disk revolutions (lower abscissa) or sliding distance (upper abscissa). The tests were performed with a load of 5 N and sliding speed of 0.05 m/s at room temperature. The curve of Fe–28Al–10Ti alloy is divided into the following four stages as designated in the figure; stage A: initial sharp increase, stage B: gradual increase, stage C: second sharp increase and stage D: plateau with nearly constant value. In the case of Fe–28Al alloy, however, the

X. Guan et al. / Materials Science and Engineering A366 (2004) 127–134

1.0

0

50

Sliding Distance, m 100 150

Coefficient of Friction

0.8

200 1.0

0.8 Fe-28Al-10Ti Fe-28Al

0.6

0.4

0.6

0.4 D

0.2

C

0.2

B

0.0

A

0

1000

(a)

2000 3000 Number of Revolution

0.0

-1

10

Load = 5 N Speed = 0.05 m/s

-2

3

Wear Rate, mm m

-1

10

-3

10

-4

10

-5

10

-6

(b) 10

Fe-28Al

Fe-28Al-10Ti

Fig. 2. Comparison of tribological behavior of Fe–28Al and Fe– 28Al–10Ti: (a) coefficient of friction, (b) wear rate.

stages B and C are missing, that is, the coefficient of friction increases directly from the initial stage to reach the final plateau values without passing through the intermediate stages B and C. It is to be noted here that the average values taken at the stage D decreases by about 10% by the Ti addition. The wear rate is also reduced by a factor of about 300 as shown in Fig. 2(b). In order to obtain better understanding of the friction behavior given above, the surface topography was conducted at each stage and the results are shown in Fig. 3. Plowing grooves and some debris were observed on the sliding surfaces at the beginning of the sliding (stage A). As the sliding continued, the number of plowing grooves decreased and the wear scar became smooth at the stage B. Deep and wide plowing grooves were found at the stage C. Moreover, the specimen suffered extensive surface damage and wear debris was gradually accumulated on the surface at this stage. At the steady stage D, the plowing grooves and some voids were formed on the wear scar. However, the wear scar was much smoother than that of the stage B. As for Fe–28Al alloy, the agglomeration of wear debris and plowing grooves

129

were observed on the sliding surface throughout all the sliding processes, though the photos were not shown here. 3.2. Effects of applied load, sliding speed and temperature Fig. 4 shows the behavior of the coefficients of friction of Fe–28Al–10Ti alloy tested in various conditions. The steady state values of the coefficients of friction showed little dependence on the applied load and on the sliding speed as shown in Fig. 4(a) and (b), though the value at the lowest load (1 N) exhibited a slightly higher one. The coefficients of friction decreased with the increase in temperature and the stages B and C disappeared at 773 K as shown in Fig. 4(c). The dependence of the wear rate of Fe–28Al–10Ti alloy on applied load, sliding speed and temperature is shown in Fig. 5(a)–(c), respectively. The wear rate increased with the increase in the applied load and a sudden great increase was observed between 5 and 7 N. The wear rate also increased abruptly at the sliding speed of 0.2 m/s. As for the temperature dependence, however, a maximum value was attained at 573 K. A series of SEM photos of the wear debris and the wear scar are shown in Fig. 6(a)–(h), respectively. The photos on the same line were taken after the sliding test conducted in the conditions of (1), (2), (3) and (4) shown in the caption. In the mild testing condition (1) the debris is in the form of fine powder and a trace of micro-ploughing is observed as shown in Fig. 6(a) and (e). However, in the severe conditions of high load (2), high speed (3) and high temperature (4), the surface appearance was quite different. The wear debris became large and flaky, the wear scars were heavily damaged, and the platelets and a trace of their detachment from the surface were observed together with the micro-ploughing. 3.3. Correlation between sliding wear behavior and mechanical properties A trial to find a correlation between mechanical and tribological properties was performed here, where the Vickers hardness, the 0.2% compressive yield strength and the strain-hardening rate were chosen as mechanical properties and the wear rate as a tribological one. The strain-hardening rate was estimated by taking the average of the slopes of the stress-strain curve at the strain levels between 0.02 and 0.2. Since it is necessary to vary the mechanical and the tribological properties as widely as possible to check the correlation, the data used in the previous works [10,11] are also included in the following figures, that is, Ti concentration was varied between 0 and 15 at.% [10] and the specimens were prepared not only by powder metallurgical method (PM) [10] but also by vacuum induction melting (VIM) [11]. As shown in Fig. 7 no clear correlation was observed between the wear rate and the Vickers hardness. The situation was the same for the correlation between the wear rate and the 0.2% yield strength (Fig. 8). However, a very

130

X. Guan et al. / Materials Science and Engineering A366 (2004) 127–134

Fig. 3. Photographs of wear scars of Fe–28Al–10Ti at various stages: (a) stage A, (b) stage B, (c) stage C, (d) stage D.

clear correlation was found to exist between the wear rate and the strain-hardening rate as shown in Fig. 9, where the correlation coefficient was as high as 0.94.

4. Discussion 4.1. Wear behavior The curve of the coefficient of friction of Fe–28Al–10Ti alloy showed four stages (Fig. 2). This implies that each stage may result from different mechanisms. At the beginning of the sliding process (stage A), plowing friction mechanism (shown in Fig. 3(a)) is considered to play the most important role in determining the initial coefficient of friction. At the stage B, the fragmentation of the ridges formed along the sides of the plowing grooves strongly adhered to the wear scar and then resulted in the smooth surface as the sliding continued. This adhesion effect may be the reason for the gradual increase in the coefficient of friction at the stage B (Fig. 3(b)). At the stage C, the debris agglomerated to form particles or flakes near the deep grooves on the sliding surface because of work hardening (Fig. 3(c)). The par-

ticles entrapped at the interface cause progressive damage to the metal surfaces by further plowing. Microscopic studies (Fig. 3(d)) revealed that the amount of agglomerated wear debris at the stage D was much less than that at the stage C and the wear scar became smooth, though some voids were observed. At this stage, plastic deformation, plowing by wear debris entrapped between the sliding surfaces and delamination seem to occur steadily. The fluctuation observed at the stage D is considered to be due to the existence of the voids. For Fe–28Al alloy, the phenomena of wear debris agglomeration and deep plowing grooves were observed from the very beginning of sliding. Adhesion effect as shown in Fig. 3(b) did not occur for this alloy. The lower coefficient of friction of Fe–28Al–10Ti at the steady state is ascribed to the smooth wear scar since the size of the wear debris of Fe–28Al is larger than that of Fe–28Al–10Ti by a factor of 10. The large wear debris existing between the contacting surfaces in Fe–28Al alloy resulted in the increase in surface roughness, which consequently increased the coefficient of friction. The great improvement in wear resistance by the Ti addition is mainly attributed to the precipitations of hard second phases including Ti. Other two factors are also thought to

X. Guan et al. / Materials Science and Engineering A366 (2004) 127–134

1.0

0

50

Sliding Distance, m 100 150

-1

10

200 10 3

Wear Rate, mm /m

0.8

(a) Sliding speed = 0.1 m/s at R.T.

-2

(a) Sliding speed = 0.1 m/s at R.T. Coefficient of Friction

131

Applied Load 1N 3N 5N 7N 10N

0.6 0.4 0.2

-3

10

-4

10

-5

10

-6

10

0.0 1.0

3

5 7 Applied Load, N

10

10

(b) Applied load = 5 N at R.T.

(b) Applied load = 5 N at R.T. -2

10

3

0.8

Wear Rate, mm /m

Coefficient of Friction

1

-1

Sliding speed 0.05 m/s 0.1 m/s 0.2 m/s

0.6 0.4

-3

10

-4

10

-5

10

0.2

-6

10

(c) Applied load = 5 N Sliding speed = 0.1 m/s

0.05

0.1 0.2 Sliding Speed, mm/s

-1

10

0.8 -2

(c) Applied load = 5 N Sliding speed = 0.1 m/s

10 3

0.6 R.T. 573 K 773 K

0.4 0.2 0.0

Wear Rate, mm /m

Coefficient of Friction

0.0 1.0

0

1000 2000 3000 Number of Revolution

-3

10

-4

10

-5

10

-6

10

293

573 Temperature, K

773

Fig. 4. Coefficient of friction of Fe–28Al–10Ti vs. the number of revolution or sliding distance in various measuring conditions: (a) effect of applied load, (b) effect of sliding speed, (c) effect of temperature.

Fig. 5. Wear rate of Fe–28Al–10Ti measured in various conditions: (a) effect of applied load, (b) effect of sliding speed, (c) effect of temperature.

contribute to the improvement. One is the ordering strengthening effect, that is, the Ti addition increases the degree of order of the compacted Fe3 Al alloy [10]. Wert et al. [12] reported that the wear rate for the disordered Cu3 Au was nearly twice as high as that for the ordered one. The other is the dispersion strengthening effect. Ti and oxide (Y2 O3 ) addition leads to the dispersion of fine particles in the matrix, which will act as obstacles for dislocation motion during deformation. The wear rate increased sharply when the applied load was larger than 5 N and the sliding speed was higher than 0.1 m/s

(Fig. 5). This implies that the wear mechanisms change with the sliding conditions. The micro-ploughing and the voids observed in Fig. 6(e) and fine debris in Fig. 6(a) indicate that the wear occurs mainly by plastic deformation in the mild sliding conditions. However, in the severe sliding conditions, the formation of surface platelets and their subsequent detachment from the surface as observed in Fig. 6(b), (c), (f) and (g) seem to be the major wear mechanisms. Similar phenomena have been observed in Al alloys [13] and a C-steel [14,15]. Archard and Hirst [13] investigated the wear behavior of aluminum alloys as a function of the

132

X. Guan et al. / Materials Science and Engineering A366 (2004) 127–134

Fig. 6. SEM photographs of wear debris (left) and wear scars (right) of Fe–28Al–10Ti taken after sliding tests in various conditions: (1) load = 5 N, speed = 0.1 m/s measured at room temperature, (2) load = 10 N, speed = 0.1 m/s measured at room temperature, (3) load = 5 N, speed = 0.2 m/s measured at room temperature, (4) load = 5 N, speed = 0.1 m/s measured at 773 K.

X. Guan et al. / Materials Science and Engineering A366 (2004) 127–134 -1

-3

10

-4

10

-5

10

-6

200

300

400

500

600

700

Vickers Hardness, HV Fig. 7. Relation between Vickers hardness and wear rate. VIM: specimens prepared by vacuum induction melting. PM: specimens prepared by powder metallurgical method.

applied load and sliding speed. They found that sliding contact at high loads and sliding speeds usually resulted in severe wear. The severe wear involved massive surface damage and large-scale material transfer to the counterface. For the mild wear (low loads and speeds), wear surface damage is less extensive and the wear scars are generally covered by wear debris. Welsh [14,15] observed that the onset of seizure occurred at a critical combination of the applied load and the sliding speed. The wear rate measured at elevated temperatures was higher than that at room temperature (Fig. 5). The wear at higher temperatures probably originates from the plastic deformation and subsequent oxide delamination as is supposed from the observation of the wear scar (Fig. 6(h)) and the wear debris (Fig. 6(d)). The occurrence of micro-ploughing 10

-1

10

-2

10

-3

10

-4

10

-5

10

-6

Fe3AlTi(VIM) Fe3AlTi(PM)

3

Wear Rate, mm /m

10

-1

10

-2

10

-3

10

-4

10

-5

10

-6

Fe3AlTi(VIM) Fe3AlTi(PM)

3

10

Wear Rate, mm /m

10

-2

Fe3AlTi(VIM) Fe3AlTi(PM)

3

Wear Rate, mm /m

10

133

0

2 4 6 Strain Hardening Rate, GPa

8

10

Fig. 9. Relation between strain-hardening rate and wear rate. VIM: specimens prepared by vacuum induction melting. PM: specimens prepared by powder metallurgical method.

and cracking was observed in the wear scars and platelets in the wear debris. These phenomena are considered to be related with the formation of oxide film on the surface at high temperatures. The oxide on the wear scar was identified to be Fe2 O3 by XRD analysis. Oxide layers that were repeatedly produced and removed would cover the wear scars. It is not certain at the present stage, however, whether the oxide layers are beneficial for the improvement of wear resistance or not [16,17]. If the oxide layer is ductile, thick and continuous and adheres to the surface as a solid lubricant, it may suppress the direct metallic contact and improve the wear resistance. On the other hand, if the oxide layer is brittle, thin and discrete, it may act as hard impurity or particle (third body) between mating surfaces and decrease the wear resistance. Oxides formed in Fe–28Al–10Ti alloy seem to correspond to the later case, because the wear rate increased at elevated temperatures. The reason for the improvement of wear resistance at 773 K in comparison with that of 573 K is not clear at the present stage. However, since the extent of oxidation should be obviously greater at higher temperature, the thicker and more continuous oxide surface observed at 773 K seems to result in the decrease in the wear rate. 4.2. Correlation between mechanical properties and wear behavior

0

500

1000

1500

2000

2500

0.2% Yield Strength, MPa Fig. 8. Relation between 0.2% yield strength from compressive test and wear rate. VIM: specimens prepared by vacuum induction melting. PM: specimens prepared by powder metallurgical method.

In spite of many trials to find a correlation between mechanical and tribological properties no clear understanding has been reached yet. This has been the most troublesome obstacles against the development of new tribo-materials and the clarification of the correlation has been highly desired. The difficulty lies in the fact that the tribological properties are not attributed to the relevant material itself but depend greatly on counter-materials, environments and other

134

X. Guan et al. / Materials Science and Engineering A366 (2004) 127–134

measuring conditions. Moreover, attention has been paid mostly to such mechanical properties as hardness and yield strength. These properties reflect only particular or instantaneous stages of deformation processes, that is, they are so-called static properties. The tribological processes, however, are continuously proceeding dynamic ones. Then more attention should be paid to the dynamic mechanical properties than to the static ones. With these suppositions in mind the experimental data including the previous ones were analyzed here and a clear correlation was found only between the wear rate and the strain-hardening rate (Figs. 7–9). Since the strain-hardening is a dynamic process consisting of plastic deformation, fatigue and micro-fracture and so on, it is regarded as one of the dynamic mechanical properties. This is to be contrasted to the static mechanical properties such as Vickers hardness and the 0.2% yield strength. The wear process is also considered to be a dynamic one where plastic deformation, fatigue and micro-fracture and so on play important roles together with material transfer and tribo-chemical reactions [18,19]. Though the wear process is not so simple as to be regarded as similar to the strain-hardening process, the experimental results found here indicates that the dynamic mechanical properties have keener relation with the wear properties than the static ones. If the similar correlation were found also in other materials, it would be very helpful for the development of new tribo-materials.

5. Conclusions Dry sliding wear behavior of Fe–28Al and Fe–28Al–10Ti was investigated to obtain the following results: (1) The Ti addition lowered significantly the wear rate and decreased slightly the coefficient of friction. (2) The wear rate increased with the increase in applied load, speed and temperature, while the steady-state values of the coefficients of friction were affected little by them. (3) Plastic deformation and the delamination or peeling-off of surface layers from the wear track were found to be the basic mechanisms in the mild sliding conditions (low load and speed) and in the severe ones (high load and speed), respectively.

(4) The wear rate was found to have a clear correlation with the strain-hardening rate but neither with the hardness nor with the yield strength. This implies that the wear properties may be related more keenly with dynamic mechanical properties than static ones.

Acknowledgements Financial support by the New Energy and Industrial Technology Development Organization (NEDO), Japan, is greatly acknowledged. Prof. Takao Araki of Ehime University is also acknowledged for his useful suggestion on the correlation between mechanical and tribological properties.

References [1] N.S. Stoloff, Mater. Sci. Eng. A258 (1998) 1. [2] S.M. Zhu, M. Tamura, X.S. Guan, K. Sakamoto, T. Horiya, K. Iwasaki, in: K.J. Hemker, D.M. Dimiduk, H. Clemens, R. Darolia, H. Inui, J.M. Larsen, V.K. Sikka, M. Thomas, J.D. Whittenberger (Eds.), Structural Intermetallics, TMS, 2001, p. 755. [3] S.M. Zhu, X.S. Guan, K. Shibata, K. Iwasaki, Mater. Trans. 43 (2002) 36. [4] X.S. Guan, S.M. Zhu, K. Shibata, K. Iwasaki, Mater. Trans. 43 (2002) 1325. [5] Y.-S. Kim, Y.-H. Kim, Mater. Sci. Eng. A258 (1998) 319. [6] Y.-S. Kim, J.-H. Song, Y.W. Chang, Scr. Mater. 36 (1997) 829. [7] J.A. Hawk, D.E. Alman, R.D. Wilson, Mater. Res. Soc. Sympo. Proc. 364 (1995) 243. [8] H.E. Maupin, R.D. Wilson, J.A. Hawk, Wear 162–164 (1993) 432. [9] S. Strothers, K. Vedula, Mod. Dev. Powder Metall. 18 (1988) 597. [10] S.M. Zhu, M. Tamura, K. Sakamoto, K. Iwasaki, Mater. Sci. Eng. A292 (2000) 83. [11] S.M. Zhu, K. Sakamoto, M. Tamura, K. Iwasaki, Scr. Mater. 42 (2000) 905. [12] J.J. Wert, F. Srygley, C.D. Warren, R.D. Reynolds, Wear 134 (1989) 115. [13] J.F. Archard, W. Hirst, Proc. Roy. Soc. A236 (1956) 397. [14] N.C. Welsh, R. Soc. London, Philos. Trans. Series A 257 (1964–65) 5. [15] N.C. Welsh, R. Soc. London, Philos. Trans. Series A 257 (1964–65) 51. [16] P.B. Madakson, Wear 94 (1983) 193. [17] N. Saka, E.M. Eleiche, N.P. Suh, Wear 44 (1977) 105. [18] Y. Kimura, in: D.A. Rigney (Ed.), “Fundamentals of Friction and Wear of Materials”, Am. Soc. Metals (1981) 187. [19] Y. Kimura, J. Phys. D: Appl. Phys. 25 (1992) A177.