Removal of Boron from Silicon by Gas refining with Water and Hydrogen

Abstract

A high demand of silicon for solar cells has lead to a search of an alternative route to solar grade silicon. A suggested method is upgrading metallurgical grade silicon. This thesis focuses on the gas refining of silicon in order to removal boron from the silicon melt. The most documented refining gas is water and has been used in this thesis in a combination with hydrogen. The aim of this work was to determine the optimum water content in combination with hydrogen for boron removal from silicon. In addition to water content investigation, two different crucible materials were examined for this aim. The temperature of the process was 1500 ºC and the furnace in question was a resistance heated alumina tube furnace. The alumina tube cracked several times during experiments and was a source of error during the experimental work. The cracking of the alumina tube is believed to be due to the high cooling and heating rates of the experiments. It was shown that water is possible oxidation agent for boron removal in Si and that addition of H2 increases the boron removal rate. However, a high water content, such as 11 vol% H2O, showed a slow boron removal rate. The best results came from experiments with 3.2 vol% H2O + 50 vol% H2. During a 1.5 hour long experiment, the boron content was decreased from its doping concentration of 170 ppmw to 1.01 ppmw. The mass transfer coefficient, k*, of this experiment, was found to be 5∙10-5 m/s. ICP-MS and resistivity measurements were conducted to determine the boron concentration of the extracted samples during experiments. The reliability of the resistivity measurements is dependent on the size of the sample. The small sample sizes might have caused some uncertainties in the results. EPMA confirmed the presence of SiC and SiO2 phases in the silicon after experiments. It is assumed that these phases may lead to a blockage of the active surface area, which may hinder the boron oxidation. SiC is formed in the silicon due to a C contamination from the graphite crucibles used. It is believed that using a quartz crucible will lower this contamination.