Quasielastic He atom scattering from surfaces: A stochastic description of the dynamics of interacting adsorbates

TL;DR

This paper introduces a stochastic model to interpret the broadening of diffusive peaks in He atom scattering from surfaces, offering an alternative to molecular dynamics simulations by using a single adsorbate approximation with two noise sources.

Contribution

It proposes a stochastic approach with Gaussian and shot noise to describe adsorbate interactions, simplifying the analysis of surface diffusion phenomena.

Findings

The stochastic model fits well with Langevin simulations for flat and corrugated surfaces.

The approach explains broadening of diffusive peaks without extensive molecular dynamics.

Analytical expressions from simple models match numerical results under Gaussian approximation.

Abstract

The study of diffusion and low frequency vibrational motions of particles on metal surfaces is of paramount importance; it provides valuable information on the nature of the adsorbate-substrate and the substrate-substrate interactions. In particular, the experimental broadening observed in the diffusive peak with increasing coverage is usually interpreted in terms of a dipole-dipole like interaction among adsorbates via extensive molecular dynamics calculations within the Langevin framework. Here we present an alternative way to interpret this broadening by means of a purely stochastic description, namely the interacting single adsorbate approximation, where two noise sources are considered: (1) a Gaussian white noise accounting for the surface friction and temperature, and (2) a white shot noise replacing the interaction potential between adsorbates. Standard Langevin numerical simulations for flat and corrugated surfaces (with a separable potential) illustrate the dynamics of Na atoms on a Cu(100) surface which fit fairly well to the analytical expressions issued from simple models (free particle and anharmonic oscillator) when the Gaussian approximation is assumed. A similar broadening is also expected for the frustrated translational mode peaks.