Purification of metallurgical grade silicon by a solar process

https://doi.org/10.1016/j.solmat.2006.02.009Get rights and content

Abstract

The purification of upgraded metallurgical silicon by extraction of boron and phosphorus was experimentally demonstrated using concentrated solar radiation in the temperature range 1550–1700 °C. The process operated with a flow of Ar at reduced pressure (0.05 atm) for elimination of P, and with a flow of H2O for elimination of B. Impurity content decreased by a factor of 3 after a 50-min solar treatment, yielding Si samples with final average content of 2.1 ppmw B and 3.2 ppmw P.

Introduction

The worldwide solar cell production reached 1256 MWp in 2004, a 67% increase over the 2003 output [1]. Crystalline silicon solar cells, particularly multi-crystalline silicon, made up of more than 90% of the world PV production [2]. Most of the solar grade silicon (SGS) used for PV applications stems from rejects of electronic grade silicon (EGS) used in the electronic industry. However, its annual supply is surpassed by its demand, urging the need for novel technological solutions for the processing of SGS at a target price of 12–15/kg and target demand of 10 g/Wp [3], [4]. Similarly to EGS production, the route for SGS processing uses metallurgical grade silicon (MGS) as starting material. Annual world production of MGS amounts to 1 million tons, obtained by quartz reduction in arc furnaces. The process reaction isSiO2+2CSi+CO,ΔH298K=695kJ.The production of 1 ton of MGS requires 2.9–3.1 ton quartz, 1.2–1.4 ton coke, 1.7–2.5 ton charcoal, 0.12–0.14 ton graphite (electrodes), and 12.5–14 MWh electricity [5]. The purity varies from 98.5% up to 99.5%, and the average price is 1.5–2.5/kg, depending on quality [3]. Some of the typical impurities are Fe, Al, Ca, Ti, Cr, B, P, O, and C. The production of EGS from MGS requires a reduction of these impurities by at least five orders of magnitude. Consequently, EGS processing occurs via distillation of silicon gaseous compounds at 35–50/kg. Purification of MGS is an optional chemical path among the routes for processing SGS. Other routes include reaction of silicon dioxide and carbon with high purity and silicon halogenide reduction with high-purity reduction materials. Purification can be accomplished using plasma techniques [6], [7]. In particular, an inductive plasma torch combined with an inductive cold crucible in a chamber was used to treat 30 kg of upgraded metallurgical silicon (UMGS) per batch. PV cells processed with the resulting material exhibited a conversion efficiency of 12.4% [7], compared with 10.12% measured on the starting material [8]. Boron removal is considered to be critical due to its rather high segregation coefficient that prevents its removal by controlled solidification. Plasma treatment with steam and directional solidification steps allowed the removal of B and other metallic impurities [9]. The main drawback of the plasma purification process is its high electricity consumption that governs the process cost (21/kg in 2002, without manpower [7]). Alternatively, concentrated solar energy can replace the plasma torch and be used as the source of high-temperature process heat. In contrast to a plasma-driven process, no excited species are formed at the reaction surface in the solar-driven process. This paper examines the solar purification of MGS, aimed at the removal of boron and phosphorus by vaporisation. Thermodynamic equilibrium computations are carried out for determining temperature and pressure requirements and product gas compositions. The experimental set-up and main experimental results are presented and discussed.

Section snippets

Thermodynamics

Thermodynamic equilibrium computations were carried out using HSC code [10]. For most calculations, the silicon feed was assumed to consist of 1 kmol silicon, 0.05 mol water, 10 ppmw phosphorus, and 6 ppmw boron mixed with a large excess of argon (100 kmol) in order to account for the renewing of the buffer gas during the experiments. Other impurities accounted were: 300 ppmw Ca, 100 ppmw Mn and Ni; 50 ppmw Fe and Al, 10 ppmw Cr and C; and 5 ppmw Ti. These concentrations correspond to levels that may

Experimental set-up

The experimental set-up is depicted in Fig. 3. The vertical-axis solar furnace of the PROMES Laboratory (Odeillo) consists of a sun-tracking flat heliostat on axis with a 2 m-diameter parabolic concentrator (Fig. 3a), capable of delivering peak solar concentration ratios of 16 000 and attaining temperatures up to 3200 °C. The reactor was placed at its focus. Concentrated solar radiation entered the reactor through a Pyrex window and heated the sample (typically 3 g) placed on a water-cooled

Experimental results

Experiments have been carried out at 0.05 atm to favour P vaporisation, using 1 l/min Ar+2.5 ml H2O to favours B vaporisation as BOH, and in the temperature range 1550–1700 °C to favour P evaporation but may result in SiO formation. Fig. 4 shows an example of temperature monitoring during a representative solar run. The measured signal was not corrected for the samples’ emissivity, which can vary between 0.7 for Si(s) and 0.1–0.3 for Si(l), and is affected by impurities [14]. At about t=763s the

Conclusion

We have performed thermodynamic equilibrium calculations to determine the gas composition over UMGS in argon over a wide temperature range. Boron may be extracted mainly as BOH(g) at low temperatures when steam is added; phosphorus may be extracted as gaseous P above 1500 °C and 1 bar. Pressure reduction further favours P elimination. We have carried out set of solar experimental runs in a solar furnace with batch samples of UMGS. The process operated at reduced pressure (0.05 atm) for elimination

Acknowledgements

Authors acknowledge RIMA for providing UMGS, Ch. Trassy (EPM Lab, Grenoble, France) for fruitful discussion; and D. Gauthier and A. Ferriere (PROMES-CNRS) for assistance during material analysis and experiments.

References (17)

  • T. Surek

    J. Crystal Growth

    (2005)
  • P. Woditsch et al.

    Sol. Energy Mater. Sol. Cells

    (2002)
  • D. Sarti et al.

    Sol. Energy Mater. Sol. Cells

    (2002)
  • C. Alemany et al.

    Sol. Energy Mater. Sol. Cells

    (2002)
  • S. De Wolf et al.

    Sol. Energy Mater. Sol. Cells

    (2002)
  • A. Yeckel et al.

    J. Cryst. Growth

    (1995)
  • M. Schmela, PHOTON International, March...
  • W. Zulehner, “Silicon”; Ullmann's Encyclopedia of Industrial Chemistry, vol. A23,...
There are more references available in the full text version of this article.

Cited by (0)

1

PROMES and ETHZ are Members of “SolLab” (Alliance of European Laboratories for Research and Technology on Solar Concentrating Systems).

View full text