FisMat2017 - Submission - View

Abstract's title: Markov State Modelling of Friction in a Realistic 2D Model
Submitting author: Martina Teruzzi
Affiliation: SISSA (Scuola Internazionale di Studi Superiori Avanzati)
Affiliation Address: Via Bonomea, 265, 34136 Trieste
Country: Italy
Oral presentation/Poster (Author's request): Oral presentation
Other authors and affiliations: Franco Pellegrini, Alessandro Laio, Erio Tosatti
Abstract
Friction, despite being a well-known and studied phenomenon, still lacks a general theory or approach, mainly due to the presence of many degrees of freedom and the difficulty in identifying few relevant collective variables. We propose a method based on Markov State Modelling, a technique aimed at reducing the dimensionality of the system, by singling out the relevant slow timescales and the observables that best describe them.
After successful application to a small toy model (1D Frenkel-Kontorova) we now apply this approach to more complicated systems, identifying few significant states that best characterize them through a rather general and automatic algorithm. This approach can provide insight into the main mechanisms of new frictional systems and, in perspective, could be enhanced by biased sampling to achieve a proper description of rare events, difficult to be sampled by standard dynamics.