It sounds to me like there's a dire need for some new plants to make poly-crystalline silicon. The question must be asked, is this lack of capacity a willful result of the industry to maintain low stocks of this in order to drive prices up considerably for a significant profit? I wish this weren't about greed, but it sure smells like it.

Supply and Demand require a time delay. Let us say that someone makes 100 units of sprockets (ah Jetsons) a month, and it takes 1 year to open another factory that will also make 100 units/month.

At the previous demand level (100 units/month demanded) one factory was enough. Suddenly everyone wants more sprockets (1000/month) and the price-elasticity of demand is perfectly inelastic (regardless of the price increases, people will still want those sprockets). After a month the factory owner notices the empty shelves, and temporarily raises prices, hoping to get some extra cash he can then invest in new factories. The price climbs, the factory owner makes beaucoup dollars and spends the extra to open new factories. He can start building a new one each month, as the extra funds come in from the rising prices.

It took the owner a month to notice, a month to get the extra cash, and it will take 1 year for the first of the factories to come online, and 9 months after that until enough factories are there to meet current demand. Capitalism just took almost 2 years to "react" even if people aren't asleep at the wheel, there aren't any raw material shortages, etc.

This was only for one stage of production. Let's say that there are other components that are necessary; they may not see the bidding war coming until after more sprockets come online, causing demand for their merchandise. So this whole above cycle happens again (but two years later) before the consumers see meaningful increases in the available quantity of sprocket-using goods.

Now owner A and B may discuss their situations, allowing them to increase production in concert (via a loan B takes out or something), but you begin to see the problems; the market does not turn on a dime.

Here are a couple of older but related references:

- Cheaper Silicon Found Effective for Solar Cells - August 14, 2005

- Solar-cell makers fear silicon shortage - April 14, 2005

- Nanotech offers solutions to solar challenges - based on the IEEE San Francisco Bay Area Nanotechnology Council May 16, 2006 event

I doubt it is some conspiracy to raise prices. As far as I knew the PV industry lived off the semi-conductor industries scraps. Now that PV's are getting more popular they need to find their own silicon.

Does this shortage affect amorphous silicon too? It doesn't sound like it. Amorphous silicon solar cells may not get the peak power density of poly-crystalline cells, but they often have higher average power production. I've believed for several years now that amorphous silicon panels are the way of the future. Shortages of the other kind reinforce that hunch.

Wow, who knew it was so hard to find sand.

C'mon, just build more plants already.
=== author's response follows ====
It might be that the capital equipment is so expensive for these factories that banks are reluctant to loan for their construction, given the changes in technology that could come from amorphous and other forms of SPV panel.

There is no conspiracy to raise prices. It takes 18-24 months to plan and build a new feedstock plant, and considerably less to plan and build a solar cell plant. That makes it pretty easy to have more solar cell manufacturing capacity than silicon refining capacity. Not to mention that the last time the silicon feedstock industry made a major expansion it got burned badly by a downturn in the microchip industry. This time around, they refused to expand until the PV industry agreed to take on some of the financial risk of expansion. The PV industry dragged its feet, plus the shortage hit about a year earlier than predicted, resulting in the mess you see now.

What the article is saying, in essence, is that new feedstock plants are being built, but not quickly enough to meet demand.

On the bright side, the shortage has given PV makers a kick in the pants when it comes to designing products that use less silicon. They're progressing faster than a lot of people would have expected, and just how thin the cells can get is sort of a wildcard in determining when feedstock supply will catch up with demand.

The shortage does not really affect amorphous silicon, even though the gases used for amorphous silicon come from a similar product stream as those used to produce crystalline silicon feedstock. It wouldn't surprise me to learn their material costs had gone up, but capital costs are the primary driver amorphous silicon PV manufacturing costs. Not so with crystalline silicon, where material costs are the primary driver, and silicon is the single largest material expenditure.

As for the price of microchips, I wouldn't worry too much about that. They use, at most, a few cents worth of silicon (remember, you can put hundreds, or even thousands, of chips on a single wafer -- but only one solar cell). A 500% increase in silicon prices is inconsequential where microprocessor prices are concerned.

"The thinking behind concentrated solar power is simple. Because energy from the sun, although abundant, is diffuse, generating one gigawatt of power (the size of a typical utility-scale plant) using traditional photovoltaics requires a four-square-mile area of silicon, says Jerry Olson, a research scientist at the National Renewable Energy Laboratory, in Golden, CO. A concentrator system, he says, would replace most of the silicon with plastic or glass lenses or metal reflectors, requiring only as much semiconductor material as it would take to cover an area the size of a typical backyard. And because decreasing the amount of semiconductor needed makes it affordable to use much more efficient types of solar cells, the total footprint of the plant, including the reflectors or lenses, would be only two to two-and-a-half square miles."

http://www.technologyreview.com/Energy/17774/page1/

I seem to be seeing more about non-silicon PV. How about an article on that?

Dyesol (titanium-dye solar) and Konarka (plastic-dye solar) are good examples of non-silicon PV (treehugger stories on both.) So far these have not made the same impact due to lower efficiencies. This PV tech is primarily focus on mobile (military) or micro (home) markets.

Silicon Shortages