Dry sliding wear of Cu–Be alloys

Dry sliding wear of Cu–Be alloys

Wear 259 (2005) 506–511 Dry sliding wear of Cu–Be alloys G. Straffelini∗ , L. Maines, M. Pellizzari, P. Scardi Department of Materials Engineering an...

497KB Sizes 0 Downloads 61 Views

Wear 259 (2005) 506–511

Dry sliding wear of Cu–Be alloys G. Straffelini∗ , L. Maines, M. Pellizzari, P. Scardi Department of Materials Engineering and Industrial Technologies, University of Trento, via Mesiano, 77-38100 Trento, Italy Received 26 July 2004; received in revised form 28 October 2004; accepted 15 November 2004 Available online 7 January 2005

Abstract The tribological behaviour of two Cu–Be alloys was investigated under dry sliding conditions against an AISI M2 steel counterface. The first alloy, Cu–Be1, was characterized by a hardness of 390 HV and a thermal conductivity of 106 W/mK, whereas the second one, Cu–Be2, had a lower hardness, 270 HV, and a higher thermal conductivity, 208 W/mK. At low loads, wear was metallic in nature. The Cu–Be1 alloy displayed the lowest wear rate, as expected. As the applied load was increased, however, the wear rate of the Cu–Be1 alloy increased, whereas the wear rate of the Cu–Be2 alloy was found to decrease and, in particular, to become lower than that of the Cu–Be1 alloy. A transition in the wear mechanism from metallic wear to tribo-oxidative wear was observed as the applied load was increased. The results were explained in terms of the characteristics of the tribological layer, which forms during sliding on the surface of the Cu–Be specimens. © 2004 Elsevier B.V. All rights reserved. Keywords: Sliding wear; Cu–Be alloys; Wear transitions

1. Introduction Cu-alloys are widely used in dry sliding applications, in particular against a steel or cast iron counterface. Their good sliding wear resistance is reported to be mainly due to their low tribological compatibility against iron alloys, their sufficiently high hardness, if properly alloyed or heat-treated, and their relatively high thermal conductivity [1,2]. Their high thermal conductivity makes these alloys very attractive in applications, like forming tools, where process or friction heat has to be removed from the surface region. In this respect, a particular role is played by Cu–Be alloys that may attain a tensile strength up to 1200 MPa, very close to the tensile strength of heat-treated steels [2]. In general, however, as mechanical strength is increased, the thermal conductivity is decreased, and a compromise is needed when both properties play an important role. The dry sliding behaviour of Cu-alloys is very similar to that displayed by ductile metals [3], although some pe∗

Corresponding author. Tel.: +39 0461 882458; fax: +39 0461 881977. E-mail address: [email protected] (G. Straffelini).

0043-1648/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2004.11.013

culiarities have to be considered. At the beginning of sliding the surface plastic deformation at the contacting asperities is favoured by a shear instability effect [4–7]. This effect gives rise to the occurrence of intense transfer phenomena. These phenomena produce the formation of a mechanically mixed layer on the surface of the Cu-alloys and sliding wear is thus given by the damage of such a layer, whereas the contribution of delamination seems to be negligible [8]. If the interaction with the environment gives rise to a surface oxidation, a tribo-oxidative wear mechanism controls the wear process, and wear is mild [9]. In the case of most engineering alloys, like steels for example, a mild tribo-oxidative wear is attained if the substrate hardness is sufficiently high to avoid the easy removal of the mechanically mixed layer rich in oxides [10,11]. In the case of Cu and Cu-alloys, however, this topic has not been investigated in detail. Aim of the investigation was to give a contribution to the understanding of the dry sliding wear behaviour of Cu–Be alloys and, in particular, to highlight the influence of the applied load and the role of the material hardness.

G. Straffelini et al. / Wear 259 (2005) 506–511

507

Table 1 Nominal chemical composition, hardness, thermal conductivity and density of the materials under study Material

Nominal chemical composition (wt.%)

Hardness (HV30 )

Thermal conductivity (W/mK)

Density (g/cm3 )

Cu–Be1 Cu–Be2

Cu–2% Be–0.5% (Co + Ni) Cu–0.5% Be–2% (Co + Ni)

389 268

106 208

8.26 8.75

Fig. 1. Microstructures of the Cu–Be1 (a) and Cu–Be2 (b) alloys (etched with ASTM E407-93 N.41).

2. Experimental procedures The dry sliding tests were carried out in a disk on disk apparatus. The disks had the following dimensions: 40 mm in diameter and 10 mm in thickness. The Cu–Be disks were cut from commercial bars, which were received in the agehardened condition. The nominal chemical compositions and representative properties of the investigated materials are reported in Table 1. It can be noted that material Cu–Be1 is characterized by a quite high hardness and a low thermal conductivity. Material Cu–Be2 is characterized by lower hardness and higher thermal conductivity than material Cu–Be1. The hardness values were obtained using a Vickers indenter and a load of 30 kg. The counterface stationary disks were made of AISI M2 tool steel and heat-treated to obtain a hardness of 880 HV. In Fig. 1(a) and (b) the microstructure of the two Cu–Be alloys is shown. The microstructure of the Cu–Be1 alloy is constituted by ␣ grains and Co–Be–Ni precipitates, known as beryllides [15]. In addition, the grain boundaries are decorated with ␥-phase. The matrix hardness is produced by a fine precipitation of metastable Cu–Be precipitates. The Cu–Be2 alloy is characterized by a lower Be-content and thus by a lower matrix hardness. On the other hand, this alloy contains a greater amount of Cu + Ni and, therefore, and increased amount of beryllides. These large precipitates induce some material hardening, because they reduce the grain size and have a relatively high microhardness (456 HV). The dry sliding tests were carried out with a sliding speed of 0.94 m/s and an applied load ranging between 50 and 200 N. During the experiments, the friction coefficient and the contact temperature (measured using a thermocouple inserted in the AISI M2 disk at a distance of 2 mm from the surface) were continuously measured. The wear evolution was recorded by periodically weighing the specimens,

using a precision balance. The experimental weight losses were converted into volume losses using the density values included in Table 1. In order to get information on the wear mechanisms, the wear tracks and worn debris were observed using both an optical and a scanning electron microscope. The phase constitution of the debris was also determined by means of X-ray diffractometry using Cu K␣ radiation.

3. Results In Fig. 2 the evolution of the mass loss with the sliding distance is shown, in the case of the tests carried out at 50 N. It can be seen that the wear evolution is linear with sliding distance and the wear rate, W, can thus be described by the following well known relation: W = kA F

Fig. 2. Evolution of the mass loss with sliding distance for the two alloys under study, under a load of 50 N.

508

G. Straffelini et al. / Wear 259 (2005) 506–511

Fig. 3. Wear coefficients of the materials under study as a function of the applied load.

where F is the applied load and kA , usually in m2 /N, is the wear coefficient. The experimental wear coefficients for the two alloys under study and the loads of 50, 100, 150 and 200 N are shown in Fig. 3. It can be noted that in the case of the Cu–Be1 alloy, the wear coefficient increases with the applied load. The Cu–Be2 alloy shows an increase in the wear coefficient as the load is increased from 50 to 100 N, and in this load range its wear coefficient is greater than that of the Cu–Be1 alloy. This is in agreement with the fact that the Cu–Be2 alloy is characterized by a lower hardness than the Cu–Be1 alloy. However, when the load is increased to 150 and 200 N a decrease in the wear coefficient is observed and, in particular, the wear coefficient becomes lower than that of the Cu–Be1 alloy. The friction records at 50 N showed that for both alloys the friction coefficient was about 0.8. As the load was increased, however, different behaviour was observed. In Fig. 4 the evolution of friction coefficient with sliding distance in the case of the Cu–Be1 alloy under a load of 150 N is shown. The reported friction evolution is typical of all the experimental conditions here investigated and is characterized by the presence of a high initial value, about 0.75, and an early sliding-distance transition with the attainment of a lower fric-

Fig. 4. Evolution with time of the friction coefficient in the case of the Cu–Be1 alloy under a load of 150 N.

Fig. 5. Variation of average steady-state temperatures as a function of load for both alloys.

tion coefficient. The steady-state friction coefficients after the transition were, for both alloys, about 0.6 for the tests at 100 N and about 0.55 for the tests at 150 and 200 N. The records of the evolution of the contact temperature with sliding distance showed that a steady-state condition was reached after the first 10 min of sliding. In Fig. 5 the average steady-state contact temperatures, are shown. It can be noted that different values were obtained for the two alloys. In particular, at 50 N the Cu–Be1 alloy displayed a greater contact temperature during sliding than the Cu–Be2 alloy. At loads higher than 50 N, however, the situation changed and the Cu–Be2 alloy displayed the highest contact temperatures. In Fig. 6 the morphology of the wear debris, collected at the end of the tests at 50 N, is shown. It can be seen that the debris have a platelike shape and that the average dimension of the plates is higher in the case of the Cu–Be1 alloy. The analysis of the XRD spectra shown in Fig. 7 highlights that the wear debris are made by metallic Cu fragments, thus showing that wear at 50 N was metallic in nature for both alloys. The morphology of the wear debris collected at the end of the test at 100 N is different than those shown in Fig. 6. As shown in Fig. 8 for the Cu–Be1 alloy, in fact, near to large platelake fragments the presence of many small equiaxed fragments can be appreciated. The analysis of the XRDspectra of the wear fragments collected at the end of the tests at 100 N, shown in Fig. 9, shows that the debris comprise both metallic and oxide fragments. The oxide fragments, in particular, are mainly made by Cu2 O oxides. The experimental observations thus suggest that the platelike fragments are metallic in nature, whereas the small equiaxed particles are oxides. In Fig. 10 the morphology of the worn surface of the Cu–Be1 alloy tested at 100 N is shown. It can be noted that the wear surface is characterized by the presence of some small scales and many small particles that are prone to leave the tribological system. Considering the characteristics and morphology of the wear debris, it can be concluded that the surface layer shown in Fig. 10 is made by oxides. The same features were observed for the Cu–Be2 alloy.

G. Straffelini et al. / Wear 259 (2005) 506–511

509

Fig. 6. Morphology of the wear debris collected at the end of the tests at 50 N. (a) Cu–Be1 and (b) Cu–Be2.

Fig. 7. XRD-spectra of the wear debris collected at the end of the tests at 50 N.

Fig. 9. XRD-spectra of the wear debris collected at the end of the tests at 100 N.

The wear debris collected at the end of the tests at 150 and 200 N are mainly oxides, as revealed by the XRD analysis. In the case of the Cu–Be1 alloy, the oxide fragments are mainly in the form of small equiaxed particles, as shown in Fig. 11(a). On the other hand, in the case of the Cu–Be2 alloy the oxide fragments have a platelike shape, as shown in Fig. 11(b). This means that during sliding the oxides are compacted to form such plates, as actually shown in Fig. 12, which shows a

detail of the wear surface of the Cu–Be2 alloy after the test at 200 N. The optical micrographs shown in Fig. 13(a) and (b) further confirm that in the case of the Cu–Be2 alloy large and protective oxide scales are formed on the surface of the alloys during sliding, whereas in the case of the Cu–Be1 alloy only small and loose scales or particles are formed on the wear surface.

Fig. 8. Morphology of the wear debris collected at the end of the tests at 100 N in the case of the Cu–Be1 alloy.

Fig. 10. Morphology of the worn surface of the Cu–Be1 alloy tested at 100 N.

510

G. Straffelini et al. / Wear 259 (2005) 506–511

Fig. 11. Morphology of the wear debris collected at the end of the tests at 200 N. (a) Cu–Be1 and (b) Cu–Be2.

Fig. 12. Morphology of the worn surface of the Cu–Be2 alloy tested at 200 N.

4. Discussion In the case of the tests at 50 N, a completely metallic surface layer was formed during sliding. The wear rate of the Cu–Be1 alloy was lower than that of the Cu–Be2 alloy, because of its higher hardness (Table 1), which rendered more difficult the removal of the surface scales. The platelike debris of the Cu–Be1 alloy were, therefore, characterized by an higher average dimension, as shown in Fig. 6, because the

surface scales stayed in its place for a longer time before being removed. The higher friction coefficient recorded during the sliding tests for both alloys, further confirm that during sliding a metal-to-metal contact was present [12]. At 100 N, metal-to-metal contact occurred only at the beginning of the sliding tests. As shown by the friction records, in fact, after a short sliding distance a transition to a different wear mechanism, characterized by a lower friction coefficient, took place. The analysis of the wear debris and tracks clearly highlights that after the transition wear became by tribo-oxidation, and the reduction in friction coefficient was, therefore, due to the lubricating action exerted by the surface oxides. The attainment of a tribo-oxidative wear mechanism was due to the increased contact temperature, which was reached during sliding. As shown in Fig. 5, in fact, in passing from 50 to 100 N the contact temperature increased by about 70 ◦ C, and this clearly favoured the occurrence of a low-sliding speed oxidative wear [13,14]. The surface oxides, however, were not able to exert a pronounced protective action, as is typical in the case of mild wear, and an increase in the wear coefficient was thus observed for both alloys. As the applied load was increased, the wear coefficient of the Cu–Be1 alloy increased, whereas the Cu–Be2 alloy showed a transition in the wear resistance, in passing from 100 to 150 N, with a decrease in the wear coefficient. In both alloys and at every applied load wear was by tribo-oxidation,

Fig. 13. Top view of the worn surface of the Cu–Be1 (a) and Cu–Be2 (b) alloys tested at 200 N.

G. Straffelini et al. / Wear 259 (2005) 506–511

as evidenced by the experimental observations of the wear debris and tracks. In the case of the Cu–Be2 alloy, the surface oxides were able to form a compacted and protective layer, whereas in the case of the Cu–Be1 alloy such a layer was not formed during sliding. Figs. 12 and 13(b) clearly show that for the Cu–Be2 alloy a surface oxide layer made by the compaction and sintering of the oxide particles was formed [10,15]. The fact that the oxide layer in the Cu–Be2 alloys covered more uniformly the wear surface of the alloy and was, therefore, able to better protect it from the tribological loading was confirmed by the results of the surface temperature measurements. As shown in Fig. 5, in fact, the contact temperature at 50 N was higher for the Cu–Be1 alloy than for the Cu–Be2 alloy. This is in agreement with the fact that the Cu–Be1 alloy is characterized by a lower thermal conductivity, as shown in Table 1. At loads higher than 50 N, however, the situation changed and this change can be only attributed to the formation of a uniform and compact oxide layer on the surface of the Cu–Be2 disk that is able to act as a thermal barrier, thus reducing the heat flow inside the rotating disk [14,16]. The different tribo-oxidative behaviour of the two alloys at 150 and 200 N can be explained by making reference to their different ability to maintain in place the surface oxide layer during sliding. Only if the oxide layer is firmly attached to the sliding surface it can get compacted and thus protect the underlying base alloy. It is, therefore, believed that in the Cu–Be2 alloy, which is characterized by a lower hardness than the Cu–Be1 alloy, the surface oxide layer can be better ‘pressed’ into the base metal and firmly attached to it. A similar explanation was proposed by Don et al. [17] to explain the better sliding resistance of soft Cu–Cu2 O and Cu–Al2 O3 alloys in comparison to hard Cu–Be alloys and by Stott and Jordan [18] in an analysis of the effect of the hardness of different steels on the maintenance of protective layers during sliding at elevated temperature.

5. Conclusions At 50 N, the lowest load here investigated, wear was metallic and the Cu–Be1 alloy, with the highest hardness (389 HV), displayed the lowest wear coefficient. At the loads of 100, 150 and 200 N, wear was by tribooxidation for both alloys. At 100 N, the surface oxide layer was not able to protect the underlying material and thus the wear coefficient was shown to increase, in passing from 50

511

to 100 N, for both materials. For the Cu–Be1 alloy such an increase of the wear coefficient was observed even as the load was increased to 150 and 200 N. On the other hand, the Cu–Be2 alloy, with the lowest hardness (268 HV), showed a transition in the wear behaviour. In passing from 100 to 150 and 200 N, in fact, the wear coefficient was found to decrease and this was attributed the characteristics of the surface oxide layer, which was able to stay in place and protect the underlying material. References [1] W.A. Glaeser, Materials for Tribology, Tribology Series 20, Elsevier, 1992, pp. 56–59. [2] A.S.M., Properties and Selection: Non-Ferrous Alloys and Pure Metals, Metals Handbook, 10th ed., vol. 2. ASM, Material Park, Ohio, USA, 1992, pp. 403–423. [3] D.A. Rigney, Transfer, mixing and associated chemical and mechanical processes during the sliding of ductile materials, Wear 245 (2000) 1–9. [4] L.H. Chen, D.A. Rigney, Transfer during unlubricated sliding wear of selected metal systems, Wear 105 (1985) 47–71. [5] D.A. Rigney, L.H. Chen, M.G.S. Naylor, A.R. Rosenfield, Wear processes in sliding systems, Wear 100 (1984) 195–219. [6] J. Shell, P. Heilmann, D.A. Rigney, Friction and wear of Cu–Ni alloys, Wear 75 (1982) 205–220. [7] R.S. Montgomery, The sliding behaviors of copper alloys, Wear 87 (1983) 339–349. [8] P. Heilmann, J. Don, T.C. Sun, D.A. Rigney, W.A. Glaeser, Sliding wear and transfer, Wear 91 (1983) 171–190. [9] M.D. Sexton, A study of wear in Cu–Fe systems, Wear 94 (1984) 275–294. [10] F.H. Stott, The role of oxidation in the wear of alloys, Tribol. Int. 31 (1998) 61–71. [11] T.F.J. Quinn, Review of oxidational wear, Tribol. Int. 16 (1983) 257–271. [12] G. Straffelini, A simplified approach to the adhesive theory of friction, Wear 249 (2001) 79–85. [13] J.L. Sullivan, S.G. Hodgson, A study of mild oxidational wear for conditions of low load and speed, Wear 121 (1987) 95–106. [14] G. Straffelini, D. Trabucco, A. Molinari, Sliding wear of austenitic and austenitic–ferritic stainless steels, Metall. Mater. Trans. 33A (2002) 613–624. [15] G. Straffelini, D. Trabucco, A. Molinari, Oxidative wear of heattreated steels, Wear 250 (2001) 485–491. [16] S. Wilson, A.T. Alpas, Thermal effects on mild wear transitions in dry sliding of an aluminium alloy, Wear 225–229 (1999) 440– 449. [17] J. Don, T.C. Sun, D.A. Rigney, Friction and wear of Cu–Be and dispersion-hardened copper systems, Wear 91 (1983) 191–199. [18] F.H. Stott, M.P. Jordan, The effects of load and substrate hardness on the development and maintenance of wear-protective layers during sliding at elevated temperatures, Wear 250 (2001) 391–400.