Solar power about to become cheaper, maybe.

Here is the text of a news article in this month's Advanced Materials & Processes. Since this magazine is not sold in bookstores and is only available to ASM members, I will pass it on here in it's totality in the hope that the powers that be over at ASM view this as "fair use".

Cheap, low-grade silicon may reduce cost of solar cells
A new technique to handle metal defects in low-grade silicon has been announced by a research team at the University of California, Berkeley. The advance could dramatically reduce the cost of solar cells. Nearly 90% of solar, or photovoltaic, cells in the world are made from a refined, highly purified form of silicon, the same material in integrated circuits. The more abundant and cheaper form of silicon is laden with metal impurities and defects, and solar cells made from this material do not perform as well. Techniques that remove impurities are expensive, negating the cost benefits of the cheaper material.
The researchers found that they were able to manipulate the distribution of the metal impurities by varying the cooling rate of the silicon. When the material is cooled quickly, the metal defects are quickly locked in a scattered distribution. By simply slowing down the cooling rate, the metal impurities diffused into large clusters.
“Using this cooling technique, we were able to improve the distance electrons could travel by a factor of four compared with dirty silicon that had been left unaltered,” said Prof. Anthony Buonassisi. “Although this is still not as efficient as ultrapure silicon, it is the proof of principle that poor-quality silicon can be easily improved. We are now looking at other techniques that could further enhance the efficiency of dirty silicon.”

For more information: Prof. Anthony Buonassisi, University of
California Berkeley, CA 94720; e-mail: buonassi@berkeley.edu; Web site: www.berkeley.edu.

What this is saying is that there are two types of PV cells:
Crystalline (expensive but more efficient) and Amorphous (cheap, but very low efficiency)

This article was written by and for Materials Scientists/Metallurgists, so there are things that are left unsaid that most people won't pick up on. As Silicon cools, it naturally wants to form crystals. Since the impurities generally don't have the proper size or electron configuration to fit into the crystal lattice properly, they act to impede electron flow. Normally, the crystal building process, if allowed to continue slowly will tend to pick up atoms that fit in the lattice correctly, but in the processing of amorphous silicon, it is cooled too quickly for that to happen. If allowed to cool slowly, the impurites will naturally tend to be concentrated in the liquidus phase between the crystals as the silicon crystals grow. Eventually they will get trapped and freeze in the pockets between the crystals. By segregating the impurities, the reserchers are basically bypassing the bottlenecks that impede electron flow, raising efficiency. The big question to me is, is the added cost associated with this processing worth the efficiency gains? That is yet to be seen.