A multiscale model for CVD growth of silicon carbide

Abstract

Thin film deposition is a complex process of chemical and physical phenomena occurring at different time and length scales. A thorough knowledge of the deposition mechanism is necessary to design chemical vapor deposition (CVD) reactors and optimize CVD processes. CVD of (111)- and (100)-oriented SiC films was studied by a multiscale approach, which was designed to describe the fluid dynamics and local composition in the reactor through the solution of the mass, energy, and momentum conservation equations with the inner doubly-iterative efficient algorithm for linked equations (IDEAL) algorithm and finite volume method (FVM). At the atomic scale, a kinetic Monte Carlo model is adopted, and the parameters are taken from the reliable sources. The coupling between the reactor and the atomic scale is realized by applying the continuity of gas phase concentration and flux at the common boundary. The predicted growth rate, film texture, and grain quality of the simulation results were qualitatively compared with corresponding experimental data. This method allows us to optimize the operating conditions of the reactor scale, which in turn can control the performance of the final film.