Lessons learned from 25 years of friction modeling

We have long known that tangential contact forces and the friction coefficient at a sliding contact arises not due to some "law" but as a result of contact geometric and material properties in combination with physical processes that are not always accessible to direct experimental verification. While tribologists have occasionally sought alternatives to the coefficient of friction as the primary means to quantify tangential forces at sliding contacts, its convenience, simplicity, long history and lack of viable alternatives assure that it will be remain in use for some time to come. Sometimes the coefficient can remain constant with normal load over orders or magnitude. One can be led to believe that the coefficient is a property of the interface or there is a "law" at work. At other times, the coefficient, usually measured, is said to vary with time, sliding distance or some other variable with no explanation as to why the variation occurs. In the present paper we discuss aspects of friction behavior and modeling, focusing on mechanics and the nature of the friction coefficient. The point of departure is a case where the adhesion theory of friction can be used to understand the mechanics of friction. One can then consider a number of situations and concepts that arise as variations of this case including average versus instantaneous friction; cases when Amontons-Coulomb relations do or don't apply; static versus dynamic friction; scale effects at rough surfaces; limits of applicability of the friction coefficient; friction measurements, thermal and velocity effects; and local versus global friction and constitutive relations.