Fast kinetic Monte Carlo simulation of strained heteroepitaxy in three dimensions

TL;DR

This paper presents accelerated 3D kinetic Monte Carlo algorithms for simulating strained heteroepitaxial growth, combining Green's function methods, surface coarsening, and efficient energy estimation to improve computational performance.

Contribution

It introduces a novel combination of Green's function formalism, surface coarsening, and energy approximation techniques for faster 3D heteroepitaxy simulations.

Findings

Significantly improved simulation speed for 3D heteroepitaxy.

Accurate modeling of surface diffusion with reduced computational cost.

Effective balance between accuracy and efficiency in atomic energy calculations.

Abstract

Accelerated algorithms for simulating the morphological evolution of strained heteroeptiaxy based on a ball and spring lattice model in three dimensions are explained. We derive exact Green's function formalisms for boundary values in the associated lattice elasticity problems. The computational efficiency is further enhanced by using a superparticle surface coarsening approximation. Atomic hoppings simulating surface diffusion are sampled using a multi-step acceptance-rejection algorithm. It utilizes quick estimates of the atomic elastic energies from extensively tabulated values modulated by the local strain. A parameter controls the compromise between accuracy and efficiency of the acceptance-rejection algorithm.