By Jennifer Kho

SolarWorld on Thursday said it established a joint venture with Scheuten Solarholding to turn dirty metallurgical-grade silicon into high-purity solar-grade silicon.

The company claims the announcement marks the first time in Germany that purification of metallurgical silicon for solar power is being implemented on an industrial scale.

But while a number of companies are pursing the purification of metallurgical-grade silicon for solar, at least one analyst says it doesn’t make sense in light of the large amounts of new polysilicon expected to hit the market in the next few years.

“There’s no reason to go to metallurgical silicon,” said Jesse Pichel, a vice president and senior research analyst of technology at Piper Jaffray. 

SolarWorld obviously thinks otherwise, as do Dow Corning, Elkem, JFE Steel, and smaller companies Solarvalue, CaliSolar, and Citizenre.

SolarWorld and Scheuter each hold a 50-percent stake in the joint venture, called Scheuten SolarWorld Solizium. (“Solizium” is probably a cross between the word “solar” and “silizium,” the German word for silicon.)  

According to SolarWorld, the joint venture will develop and build a manufacturing plant to produce, initially, 1,000 tons of solar-grade silicon from metallurgical-grade silicon. The Bonn-based company didn’t say when it expects the plant to come online, but did say the joint venture is “future-oriented.”

SolarWorld bought a 20,000-square-meter plot of land, including an administrative building and a lab, near its Freiberg technology facilities. In other recent news, Dow Corning announced in September it developed a process to blend metallurgical- and solar-grade silicon without losing efficiency, Solarvalue in Berlin said last month it is converting its metallurgical plant to turn metallurgical silicon into solar silicon, and solar startup Citizenre CEO David Gregg said last week the company also plans to take that route.

The Wilmington, Delaware-based company plans to install residential solar systems at no upfront cost, instead charging customers for the electricity at 2005 rates. Mr. Gregg said the company plans to make its own solar modules, and to get the silicon by purifying metallurgical silicon into solar silicon. 

The reason for all this activity is a worldwide shortage of solar-grade silicon that is constraining silicon supply and raising the price (see Solar’s Silicon Shakeup).

According to Michael Rogol, managing director for Photon Consulting, the average of global prices for solar-grade silicon has grown from $35 per kilogram in 2004 to $74 per kilogram this year, and is expected to reach $95 per kilogram next year.

Those prices finally make metallurgical-silicon purification make sense, Mr. Gregg said. “With long-term contract prices for solar silicon still $80 to $90 a kilogram, and spot prices more than double that, it’s not too enticing to try to negotiate for those contracts,” he said. “We absolutely feel this would be cheaper.” 

But not everyone thinks so. Mr. Pichel said he doesn’t think metallurgical silicon will ever work.

“The quality is worse than scrap,” he said. “The semiconductor industry produces an abundance of scrap used by the solar industry, and I think that’s probably more interesting than metallurgical silicon.”  

Metallurgical silicon must be mixed with high-purity silicon, called polysilicon, and it lowers the efficiency of a solar cell, in turn lowering companies’ revenue opportunity, he said. “That’s just not what solar companies want to do,” he said.

Also, the silicon shortage is only a temporary problem, with a “great amount” of polysilicon coming online in 2008 and 2009, he said. 

It’s true that plenty of solar-grade silicon producers have announced they are expanding capacity, and analysts expect the shortage will ease in the next few years.

REC, Wacker Chemie, and Joint Solar Silicon—another SolarWorld joint venture, this one with Degussa—have also announced big expansions in the last few months, among others.

Mr. Pichel said it doesn’t make sense to put resources into developing this new metallurgical technology when there’s nothing wrong with the current method.

“I don’t think you have to reinvent the wheel,” he said. “Polysilicon only costs $25 a kilo to make, and that’s plenty cheap. The problem is not enough capacity. So to try to put a different technology approach to silicon in kind of stupid when all you need to do is put up more polysilicon plants.”

More interesting than a new metallurgical refining process, he said, is next-generation polysilicon technology that promises to make polysilicon for $15 to $20 a kilo.  

Still, he understands the attraction of metallurgical silicon, he said.

“It’s a very seductive technology because the hope is lower-cost silicon,” he said. “But it’s like fool’s gold.”Contact the Writer: jkho@redherring.com