SEPTEMBER 6, 2004

Another Dawn For Solar Power

Tech breakthroughs and high energy prices are rekindling the industry

By Otis Port in New York

On a sunny june day in 1979, President Jimmy Carter held the first and only press conference on the White House roof. Atop the West Wing, he unveiled a $28,000 solar cell system that captured the sun's energy to provide hot water for the White House. He also launched a sweeping drive aimed at harnessing the sun, the wind, and other renewable resources to generate 20% of America's electricity by 2000.

It didn't happen, of course. The share of electricity produced by solar cell technology in the U.S. last year was a mere 0.07%. Carter's solar water-heating system was removed in 1986 so a leak in the roof could be fixed. The solar panels were supposed to be reinstalled but they never were.

Now, several startups aim to revive Carter's dream of harvesting the sun's rays. Even the optimists admit it will take years to build enough solar energy capacity to make a dent in coal and natural-gas consumption. But with better technology, new solar systems could start plugging the power supply gap during periods of peak demand in some regions by 2007.

There are two basic concepts for tapping the sun's energy: collect its heat or convert its light. The solar-thermal approach uses mirrors to reflect the heat energy from a large area onto a small space, such as a pipe filled with a fluid like molten salt. Once the fluid's temperature has been raised to hundreds of degrees, it can be used to boil water and produce steam for a conventional generator.

With the other approach, called photovoltaics, a semiconductor -- typically silicon -- absorbs the photons streaming from the sun and reacts by giving off a flow of electrons, or electricity.

HOT OVERSEAS

Both techniques have been around for ages, but they have made little headway in the U.S. because of cheap fossil fuels. In Japan and Europe, where such fuels aren't so cheap, unit sales of solar energy systems have been mushrooming 35% annually since the mid-1990s. Now, the combination of rising fuel costs and engineering advances that boost the efficiency of solar power systems are stimulating fresh U.S. investment.

The solar-thermal strategy is roughly 30% efficient at turning the sun's heat into electricity -- about double the efficiency of photovoltaics. As a result, the thermal technique enjoys a pricing advantage. The giant solar-dish mirrors designed by Stirling Energy Systems Inc. could generate electricity for less than 8 cents per kilowatt hour (kwh) -- maybe even 6 cents, asserts David J. Slawson, CEO and founder of the Phoenix startup. Unlike photovoltaic panels, though, solar dishes aren't practical for homes. To work effectively, they need to be large -- too big to plant on a roof or even in a backyard. Stirling Energy's dishes are 38 feet across and 40 feet tall and generate 25 kilowatts, or enough juice for five or more homes.

The dishes are based on technology pioneered at the Solar Two "farm" in California's Mojave Desert, sponsored by the Energy Dept. Mothballed in 1999, Solar Two used 1,800 solar dishes in concentric arcs to reflect the sun's heat onto a central "power tower." That layout required a massive $150 million investment for 10 megawatts of capacity -- or $15 per watt. To reduce costs with a modular approach, Slawson has scrapped the tower. Instead, he mounts a miniature generator at the focal point of each dish.

When these power dishes enter volume production, expected around the end of 2006, Slawson predicts costs will tumble 90%, to $25,000 per 25-kilowatt dish. That would put the capital cost of a 10-Mw plant at $10 million, or $1 per watt. But Arizona Public Service Co., which is under a state mandate to generate 1.1% of its electricity through renewable resources by 2007, isn't waiting. It will install 10 dishes next year. And utilities in Nevada and California are haggling for 40.

On the photovoltaics front, upstart Solaicx Inc. in Los Gatos, Calif., predicts that residential and commercial solar panels made with its silicon material will soon compete with conventional fossil-fuel generators in markets where electricity costs at least 10 cents per kwh. But for that to happen, the capital cost of solar cell systems needs to reach the same magic number of $1 per watt of generating capacity. That up-front investment, along with operating efficiency and equipment depreciation, determines the price at which kilowatt-hours of output can be sold. Today, the installed cost of high-efficiency silicon solar panels starts at $3 per watt. "Our customers should get to $1 a watt by 2007," declares Robert S. Ford, CEO of Solaicx. Around 2010, the Energy Dept. expects improved solar cells to generate electricity at 6 cents per kwh.

CHEAPER CELLS

Solaicx won't begin shipping silicon from its new factory until October, but Ford claims that two top producers of solar panels -- he won't say which ones -- have already signed up to buy the silicon platters. So has SunPower Corp. in nearby Sunnyvale, Calif. A solar panel typically has 36 solar cells carved from compact-disk-size silicon wafers. Solaicx will churn out the wafers with equipment similar to that used to make raw silicon for the semiconductor industry -- but tailored from scratch for high-volume production. "It's just good old-fashioned silicon done right," says John T. Sedgwick, co-founder of Solaicx.

Solaicx uses two tricks to produce better solar cells: make the silicon wafers 40% thinner and thus cheaper, and increase something called carrier lifetime. When photons from the sun hit the surface of a solar cell, they dislodge electrons from silicon atoms. The longer these electrons remain free -- the carrier lifetime, measured in milliseconds -- the greater the chance that they will stream off the solar cell "and into your toaster or light bulb," says J. William Yerkes, chief technology officer of Solaicx and a photovoltaics pioneer for three decades.

With today's silicon, roughly 16 of every 100 bumped-off electrons make it out into the real world, Yerkes says. That means the cells are 16% efficient at turning sunlight into electricity. By boosting carrier lifetime, Solaicx' lower-cost wafers have yielded prototype solar cells that are 21% efficient.

Silicon-based solar cells represent 92% of today's photovoltaics market, but some producers are betting on different, much cheaper materials for the future. One is a conductive polymer salted with nano-size carbon molecules dubbed buckyballs. Last year, researchers at Germany's Siemens Solar Group (now Shell Solar) applied a very thin coating of the buckyball mixture on a plastic film and produced a solar cell that topped 5% efficiency -- the best yet for an organic solar cell. The cost of these solar cells would be a fraction that of their silicon cousins, so a homeowner could buy five or ten times the surface area and still save money.

Despite its modest cost, the low output may limit initial applications to fold-up recharging pads for laptop computers and other portable gadgets. But the researchers are confident that efficiency can be hiked to 7%, or perhaps 10%. Then the solar coating applied to a roof could supply all the required electricity to a home.

Turning rooftops into power plants is also a focus of newcomers on opposite coasts. Nanosys and Nanosolar, both in Palo Alto, Calif., and Konarka Technologies in Lowell, Mass., are developing liquid-plastic compounds that can be applied to most surfaces. But instead of buckyballs, they spice the mixture with nano-size semiconducting wires or pinheads called quantum dots.

Japanese giant Matsushita Electric Industrial Co. (MC) has teamed up with Nanosys to develop solar coatings that could be painted on roofs and walls. Commercial products are still a couple of years off. Konarka echoes that timetable for its solar roof technology, but recharging pads and solar energy coatings for Army tents could be ready next year.

A Midwest company, Energy Conversion Devices Inc. (ENER)'s United Solar Ovonic unit in Auburn Hills, Mich., grabbed the early lead in solar roofs. Since 1997, it has been applying silicon coatings on roofing materials. Its solar film can also be supplied as peel-and-stick rolls, and these will be used on the roof of Beijing New Capital Museum in China.

Thanks to the revolution in materials science wrought by nanotechnology and conducting plastics, organic solar coatings are also under development at such industrial stalwarts as General Electric Co. (GE) and IBM (IBM). And organic solar cells are the focus of a new 4.6 million euro, 30-month research effort in Europe. Called Molycell, its goal parallels what Solaicx and Stirling Energy promise: solar power systems with a capital cost of $1 per watt. France's atomic energy agency heads the Molycell team. Konarka, Siemens, and a half-dozen universities and energy research centers are participating.

While the U.S. is still a major player in solar research, it has fallen behind in reaping profits from solar cells. Japan is way ahead of every other country. In 2001, its annual capacity was nearly four times that of America's 167.8 Mwp -- the p means peak, or no cloudy days -- and Germany was a solid No. 2, with 260.6 Mwp. Last year, Japan generated half of all the world's solar power, built 44% of all new solar energy equipment, and installed five time as much new solar power capacity as the U.S. One company, Sharp Corp. (SHCAY ), accounted for 27% of all new solar panels, according to market researcher PV Energy Systems Inc. in Warrenton, Va.

Worldwide, unit shipments of solar systems have jumped 35% or more for each of the last eight years. Include installation costs, and market researcher Clean Edge Inc. pegs solar industry sales at $4.7 billion in 2003. For 2013, the Oakland (Calif.) company sees solar topping $30.8 billion.

However, that projection is based on installed per-watt costs remaining way above $1. If Solaicx, Stirling Energy, and Molycell deliver solar generation for $1 a watt by 2007, the business of harnessing the sun's energy could go supernova.