Although particle detachment is a common phenomenon associated with most tribological processes, it seldom occurs that each piece of elemental debris can be considered as the result of a single event. Such an association has been revealed by the systematic study of a specific system, where a pin of graphite is made to rub against thoroughly polished steel. While the discontinuous nature of the transfer film allows a quantitative assessment of the volume of transfer h e to be made by 3D optical-profilometry, the linear dependence of the rate of particle detachment d h e /d n ( n =number of rubbing cycles) with the logarithm of sliding speed v strongly suggests the existence of a particular type of stick–slip, where each stick may lead to the detachment of a debris particle. The variations in size of these debris with environment as revealed by AFM, further suggest that the global rate of particle detachment is of the form: d h e /d n = Nx ε i , where N is the number of stick–slip events per rubbing cycle, x the proportion of stick events leading to a cohesive rupture, and ε i the mean volume of an elemental particle. While this relation is apparently supported by most experimental results, its actual validation can only be made by experiments at the level of single (nanoscale) asperities, carried out under well-controlled experimental conditions.

Material transfer is a common phenomenon in dry friction, generally associated with a change in surface topography and in frictional interactions. In previous work, the actual influence of adhesion, cohesion and interfacial shear stresses on the extent of transfer (the transfer being assessed experimentally by image analysis) was established both on a theoretical and experimental basis, but at constant sliding speed. In this work emphasis is put on the influence of sliding speed. Both the frictional forces and the transfer characteristics (i.e., the area fraction as well as the particle density), as determined after identical sliding distances, decreased linearly with the logarithm of the sliding speed. The possible existence of stick-slip phenomena suggested by these experimental relationships was investigated further by studying the influence of various parameters affecting either the mechanical characteristics of the tribometer (stiffness, inertial moment) or the physicochemical characteristics of the compacted graphite (granulometry of the graphite powder prior to compaction, crystallinity of graphite, relative humidity during sliding, etc.). The results are interpreted by introducing a complex time of stick accounting for the coexistence of both macrostick-slip and microstick-slip, and by introducing this time dependence into the various adhesive interactions involved in the previous model of transfer stability. (C) 1997 Elsevier Science S.A.