A Markov state modeling analysis of sliding dynamics of a 2D model

Martina Teruzzi, Franco Pellegrini, Alessandro Laio, Erio TosattiThe Journal of Chemical Physics 2017 · DOI · arXiv

Abstract

Non-equilibrium Markov State Modeling (MSM) has recently been proposed by Pellegrini et al. [Phys. Rev. E 94, 053001 (2016)] as a possible route to construct a physical theory of sliding friction from a long steady state atomistic simulation: the approach builds a small set of collective variables, which obey a transition-matrix-based equation of motion, faithfully describing the slow motions of the system. A crucial question is whether this approach can be extended from the original 1D small size demo to larger and more realistic size systems, without an inordinate increase of the number and complexity of the collective variables. Here we present a direct application of the MSM scheme to the sliding of an island made of over 1000 harmonically bound particles over a 2D periodic potential. Based on a totally unprejudiced phase space metric and without requiring any special doctoring, we find that here too the scheme allows extracting a very small number of slow variables, necessary and sufficient to describe the dynamics of island sliding.

Introduction

Friction is a pervasive physical phenomenon, yet it lacks a clean theoretical formulation. When two surfaces slide against one another, the physics is governed by the interactions of roughly $10^{23}$ atoms. Tracking all of them is impossible, and identifying the handful of "collective variables" that actually matter remains a significant challenge.

Recently, researchers proposed using Markov State Modeling (MSM)—a probabilistic technique usually used for equilibrium systems like protein folding—to model the non-equilibrium dynamics of friction. However, initial proofs were limited to small 1D toy models. This work investigates whether the MSM approach can scale to larger, more realistic 2D systems without a prohibitive increase in computational complexity.

Methods

The study utilizes a 2D Frenkel-Kontorova model consisting of a hexagonal "island" of 1,027 harmonically bound particles dragged across a periodic, 2D triangular substrate. A long, steady-state Molecular Dynamics (MD) simulation provides the raw data.

Figure 1: (a) The 2D hexagonal island model. (b) Mean friction force vs. stiffness ratio $K/U_0$. (c) Implied timescales $\tau_i$ extracted from the MSM. (d) Center-of-mass motion showing the transition from smooth sliding to stick-slip.

The MSM pipeline involves several key steps: a "blind" metric using geometric distance between configurations (modulo the lattice spacing), clustering into microstates via the Density Peak algorithm, and the construction of a coarse-grained Transfer Matrix. By extracting the eigenvalues and eigenvectors of this matrix, the authors identify the "slow modes" that govern fundamental frictional events.

Key Insights and Physical Mapping

A major finding is that a completely naive metric works effectively for 2D friction, automatically filtering out fast atomic vibrations to reveal macroscopic sliding dynamics. These mathematical "excitations" (eigenvectors) map directly to real physical phenomena.

Figure 4: (a) Steady-state density distributions and (b) the first three excited states for particle positions across three stiffness regimes. The patterns reveal the creation and flow of solitons (ripples) in the soft island regime.

For a stiff island ($K/U_0 = 2$), the variables describe smooth, whole-island sliding (superlubricity). Conversely, for a soft island ($K/U_0 = 0.2$), the variables capture the creation and flow of "solitons"—ripples in the lattice that cause jerky stick-slip friction. The MSM is sensitive enough to detect marginal stick-slip behavior even in the superlubric regime, caused by edge pinning.

Key Results

The model successfully identified the primary implied timescale of the system at approximately $6 \times 10^4$ timesteps, corresponding to the physical time required for the external force to move the island halfway across a substrate lattice spacing. By varying the stiffness ratio ($K/U_0$), the MSM captured the transition across three distinct frictional regimes: superlubric, intermediate, and pinned/stick-slip. Furthermore, the first three excited states were sufficient to reconstruct the probability distributions of physical observables such as harmonic energy and work distribution, validating the choice of slow variables.

Conclusion and Future Work

While the 2D Frenkel-Kontorova model is a simplified representation, these results demonstrate that MSM can effectively reduce the dimensionality of complex frictional systems without prior physical prejudice. The successful scaling from 1D to 2D suggests that this framework can be applied to full 3D atomistic simulations of real sliding interfaces, moving toward a predictive physical theory of friction.

Summary updated

Abstract

Non-equilibrium Markov State Modeling (MSM) has recently been proposed by Pellegrini et al. [Phys. Rev. E 94, 053001 (2016)] as a possible route to construct a physical theory of sliding friction from a long steady state atomistic simulation: the approach builds a small set of collective variables, which obey a transition-matrix-based equation of motion, faithfully describing the slow motions of the system. A crucial question is whether this approach can be extended from the original 1D small size demo to larger and more realistic size systems, without an inordinate increase of the number and complexity of the collective variables. Here we present a direct application of the MSM scheme to the sliding of an island made of over 1000 harmonically bound particles over a 2D periodic potential. Based on a totally unprejudiced phase space metric and without requiring any special doctoring, we find that here too the scheme allows extracting a very small number of slow variables, necessary and sufficient to describe the dynamics of island sliding.
BibTeX
@article{teruzzi2017markov,
  author = {Martina Teruzzi and Franco Pellegrini and Alessandro Laio and Erio Tosatti},
  title = {{A Markov state modeling analysis of sliding dynamics of a 2D model}},
  journal = {The Journal of Chemical Physics},
  year = {2017},
  doi = {10.1063/1.4986458},
}
Figure