Theoretical adjustment of metalorganic chemical vapor deposition process parameters for high-quality gallium nitride epitaxial films

Abstract

Semiconductor thin films for electronic devices are usually produced through processes such as chemical vapor deposition, which requires careful control of the gas flow, heat distribution, and concentration distribution over the substrate to ensure a uniform deposition rate and thickness. Herein, a systematic method is proposed for the theoretical adjustment of metalorganic chemical vapor deposition (MOCVD) process parameters. To this end, a GaN-MOCVD reactor with a vertical injection structure was simulated based on computational fluid dynamics to analyze the stable flow under a fixed flow rate. The orthogonal experimental design was used to analyze the influence of process conditions on film quality. A neural network and genetic algorithm were used to optimize the inlet flow under the stable flow state to render the coefficient of variation <3%. Under these conditions, the flow field in the reactor was stabilized to ensure a uniform thickness for the deposited film. This study provides not only an effective solution for high-quality epitaxial growth but also a theoretical basis for subsequent experiments and equipment improvement.