Speak of cutting the cost of solar cells and most people think of reducing the amount of silicon the cells include or increasing the amount of sunlight they can convert into electricity.

But technology advances related to the gases used to make the cells and panels - by far a more obscure approach - potentially could cut costs up to 13 cents per watt, said Anish Tolia, market development manager for solar for gas supplier Linde Electronics, at Intersolar North America last week.

That might seem like small change, but as companies begin building plants that can produce up to a whopping gigawatt of solar cells per year, those savings could amount to as much as $130 million per year, he said.

And that could give gases an important role in bringing costs down to the industry mecca of less than $1 per watt.

"The price of materials and gases in particular are so significant - much more than other industries - that [companies] can't reach those targets unless the price comes down," Tolia said.

The "gigafab," as he calls it, is an important trend because it brings the economies of scale to the point where they can significantly cut costs, he said.

The typical solar factory has grown by an order of magnitude in the last five years to produce an average of 100 to 200 megawatts, he said. And, in as little as another six months, the industry will make another order-of-magnitude leap to 1-gigawatt facilities, he said.

"From a delivery and logistics point of view, this is a big transition," he said.

Sharp Corp., for example, in November said it is building a 1-gigawatt thin-film plant in Japan and the Nikkei reported that Showa Shell Sekiyu, a subsidiary of Royal Dutch Shell, is building another 1-gigawatt thin-film plant (see Thin Films Lead U.S. Solar Production, Thin-Film Solar Production to Leap Forward and 9 Big Trends: Larger is Better).

Although Tolia said he doesn't believe all the press-release announcements will come to fruition, all together companies have announced plans to expand to 50 gigawatts of capacity by 2012.

A large part of the investment in expanding this capacity is going into thin films made of amorphous silicon, which Tolia called "by far the most easily scaleable technology out there."

"You can't really [reach gigawatts of capacity] with crystalline silicon because you won't get the silicon, and you can't do it with cadmium-telluride because you don't know how to do it - it will take 10 years because that's how long it took First Solar," he said. "So [thin-film using silicon] is what you would do if you had money but no particular know-how, and that's I think what we're seeing."

Amorphous-silicon companies face some of their own challenges, such as questions about efficiency, and haven't yet reached large volumes in spite of years of development.

Travis Bradford, president of the Prometheus Institute, a Greentech Media partner, expects amorphous-silicon films - such as those made by Applied Materials and Oerlikon - to stumble slightly in 2008 and 2009, as technologies get debugged and verified, before taking off in 2010 (see Thin-Film Solar Has Bright Future).

By 2012, the institute forecasts that amorphous silicon will make up the largest chunk of the thin-film pie with 4.5 gigawatts of production, based on the huge number of orders Applied Materials and Oerlikon already have announced.

Gases make up about 17 percent of the cost of silicon-based thin-film manufacturing, and two gases -- silane and nitrogen trifluoride -- make up the largest piece of the pie, followed by hydrogen, nitrogen and other gases, Tolia said.

A "gigafab" would require about 300 to 800 tons of silane - used to deposit the active layer in silicon-based thin film -- per year and 500 to 800 tons of nitrogen trifluoride -- used to clean the process chamber -- per year, Tolia said.

It would use 10,000 to 15,000 cubic meters of nitrogen per hour, and 1,000 to 4,000 cubic meters of hydrogen per hour, he said.

Silane is of particular interest, as some industry watchers have suggested that a shortage of the gas is on the way. Already announced thin-film production plans are expected to require more silane than is produced today, so gas companies such as Linde are taking steps to produce more, Tolia said.

Earlier this month, the Linde Group announced it would build a new silane plant, expected to begin production in early 2009, in partnership with Schmid Silicon Pilot Production.

The company also is working to recycle gases, including silane, to reduce material costs, Tolia said, adding that the technology to recycle silane will take two or more years of development and testing to be ready for commercial use.

Another advancement is more market-ready: the potential to replace the cleaning gas, nitrogen trifluoride, with fluoride.

After all, cleaning gas represents 30 to 40 percent of the total gases used on the line, Tolia said. And fluorine is the part of the molecule that does the cleaning, while the nitrogen has no effect, he said.

Fluorine would be cheaper and more efficient, and would clean up to three times faster, resulting in up to 10 percent more production from the same line, he said.  

It also would be a timely switch, as some scientists have theorized that nitrogen trifluoride could be contributing to global warming (see Los Angeles Times story).

"NF3 is not a popular molecule," Tolia said. 

But switching isn't quite as simple as it might seem. Transporting fluorine comes with safety issues and is "not trivial," he said.

Linde expects to avoid the issue by making the gas on the spot instead. Fluorine generators already are commercially available today and used in liquid crystal displays, for example.

The company is working with customers and manufacturers to test and improve the process for using fluorine for solar, and expects to bring fluorine to solar manufacturing within the next year.

"The solution exists and the cost impact is immediate," he said.

On-site hydrogen and nitrogen generators also exist now and are based on well-understood technology, he said.

The challenge is to match the supply of these gases with the demand as a large factory ramps up, he said. Customers want to start with a smaller generator, then grow into a larger one, and Linde is figuring out how it will manage these programs now.

The company also is developing new chemicals and additives, including a silane additive, to increase efficiency and throughput of manufacturing lines, as well as a way to replace costly helium, which is used for cooling. If helium can't be designed out of the process entirely, the well-understood technology to recycle helium also could be added to some factories, he said.

Tolia said the company has a long-term joint development with customers in Europe to test some of these more complicated technologies, with the hope of commercializing them "more in the two- to three-year horizon."

"Cleaning is the low-hanging fruit," he said. "Anything that touches the process will take longer and more study."

All together, all of the gas-related improvements could directly cut costs by 8 cents per watt, with a throughput impact - savings resulting from increased speed of production -- of 5 cents per watt, he said.

"Even if it's a few cents per watt, at the 1-gigawatt scale, we're talking about real money," he said. "We're ready to play our part."