A hybrid scheme is used to model the vapor phase deposition of thin films at the feature scale. The transport of the chemical species to the substrate surface is modeled with a Collisionless Direct Simulation Monte Carlo (DSMC) method. The Level Set Method is used to model the growth of the thin-film on the substrate. The convergence criteria for these methods were not found in literature. The governing equations for the Level Set Method are, in general, non-linear partial differential equations. The coupling of the DSMC Method with the Level Set Method results in a set of non-Gaussian stochastic non-linear partial differential equations. Developing general convergence criteria proved exceedingly difficult, and only qualitative results are presented to support our convergence criteria. Simulation results are in qualitative agreement with experiments and other results from literature.

In this paper we discuss the use of atomistic Monte Carlo simulations to predict film microstructure evolution. We discuss physical vapor deposition, and are primarily concerned with films that are formed by the nucleation and coalescence of 3D islands. Multi-scale modeling is used in the sense that information obtained from molecular dynamics and first principles calculations provide atomic interaction energies, surface and grain boundary properties and diffusion rates for use in the Monte Carlo model. In this paper, we discuss some fundamental issues associated with thin film formation, together with an assessment of the sensitivity of the film morphology to the deposition conditions and materials properties.