Open Collections

Taposhe, Golam Ismot Ara

University of British Columbia

Solar-grade silicon with a purity of 6-7N (99.9999-99.99999) is generally produced via chemical vapor deposition. Considering the high energy requirements of the process, developing an alternative route for producing solar-grade silicon is critical. Boron (B) and Phosphorus (P) are two common impurities in silicon that are difficult to remove. In this research, a combination of slag (CaO-SiO₂-Al₂O₃-Na₂O) and solvent refining (Si-20wt%Fe alloy) was employed to remove B and P from silicon at 1300˚C. The partition ratios of B and P between the slag and alloy were investigated at various CaO/SiO₂, SiO₂/Al₂O₃, and wt% Na₂O. The findings demonstrated that the partition ratios of B and P were strongly influenced by both basicity and oxygen potential. The addition of 10 wt% Na₂O increased the partition ratio of B and P. Thermodynamic evaluations showed the borate and phosphate capacities have a linear relation with the optical basicity of slag. Slag refining experiments were done at 1250-1400˚C to examine the effect of temperature. The partition ratio of B increased with increasing temperature. However, no significant change was observed in the partition ratio of P. The thermodynamic analysis indicates that the borate and phosphate capacity values are higher at lower temperatures. The Si-Fe alloy and purified silicon were separated to examine the role of solvent refining. The apparent distribution coefficients of B and P between alloy and purified silicon were determined. A lower apparent distribution coefficient was found for B and P at 1250˚C which is beneficial for B and P removal. Upon examining the purified silicon samples, small alloy particles were found to be entrapped within the purified silicon crystals. An attempt was made to decrease the entrapment of alloy particles by solidifying silicon at a reduced cooling rate. Kinetic analysis of B and P removal was carried out with the optimum slag composition at 1300˚C. The mass transfer coefficient was determined for both B and P in the alloy and the slag phase. The result shows that mass transfer in the slag is the rate-limiting step for B removal while for P, mass transfer in the alloy phase is the rate-limiting step.

Text

2024-12-20

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/89995

Materials Engineering

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/