Sunday, September 10, 2006

Pictures of Metallurgical Silicon Refining

This is a pictoral description of the typical industrial scale process used for metallurgical polysilicon production by Carbothermic Arc Reduction. Most common application of the output of arc reduction of silica to form polysilicon, is for aluminum and steel metal alloying feedstock.

But with care in the selection of the quartz silica and carbon raw material purity and in details of the furnace operation and materials postprocessing, this process is used for the feedstock for subsequent refining to produce solar grade polysilicon. It is apparent that Arc Reduction of silica, is an an energy hog, as one can observe in the furnace picture below.

Solar grade silicon, although presently capacity constrained, is not subject to exactly the same hard limits as electronics grade poly, especially since the logistics of capacity increases for solar grade are not nearly as arduous nor as capital intensive as electronics grade polysilicon capacity. Although solar grade silicon still requires unusual care in the materials production, it is vastly simpler than electronics grade polysilicon production.

This - carbothermic arc reduction of silica - is how the bulk of solar grade polysilicon used by Qcells - the worlds largest manufacturer of silicon photovoltaics, is produced and similarly solar grade polysilicon for the up and coming Evergreen Solar, a new strategic partner with Qcells .

As a process engineer, one can be curious as to how one might improve energy efficiency of this relatively old method of electrical arc chemical reduction of silica to form polysilicon might be improved upon ??? Evergereen is doing some of this, but are there even better ways??

One the face of it, one can observe that the repeated heating and cooling from the reduction remelting, followed by cooling, then pulverizing, then recasting (heating and cooling) then slicing and then diffusion doping with heat once again to make actual silicon solar cells, is not the best EROI one might envision in an optimized silica to solar cells process.

This is a picture of a massive CarboThermic Arc Reduction Furnace owned by Elkem Silicon of Norway. Input is silica quartz, oxygen and ultrapure carbon and graphite. Apparently there are a number of arc reduction furnaces of comparable size in Elkem's inventory.

The quartz silica feedstock for Elkem silicon's efforts must be unusually pure to permit the use of the relatively crude metallurgical grade arc reduction process to be successfully used for subsequent refining of solar grade silicon. Elkem is a major provider of metallurgical grade silicon to the PV and electronics industries, but it is all further refined by companies like REC, Wacker, and Hemlock before moving on to electronic products.

Q-Cells has purchased its silicon primarily from REC. Evergreen and EverQ also purchase most, if not all, of their silicon from REC (though Evergreen was buying from MEMC until last spring, when MEMC pulled the rug out from under them). Considering that REC holds a 15% stake in EverQ, this is likely to be a durable supply relationship.

The best in the world for ultrapure mined silica feedstock suitable for solar grade polysilicon production seems to be mined at this very beautiful location on the Tysfjord in Norway. The mining operation, with its low cost chemical purification of the raw mineral silica feedstock, is impressive for its attention to detail and the purity of the large deposit of excellent quality quartz pegmatite.

This page shows some interesting pictures of the mine and the manufacturing of the solar grade silica powder. It does not describe methods to purify the powder, which despite the excellent mineralogy of the TysFjord Pegmatite Silica ore, is likely to be chemically treated to optimze the reduction of impurities needed for best in class solar grade polysilicon, at very modest cost increase.

One can envision that this mine provides most of the raw material for Elkem's furnaces, which then ships massive quantities of polysilicon to silicon refiners for solar grade and electronics aspplications.

NOTE updated with input from ALAN on Sept 11 2006

Addendum Dec 9th 2006 - Here are some links to the vistas of the Tysfjord - mountains, maps, and unusual pictures of Orca killer whales inhabiting the fjord itself. Beautiful amazing pictures, not far from the quartz pegmatite mines.

a suspended walkway over a stream

pictures from an Orca whale watching tour ship

amateur photos from a snorkeling tour of the fjord

more of these tour pics with incredible closeups of Orcas in the Tysfjord

maps and more local pictures of a fishing village in summer and winter


Blogger clem clements said...

Thanks for this closeup of the quartz process. I was in Norway this summer and noticed several open-pit mines, but never knew about this underground mine.

BTW, a "solar cell biz trip" to Norway would be pretty fun all-around...

8:14 PM  
Blogger Mark Wendman said...

I was fascinated as I found out about this, as it is apparent how large an industrial endeavor it seems to be, and how mostly it seems to feed only a small number of factories in other EU countries. The process picture is likely showing configuration of a true metallurgical grade silicon furnace, and the photograph seems to have a different furnace set up at first glance. The mine set on the fjord - that is quite picturesque...

12:53 AM  
Anonymous Alison Chaiken said...

one can be curious as to how one might improve energy efficiency of this relatively old method of electrical arc chemical reduction

By figuring out which impurities need to be removed and measuring their diminishing concentrations in real-time instead of over-purifying to make certain they're gone? Of course, how could you measure their ppm concentrations without accelerator mass spectrometry? I wonder what concentration can be detected by flame-induced fluorescence. Depends on the element for sure.

6:53 PM  
Anonymous Anonymous said...

I'm writing a paper on the photovoltaics industry and I'm so glad I found your blog because I've been having trouble finding out where the quartz was being mined for solar cells. Do you have any more information about quartz pegmatite mining? What sources did you use in researching how solar grade silicon is produced? Thanks!

11:20 AM  
Anonymous Anonymous said...

Posted incorrectly via the MRSA article were two attempts at comment posting apparently from the same fellow in NYC.

#1 -
to me

Jun 14 (17 hours ago)
Anonymous has left a new comment on your post "Possible Light at the End of the MRSA tunnel ?":

"You have pictures and describtion of the elkem solor silicon furnance. Elkem invested over 1 billion in that plant. And is quality or capacity will not be known until at least 1 year after it starts up. Dow Corning's also failed. has a series of articles about a UMG scam company called Timminco. its a Canadian penny stock that is being promoted by media that doesnt know UMG tech or the background of the individuals that fromed it. is run by manuel asensio who was the first to reported on several very similiar scams.

Posted by Anonymous to WENDMANs VIEWS on NANOTECH at 8:41 PM

#2 -
Anonymous to me

show details Jun 14 (17 hours ago)
Anonymous has left a new comment on your post "Possible Light at the End of the MRSA tunnel ?":

Pat Ryan of Ryan Notes, the silicon industry's primary market price and size reporter, has not found a single source of UMG. What is it with Elkem? they may be real, or they may fail. but there are dozens of scams being promoted like Timminco Limited. can you comment?

Posted by Anonymous to WENDMANs VIEWS on NANOTECH at 8:44 PM

2:11 PM  
Blogger Mark Wendman said...

To the anonymous fellow in NYC who was posting on his take on UMG ( Upgraded Metallurgical Grade ) processes for solar polysilicon feedstock.

1 It is indeed true several firms have tried to develop lower cost Upgraded Metallurgical Grade Silicon, and not succeeded.

2 It is false to assume because of prior failures in large investments from well funded firms, that success in UMG process development is impossible.

3 You claim Timminco's effort for UMG Silicon is, as you call it a scam, and that Manuel Asenio understands this. I will not answer this, as I am merely a technical person, and do not trade in Timminco.

Personally, I think that Timminco's technology advance is real, and has likely succeeded in developing a successful UMG Solar Grade Silicon purification process, not requiring costly gas phase distillation, for solar cell uses.

I do not have first hand knowledge of the apparent success, nor the specifics of their technology, but there is evidence in the patent literature that UMG via metallurgical methods of purification can work in this application. It is not easy in the slightest, that is for certain.

It is apparent that for UMG to succeed, very careful process controls must be used to ensure both in development phase and manufacturing scale up that purification adequate for solar cells is attained. It is indeed tricky to succeed at, but Timminco has run pilot scale processing for about a year or more prior to get qualified by approximately 4 long term customers, using output from a pilot plant.

Q-Cells - the worlds largest manufacturer of silicon solar cells has committed to a long term purchase contract with Timminco apparently, the details I do not know beyond the press release from Timminco.

I think it wrong to bet against the possibility that UMG from Timminco will succeed. I personally do not know details first hand, but it looks like the work was fairly credible and methodical. Not a 100% guarantee of success, but not a good idea to bet against an innovation that has considerable long term contract dollars closed, and plants being built, and the most prestigious, largest solar cell manufacturer signed as a customer. I'd trust that Q-Cells did a decent analysis and sample testing of the Timminco process, but I do not know for certain. Q-Cells is a very large and successful firm, so I'd tend to trust them more than a stock speculator such as Mr. Asenio.

I'd place my money on innovation when the work is credible, as it seems with Timminco, but again, I do not have first hand knowledge of Timminco's technology.

2:34 PM  
Anonymous Jim Thomas said...

Jim Thomas said...

Thanks for the info on these quartz product. I am interested to find out what does the purity of the raw quartz need to be to make this poly silicon?

12:32 PM  

Post a Comment

Links to this post:

Create a Link

<< Home

Enter your email address:

Delivered by FeedBurner

Subscribe to Wendmans Views on Nanotech by Email