Optimization of the chemical vapor deposition process for gallium nitride

Simulation and optimization of the Metalorganic chemical vapor deposition (MOCVD) process for the deposition of Gallium Nitride (GaN) in a rotating-disk reactor is studied. Precursors, trimethylgallium (TMGa) and ammonia (NH₃) are carried by hydrogen (H₂). The focus of the study is on the rate of deposition and on the uniformity of the thin film. The level of uniformity needed depends on the intended application, with electronic and optical materials imposing the most stringent demands. Large area film thickness and composition uniformity are achieved by proper control of the governing transport processes. This study is broadly divided into three parts. Initially, thin film deposition is simulated by the Computational Fluid Dynamics (CFD) model. The chemistry model has 17 gas phase and 23 surface species participating in 17 gas phase and 52 surface reactions. These numerical simulations are used to determine the effects of important design parameters and operating conditions on the deposition rate and film uniformity. Two design variables, inlet velocity and inlet precursor concentration ratio, which have a significant effect on the deposition rate and uniformity of the film are identified. Inlet precursor concentration ratio is defined as the ratio of the volume flow rate of ammonia to the volume flow rate of trimethylgallium. In the second part, response surfaces for deposition rate and uniformity as a function of inlet velocity and inlet precursor concentration are generated. Compromise response surface method (CRSM) is used to develop the response surfaces. Lastly, the response surfaces are used to generate the Pareto front for the conflicting objectives of optimal deposition rate and uniformity. The trade-off between deposition rate and uniformity is captured by the Pareto front.