SOME NEW EXPERIMENTAL STUDIES OF FRICTION*)

Power losses caused by the presence of friction reach 20% in an automobile, 9% in an aircraft engine, and \(1 \tfrac{1}{2}\)—2% in a turbine. No less important is the wear of rubbing parts. In any case, both these reasons fully justify the attention devoted to studying the mechanism of friction.

One of the essential problems in this field is the question of the contact surface of two solid bodies. This question is most simply solved by examining under a microscope sections perpendicular to the contact surface. The effectiveness of this method increases sharply if the sections are made at small angles; irregularities in the vertical direction can thereby be increased, thus

) F. P. Bowden, Nature 166*, 330 (1950).

Thus, by more than 10 times. In such microphotographs the horizontal and vertical scales will be different.

Figure 1 shows a section of this kind of a polished copper surface. The vertical magnification is 1800, and the horizontal is 175. The height of the surface “teeth” is about 50,000 angstroms.

Of course, to obtain a microphotograph of a section one may use an electron microscope, and surface irregularities of up to several tens of angstroms may be detected. It has been shown that even the best polishing creates irregularities of the order of several hundreds of angstroms in height.

Measurement of surface profiles in microphotographs makes it possible only to evaluate qualitatively the contact surface of two bodies. Quantitative measurements carried out by the method of electrical resistance show that the true area of contact has

Fig. 1.

a magnitude on the order of 0.001 of the visible surface. At the same time, the true area of contact is almost independent of the size of the surfaces, but is, however, strictly proportional to the load.

The conclusion that plastic flow of metal occurs at the points of contact is unquestionable. This flow occurs already at the very smallest load. At the points of contact there also occurs “pressure welding.” This phenomenon can be studied by the method of radioactive isotopes.

Figure 2 shows a photograph of the surface of copper over which a piece of radioactive copper was slid (in the presence of lubrication). The dark spots show the places into which the radioactive copper penetrated. The surface concentration is approximately \(10^{-8}\) g/mm\(^3\).

Of considerable interest are investigations of the true temperatures that develop at the points of contact of surfaces during rapid sliding. Measurements of this type can be made using rubbing metals as a thermocouple.

The temperature is determined by the pressure, the sliding velocity, and the thermal conductivity of the metals. The temperature changes very sharply in

time of sliding, and to record the temperature curve it is necessary to use a cathode oscillograph.

In low-melting metals, during friction the melting temperature is reached; in high-melting metals it reaches up to 500—1000°C. These temperatures may occur in a very thin layer and may exist for a time on the order of ten-thousandths of a second. Fig. 3 shows such a temperature curve (time horizontally in milliseconds, temperature in °C). The load (500 g) was a constantan slider, which slid (at a speed of 0.3 m/sec) over a well-polished steel surface. In the friction of nonconductors one may expect

Fig. 2.

Fig. 3.

higher temperatures. For measuring temperature here the thermoelectric method is no longer applicable. However, another way is possible for studying the phenomena. If, in complete darkness, one observes the friction of transparent polished solid bodies (glass, quartz), local flashes may be detected on the surface of the rubbing bodies. At low sliding speeds the flashing points are red; with an increase in

speed they glow and become brighter. These “hot” spots on the contact surfaces of solid bodies can easily be photographed. Photoelectric cells were also used to record these flashes.

Recently the coefficients of friction of new materials belonging to the family of plastics have been measured.

It is interesting to compare the following figures for the coefficients:

\[ \begin{array}{lll} \text{Teflon} & \left[-\mathrm{CF}_{2}-\mathrm{CF}_{2}-\right]_n & 0.04, \\[1.2em] \text{polyethylene} & \left[-\mathrm{CH}_{2}-\mathrm{CH}_{2}-\right]_n & 0.1, \\[1.2em] \text{polystyrene} & \left[-\mathrm{CH}_{2}-\underset{\mathrm{CH}_{3}}{\overset{\mathrm{C}_{6}\mathrm{H}_{5}}{\mathrm{CH}}}-\right]_n & 0.5, \\[1.2em] \text{Perspex} & \left[\mathrm{CH}_{2}-\underset{\mathrm{COOMe}}{\overset{\mathrm{CH}}{\mathrm{C}}}-\right]_n & 0.8. \end{array} \]

As is evident from this table, polystyrene and Perspex, whose polymer molecules have side chains, possess high coefficients of friction. It is also possible to understand why the coefficient of friction of Teflon is less than the coefficient of friction of polyethylene. The effect of replacing hydrogen atoms by fluorine atoms may be imagined as follows. Fluorine atoms are larger than hydrogen atoms and have an effective negative charge; they completely shield the carbon atom. During sliding, contact occurs only between fluorine atoms. At the same time, in polyethylene, along with intermolecular hydrogen—hydrogen contacts, contacts may occur between carbon and hydrogen atoms.

Teflon also has the remarkable property that it retains the value of its coefficient of friction unchanged up to temperatures on the order of 300° C. The practical use of Teflon is hampered by its low thermal conductivity and large coefficient of expansion. If, however, a composite material is made—namely, if a thin layer (1 mm) of Teflon is pressed into a piece of porous copper—then a very interesting material is obtained, possessing all the properties of copper but having a coefficient of friction of 0.05 instead of 1.0 up to a temperature of 250° C. It should be emphasized that the value 0.05 is the minimum value of the coefficient of friction in the presence of good lubrication.

A. K.