Lattice gas cellular automaton modeling of surface roughening in homoepitaxial growth in nanowires

Our research addresses the problem of bridging large time and length scale gaps in simulating atomistic processes during thin film deposition. We introduce a new simulation approach based on a discrete description of atoms so that the unit length scale coincides with the atomic diameter. The interaction between atoms is defined using a coarse-grained approach to boost the computation speed. This approach does not heavily sacrifice the atomistic details in order to study structural evolution of a growing thin film on time scales in the order of seconds and even minutes. Our approach is inspired by lattice gas cellular automata models for chemically reacting systems, where individual particles interact with surrounding through assumed local driving forces. For homoepitaxial thin film deposition, the local driving force is the propensity of an atom to establish as many chemical bonds as possible to the underlying substrate atoms when it executes surface diffusion. Simulation results of Si layers deposited on a flat Si(001) substrate are presented.