Experimental Study of Silicon Refining by Slag Treatment: Distribution of Boron and Phosphorus

¹ Fluid and Process Dynamics Research Group, Department of Mechanical and Product Design Engineering, Swinburne University of Technology, Melbourne, VIC, 3122, Australia.
² School of Chemical Engineering, The University of Queensland, Brisbane, Queensland, 4072, Australia.
³ CSIRO Mineral Resources, Clayton South, Victoria 3169, Australia.

^* Corresponding author: kavarmaa@swin.edu.au; arhamdhani@swin.edu.au

Abstract

Current production of metallurgical-grade Si relies on an energy-intensive process based on complex solid-liquid-gas reactions in an electric submerged arc furnace using graphite electrodes. This is followed by a slow Siemens process to grow ultra-high purity solar-grade silicon. Due to increasing decarbonization and environmental pressures that the metal industry is facing, new sustainable process technologies are required. This study is part of a broader investigation on pyrometallurgical electrorefining of silicon in the presence of slag and provides the basis for impurity behaviour in Si-slag system without an imposed electromotive force. For the first time, experimental technique of drop-quenching followed by direct phase analysis techniques of electron microprobe analyzer (EPMA) and laser ablation inductively coupled plasma mass spectrometer (LA-ICP-MS) were employed in silicon processing environment. The experiments were carried out to define the distribution coefficients of minor elements (B, P) between Si and SiO₂-CaO-Al₂O₃ at silica and alumina crucibles under inert Ar gas at 1500 °C. The results presented are highly novel and show the influence of slag composition and process conditions on the element properties. The experimental results were compared to the computational simulation results by FactSage© using its FToxid and FSupsi databases. The results can be utilized to design Si–slag process for primary ore resources and for secondary silicon-rich scraps in pyrometallurgical smelting environment, as well as to potentially improve the thermochemical databases of computational simulation programs.