CFD modeling and optimal design of SiC deposition on the fuel combustion nozzle in a commercial CVD reactor

Abstract

This work presents a new CFD model developed to investigate the SiC film growth over a fuel combustion nozzle in the commercial hot-wall CVD reactor. A detailed 3D model consisting of the compressive transport models and the reaction mechanism to account for the chemistry of the gas-phase and surface reactions are developed. The velocity, temperature, and concentration profiles inside the reactor are predicted numerically. The multispecies transport and its interplay with the gas and surface reactions are understood to operate the reactor effectively. The natural convection and hydrodynamics stability of the flow is investigated using various dimensionless groups (Re, Pr. Pe, and Gr/Re2). It has been observed that the buoyancy-driven flow is dominant at a large Reynolds number (Re) and Gr/Re2 ratio. This results in flow recirculation, which ultimately deteriorates the film uniformity. A sensitivity analysis is performed to identify how critical parameters affect the deposition process. Besides, the optimisation of the CVD reactor is also served by combining the support vector machine, a versatile supervised machine learning method, and the Nelder-Mead algorithm to improve the film quality. Our results demonstrate that the present methodology effectively obtains optimal process conditions, significantly enhancing the film performance.