Optimization of the thermal field of 8-inch SiC crystal growth by PVT method with “3 separation heater method”

Abstract

Silicon carbide (SiC) has important application prospects in power and radio frequency (RF) devices. However, preparing large-diameter, high-quality SiC crystals is challenging. The major technique for growing large-diameter SiC crystals is the physical vapor transport (PVT) method and the key point for adjusting conditions is to find an optimal thermal field. In this study, we propose a resistance heating method called the “3 separation heater method” with 3 heaters separated by the graphite foam insulations so that 3 important parts (the SiC source powder area, SiC seed area, and seed holder) of the growth cell can be heated independently through thermal radiation from each separated heater. Further, the insulation separators play an important role as a switch of the heat flux. The numerical studies of the thermal field design for this model are conducted with the help of the numerical simulation software COMSOL Multiphysics and its appropriate parameters are explored. At last, the relationship between the parameters and crystal growth conditions, such as the radial temperature difference (RTD) and edge temperature gradient (ETG), is expressed by the back-propagation (BP) neural network and the optimal parameter values for crystal growth conditions have been confirmed using the non-dominated sorting genetic algorithm (NSGA-II).