W&ZV Elsevier Sequoia
in the Netherlands
M. MATSUNAGA Institute of Industrial Science, University of Tokyo, Roppongi, Tokyo (Japan) K. HOSHIMOTO Nationaf
Research institute for Metals, ~egur~ku,
Y. UCH~AMA Faculty of Te~hnalogy, Kanazawa Unjversiry, Kanazawa (Japan) (Received
May 8, 1972)
The coefficient of friction and stop time effect of molybdenum disulphide in high vacuum were investigated. It was found that two kinds of stop time effect existed. In the high coefficient of friction region, friction decreased during stopping. On the contrary, in the low coefficient of friction region friction increased during stopping. These phenomena can be explained by vapour adsorption. It was also found by electrical contact resistance measurements between a slider and a base covered with a molybdenum disulphide film that the wear mechanisms of the film were different in air and in high vacuum.
It is believed that the coeflicient of friction (& usually rises in higher vacuum in the case of solids which are bound by a metallic or covalent bond, because the surface film prevents inter-atomic binding. Many experimental observations support this view. However, it is possible that adsorbed vapours strengthen inter-atomic binding as, in solids which are bound by van der Waals’ forces, especially when the anion is exposed on the surface, it is observed that j& decreases in higher vacuum. Lamellar solids, especially molybdenum disulphide (MO&) have such properties’. It has been reported that the j.&of MoS, decreases in higher vacuum’ and that friction is affected by water vapour2. This report describes the authors’ experiments in this field. APPARATUS
The apparatus used is shown in Figs. 1 and 2 and consisted of a metal system bakable to [email protected]
evacuated by a sorption pump (2) and an ion pump (1) to ensure freedom from contamination. The equipment was used in a vacuum system at presWear, 22 (1972)
Fig. 1. Friction apparatus (schematic) (1) Ion pump (2) Sorption pump (3) Valve (4) Strain gauges (5) Plate for measuring frictional force (6) Slider (7) Vacuum gauge (8) Connecting lever (9) Bellows (10) Bearing (11) Crank (12) Speed changer (13) Gear box (14) Electric motor.
Fig. 2. Apparatus.
sures down to 10m9 Torr. To ensure high vacuum, penetration of the moving part through the vacuum wall was avoided, and movement of the frictional apparatus was effected by means of bellows. The reciprocating stroke was 15 mm and the reciprocation period was 3 seconds. The load on the slider was 110 g or 95 g. Electrical contact resistance between Wear, 22 (1972)
FRICTIONAL BEHAVIOUR OF MoS, IN HIGH VACUUM
Fig. 3. Electron micrograp~ of the MoS, used.
the slider and the plate was measured at a low voltage, according to a method developed by Matsunaga3. Figure 3 shows an electron micrograph of the MO& used. The electron beam accelerated by 80 kV could penetrate the thin flakes. The size of the flakes was of the order of 1OOpm. Lumps of about I pm diameter which appeared to be aggregates were observed. No diffraction pattern other than that of MoS, (ASTM X-ray card No. 6-0097) was detected. The main object of the diffraction study was to detect if MOO, reported as the most sensitive compound for vapour adsorption and additives were present. However, neither was detected. Tsuya’s method4, was adopted to make MoS, adherent to a copper plate. MoS, was suspended in ethyl alcohol, floated on water and scooped on to a copper plate. Adherent MoS, was measured as 1 pm by weighing. TWO KINDS OF STOP TIME EFFECT
It was reported by Halter’ that friction increased after the experiment was stopped and the apparatus allowed to stand for varying periods. The phenomenon of varying friction after stopping is called the “stop time effect”. In the present work it was found that the stop time increased friction in some instances but decreased it in others. Measurements were first made on a sedimentary film MoS, about 1 pm thick. The load was 95 g. The coefficient of friction and the electrical contact resistance R are shown in Fig. 4. At the initial stage of the friction test, the coeffjcient of friction was relatively high, about 0.08. From 400 to 2600 strokes friction decreased during stopping. This decrease in friction during stopping has not been reported previously. The change of friction on re-starting is shown in Fig. 5. After 3800 strokes, friction increased during stopping, as previously reported’. The stationary coefIicient of friction decreased to a value of 0.04. The transient friction and the electrical contact resistance are shown in Fig 6. After longer periods of standing, a higher coefficient of friction was obtained upon resumption of sliding followed by a gradual decrease to the steady-state. The initial value of g,=O.OS was similar to that at the initial stage of run-in in Fig. 4. It Wear, 22 (1972)
1 f 0.
Fig. 4. Coefticient of friction and electrical strokes in a vacuum of 5 x 10m9 Torr.
as a function
of the number
1 5 X10-‘torr. 50
Fig. 5. Friction and electrical contact resistance during stopping in a vacuum of 5 x 10M9 Torr. Wear, 22 (1972)
for MoS, on re-starting
when friction decreased
was reported by Haltner’ that the stop time effect was due to the adsorption of gases on the MoSz surface. He postulated that the intrinsic pk of clean MoS, was low and that the adsorbed vapors increased pk In the present experiments the lower pk was found in a higher vacuum of lo-’ Torr, in accordance with Haltner’s discussion. Furthermore, it was found that the stop time effect was more prominent in a vacuum of 1.1 x 10m7 Torr than in one of 5 x lo-’ Torr (e.g., the friction transient
Repeats of reciprocations Fig. 6. Friction and electrical contact resistance transients during stopping in a vacuum of 5 x 10m9 Torr.
for MoS, on re-starting
1.1 X lo-’
when friction increased
Fig. 7. Friction and electrical contact resistance transient during stopping in a vacuum of 1.1 x lo-’ Torr. Wear, 22 (1972)
0 lhr “30 min n 10 min 0 4 min @ 2 min
when friction increased
in 1.1 x 10e7 Torr is indicated in Fig. 7) that the stop time effect became more significant after a longer period of standing, and that the stop time effect did not appear, as reported by Tsuya’, if out-gassing of MoS, was performed thoroughly. These phenomena are in reasonable agreement with the postulate of gas adsorption. The difference in stop time effect due to film thickness can be explained by vapour adsorption in the substrate. With a thick film, vapour adsorbed under the surface was not evacuated from the film even in high vacuum. During the friction test vapour diffused to the surface and the friction was high. As equilibrium was attained after stopping, friction became relatively low on resumption of sliding. However, with a thick film, vapour diffused from the substrate to the surface and friction increased in a short period. With a thin film, frictional heating expelled the adsorbed vapour and the friction was low. Friction increased by vapour adsorption during stop times. As the adsorbed quantities of vapour are in a state of equilibrium, the friction just after the resumption of sliding was the same regardless of the film thickness and depended only on the vacuum. To determine the nature of the vapour which influenced friction it would be necessary to analyze the vapour by mass spectrometry. T~NSITION
The variation of friction compared with pk on re-starting the friction test was investigated under a load of 200 g. The film was out-gassed at 250°C for 8 hours and the vacuum was 4 x lo-” Torr and 2 x loss Torr. The experiment was performed under relatively thin film conditions when the friction increased after stopping.
Fig. 8. The variation Wear, 22 (1972)
of friction on re-starting.
A friction test in a vacuum
of 4 x lOa
The variation of ,ukon restarting in a vacuum as low as 4 x 10e6 Ton: is shown in Fig. 8. The stable yk was 0.032 and the ordinate indicates p-pk, where p is the coefftcient of friction at each instant. As shown, p-pk decreased exponentially when the stop time was shorter than 5 minutes. When the stop time was longer than 5 minutes, the friction remained at a constant value for a time after re-starting and the transient curve has a.plateau. For instance, a plateau occurs at p-pk = 0.058 in a vacuum of 4 x 10m6 Torr as shown in Fig. 8. The value of the plateau was different depending on the specimens, but it was reproducible for each specimen. It was assumed that the part of the curve which decreased exponentially was caused by desorption of the adsorbed vapour and that the plateau was caused by chemisorption. WEAR MECHANISM
OF THE FILM
Comparison of the wear mechanisms of &loS, films in high vacuum and in air was.made by electrical contact resistance measurements during the friction test. The transitions of the distribution of the electrical contact resistance were measured during the friction tests, in a vacuum of 1 x lo-’ Torr under a load of 200 g. An experimental example is shown in Fig. 9. The abscissa indicates the distribution ratio or percentage of the part having resistance exceeding the value of the ordinate. As shown the curve obtained in high vacuum is nearly flat, indicating that the distribution of resistance
Fig. 9. Transition of the distribution curve of the electrical Torr. The specimens were baked for 8 hours at 250” C. Fig. 10. Distribution is worn by removal Wear, 22 (1972)
in a high vacuum
curves in air which are different. The curves are nearly vertical, of large pieces.
of 1 x low9
that the lilm
values was almost constant at a definite resistance, that the value decreased gradually in the course of the friction test and that the MoS, film was worn in thin flakes. Figure 10 shows similar curves for air. It is assumed that the adsorbed vapours strengthen the binding force between MoS, layers. On the other hand, the binding force between the metal surface and the MoS, layer would be weaker in air, since, in air, the metal surface is covered with a protective film. This caused easy sliding between the MoS, aggregates and metal surfaces, and wear of the MoS, film by removal of large pieces. In high vacuum, adhesion of MoS, to a metal surface is strong because the metal surface is free from a protective film. Therefore, wear proceeds by sliding between the MoS, layers and by flaking of the layers. Although the surface layer of the MoS, film was removed the substrate layers would remain to prevent seizure. ACKNOWLEDGEMENT
The authors are indebted to Mrs. Y. Tsuya, Mechanical Engineering Laboratory, for many worthwhile and interesting discussions during this work. REFERENCES 1 2 3 4 5
A. J. Haltner, Wear, 7 (1964) 102. C. Pritchard and J. W. Midgley. Wear, 13 (1969) 39. M. Matsunaga, Rep. Inst. Ind. Sci., Tokyo Univ., 7 (1958) 225. Y. Tsuya, Proc. ASLE Intern. Conf. on Solid Lubrication, Denver, Cola., 1971, P. 59. Y. Tsuya, personal communication.
Wear, 22 (1972)