IBM muscles into CIGS solar-cell market

Add IBM to the hordes of companies trying to build a better solar cell.

The computing giant on Monday is expected to announce a deal with a Japanese semiconductor equipment manufacturer to make thin-film solar cells from CIGS, a combination of copper, indium, gallium, and selenide.

Neither IBM nor its partner, Tokyo Ohka Kogyo (TOK), plan to manufacture cells themselves. Rather, they will develop technology that can be licensed to solar companies in two or three years, said Supratik Guha, lead scientist for photovoltaics at IBM Research.

IBM has already built a prototype device. Once made at large volumes on a glass substrate, the cells are expected to deliver electricity at less than one dollar per watt at peak times--a long-held target of many solar outfits.

"We have the skills that we have developed in other areas--standard silicon semiconductors, materials chemistry--and we're looking to utilize those skills in the photovoltaic space and develop IP (intellectual property) and know-how that other people don't have," Guha said.

Traditional solar cells are made from silicon, but alternative thin-film materials are becoming a larger share of the market. Thin-film cells are less efficient at converting sunlight to electricity than silicon, but they require much less material to produce a cell, making them cost-competitive. Solar high-flier First Solar sells thin-film cells from cadmium telluride.

CIGS is a material that a number of companies are betting on, including Nanosolar, Global Solar Energy, Miasole, and Heliovolt.

These companies are not producing cells at large volumes yet, but use of CIGS is expected to catch on quickly next year as their factories come online.

"CIGS will be the big story of 2009 because we know how many companies are putting in multimegawatts of CIGS (production capacity) in 2009," predicted solar expert Travis Bradford, president of the Prometheus Institute, who spoke at a recent Greentech Media solar briefing.

15 percent efficiency goal
IBM's CIGS manufacturing technique came out of research IBM had done about 10 years ago in flexible electronics.

It's a break with the most common CIGS manufacturing process, called co-evaporation, in which active chemicals are immersed in a solution that gets removed in a vacuum.

IBM's "solution-based processing" calls for the chemicals to be dissolved in a liquid and then dried. It does not require a vacuum, doesn't require as much energy to run, and can be done faster than co-evaporation, Guha said.

IBM is also looking to leap-frog existing CIGS manufacturers on efficiency with a target of about 15 percent.

The efficiency of the CIGS cells on the market now is at about 9 percent or 10 percent. HelioVolt recently announced that it hit 12.2 percent efficiency with a process that is faster than co-evaporation. Global Solar said it expects to get to 14 percent, eventually.

The record for efficiency was done by the U.S. National Renewable Energy Research Laboratory (NREL) earlier this year, which reached 19.9 percent efficiency through a co-evaporation process.

In about a month, IBM intends to provide more technical detail of its solutions-based process in an advanced-material paper, Guha said.

CIGS is not the only area of the solar industry where IBM is investing.

Again by modifying chip-manufacturing technology, IBM created a cooling method for a solar concentrator that it hopes to license to others.

It is also doing work on techniques to manufacture silicon solar cells on glass using very little silicon, Guha said.

Rival Hewlett Packard, meanwhile, earlier this month licensed transparent electronics technology to Xtreme Energetics to make a solar concentrator.

[Marcus Westrup]

What I want to see, is technology that doesn't rely on rare materials like indium and gallium. The world supply of these metals is so limited, large scale production of solar cells, LED's, super conductors etc. will not be possible, and thus won't make a real difference to the green economy.

[Barry_]

There is a new solar technology that uses quantum dots which are not made from rare or toxic materials...soon to be on the market

[Remo_Williams]

Correct, reliance on iridium would be a mistake. I'd rather see more work in tech that effectively traps the light in nanotubes. The move should be to 35-50% efficiency of conversion so the process can become cost-effective at scale without rare metals.

-R

[mlamonica]

Cadmium (which I believe is mined with zinc) can also be a human toxin: http://en.wikipedia.org/wiki/Cadmium#Toxicity

[mlamonica]

I should make clear: CIGS does not use cadmium but it is used in other thin-film solar cells (ie Cadmium telluride)

[OneOfTheToolsInTheShed]

IBM doesn't have all that much to contribute here, frankly. I'm not dissing Big Blue in any way but while they undoubtedly know a lot about what a good multi-phase IAJ CIGS layer looks like on the charac. tools, there are solar-specialized companies out there with hundreds of man-years of deposition experience.

The problem is really NOT to make small-scale NREL-busting champion devices, but in scaling-up something that can do >15-16% to a >1m² substrate.

And just to correct Martin, the best CIGS cells DO use Cadmium, but only in very small quantities, for the CdS 'buffer' layer. You can avoid it if your really want to, but it generally is used.

[iosono5466]

Thank you