Despite studies from several perspectives, friction still lacks a general theory or approach, mainly due to the complexity in identifying a few relevant variables for the frictional dynamics, among a macroscopic number of degrees of freedom, suitable for the description of the main slow modes of the system. The main goal of the following PhD research project is to try to fill this gap using statistically-based techniques developed in the fields of data science and biomolecular simulations, like clustering algorithms and Markov State Modeling, here extended to non-equilibrium phenomena. Their combined effect is to reduce the dimensionality of the system, by singling out a few, slow, and most relevant time scales and the observables that best describe them. The method we have developed comprises three main steps: a long atomistic Molecular Dynamics (MD) simulation of steady-state frictional sliding, a dimensional reduction and finally an analysis of the time scales that underlie the transitions between the main events. To develop the new method, systems of increasing complexity were successfully considered, from a 1D model to a realistic three-dimensional (3D) system of an island of graphene sliding on a rough substrate of gold.
|Titolo:||Markov State Modeling of 2D Nanofrictional Sliding, on Smooth and Rough Surfaces|
|Data di pubblicazione:||25-ott-2019|
|Appare nelle tipologie:||8.1 PhD thesis|