Microstructure and dry-sliding wear properties of DC plasma nitrided 17-4 PH stainless steel

Microstructure and dry-sliding wear properties of DC plasma nitrided 17-4 PH stainless steel

Available online at www.sciencedirect.com NIM B Beam Interactions with Materials & Atoms Nuclear Instruments and Methods in Physics Research B 266 (...

2MB Sizes 0 Downloads 42 Views

Available online at www.sciencedirect.com

NIM B Beam Interactions with Materials & Atoms

Nuclear Instruments and Methods in Physics Research B 266 (2008) 1964–1970 www.elsevier.com/locate/nimb

Microstructure and dry-sliding wear properties of DC plasma nitrided 17-4 PH stainless steel Gui-jiang Li a, Jun Wang a, Cong Li b, Qian Peng b, Jian Gao c, Bao-luo Shen a,* a

College of Materials Science and Engineering, Sichuan University, Chengdu 610064, China b Nuclear Power Institute of China, Chengdu 610041, China c Chengdu Tool Institute, Sichuan, Xindu 610051, China Received 10 October 2007; received in revised form 19 December 2007 Available online 13 March 2008

Abstract An attempt that the precipitation hardening steel 17-4PH was conducted by DC plasma nitriding (DCPN) is made to develop a kind of candidate material for nuclear reactor. Nitriding process performed at temperature 6 400 °C takes effect on creation of the layers composed of S-phase (expanded austenite) and a0N (expanded martensite). Up to the temperature of 420 °C, the S-phase peaks disappear due to the transformation occurrence (S-phase ? a0N + CrN). For the samples nitrided at temperature P 450 °C, no evidence of a0N is found owing to a precipitation (a0N ! a þ CrN) taking place. For the 480 °C/4 h treated sample, it is the surface microhardness that plays the lead role in the wear rate reduction but the surface roughness; while for the 400 °C/4 h treated sample, it is both of the surface roughness and the S-phase formation. Dry sliding wear of the untreated 17-4PH is mainly characterized by strong adhesion, abrasion and oxidation mechanism. Samples nitrided at 400 °C which is dominated by slight abrasion and plastic deformation exhibit the best dry sliding wear resistance compared to the samples nitrided at other temperatures. Ó 2008 Elsevier B.V. All rights reserved. Keywords: 17-4PH stainless steel; DC plasma nitriding; Microstructure; Wear resistance

1. Introduction It is reported that 17-4 precipitation hardening stainless steel (17-4 PH) has been increasingly used in a variety of applications including aircraft fittings [1], gears, fasteners, compressor impeller [2,3], nuclear reactor components [4– 6]. However, the wider applications are restricted by their relatively low hardness and poor tribological properties, for example, the components made from 17-4PH stainless steel for nuclear reactor are unable to service due to the mild wear failure. Under this background, it is required to try to take advantage of a certain surface engineering technique to address the problem in order to improve the anti-wear property of 17-4 PH. Plasma nitriding is estab*

Corresponding author. Tel.: +86 13688456619. E-mail address: [email protected] (B.-l. Shen).

0168-583X/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2008.02.073

lished to be an effective way for the wear resistance improvement of stainless steel [7,8], consequently, it can be employed to make an attempt to develop a kind of candidate material for nuclear reactor. It is found that some preliminary studies have been conducted to explore the possibility of enhancing the surface hardness of 17-4 PH stainless steels by plasma nitriding; however, little or no attention has been paid to the correlation of microstructure with dry-sliding wear behavior of the plasma nitrided material [9–11]. Probably, new insights into the dry-sliding mechanism of the plasma nitrided 17-4 PH stainless steel can be probably provided by the ring-on-block contact configuration used in the test. Therefore, the aim of this paper is to investigate the influence of microstructure on the dry-sliding wear resistance of 17-4 PH stainless steel by DC plasma nitriding temperature, as well as the wear mechanism.

G.-j. Li et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1964–1970

1965

2. Experimental 2.1. Material and treatments The samples were prepared from grade 17-4 PH stainless steel with the following composition (wt.%): 0.04% C, 16.39% Cr, 4.32% Ni, 3.40% Cu, 0.30% Mn, 0.60% Si, 0.023% P, 0.36% Nb, trace Mo and balance Fe. The samples were cut from a hot rolling bar and then machined into a size of 10 mm  10 mm  10 mm. The flat surfaces of the block sample were manually grounded, using sandpapers (180, 320, 400, 600 and 1200 mesh), polished with diamond slurry in a polishing machine to achieve a fine finish (Ra  0.1 lm) and then ultrasonically cleaned with acetone, alcohol and distilled water in succession before plasma nitriding. Plasma nitriding was carried out using a Type CTI-7C400M DC plasma nitriding unit in a 20% N2 + 80% H2 atmosphere at a pressure of 6  103 bar. A wide range of processing temperatures from 350 to 480 °C and time of 4h was used (Table 1). Sputtering pre-treatment was carried out during the heating up stage for a period of about 2 h using the same gas composition. 2.2. Characterization The phase composition of nitrided layers was characterized by Type Dmax-1400 X-ray diffraction analyzer (XRD) using Cu Ka radiation (k = 0.154056 nm). Micro-hardness testing was performed on the surface of the nitrided samples on a type of HV-1000 tester. Loads of 100 g and dwell times of 15 s were employed. For the surface micro-hardness value of each sample, five indentations at different positions of the treated layer were used to evaluate the micro-hardness. The cross-section of sample was accomplished by carefully removing consecutive layers by mechanical polishing. Information about the nitrogen cross-sectional profiles at different temperatures is obtained by glow discharge optical spectroscopy (GDOS). The surface roughness Ra was measured using a Talysurf 10 surface profilometer. The cross-sectional microstructure and the worn surface morphologies of the nitrided layers were evaluated with Type JSM5910-LV scanning electron microscopy (SEM). Dry sliding tribological tests were performed using a rotational speed of 200 rpm (equating to a linear sliding velocity of 9.59 cm/s) for duration of 3600 s. A normal operating load of 30 N was used. For the counter body, 50-mm-diameter GCr15 rings (HRC 52) were selected. The ring-on-block contact configuration

Fig. 1. Schematic illustration of the wear test device.

employed to access the sample’s tribological properties was shown in Fig. 1. Wear loss was measured by Type TG328A analytical balance with an accuracy of 0.1 mg. The wear rates of the original substrate and all the nitrided layers were calculated using the equation of K = W/S where W is the wear weight in mg, S is the total sliding distance in km. 3. Results and discussion 3.1. Microstructure analysis SEM micrographs of cross-section of the nitrided layers of the sample are shown in Fig. 2. In order to evaluate the effect of temperature on the resulting microstructure, plasma nitriding of the grade 17-4 PH was carried out at various temperatures. As can be seen from the cross-sectional micrograph shown in Fig. 2(a), a bright layer of approximately 5lm is observed for the 350 °C/4 h treated sample. However, for both of the 420 °C/4 h and 480 °C/ 4 h treated samples, a relatively dark layer of approximately 10 lm is observed (Fig. 2(b), (c)). This indicates that the type of layer (from temperature above 400 °C processes) is susceptible to etching in nital contrary to layers created in lower temperature nitriding process. Additionally, since thickness differences are comprehensible in diffusion velocity context, it is observed from the Fig. 3 that the thickness of nitrided layers on 17-4 PH steel obtained at temperature higher than 400 °C increases dramatically with temperature increasing.

Table 1 Different nitriding processes of 17-4 PH Process

Temperature/ °C

Pressure/ Pa

Time/h

Voltage/V

(N2:H2) /ml min1

Current density /mA cm2

1 2 3 4 5

480 450 420 400 350

600 600 600 600 600

4 4 4 4 4

DC-520 DC-520 DC-520 DC-520 DC-520

60:240 60:240 60:240 60:240 60:240

0.9 0.9 0.9 0.9 0.9

1966

G.-j. Li et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1964–1970

Fig. 2. SEM micrograph of cross sectional microstructure of samples nitrided at various temperatures (a) 350 °C, (a) 420 °C (c) 480 °C.

Fig. 3. Case depths of 17-4 PH plasma nitrided at various temperatures.

3.2. EDX analysis Nitrogen concentration profile measured by EDX across nitrided layers is shown in Fig. 4. At high temperatures, a reduction in the surface stress due to defect recombination (Frenkel pairs) occurs and nitrogen diffusion is favored. Accordingly, the highest nitrogen contents of these nitrided

Fig. 4. Nitrogen concentration profile obtained at different temperatures for 4 h.

layers measured using EDX microanalysis amount to 12 wt.% at 480 °C. It is well known that the higher concentration values and larger diffusion depths correspond to higher temperatures. Therefore, the layers nitrided at 480 °C contain relatively higher nitrogen concentration at

G.-j. Li et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1964–1970

the surface and can have a deeper nitrogen penetration depth, compared with that of the layers nitrided at temperature < 480 °C. Moreover, nitriding depths of more than 10 lm can be achieved at correspondingly increased temperature. 3.3. XRD analysis The XRD diffraction patterns of phase structure of the layers are shown in Fig. 5. It can be clearly observed that a set of alpha (a) peaks for the unnitrided sample (donated as ‘000’) is sharp. Since nitrogen is an ‘austenite stabiliser’ and the nitrogen rich layer tends to have an f.c.c. structure rather than a b.c.c. structure at temperature 6 400 °C, the alpha (a) grains are prone to transforming to f.c.c. grains during the plasma nitriding treatment [12]. When the nitrogen diffuses inwards, the f.c.c. grain lattice is supersaturated by nitrogen to such extent that the transformation of alpha (a) into S-phase occurs [13]. Therefore, at 350 °C and 400 °C, a broad peak at the diffraction angle 2h of 42°, which has been associated with a metastable phase called ‘S-phase’ (expanded austenite, cN) [14–18], is observable together with peaks of the substrate alpha (a) and the a0N (nitrogen-containing martensite supersaturated by nitrogen) [19]. However, the S-phase peaks disappear at temperatureP420 °C. This is probably due to the fact that the precipitation of CrN depletes the expanded phase of chromium at 420 °C, favoring the formation of a mixture layer of a0N and CrN (S-phase ? a0N + CrN). It is reported that Cr-nitrides formation is favored due to their high negative enthalpy and low Cr diffusivity in the matrix at temperature higher than 450 °C [20]. Therefore, for t0t temperature P 450 °C, no evidence of a0N but CrN is found. This is possibly due to the decomposition of a0N into alpha (a) and CrN. Furthermore, the alpha (a) peak sharpens and returns to the unnitrided diffraction angle at temperature P 450 °C. A structure relaxation produced by the precipitation (a0N ! a þ CrN) is possibly the reason. The diffraction peak at 2h of 45° is broadened considerably at temperature 350, 400 and 420 °C. This is possibly due to the substrate alpha (a) and the a0N diffraction peak overlapping.

Fig. 5. XRD patterns of nitrided and unnitrided 17-4 PH.

1967

3.4. Hardness The microhardness of nitrided layers as a function of temperature is shown in Fig. 6. According to the figure, it is clearly observed that microhardness of the layers depends mainly on nitriding temperature and varies between 950 HV0.1 for temperatures about 350 °C and 1270 HV0.1 for temperatures about 480 °C. The maximum values measured from the treated surface is observed to be 1270 HV0.1 at 480 °C, which is about 3.5 times as hard as the untreated material (362 HV0.1). In general, the microhardness values are influenced by the core material and the thickness and structure of the layer. The microstructure is considered to be closely associated with the properties, so a dramatic increase in hardness at temperature 420 °C due to the formation of Cr nitrides can be observed. In fact, it is possible that the microhardness of the nitrided layer is partly influenced by layer thickness increase. Increasing the temperature, the modified region thickness increases and consequently the substrate affects less the near surface hardness. Therefore, the microhardness shows mild increase as the temperature increasing higher than 420 °C. 3.5. Wear behavior Dry-sliding wear rate data of the unnitrided 17-4PH steel and nitrided 17-4PH steel layers is shown in Table 2. These results represent the effect of superficial roughness, layer structure phases and surface microhardness on the wear behavior of plasma nitrided samples. The superficial roughness can be confirmed through images obtained by SEM, as indicated in Fig. 7(a) and (b). It can be seen that the surface morphology of sample nitrided at 480 °C is much rougher than that nitrided at 350 °C. This fact occurs due to the interaction between ions and the sample surface, mainly through the superficial sputtering which is expected to be present at the cathode surface. An increase in treatment temperature causes an increase in the arrival energy of the species at the sample surface, leading to action of

Fig. 6. Surface micro-hardness of the 17-4 PH nitrided at various temperatures.

1968

G.-j. Li et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1964–1970

Table 2 Wear rate of 17-4 PH performed at different parameters Process parameters

No. 0 unnitrided

No. 1 480 °C/600 Pa/4 h

No. 2 450 °C/600 Pa/4 h

No. 3 420 °C/600 Pa/4 h

No. 4 400 °C/600 Pa/4 h

No. 5 350 °C/600 Pa/4 h

Phase composition Microhardness (HV0.1) Roughness Ra (lm) Roughness Rmax (lm) Wear rate (mg km1)

a (Matrix) 362 0.05 0.1 10.35

CrN + a 1278 0.25 2.4 1.94

CrN + a 1187 0.21 2.3 1.85

CrN + a0N 1140 0.18 2.1 1.49

S-phase + a0N 1022 0.11 1.32 0.85

S-phase + a0N 973 0.10 1.12 1.27

Fig. 7. Surface morphology of the 17-4 PH nitrided at temperatures: (a) 400 °C (b) 480 °C.

etching and consequent surface roughening [21]. For samples nitrided at 400 °C, the resulting roughness (measured by surface profilometer) was Ra  0.11 lm, while the samples nitrided at 480 °C a Ra  0.25 lm. As can be confirmed in Fig. 7(a), it presents a smoother surface and a dense structure in top view in comparison with Fig. 7(b), resulting in smaller wear rate. All the dry sliding wear rates of the plasma nitrided samples in temperature range of 350–480 °C exhibits much lower wear rate in comparison to that of the unnitrided samples. But small difference among the dry sliding wear rates of the plasma nitrided samples can be clearly observed. Probably the surface roughness plays an import role in this. On the other hand, the dry sliding wear rate of the 400 °C nitrided sample is significantly lower than that of 350 °C, 420 °C, 450 °C, 480 °C, indicating that the most excellent dry sliding wear resistance can be obtained at temperature 400 °C. It is expected that the 350 °C/4 h treated sample should provide the same wear resistance as the 400 °C/4 h treated sample, however, the wear resistance of the 400 °C/4 h treated sample is higher that the 350 °C/4 h treated one. In fact, thickness and hardness of the nitrided layer plays an important role in improving the dry sliding wear resistance of samples. A thicker and harder layer can be obtained at temperature 400 °C in comparison to that at 350 °C. Therefore, samples nitrided at 400 °C, which exhibit the best dry sliding wear resistance compared to the samples nitrided at other temperatures, have better

dry sliding wear resistance than that at 350 °C. Besides, it seems that the formation of S-phase in the samples can enhance their wear resistance. SEM observations of worn surfaces of the nitrided and unntrided samples dry-sliding against the GCr15 at room temperature are shown in Fig. 8 to help in understanding the mechanisms governing the wear of the studied samples. For the unnitrided samples, due to its low surface hardness, it can be observed that the worn surface is severely deformed and scored under the dry sliding wear condition, with the formation of microcracks (see Fig. 8(a)). This is due to the fact that during the sliding course, a plastic deformation mechanism which is typical of the adhesive wear is found to occur superficially resulting in superficial material detachment and debris creation [22]. The detached material and wear debris between slider and sample are then facilitated to adhesive and abrasive wear on the unntrided samples as sliding distance increasing. Therefore, wear rate of unnitrided samples is much higher than that of the nitrided samples. On the other hand, under such test condition as that the samples are performed on the ring slider under the applied normal load of 30 N in dry-sliding wear test, very high contact stress would occur [23], thus asperity fracture tend to take place at the beginning of the wear process. When the contact surfaces are smeared and kneaded under sliding motion, small amounts of wear particles produced via asperity fracture are finally created as wear debris, which is responsible for parallel fine and

G.-j. Li et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1964–1970

1969

Fig. 8. Typical appearance of worn surfaces of 17-4 PH: (a) unnitrided (b) 480 °C (c) 400 °C.

shallow scratches on the sliding surface (see Fig. 8(a)). This is probably another reason for the fact that wear rate of unnitrided samples is much higher than that of the nitrided samples. Additionally, evidence of oxide particles (white tiny particles in Fig. 8(a)) is observed from the SEM micrograph, indicating that the oxidation wear also takes place simultaneously during the sliding progress. Oxidative wear is due to the local surface material removal by the slider followed by the oxidation of the exposed metal surface. As sliding distance increasing, large amount of oxide debris in the untreated sample are produced. This indicated that the surface degradation of the sample can be also attributed to oxidative wear. SEM micrographs of the worn surface of the plasma nitrided samples are shown in Fig. 8(b), (c). A different mode of wear of the nitrided surfaces under the present testing conditions is observed, compared with the unnitrided surface described above. As has been shown in Fig. 8(b), microcracks and a certain amount of flake-like debris with slight plastic deformation are observed on the worn surface of the 480 °C/4 h treated sample, indicating that a mild adhesive wear and abrasive wear occurred. Therefore, the improved dry sliding wear resistance of the treated samples

is obtained. The 480 °C/4 h treated sample has a nitrided layer of thickness 10lm and roughness Ra = 0.25 lm; this reduces the contact area compared to the untreated sample (Ra = 0.05 lm), increasing the pressure on the asperities. Therefore greater local plastic deformation leads to flattening of the asperities [24]. Consequently the contact area increases reducing the contact pressure, leading to a mild adhesive wear and abrasive wear mechanism. However, despite of the high surface roughness compared to the untreated surface, its wear rate is lower than that of the unnitrided samples. This is due to the fact that a single CrN nitride layer which is hard and rigid is supposed to form on the 480 °C/4 h treated sample surface. Thereby, it is the surface microhardness that plays the lead role in the wear rate reduction but the surface roughness under this test circumstances. In addition, it is proposed that the brittle CrN nitrided layer can be pressed into the substrate in such test mentioned above, thus the micro-cracks formation is produced. What is more, the oxidative film serves as lubricant to reduce the friction between slider and sample surface since oxidation wear is considered to be the dominant mechanism in dry sliding wear tests, consequently, its dry sliding wear rate is remarkably reduced.

1970

G.-j. Li et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1964–1970

Fig. 8(c) shows the SEM micrograph of worn surface of the 400 °C/4 h treated samples, which is thought to be produced by the influence of asperity along the wear direction. According to the figure, smooth worn surface without micro-cracks is observed. Since different wear resistance of the nitrided samples can be attributed to the difference on the wear mechanism, the S-phase layer which is dominated by slight abrasion and frictional polishing [25] is supposed to exhibit good wear resistance in comparison to the unnitrided ones. For the 400 °C/4 h treated layer that is composed of S-phase and a0N , the formation of S-phase by plasma nitriding contributes to its wear resistance increasing due to the ductile behavior of S-phase in scratching and in abrasion [26]. So it can be concluded that the observation of the smooth worn surface without microcracks is responsible for the low wear rate. On the other hand, its surface has roughness of Ra = 0.11 lm, with high peak-to-valley height (1.32 lm) which increases the contact area of the GCr15 counterface even more in comparison to the 480 °C/4 h treated sample, reducing initial asperity deformation and polishing. Thereby, dry sliding wear resistance of the 400 °C/4 h treated sample is exhibited to be better than that of the 480 °C/4 h treated sample. In general, the wear resistance of a material is positively proportional to its hardness [27]. However, the 400 °C/ 4 h treated sample, with lower surface microhardness than that of the 480 °C/4 h treated sample, exhibits the lower wear resistance than that of the 480 °C/4 h treated sample. This is most probably because in such case of unlubricated sliding between steels, work hardening, frictional heating, phase transition and oxidization can take place at the same time in a complicated (hardly controllable) way [28], so that the wear resistance of the nitrided steels is not anymore simply related to surface hardness. 4. Conclusion Microstructure and dry sliding wear properties of the DCPN nitrided layers are investigated in this paper. It is certain that layers of different structures are produced. Nitriding process performed at temperature 6 400 °C take effect on creation of the layers composed of S-phase and a0N . Up to the temperature of 420 °C, the S-phase peaks disappear due to the transformation occurrence (S-phase ? a0N + CrN). For the samples nitrided at temperature P 450 °C, no evidence of a0N is found owing to a precipitation (a0N ! a þ CrN) taking place. The surface morphology of sample nitrided at 480 °C is much rougher than that nitrided at 350 °C, due to the increased arrival energy of the species independent of temperature increase at the cathode surface. For the 480 °C/4 h treated sample, it is the surface microhardness that plays the lead role in the wear rate reduction but the surface roughness; while the 400 °C/4 h treated sample, it is both of the surface roughness and the S-phase formation. Wear of the untreated 17-4PH is dominantly characterized by strong adhesion, abrasion and oxidation mecha-

nism. Samples nitrided at 400 °C which is dominated by slight abrasion and plastic deformation exhibit the best dry sliding wear resistance compared to the samples nitrided at other temperatures. Therefore, the dry sliding wear resistance of 17-4PH stainless steel is effectively improved by DC plasma nitriding. Acknowledgement The authors gratefully acknowledge support of this work by Nuclear Power Institute of China and Coating Technology development Department of Chengdu Tool Institute. References [1] Wen-Tung Chien, Chung-Shay Tsai, J. Mater. Proc. Technol. 140 (2003) 340. [2] Ping Li, Qizhou Cai, Bokang Wei, Xianzhong Zhang, J. Iron. St. Res. 13 (2006) 73. [3] P. Kochman´ski, J. Nowacki, Surf. Coat. Technol. 200 (2006) 6558. [4] Jun. Wang, Hong Zou, Cong Li, Yanhua Peng, Shaoyu Qiu, Baoluo Shen, Nucl. Eng. Design 236 (2006) 2531. [5] Jun Wang, Hong Zou, Cong Li, Shaoyu Qiu, Baoluo Shen, Mater. Charact. 57 (2006) 274. [6] F. Christien, R. Le Gall, G. Saindrenan, Scripta Mater. 48 (2003) 11. [7] X.M. Zhu, M.K. Lei, Surf. Coat. Technol. 131 (2000) 400. [8] Y. Sun, T. Bell, G. Wood, Wear 178 (1994) 131. [9] M. Belgrado, A. Delpuglia, B. Tesi, A. Fusi, Heat. Treat. XV (11) (1983) 30. [10] M. Belgrado, B. Tesi, E. Casci, G. Tonini, VI Congr It del Vuoto (1979) 119. [11] Y. Sun, T. Bell, Mater. Sci. Eng. A 140 (1991) 419. [12] C. Blawert, A. Weisheit, B.L. Mordike, R.M. Knoop, Surf. Coat. Technol. 85 (1996) 15. [13] B. Larisch, U. Brusky, H.J. Spies, Surf. Coat. Technol. 116–119 (1999) 205. [14] S. Parascandola, R. Gu¨nzel, R. Gro¨tzschel, E. Richter, W. Mo¨ller, Nucl. Instr. and Meth. B 136–138 (1998) 1281. [15] S.P. Hannula, P. Nenonen, J.P. Hirvonen, Thin Solid Films 181 (1989) 343. [16] T. Bacci, F. Borgioli, E. Galvanetto, G. Pradelli, Surf. Coat. Technol. 139 (2001) 251. [17] C.X. Li, T. Bell, Corros. Sci. 46 (2004) 1527. [18] F. Borgioli, A. Fossati, E. Galvanetto, T. Bacci, Surf. Coat. Technol. 200 (2005) 2474. [19] S.K. Kim, J.S. Yoo, J.M. Priest, M.P. Fewell, Surf. Coat. Technol. 163–164 (2003) 380. [20] C.E. Foerster, F.C. Serbena, S.L.R. da Silva, C.M. Lepienski, C.J.deM. Siqueira, M. Ueda, Nucl. Instr. and Meth. B 257 (2007) 732. [21] A.L. Yerokhin, A. Leyland, C. Tsotsos, A.D. Wilson, X. Nie, A. Matthews, Surf. Coat. Technol. 142–144 (2001) 1129. [22] M. Vardavoulias, M. Jeandin, F. Velasco, J.M. Torralba, Tribol. Int. 29 (1996) 499. [23] C.X. Li, T. Bell, Wear 256 (2004) 1144. [24] C. Tsotsos, A.L. Yerokhin, A.D. Wilson, A. Leyland, A. Matthews, Wear 253 (2002) 986. [25] Xiao-Qin Zhao, Hui-Di Zhou, Jian-Min Chen, Mater. Sci. Eng. A 431 (2006) 290. [26] M.K. Lei, X.M. Zhu, Surf. Coat. Technol. 193 (2005) 22. [27] Koji Kameo, Kazuaki Nishiyabu, Klaus Friedrich, Shigeo Tanaka, Toshio Tanimoto, Wear 260 (2006) 674. [28] B. Bhushan (Ed.), Modern Tribology Handbook Volume One Principles of Tribology, CRC Press, Boca Raton, 2001.