Wear, 149 (1991) 353-374
Sliding friction and wear of metals in vacuum* T. Akagaki+ and Materials
D. A. Rigney and Engineering,
April 30, 1991)
Abstract Sliding friction and wear in vacuum were studied for 16 combinations of four pure metals (copper, nickel, iron and molybdenum) from the points of view of adhesion energy and hardness ratio H,/H, (ratio of hardness of disk to hardness of pin) before and after the test. It was found that severe wear usually occurred when the range of hardness ratio Hd/ H, after the test included values below about 1.0. Mild wear typically occured when this ratio included only values above about 1.0. It was also found that friction and wear behaviour were dependent on the adhesion energy and the crystal structure. Mutual transfer was observed in both wear modes for every combination of different metals. Transfer ranged from widely scattered local transfer to relatively uniform transfer. Three types of wear debris were observed in severe wear: plate-like, wedge-like and cylindrical-spherical debris. Although wear debris consisted mainly of disk material, it also commonly contained pin material existing as small scattered pieces. The size of wear debris was strongly dependent on the hardness of the disk material and the adhesion energy. The data from self-mated tests were consistent with those using two different metals.
It is well known that transfer of material between contacting surfaces can occur during sliding [l-6]. Transfer strongly affects both friction and wear behaviour. There is clear evidence that wear debris is commonly generated from surface material which contains components from both mating materials [6-81. In some cases contact is made through transfer fragments during sliding . On the basis of adhesion theory, Chen and Rigney have shown that the preferred direction of transfer can be predicted for a given material combination [lo]. Cocks [ll] and Antler [12, 131 have suggested that the dominant direction of transfer may depend not only on the material combination but also on the geometry of the sliding system. Kato and Hokkirigawa have shown that the wear mode can be predicted according to the attack angle, which can be related to the degree of penetration and the shear strength at the interface in abrasive wear [14, 151 and the attack angle and the hardness ratio in sliding wear . These results suggest that material combination (adhesion energy and hardness ratio of mating surface) and contact geometry (attack angles of harder asperities) are *Paper presented at the International Conference on Wear of Materials, Orlando, FL, U.S.A., ApriI 7-11, 1991. ‘Present address: Department of Mechanical Engineering, Hachinohe National College of Technology, Hachinohe, Aomori 039-11, Japan.
0 1991 -
important for understanding in a sliding system become
wear and transfer
Both wear surfaces
harder than the original surface by work hardening in the Wear debris may also become harder when it is pressed. A simple
way to correlate sliding behaviour with these effects is to use the ratio of hardness of disk to hardness of pin before and after sliding. Sliding friction, wear and transfer were studied for 16 combinations of four pure metals. The effects of metal combination and hardness ratio on the friction, wear and transfer have been studied with the aid of scanning electron microscope (SEM)
observations and wavelength-dispersive and wear debris. The results indicate mild or severe.
X-ray (WDX) spectroscopy of mating surfaces how these factors determine whether wear is
The experiment was conducted with a pin-on-disk test machine similar to that used by Chen and Rigney . The system was enclosed in a vacuum chamber to provide a reproducible environment with reduced oxidation and contamination. A schematic diagram is shown in Fig. 1. The disk was fixed on the rotating shaft and the pin was fixed on a lever which was used to apply a normal load and measure frictional force. The pin was a cylinder 3.2 mm in diameter and 2 mm long. Its curved surface contacted the flat surface of the disk and it was aligned with its axis parallel to a disk radius. The center of the pin contact was at a disk radius of 6.5 mm. The load was applied to the vertical flat surface of the disk by using a lever loading system. The materials used in this study were copper (purity 99.9%), nickel (99.7%), iron (99.5%) and molybdenum (99.95%). The copper, nickel and iron were prepared by annealing as described in Table 1. Vickers hardness values are also given in Table 1. Both the pin and disk specimens were prepared by mechanical polishing with 0.05 pm diamond compound for the final stage and their surface roughness R,, was less than 0.1 pm. Material combinations (A-B, pin-disk) and hardness ratios (Z&/H,) are summarized in Table 2. To provide information on geometrical effects, the experiments were conducted with both metal A sliding on metal B and metal B sliding on metal A for each combination. The hardness ratio H,/H, before testing was between 0.20 and 4.13. Sliding conditions are summarized in Table 3. The sliding velocity and applied load were 13.8 mm s-l and 4.6 N respectively. The vacuum condition was Vacuum
Load _~_ 3 Disk Holder
Fig. 1. Schematic diagram of experimental apparatus.
Disk Hardness (HV)
400 “C, 1 h 600 “C, 1 h 600 “C, 1 h As received
56.3(56 f 2) 103.9(104*3) 133.1(133*3) 242.3(242 f 2)
400 “C, 1 h 700 “C, 1 h 600 “C, 1 h As received
48.3(48 It 1) 95.0(95 f 4) 130.8(130*5) 232.4(232 f 4)
2 and initial hardness
(Hd Vickers hardness
of disk, HP, Vickers
0.86 1.69 2.32 4.13
0.47 0.91 1.26 2.24
0.36 0.71 0.98 1.75
0.20 0.39 0.54 0.96
Sliding velocity (mm s-‘) load (N)
Sliding distance (special case) Vacuum
Ni Fe MO
(load 4.90 N)
Material combinations hardness of pin)
values of specimens
Ni Fe MO
and Vickers hardness
13.8 4.6 2.7, 100 (7.5) (4-10) x 1o-3
(4-10)x 10e3 Pa. Two steps of experiments were carried out for each material combination. Sliding distances were 2.7 m (pre-steady state) and 100 m (steady state). In some cases a sliding distance of 7.5 m was also used. All specimens were cleaned using acetone in an ultrasonic cleaner. Frictional force was measured with a strain gauge bonded to the lever; it was recorded continuously with a chart recorder. Wear losses of pin and disk were measured by weight changes to an accuracy of 0.01 mg. After each test the hardness of the wear surface was measured with a Vickers microhardness tester. Surface profiles of the wear scars of pins and disks were recorded with a profilometer, both perpendicular and parallel to the sliding direction. Wear scars of pin and disk and wear debris were observed with an SEM and analysed with a WDX analysis facility.
3.1. Friction and wear characteristics 3. I.1. Hardness ratio Three types of friction curve observed in this study are shown in Fig. 2. When the initial hardness ratio (HJH,,) < 1.0 (Cu-Cu, Ni-Cu, Ni-Ni, Fe-Cu, Fe-N;, Fe-Fe, Mo-Ni, MoFe and Mo-MO), severe wear occurs soon after the start of testing (Fig. 2(a)). The sliding distance for transition from mild to severe wear is extremely short, less than a few centimeters. The only exception was the Mo-Cu combination (H,/H,=O.20). In this case the transition did not occur and mild wear was maintained. During severe wear the coefficient of friction is large and erratic (maximum range 0.1-2.4). When the initial hardness ratio H,/H,> 1.0, two types of friction curve are observed. For Cu-Ni, Ni-Fe and Ni-Mo, the transition occurs after a relatively long sliding distance (Fig. 2(b)). This distance becomes longer when H,H, increases. These effects are summarized in Fig. 3. For Cu-Fe, Cu-MO and Fe-MO the transition does not occur and mild wear is maintained as shown in Fig. 2(c). In mild wear, the coefficient of friction is small and almost constant (0.24.4). Thus the friction mode and wear mode depend on the hardness ratio HJH,. Figure 4 shows the relationship between the specific wear rate of the disk at steady state, W,, and the hardness ratio after the test (HJH, after 100 m of sliding). The specific wear rate is in the range 10-3-10-2 mm3 N-’ m-r in the severe wear mode and 10-5-10-4 mm3 N-r m-r in the mild wear mode. The specific wear rate of the pin was always smaller than that of the disk. It was of the order of 10-5-10-4 mm3 N-’ m- ‘for severe wear and (l-4) x lo-’ mm3 N-r m-’ for mild wear. Figure 4 shows that severe wear occurs when the hardness ratio after the test includes values below 1.0 and mild wear occurs when all values are above 1.0. The only exception is the MoCu combination mentioned earlier. Thus the hardness ratio after the test can be used to predict the predominant wear mode. This suggests that processes such as work hardening and mechanical mixing of materials in the interface region play an important part in friction and wear behaviour.
E I EiI
Ni Fe MO
83 13 01
Hd/Hp Fig. 2. Three types of friction curve. Fig. 3. Relationship between sliding distance for transition from mild to severe wear and initial hardness ratio HJH,. To aid comparison, data points for self-mated pairs are shaded. Metal elements denote pm material.
Cu Disk Ni 31
10-S - 10-4 25
Fig. 4. Relationship between specific wear rate of disk at steady state, W,, and hardness ratio after test, Hd/Hp at 100 m sliding. To aid comparison, data points for self-mated pairs are shaded. Metal elements denotes pin material. Except for Mdu, mild wear occurs when all values of H,/H, are greater than about 1.0 and severe wear occurs when H,,IH, includes values less than about 1.0.
between Vickers hardness of wear scars and sliding distance for Cu pin-Ni disk combination: (a) pin; (b) disk.
Fig. 5. Relationship
3.1.2. Hardening of mating s&aces
Figure 5 shows the relationship between the sliding distance and the hardness of the wear scar for a copper pin and nickel disk. The process corresponds to the friction curve shown in Fig. Z(b). Figure 6 shows the hardness ratio at each stage of the test. In the mild wear regime (at 2.7 m sliding) the hardnesses of both surfaces are almost the same as the initial hardness. Just after the transition to severe wear (at 7.5 m sliding) both hardnesses increase suddenly and then more gradually. Thus the hardness ratio HJH, has a wide range including values below 1.0. After sliding for 100 m, each combination for which severe wear predominates has a wide range of hardness ratio including values below 1.0 (0.2-3.0). In contrast, the surface hardening is much less in mild wear and the hardness ratio does not
358 3 T
Fig. 6. Hardness ratio H,,/H, at each stage of test for Cu pin-Ni disk combination.
(Hd / Hp) before Fig. 7. Relationship between hardness ratio HJH, before and after test. Metal elements denote pin material. change appreciably. Therefore the hardness ratio after the test does not include values below 1.0. These results are summarized in Fig. 7.
3.1.3. Adhesion energy Figure 8 shows the relationship between the adhesion energy and the specific wear rate of the disk. The adhesion energy was calculated from the formula used by Chen and Rigney [lo]. The specific wear rate in f.c.c. metals (copper and nickel) decreases suddenly as the adhesion energy increases. In contrast, the specific wear rate in b.c.c. metals (iron and molybdenum) increases suddenly as the adhesion energy increases. Thus wear behaviour also depends on the crystal structure.
-o- vs. Cu Disk
Fig. 8. Relationship To aid comparison, material.
between specific wear rate at steady state, W,, and adhesion energy, AEAB_ data points for self-mated pairs are shaded. Metal elements denote pin
3.2. Wear mode and transfer mode Figures 9-11 show the results of SEM observations and X-ray analyses of copper pins and nickel disks at each stage of sliding. These results correspond to Fig. 2(c) and Figs. 5 and 6. The initial hardness ratio is 1.69. At the early stage of friction (2.7 m sliding, mild wear regime) the wear scars of the pin and disk are smooth and small as shown in Fig. 9(a) and 9(e). Wear loss is extremely small and undetectable. The hardness ratio (H&Z,,) decreases a little and is in the range 1.0-2.1 as shown in Fig. 6. Many mild plowing grooves are observed on the wear scars of the pin and disk as shown in Fig. 9(b) and 9(f). At the exit side of the pin wear scar, plate-like layers are extruded by plastic flow of pin material. Transfer fragments (less than about 15-20 pm) from the disk (nickel) are widely scattered on the pin surface, especially along grooves formed by plastic deformation, as shown in Fig. 9(b) and 9(c). They are embedded in the softer pin surface as shown in Fig. 9(d). Transfer fragments (less than about 15 pm) from the pin (copper) are also observed on the disk surface. Again, the transfer material tends to lie along the grooves as shown in Fig. 9(f) and 9(g), but in this case the transfer material remains on the harder disk surface as shown in Fig. 9(h). Thus mutual transfer occurs even when the coefficient of friction is small and mild wear persists. Mutual transfer was also observed with other combinations (Cu-Fe, Cu-MO, Fe-M0 and Mo-Cu) in which mild wear predominated. Figure 10 shows the results just after the transition from mild to severe wear (7.5 m sliding, severe wear regime). At this stage the coefficient of friction was large and fluctuated over a large range (0.3-2.0) as shown in Fig. 2(b). The specific wear
Fig. 9. SEM images and results of WDX analysis of wear scars of pin and disk at 2.7 m sliding (mild wear regime) for Cu pin-Ni disk combination: (a) overview of wear scar of pin; (b) enlargement of region indicated by arrow tip in (a); (c} Ni Ka mapping of (b); (d) cnlargcment of region indicated by arrow tip in (b); (e) overview of wear scar of disk; (f) enlargement of region indicated by arrow tip in (e); (g) Cu Ka mapping of (f); (h) enlargement of region indicated by arrow tip in (f). The arrow indicates the relative direction of motion of the countcrfacc.
Fig. 10. SEM images and results of WDX analysis of wear scars of pin and disk at 7.5 m sliding (severe wear regime) for Cu pin-Ni disk combination: (a) wear scar of pin; (b) Ni Ka mapping of (a); (c) wear scar of disk; (d) enlargement of region indicated by arrow tip in (c); (e) Cu Ka mapping of (d). The arrow indicates the relative direction of motion of the counterface.
rates also increased suddenly to 5.8X lo-’ and 1.1 x lo-’ mm3 N- ’ m-l for the pin and disk respectively. The wear scars of the pin and disk are large and rough as shown in Fig. 10(a) and 10(c). A large amount of nickel from the disk has transferred to the pin as shown in Fig. 10(a) and 10(b). It is in the form of large and plate-like deposits. Also, a large amount of copper has transferred to the disk as shown in Fig. 10(d) and 10(e). Thus mutual transfer is also observed in the severe wear regime and the amount is much greater than for the case of mild wear. A similar correlation with the amount of transfer material has been noted previously by Rigney et al. (171. Figure 11 shows the results at steady state (100 m sliding, severe wear regime). The specific wear rates are 1.8~ 10m5 and 6~ low3 mm3 N-’ m-l for the pin and disk respectively. The coefficient of friction p is high and shows large fluctuations (~=0.4--2.1). The hardness ratio is in the range OS-l.4 as shown in Fig. 6. The wear scars of the pin and disk are very large as shown in Fig. 11(a) and 11(e) and the pin surface is almost uniformly covered with transfer metal from the disk (nickel) as shown in Figs. 11(b) and 11(d). The transfer material consists of relatively smooth plate-like layers as shown in Fig. 11(a) and 11(c). Many large fragments consisting of plate-like layers are observed on the disk surface as shown in Fig. 11(e) and 11(f). They are rich in disk material nickel as shown in Fig. 11(h), but they also contain the pin material (copper) as shown in Fig. 11(g). Therefore mechanical mixing has occurred during sliding. Similar mutual transfer and mechanical mixing were observed with other combinations for which severe wear predominated (Cu-Cu, Ni-Cu, Ni-Ni, Ni-Fe, Ni-Mo, Fe-Cu, Fe-Ni, Fe-Fe, Mo-Ni, Mo-Fe and Mo-MO). Thus after 100 m of sliding in the severe mode the pin surface is largely covered by a layer of mixed material whereas the disk surface is less completely covered. In mild wear neither surface is covered by much of the mixed material. 3.3. SEM observation of wear debris Figures 12-14 show the SEM images and X-ray analysis results of wear debris generated in severe wear. Wear debris was not collectable in mild wear. Three types of wear debris were observed in this study, i.e. irregular plate-like debris with lamellar structure (Fig. 12), wedge-like debris with irregular shape (Fig. 13) and cylindrical and spherical debris containing clear evidence of formation by rolling (Fig. 14). Although these debris particles consist mainly of disk material, they also contain pin material, some of it partially mixed and some in the form of scattered small pieces, as shown in Figs. 12-14. Thus the debris and surface layer have similar mixed composition. Figure 15 shows the relationship between the maximum particle size and the Vickers hardness of the disk. Except for the mild wear combinations such as Mo-Cu, Cu-MO, Cu-Fe and Fe-MO, there is a good correlation. The particle size was measured from the SEM images and the number measured was 30-40 for each material combination. The maximum particle size was the average value of five wear debris particles measured in order of size. With increasing hardness of the disk the maximum size of wear debris decreases almost linearly. The order of maximum particle size is as follows: Cu > Ni > Fe > MO. This is the same order as the ductility. Figure 16 shows the relationship between the maximum particle size and the adhesion energy. The trend is almost the same as in Fig. 15. 3.4. SEM observation
of wear surfaces
3.4.1. Mild wear
As shown in Fig. 9, the wear scars of the pin and disk in mild wear are smooth and small. Wear debris was not collectable. Many small grooves and scattered pieces
Fig. 11. SEM images and combination: (a) wear scar (c); (e) wear scar of disk; direction of motion of the
results of WDX analysis of wear scars of pin and disk at 1130 m sliding (severe wear regime) for Cu pin-Ni disk of pin; (b) Ni Kcx mapping of (a); (c) enlargement of region indicated by arrow tip in (a); (d) Ni Ka mapping of (f) large transfer fragments; (g) Cu Kcr mapping of (f); (h) Ni Ka mapping of (f). The arrow indicates the relative counterface.
Fig. 12. SEM images of plate-like wear debris and results of WDX analysis: (a) plate-like wear debris with lamellar structure generated in Ni-Mo combination; (b) Ni Ka mapping of (a); (c) enlargement of region indicated by arrow tip in (a).
of transfer material are evident on both surfaces. There are several possible explanations for the deformation of the harder material. One is related to variations in surface topography. Kayaba et al. showed that a harder asperity contacting a softer material can be deformed plastically and fractured below a certain asperity angle 118, 191. Another explanation is based on the local nature of deformation and work hardening, so that a local region of the “soft” material can be temporarily harder than the “hard” material, especially during the early stages of sliding. Yet another depends on transfer and mechanical mixing, which creates hard regions on the softer material. 3.4.2. Severe wear Figure 17 shows SEM images of the wear scar of a molybdenum pin after sliding on an iron disk. In the severe wear regime the pin surface is nearly completely covered with material rich in iron as shown in Fig. 17(a) and 17(b). Transfer material, which consists of plate-like layers and has some cracks, is observed at the inlet position of the scar as shown in Fig. 17(c). Near the center of the wear scar, separation of transfer material is observed as shown in Fig. 17(d). Figure 18 shows SEM images of the wear scars of nickel and copper disks. Many large fragments consisting of plate-like layers (lamellar structure) are observed as
-o- vs. Cu Disk -ANi 11 -aFe 11 -vMO ‘1
0 1.0 -\
2 E OS-
u 8 200
Fig. 15. Relationship between maximum particle size D,,, and Vickers hardness of disk, Hd. To aid comparison, data points for self-mated pairs are shaded. Metal elements denote pin material.
AEAB (J/m21 Fig. 16. Relationship between maximum particle size D, comparison,
data points for self-mated
pairs are shaded.
and adhesion energy AE,,. To aid Metal elements denote pin material.
shown in Fig. 18(a), 18(b) and 18(d). They probably result from back transfer and are common features on the wear scars of disks. The amount of pin material is relatively large in these fragments as shown in Figs. 10, 11‘ and 18(c). Shear dimples showing ductile fracture are also visible as shown in Fig. 18(e) and 18(f). Given these surface features, it seems likely that wear debris is generated mainly from these back-transfer fragments on the disk and from the material with lamellar structure on the pin. Figure 19 shows SEM images of the wear scar of a molybdenum disk after sliding against a nickel pin. Transfer and mixing have occurred, but the transfer fragments
Fe Ka d
Fig. 17. SEM images of wear scar of pin (molybdenum) generated in Mo-Fe combination in severe wear regime: (a) overview; (b) Fe Ka mapping of (a); (c) enlargement of region indicated by lower arrow tip in (a); (d) enlargement of region indicated by upper arrow tip in (a), The large arrow indicates the relative direction of motion of the counterface.
are not large. The scar is uniformly covered with irregular lameliar structures which have a scale-like appearance (Fig. 19(a) and 19(b)). At higher ma~ifi~t~on, finer scale-like features are visible (Fig. 19(c)). Debris particles were smaller than for cases of severe wear having large amounts of transfer.
4. Discussion As shown in Fig. 4, severe wear occurred when the hardness ratio after the test included values below 1.0, even if the initial hardness ratio was above 1.0. The only exception was the Mo-Cu wmbination. Mild wear occurred when the hardness ratio after the test was con$istently above 1.0. Therefore it is possible to predict roughly the wear mode by considering the hardness after the test. This result suggests that severe wear occurs when hard asperities and transferred fragments on the pin can penetrate into the disk surface. In severe wear, attack angles of transferred fragments on the pin and of the pin itself were less than about 30-W. These values correspond to the regions of plowing and wedge formation in an abrasive wear mode diagram
Fig. 19. SEM images of wear Scar of disk (molybdenum) generated in Ni-Mo combination in severe wear regime: (a) overview; (b) characteristic scale-like wear aear; (c) enlargement of region indicated by arrow tip in (b); (d) enfargement of region indicated by arrow tip in (a); (c) Ni Ku mapping of (d). The arrow indicates the relative direction of motion of the counterface.
371 [14, 151. In mild wear the attack angle of the pin itself was less than 6” and corresponded to the region of plowing in.such a diagram. In severe wear, wedge-like wear debris was commonly observed in some systems as shown in Fig. 13. As shown in Fig. 3, severe wear occurred very soon after the start of the test when the initial hardness ratio was below 1.0. On the other hand, when the hardness ratio was above 1.0, any transition from mild wear to severe wear occurred only after sliding a certain distance. The sliding time before the transition can be considered as the preparation period for work hardening of the pin surface and for producing a critical amount of transfer material. These results also provide some insight into the sliding behaviour with self-mated tests. Figures 3, 4, 7, 8, 15 and 16 indicate that self-mated tests are not really different from other cases from a mechanical point of view, i.e. they follow the same trends. This is not surprising. If Z&=HP at the start of a test, HJH, will become less than 1.0 because of the duty cycle (the pin hardens faster because of more contact and more deformations. This should mean that different duty cycles, easily obtained by changing the test geometry or load, will give different sliding distances for transitions. Also, if Hd is made high at the beginning by suitable specimen preparation, the ratio H,/H, will approach 1.0 but not fall below it (in vacuum). At long times, even if Hd/ HP falls below 1.0, it should reverse direction eventually and approach 1.0. However, it can never get to the “safe” level above 1.0 at long times. Thus, for the self-mated sliding of simple metals in vacuum, severe wear is expected from application of a simple hardness ratio rule. Metallurgical techniques for modifying the pin hardness wouid generally make things worse; however, various techniques for modi~ng the disk should postpone or inhibit severe wear. In this discussion the critical value of HJH, has been taken as 1.0. However, Bowden and Tabor have reported some interesting results which indicate that a critical value H,,/H,=1.2 might be preferred . A piece of steel was heat treated so as to obtain a hardness gradient and a slider of intermediate hardness was moved along the steel flat. The result was a kind of scratch hardness test. Friction was high on the soft part but there was a sharp transition to low friction when the slider was 1.2 times the hardness of the steel sample (HP= l.Wd). Bowden and Tabor then suggested a “natural” hardness scafe based on the factor of 1.2 instead of the closely related Mohs scratch hardness scale which used a value of about 1.6. Thus there is some basis for expecting a critical value of Hd/Hp near 0.8 rather than 1.0. Our data are not sufficient to distinguish between these two criteria. Rice and Wayne [2I] have reported that the friction and wear behaviour change when the pin and disk materials are interchanged. Chen and Rigney  reported similar behaviour for G-Fe and Fe-MO systems. These results can be understood in terms of changes in f&/H,, when the materials are interchanged. In the current study the wear mode is changed by interchanging materials for Fe-MO and Cu-Fe. In other combinations such as Cu-Ni, Ni-Fe and Ni-Mo only severe wear occurred. For Cu-MO only mild wear occurred. The original work of Chen and Rigney  suggested a good correlation of sliding behaviour with adhesion energy for metals in vacuum. The current results place more emphasis on the local hardness ratio, yet Figs. 8 and 16 show interesting trends with adhesion energy. One of the problems here is that adhesion and deformation effects are intimately related when materials contact each other . It is difficult to have one without the other. For example, it has been shown that unlubricated sliding wear of metals commonly involves a complex sequence of events which may be outlined as follows. Local contacts cause large plastic strains, followed by transfer, mechanical
mixing and generation of debris from the mixed material 1231. Adhesion influences the initial transfer events, the integrity of the mechanically mixed material and the chemical composition and volume fractions of phases in the mechanically mixed material and in the debris. These in turn affect the hardness, yield strength, ductility and fracture characteristics of material in the interface region. Therefore the hardness ratio can be affected by adhesion energies. The case of Mo-Cu remains anomalous. Hiratsuka and Sugahara have reported similar results for MoCu in vacuum . They suggest that the low mutual solubility in this system is responsible for the sliding behaviour. The importance of mutual solubility was also emphasized in the work of Goodzeit and coworkers [25, 261. This idea is closely related to the “compatibility” suggestion of Rabinowicz ; however, there are many exceptions to the simple compatibility rule as originally proposed. Also, as pointed out by Miedema and den Broeder , adhesion energy is dominated by the surface energy contributions, even for systems such as Mo-Cu in which the chemical interaction term is larger than usual. This means that “any two metals can be bonded strongly, provided that the initial surfaces are clean” . Miedema and den Broeder also supply references for experimental results which support this claim. The low solubilities of molybdenum and copper may affect sliding behaviour at high sliding speeds or elevated temperatures where diffusion is appreciable, but it is not clear how this factor could influence sliding behaviour near room temperature. Perhaps the limited ductility of molybdenum has an important influence on sliding with various counterface materials. It is clear that sliding systems with molybdenum, and Mo-Cu in particular, need further work. 5. Summary
(1) Both mild wear and severe wear were observed. When the range of hardness ratio HJH, after the test included values below about 1.0, severe wear occured. When it included only values above about 1.0, mild wear occurred. The Mo-Cu combination was an exception: mild wear occurred even though the hardness ratio was below 1.0. (2) A transition to severe wear can occur after a certain sliding distance in the mild wear mode. If the initial hardness ratio is less than about 1.0, the critical sliding distance is a few centimeters (l-2 rev). For the initial hardness ratio I;I,/H,>_l.O the critical distance was longer for larger initial hardness ratio. (3) The H,/H, ratio is also useful for understanding the sliding behaviour of selfmated systems. (4) Mutual transfer was observed in both wear modes for every combination. (5) With increasing adhesion energy the specific wear rate of the disk decreased for f.c.c. metals (copper and nickel) and increased for b.c.c. metals (iron and molybdenum). (6) Uniform transfer was observed on the pin surface and scattered transfer on the disk surface in the severe wear mode. Scattered transfer was observed on both surfaces in mild wear. (7) Three types of wear debris particles were observed in severe wear: irregular plates with lamellae, irregular wedge-like debris and debris produced partly by rolling (cylinders, spheres). Although the debris consisted mainly of disk material, it often contained the pin material in the form of small pieces. (8) The maximum size of wear debris was strongly dependent on the disk material (hardness) as well as the adhesion energy. The order for the disk material was as follows: Cu > Ni > Fe > MO.
We are pleased to acknowledge helpful discussions with Professor Koji Kato, Tohoku University, Sendai, Japan. Also, the support of the National Science Foundation (DMR-8718834), the support of the Japanese Government and a grant from TRIBOTEX Company, Nagoya, Japan are appreciated.
References 1 F. P. Bowden, A. J. W. Moore and D. Tabor, The ploughing and adhesion of sliding metals, J. Appl. Phys., 14 (1943) 80-91. 2 E. Rabinowicz and D. Tabor, Metallic transfer between sliding metals: an autoradiographic study, Froc. R. Sot. Lond. Ser. A, 208 (1951) 455-475. 3 M. Kerridge and J. K. Lancaster, The stages in a process of severe metallic wear, Proc. R. Sot. Lond. Ser. A, 236 (1956) 250-264. 4 D. H. Buckley, Friction, wear and lubrication in vacuum, NASA Spec. Pz& 277, 1971. 5 T. Sasada and S. Norose, The formation and growth of wear particles through mutual transfer, in T. Sakurai (ed.), Proc. JSLE-ASLE Int. Lubrication Conf, Tokyo, 197.5, Elsevier, Amsterdam, 1976, pp. 82-91. 6 L. H. Chen and D. A. Rigney, Transfer during unlubricated sliding wear of selected metal systems, Wear, 105 (1985) 47-61. 7 T. Sasada, S. Norose and Y. Shimura, Composition of wear particles produced under sliding friction of different metal combinations, Proc. 18th Jupan Congr. on Muferiul Research, Society of Materials Science, Kyoto, 1975, pp. 77-81. 8 P. Heilmann, J. Don, T. C. Sun, D. A. Rigney and W. A. Glaeser, Sliding wear and transfer, in K. C. Ludema (ed.), Proc. Inr. Conf: on Wear of Materials, Resron, VA, I983 American Society of Mechanical Engineers, New York, 1983, pp. 415-425; Wear, 91 (1983) 171-190. 9 T. Sasada, S. Norose and H. Mishina, The behaviour of adhered fragments interposed between sliding surfaces and the formation process of wear particles, J. Lubr. Technol., IO3 (1981) 195-202. 10 L. H. Chen and D. A. Rigney, Adhesion theories of transfer and wear during sliding of metals, Wear, 136 (1990) 223-235. 11 M. Cocks, Interaction of sliding metal surfaces, /. Appl. Phys., 33 (1962) 2152-2161. 12 M. Antler, Processes of metal transfer and wear, Wear, 7 (1964) 181-203. 13 M. Antler, Tribological properties of gold for electric contacts, IEEE Trans. Parts, Hybrids Puckug., 9 (1) (1973) 4-14. 14 K. Kato and K. Hokkirigawa, Abrasive wear diagram, Proc. 4th Znt. Congr. on Tribologv ([email protected]
& ‘SS), Ecu&, 1985, Elsevier, Amsterdam, 1985, paper 3. 15 K. Hokkirigawa and K. Kato, An experimental and theoretical investigation of ploughing, cutting and wedge formation during abrasive wear, Tribal. Int., 21 (1988) 51. 16 Y. C. Chiou and K. Kato, Wear mode of microcutting in dry sliding friction between steel pairs (part. 1); effect of attack angle of specimen, JSLE Int. Edn., (9) (1988) 11-16. 17 D. A. Rigney, L. H. Chen and M. Sawa, Transfer and its effects during unlubricated sliding, in H. D. Merchant and K. J. Bhansali (eds.), Metal Transfer and Culling in Metallic System, The Metallurgical Society, Warrendale, PA, 1986, pp. 87-102. 18 T. Kayaba, K. Kato and K. Hokkirigawa, Theoretical analysis of the plastic yielding of a hard asperity sliding on a soft flat surface, Wear, 87 (1983) 151. 19 T. Kayaba, K. Hokkirigawa and K. Kato, Experimental analysis of the yield criterion for a hard asperity sliding on a soft flat surface, Wear, 96 (1984) 255. 20 F. P. Bowden and D. Tabor, The Friction and Lubrication of Solids, Vol. 2, Oxford University Press, Oxford, 1964, pp. 346-348. 21 S. L. Rice and S. F. Wayne, Specimen material reversal in pin-on-disk tribotesting, Wear, 88 (1983) 85-92.
374 22 D. Landbeer, A. J. G. Dackus and J. A. Klostermann, Fundamental aspects and technological implications of the solubility concept for the prediction of running properties, Wear, 62 (1980) 255-286. 23 D. A. Rigney, L. H. Chen, M. G. S. Naylor and A. R. Rosenfield, Wear processes in sliding systems, Wear, 100 (1984) 195-219. 24 K. Hiratsuka and A. Sugahara, Role of atmospheric oxygen in friction and wear between dissimilar pure metals, 1. Tribal., 34 (11) (1989) 19-26 (in Japanese). 25 C. L. Goodzeit, R. P. Hunnicutt and A. E. Roach, Frictional characteristics and surface damage of thirty-nine different elemental metals in sliding contact with iron, Trans. ASME, 78 (1956) 1669-1676. 26 A. E. Roach, C. L. Goodzeit and R. P. Hunnicutt, Scoring characteristics of thirty-eight elemental metals in high speed sliding contact with steel, Trans. ASME, 78 (1956) 1659-1667. 27 E. Rabinowin, The influence of compatibility on different tribological phenomena, ASLE Trans., 15 (1972) 206212. 28 A. R. Miedema and F. J. A. den Broeder, Z. Metullk., 70 (1979) 14-20.