Abstract：The fluidized bed reactor (FBR) process for polysilicon production using silane suffers from issues such as low gas?solid transfer efficiency, poor deposition uniformity, and difficulty in suppressing homogeneous nucleation side reactions. To address these problems, this study established a multiscale computational fluid dynamics framework integrating the Euler-Euler two-fluid model, kinetic theory of granular flow, and population balance model. This framework systematically revealed the cross-scale coupling mechanisms governing gas-solid flow, heat transfer, reaction, and deposition processes. The results indicated that approximately 78% of silicon deposition occurred at the dynamic interface between the dense and bubble phase, where the granular temperature reached up to 1.5×10-3 m2/s2 and the surface renewal frequency was 3.2 times that in the dense core region, which significantly enhanced interfacial mass transfer and active site regeneration. The heterogeneous SiH2 pathway contributed about 10% of the total deposition, whereas when the temperature exceeded 873 K or the silane mole fraction was above 0.5%, the reaction pathway shifted towards homogeneous nucleation, sharply increasing the fine silicon powder yield to over 15%. Through multi-objective optimization, the optimal operating conditions were as follows: superficial gas velocity was 0.50 m/s, the temperature was 873 K, the silane mole fraction was 0.4 %. Under these conditions, the silane conversion rate reached 82.3%, while the fine silicon powder yield was effectively suppressed below 5%. This work clarified the kinetic boundaries governing deposition efficiency and by-reaction competition, providing a reliable theoretical basis for the structural design and intelligent control of high-performance FBR systems.
LI Lianping,LIANG Bo,LI Yin et al. CFD Simulation and Process Optimization of Polysilicon Deposition in a Silane Fluidized Bed Reactor[J]. Chemical Reaction Engineering and Technology, 2026, 42(2): 106-114.
