This thesis describes the key steps in the production of high purity (solar-grade) silicon from metallurgical-grade silicon for use in the production of photovoltaic cells as alternative renewable, environmentally benign and cheap energy source. The initial part of the project involves the development and optimization of a small chemical production platform system capable of producing alkoxysilanes from metallurgical-grade silicon as green precursors to solar-grade silicon production. Specifically, the main aim of the study was to synthesize trialkoxysilanes in continuous flow mode, although the synthesis on monosilane was also done in batch mode. The alkoxylation reaction was carried out in a traditional slurry phase batch reactor, packed bed flow tubular reactor and also attempted in a continuous flow falling film tubular reactor. The effect of key parameters which affect the silicon conversion and selectivity for the desired trialkoxysilane were investigated and optimized using ethanol as a reagent model. The synthesis was then extended to the other alcohols namely methanol, n-propanol and n-butanol. Copper catalysts which were tested in the alkoxylation reaction included: CuCl, Cu(OH)2, CuO and CuSO4. CuCl and Cu(OH)2 showed comparable activity in the batch mode but the former was more efficient in the packed bed flow tubular reactor. Cu(OH)2 could be used as a non-halide catalyst but its activity is limited to short reaction cycles (<10 h). The uncatalysed reaction resulted in negligible reaction rates in both types of reactors. High temperature catalyst pre-heating (>500 oC) resulted in a lower rate of reaction and selectivity than when slightly lower temperatures are used (<350 oC) in both reactors, although much difference was noticed in the packed bed flow tubular reactor. Synthesis in the batch reactor needed longer silicon-catalyst activation time, higher pre-heating temperature and higher catalyst amounts as compare to the packed bed flow tubular reactor. Reaction temperature and alcohol flow rate influenced the reaction in both methods. The optimum reaction temperature range and alcohol flow rate was comparable in both reactors (230 to 240 oC) and 0.1mL/min respectively. The effect of alcohol R-group (C1 to C4) on the reaction revealed that conversion and selectivity generally decrease with an increase in carbon chain length in both methods. Ethanol showed highest selectivity (>95% in batch and >97% in flow) and conversion (about 88% in batch and about 64% in flow) as compared to all other alcohols studied showing that it could be the most efficient alkoxylation alcohol for this reaction. Overally, the packed bed flow tubular reactor resulted in higher selectivity to trialkoxysilanes than the batch system. Performing the reaction under pressure resulted in increased conversion but selectivity to the desire trialkoxysilane diminished. Synthesis in a continuous flow falling film tubular reactor was not successful as it resulted in very poor conversion and selectivity. Monosilane was successfully synthesized from the disproportionation of triethoxysilane using homogeneous and heterogeneous catalysts in batch mode. The results obtained from homogeneous catalysis showed that the reaction can be conducted at room temperature. The heterogeneous catalysis method resulted in slow conversion at room temperature but mild heating up to 55 oC greatly improved the reaction. Conducting the reaction under neat conditions produced comparable results to reactions which were carried out using solvents. The disproportionation reaction was best described by the first order kinetic model. The results obtained in this research indicate that the packed bed flow tubular reactor can be utilized with future modifications for continuous flow synthesis of alkoxysilanes as feedstock for the solar-grade silicon production.

Not available / text