Hostname: page-component-cd9895bd7-jkksz Total loading time: 0 Render date: 2024-12-29T14:49:55.265Z Has data issue: false hasContentIssue false

Commercial Production of Silicon Solar Cell Feedstock by Upgrade of Metallurgical Grade Silicon

Published online by Cambridge University Press:  31 January 2011

John R Mott
Affiliation:
jrm@ohsoen.com, Silicon Forge, Alliance, Ohio, United States
Julio A Bragagnolo
Affiliation:
jab@ohsoen.com, Silicon Forge, Alliance, Ohio, United States
Michael P. Hayes
Affiliation:
mph@ohsoen.com, Silicon Forge, Alliance, Ohio, United States
Get access

Abstract

The relation between impurity content in Solar Grade Silicon (SGS) and solar cell quality is the subject of intensive research. The PV industry has developed around the use of silicon made by the Siemens process for the semiconductor industry, with impurity levels typically in the parts per billion by weight (ppbw) range. There is a growing consensus that SGS with impurities in the parts per million range (ppmw) can be obtained cost effectively from Metallurgical Grade Silicon (MGS) and used to yield solar cells with comparable performance (see for example ‘Beneficial Effects of Dopant Compensation on Carrier Lifetime in Upgraded Metallurgical Silicon’ by S. Dubois et al. in the 23rd European Photovoltaic Solar Energy Conference, Valencia, September, 2008). This provides insight on the success encountered by Timminco, an early SGS market entrant, in commercializing silicon material with [P] levels of the order of 2 ppmw. Current Work We have successfully reduced P to about 2 ppmw, a level that appears acceptable for solar cell fabrication, by application of a novel unidirectional solidification (UDS) technique at a 50% material yield. This is important as UDS, by its nature, implies a loss of silicon, while little or no silicon is lost in B reduction, partially achieved in this furnace using a glass slagging process. Figure 1 shows [P] data from 16 UDS runs on samples taken from the melt, before and after UDS, and a solid sample taken from the silicon frozen on the cold silicon collection surface. The error bars represent a standard 20% error value. We note that the average values of [P] in the molten silicon samples increase from 11.9 ppmw before UDS to 15.9 ppmw after UDS. The average value of [P] in the solid silicon sample is 4.9 ppmw. The average value of the solid silicon, 4.9 ppmw P, taken with the average value of the starting silicon, 11.9 ppmw P, demonstrates an effective refining ratio of 0.41, even at a 50% solid fraction. Performing a second UDS on silicon obtained from runs in Figure 1, yields [P] around 2 ppmw (Figure 2).In addition to P and B reduction, in this paper we also discuss the hardware designed to implement this process in commercial production in volumes exceeding 4,000 MT per year. MB Scientific, the original process developer, and NC Consulting, an engineering company, have developed a plant design that can produce SGS at an estimated cost that will allow for profitable large scale production, and have joined in a new company, Silicon Forge, to commercialize the large-scale production technology.

Type
Research Article
Copyright
Copyright © Materials Research Society 2010

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

[1] Dubois, S. et al., 23rd EU PVSCE Conference, 1-5 September 2008 Valencia, Spain.Google Scholar
[2] Ceccaroli, Bruno et al., Handbook of Photovoltaic Science and Engineering; edited by Luque, A. and Hegedus, S., John Wiley & Sons, Ltd. (2003), p. 177.Google Scholar
[3] Timminco Annual Report 2008, Value Proposition p. 12.Google Scholar