Search

Hydrogenation of silicon tetrachloride(STC) in the presence ofsilicon was investigated both thermodynamically and experimentally.The thermodynamic results revealed that the reaction is thermodynamicallyconstrained and the STC equilibrium conversion is nearly temperature-independentbut increases with increasing pressure and H2/STC feedratio. Trichlorosilane (TCS) was calculated to be the dominant product,with trace amounts of dichlorosilane (DCS) and HCl as feasible byproducts.Experimental tests were conducted in a fixed-bed reactor loaded withsilicon. The results indicated that the reaction is kinetically restrictedin the absence of a catalyst. Adding CuCl to the silicon mass bedwas found to result in a dramatic increase in the STC conversion,and Cu3Si was determined to be the active species. In allof the tests runs, TCS was determined to be the dominant product,and the STC reaction rate was found to increase markedly as the temperatureand STC partial pressure increased but not to depend on the siliconparticle size. [ABSTRACT FROM AUTHOR]

TheCuCl-catalyzed hydrogenation of silicon tetrachloride (STC)in the presence of silicon was conducted in a fixed-bed reactor. Thesolid mass from different reaction stages was collected and characterizedusing X-ray diffraction (XRD) and scanning electron microscopy coupledwith energy-dispersive X-ray spectroscopy (SEM-EDX). The characterizationresults were used to develop a novel reaction mechanism. This mechanismis remarkable in that the Cu–Si surface species acted as botha catalyst and a solid reactant and could be regenerated via Cu diffusioninto the bulk silicon phase. Different kinetic models were formulatedby employing the Langmuir–Hinshelwood (L–H) and Eley–Rideal(E–R) approaches. Model discrimination was used to select akinetic model based on the E–R mechanism, in which the surfacereaction between the adsorbed STC and gas-phase H2overthe Cu3Si active sites was assumed to be the rate-determiningstep (RDS). The kinetic parameters were estimated using nonlinearregression, and the model predictions were in good agreement withthe experimental data. [ABSTRACT FROM AUTHOR]