Numerical modeling of gas-phase nucleation and particle growth during chemical vapor deposition of silicon

A numerical model was developed to predict gas-phase nucleation of particle s during silane pyrolysis. The model includes a detailed clustering mechani sm for the formation of hydrogenated silicon clusters containing up to ten silicon atoms. This mechanism was coupled to an aerosol dynamics moment mod el to predict particle growth, coagulation, and transport. Both zero-dimens ional transient simulations, at 1-2 atm pressure, and one-dimensional stead y-state stagnation-point flow simulations, at 1-2 Ton pressure. were conduc ted. The effects of carrier gas, temperature, pressure, silane concentratio n, and flow rate were examined. The results predict that hydrogen as carrie r gas, compared to helium, suppresses nucleation, and that particle formati on for the case of hydrogen carrier gas increases strongly with increasing initial silane-to-hydrogen ratio. For the conditions examined, predicted pa r tide nucleation rates increase dramatically with increasing temperature. The effect of total pressure depends on the pressure regime: at 1-2 atm pre ssure particle formation is predicted to be insensitive to pressure, wherea s at 1-2 Ton particle formation is predicted to increase strongly with incr easing pressure. The predicted effects on particle formation of temperature , pressure, carrier gas, and silane concentration are all qualitatively con sistent with published experimental results. In the stagnation-point flow s imulations the flow rate is found to affect particle dynamics because of th e opposed effects of convective transport toward the heated water and therm ophoretic transport away from the wafer. (C) 2000 The Electrochemical Socie ty. S0013-3651(00)02-051-6. All rights reserved.