Systematic analysis of Persson's contact mechanics theory of randomly rough elastic surfaces

TL;DR

This paper critically evaluates Persson's contact mechanics theory for rough elastic surfaces, identifying key assumptions, quantifying deviations through simulations, and suggesting avenues for systematic improvements.

Contribution

It systematically tests and quantifies the assumptions of Persson's theory using high-resolution simulations, revealing significant discrepancies and potential for theory refinement.

Findings

Diffusion coefficient depends on pressure at small p

Reentrant contact points are significant

Error cancellations lead to accurate overall predictions

Abstract

We systematically check explicit and implicit assumptions of Persson's contact mechanics theory. It casts the evolution of the pressure distribution ${\rm Pr}(p)$ with increasing resolution of surface roughness as a diffusive process, in which resolution plays the role of time. The tested key assumptions of the theory are: (a) the diffusion coefficient is independent of pressure $p$, (b) the diffusion process is drift-free at any value of $p$, (c) the point $p=0$ acts as an absorbing barrier, i.e., once a point falls out of contact, it never reenters again, (d) the Fourier component of the elastic energy is only populated if the appropriate wave vector is resolved, and (e) it no longer changes when even smaller wavelengths are resolved. Using high-resolution numerical simulations, we quantify deviations from these approximations and find quite significant discrepancies in some cases. For example, the drift becomes substantial for small values of $p$, which typically represent points in real space close to a contact line. On the other hand, there is a significant flux of points reentering contact. These and other identified deviations cancel each other to a large degree, resulting in an overall excellent description for contact area, contact geometry, and gap distribution functions. Similar fortuitous error cancellations cannot be guaranteed under different circumstances, for instance when investigating rubber friction. The results of the simulations may provide guidelines for a systematic improvement of the theory.