Self-learning kinetic Monte Carlo simulations of self-diffusion of small Ag islands on the Ag(111) surface

J Phys Condens Matter. 2016 Jan 20;28(2):025001. doi: 10.1088/0953-8984/28/2/025001. Epub 2015 Dec 11.

Abstract

We studied self-diffusion of small two-dimensional Ag islands, containing up to ten atoms, on the Ag(111) surface using self-learning kinetic Monte Carlo (SLKMC) simulations. Activation barriers are calculated using the semi-empirical embedded atom method (EAM) potential. We find that two- to seven-atom islands primarily diffuse via concerted translation processes with small contributions from multi-atom and single-atom processes, while eight- to ten-atom islands diffuse via single-atom processes, especially edge diffusion, corner rounding and kink detachment, along with a minimal contribution from concerted processes. For each island size, we give a detailed description of the important processes, and their activation barriers, responsible for its diffusion.