Power Revolution

Thanks to Silicon Valley's money and ideas, solar and other alternative technologies finally pay off


Although solar PV arrays have been crowning more and more American buildings—the 1.6-megawatt project opened this year at Google's headquarters in Mountain View, Calif., is the largest—they're expensive. Developments are underway to bring down the price by reducing the silicon from its usual wafer form to an ultrathin film deposited on glass. But at this point, PV cost estimates span from an uncompetitive 23 to 32 cents per kilowatt-hour, while residential electricity prices in this country range from 5.8 to 16.7 cents.

With current technology, "concentrating solar power" would cost about 40 percent less than PV—tantalizingly close to competitive in areas like California with high energy prices. Exactly how the mirrors will be configured could bring the cost down more. Rows of curved reflectors work well; since the 1980s, a dazzling "parabolic trough" display has provided reliable power to California, the only operating concentrating solar power project in the country. But Spanish firm Abengoa this summer aimed for greater efficiency by focusing circles of mirrors onto a central "power tower" near Seville for the first commercial European Union CSP plant. The company would like to do the same in the U.S. Southwest. One firm that has been working with PG&E is Ausra, which recently relocated from Sydney to Silicon Valley and received $40 million in backing from investors led by Khosla Ventures to promote its big idea. Ausra thinks it can drive CSP costs down with simpler, flat mirrors and turbines that run at a cooler temperature, like those in nuclear power plants. Ausra also aims to engineer into the system the ability to store thermal energy so it can provide power when the sun stops shining.

Engineering firm Black & Veatch of Overland Park, Kan., estimates that outside of environmentally sensitive areas, there's enough available flat "high solar resource" land—also known as desert—in California to provide six times the power that the Golden State uses today.

Deep geothermal

One of the most promising renewable-energy wellsprings is underground.

Geothermal is lower profile than a range of other alternative energy technologies, even though many homes—including President Bush's Crawford ranch—have heat pumps that tap into the Earth's steady, reliable warmth. But few realize that the United States is the world's biggest mass producer of geothermal power, with long-running plants in western hot springs and geyser areas that generate more electricity than all U.S. wind and solar energy combined. It was long thought that big-scale geothermal had reached its natural limit. Few locales are graced with steamy water reservoirs close enough to the surface (less than 2 miles under) to be easily tapped to run electric turbines on the ground above. And many hot spots happen to be beneath scenic treasures like Yellowstone National Park or on American Indian reservations.

But recent study shows deep-drilling and seismic-exploration techniques developed in the oil industry could be exploited to draw out the geothermal energy found 3 or more miles underground, locked in dry rock that's more than 300 degrees Fahrenheit. A developer could drill a well and use high-pressure water to open fractures in the rock. Then, injection wells would be drilled to circulate the water in the man-made reservoir and extract steam to the surface to run electric turbines. This year, a government-sponsored study led by Massachusetts Institute of Technology concluded that these "heat mining" methods could offer access to a staggering amount of energy. Just 2 percent of the U.S. geothermal resource base could yield nearly 2,000 times the power that the nation now consumes each year.

The limiting factor is the cost. The old-fashioned geothermal sites now operating in geyser areas aren't any more pricey than old coal plants, at about 3.5 cents per kilowatt-hour. MIT looked at a half-dozen potential enhanced geothermal sites across the country and came up with estimates ranging from a potentially competitive 10.3 cents per kilowatt-hour to a sky-high $1.05 per kilowatt-hour. However, if a few well-defined technical problems were tackled to boost what's known as the "fluid production rate," MIT said costs would plummet to 3.6 to 9.2 cents per kilowatt-hour.