Analysis and optimization of surface quality of polysilicon in a 45-pair rods reduction furnace

Abstract

To obtain dense polysilicon, the deposition rate of polysilicon needs to match the surface migration rate of silicon atoms. For this purpose, a numerical model has been established to simulate the deposition rate and the surface temperature of the polysilicon rods in a 45-pair rods reduction furnace. Meanwhile, the distributions of dense and non-dense polysilicon on the rod surface have been obtained based on the production data from the same reduction furnace. Further, by combining the numerical results and production data, a ratio Φ of deposition rate to temperature has been proposed. The results reveal that Φ of 8.0 × 10^-7 kg·m^-2·s^-1·K^-1 can be considered as a threshold to ensure the generation of dense polysilicon. Moreover, using the established numerical model, the operation conditions have been optimized. On the basis of existing operation conditions, reducing the surface temperature of the rods by 10 % and the trichlorosilane (TCS) feed rate by 10 % is beneficial for the generation of dense polysilicon. The results show that the proportion of non-dense polysilicon can be decreased to 2.3 % with a 4.3 % reduction in the deposition rate under the optimized operation conditions.