This is the second of two articles by David Rotman on technology and the federal stimulus package. The first, “Can Technology Save the Economy?,” appeared in the May/June 2009 issue and examined the economic consequences of the U.S. government’s plans to spend $100 billion on technology.

The abandoned industrial site on the far edge of Chicago’s South Side is an unlikely location for a large solar power plant. For one thing, Chicago is not a very sunny city. And the land itself, once a center of postwar manufacturing, has been vacant for 35 years and is now overgrown with trees and bushes, surrounded by a gritty neighborhood of aging houses. But Exelon, one of the country’s largest electric utilities, says that by the end of the year it hopes to turn a 39-acre lot into the nation’s largest urban solar plant. If it succeeds, row after row of nearly 33,000 silicon solar panels built and installed by SunPower, a photo­voltaics manufacturer based in San Jose, CA, will cover the lot to produce 10 megawatts of power–enough for about 1,200 to 1,500 homes.

But there is a big if in this scenario of urban transformation. It will happen only if Exelon receives the generous loan guarantees for renewable-energy projects promised in this year’s federal stimulus bill–funds that in this case would cover 80 percent of the project’s costs. Barely viable with the loan guarantees and a handful of other federal and state subsidies, the $60 million solar plant would not be possible without such government support. Speaking from the 48th-floor offices of Exelon, nearly 20 miles away in downtown Chicago, Thomas O’Neill, the utility’s senior vice president for new business development, is blunt about the economics of the solar plant. “If we can’t secure the loan guarantee, we can’t go forward with the project,” he says.

Even with the federal subsidies, says O’Neill, the solar plant won’t offer the double-digit returns usually required by investors in large energy projects. It would cost $6 a watt to build, whereas wind and natural-gas plants cost roughly $2 a watt and $1 a watt, respectively. And its 10 megawatts will contribute an insignificant amount of electricity to Exelon’s vast generation capacity of 36,000 megawatts. But, says O’Neill, the project is “tailor-made” for some of President Obama’s goals in the stimulus package. It would create jobs (250 people would be needed to construct it), and it would demonstrate that solar power can be “brought to the Midwest and to the inner city.”

The proposed plant in Chicago is just one of the many renewable-­energy projects that could get built because of the federal stimulus bill passed in mid-February (see “Can Technology Save the Economy?” May/June 2009). The U.S. Department of Energy is still in the process of choosing the projects that will receive loans and deciding how other newly available subsidies will be spent. But the potential windfall is already jump-starting plans for wind farms in the Midwest, massive solar plants in the deserts of southwest Nevada and southeastern California, and geothermal power plants in the Northwest. According to a recent analysis by the Energy Information Administration, an independent agency within the DOE, the stimulus bill will increase the amount of generating capacity from renewable sources to 156 gigawatts in 2015, up from 114 gigawatts today; renewable capacity would increase only to 118 gigawatts without the legislation.

The EIA report also points to a troubling reality, however: this increased use of renewable energy will have only a slight long-term effect on carbon dioxide emissions (see “Powering Up”). Even 156 gigawatts would satisfy only a small fraction of U.S. energy needs. And as Secretary of Energy Steven Chu has frequently argued, existing renewable-energy technologies cannot provide the large amounts of cost-competitive energy required to significantly reduce the country’s reliance on greenhouse-gas-emitting fossil fuels.

Both the research funding and the subsidies aimed at existing technologies could be particularly critical for the solar industry. A number of physicists and chemists argue that finding more efficient ways to use the sun’s energy offers the only feasible long-term option for replacing fossil fuels and significantly decreasing production of greenhouse gases. “We’re bathed in these quantum particles that rain down on us from the sun, each of them carrying about two electron-volts of energy,” says Paul Alivisatos, interim director of the Lawrence Berkeley Laboratory and head of its solar research center. “That’s where the energy is.” But solar power now accounts for a fraction of 1 percent of the total U.S. electricity capacity of 1,000 gigawatts. The main reason is cost.

Silicon cells like the ones Exelon would use, which are made from the type of high-grade silicon used in computer chips, represent the vast majority of installed photovoltaic capacity but are still about five times too expensive to compete with conventional sources of electricity. Newer types of solar cells that replace single-­crystal silicon with thin films of semiconducting materials could be cheaper to make but are less efficient. Concentrated solar thermal power, in which large arrays of mirrors are used to collect sunlight and create steam that drives turbines, could come closer to fossil fuels in cost, but the facilities are expensive to build and require large areas of land in extremely sunny spots. In fact, no existing solar technology is currently competitive without help from government subsidies. That means the fate of solar power is especially vulnerable to the vagaries of government policy and the choices of those who make it.

The Sun King
Arnold Goldman knows how profoundly state and federal energy decisions affect the solar-power industry. In the early 1980s, his company, Luz International, built nine large solar thermal plants, with a combined capacity of 354 megawatts, in the middle of California’s Mojave Desert. At the time, the Luz facilities supplied 90 percent of the world’s solar-generated electricity. The technology they used was based on an ingenious design in which hundreds of thousands of mirrors, spread out over the ground, concentrate sunlight on a network of overhead tubes containing a synthetic oil; the hot oil heats water to create steam that then drives turbines to generate electricity.

The solar facilities, the first of which came online in 1984, were economically possible because of generous incentives from both the federal and state governments. In 1979, President Carter had set a goal that 20 percent of U.S. electricity should come from renewable energy by 2000 (today the figure is still only about 2 percent). Carter and Congress passed hefty tax credits for investors in renewable-energy projects, and a federal law called the Public Utility Regulatory Policies Act, passed in 1978, offered further incentives to producers of alternative energy.

Then, in late 1990 and early 1991, it all collapsed, recalls ­Goldman. The tax credits put in place by the Carter administration had “deteriorated” during the presidency of Ronald ­Reagan, he says. But the final blow came from a seemingly esoteric change to California’s tax code. The state had exempted renewable-energy facilities from paying property taxes, and because Luz’s largest facilities were valued at more than $1 billion apiece, that exemption was worth as much as $20 million to $30 million per plant. Toward the end of 1990, California’s governor vetoed an extension of the property-tax exemption. But a new governor was taking office in January, and Luz, which was spending $20 million a month to build its 10th plant, gambled that he would quickly reverse the action. When the new governor didn’t immediately reinstate the tax exemption, Luz lost its bet. “We miscalculated,” says Goldman. “We ran out of money and closed down operations.”

But as with the solar plant in Chicago, the stimulus money will be critical to the viability of the project, which will cost roughly $2 billion to build. The federal loan program, which provides for direct lending from the U.S. Treasury, could cover 60 percent of the cost. That would require the company to raise only $800 million from investors, who then would be eligible for $600 million in the form of refundable tax credits. (The investment tax credit for renewable energy existed before the stimulus package passed, but the legislation made a key change: it now gives investors an option to receive a direct grant equivalent to 30 percent of their investment, whereas previously they had to apply the credit toward any tax liability they might have.)

Jack Jenkins-Stark, Brightsource’s CFO, is responsible for making the numbers work. “It’s all about capital,” he says. The cost of operating the plant will be minimal–“maybe $20 million a year.” But finding a $2 billion loan to cover the construction costs became nearly impossible after the commercial lending and debt markets collapsed last fall. The only practical way to find such financing these days, he says, is by pairing the federal loan with financing from investors encouraged by the government incentives.

But the federal money, as Jenkins-Stark is quick to point out, comes with plenty of risks. The loan, of course, will have to be paid back. And though finding investors willing to risk several hundred million dollars to build a giant solar plant using new technology is “much easier” with federal incentives in place, he says, “it is still very hard.”

It’s a challenge that Arnold Goldman, for one, is happy to take on. Undeterred by the bankruptcy of his earlier solar empire, ­Goldman now envisions massive solar thermal plants across a wide swath of Nevada, California, New Mexico, and Arizona. This time, though, Goldman could be part of something even bigger. Nearly six gigawatts’ worth of new solar thermal capacity is planned in California alone. But, says Goldman, “we need a predictable policy environment.”

Catching Some Rays
About the same time in the mid-1980s that Arnold Goldman was filling the Mojave with mirrors, Richard Swanson, then a professor of electrical engineering at Stanford University, founded his company, SunPower. Both men had visions of large solar plants spread across the desert. But while Goldman was intent on making electricity by turning the sun’s energy into steam, Swanson–an expert on semiconductors and microelectronics–envisioned using photovoltaic cells built from precisely manufactured silicon wafers.

In a dusty field in back of SunPower’s headquarters in San Jose, Swanson shows off the technology that, if all goes well, Exelon will install in Chicago. A row of large solar panels, mounted on a tracking apparatus, tilts imperceptibly every few minutes so the panels can follow the sun; each panel holds dozens of high-­efficiency solar cells of the type that Swanson developed at Stanford. A solar-powered motor wheezes slightly every time it moves the panels. At night, the motor will swing the panels toward the east, waiting for the next day’s rising sun.

Swanson’s cells are among the most efficient commercially available forms of photovoltaic technology; they convert around 22 percent of the sunlight hitting them into electricity. (The solar panels in Chicago will produce about two-thirds as much power as they would in a sunnier location.) But the panels and wheezing motor are also a stark reminder of just how difficult it has been to make silicon photovoltaics cheap enough to compete with more conventional sources of electricity.

Right now, with the 30 percent investment tax credit, the cost of energy from a photovoltaic plant in a sunny region is competitive with electricity produced by fossil fuels during peak hours, says Swanson. But that is the best-case scenario for solar. In less sunny regions and at times other than the middle of the day, when electricity prices are high and solar cells are most efficient, the power produced by photovoltaics is still far too expensive.

Dozens of startups have formed in recent years to pursue technologies that their founders hope will be more cost effective. To Swanson’s mind, however, the attention given to these efforts is misplaced. The cost of electricity from silicon photovoltaics is decreasing by 5 to 8 percent a year as the industry grows at a rapid pace, he says; within five years, as the existing technology improves and manufacturers realize economies of scale, it will be competitive without federal incentives.

“We don’t need a breakthrough,” Swanson says. “Waiting for the next big breakthrough [in photovoltaics] will do nothing but cause you to grow moss underneath your feet.” He adds, “We have a road map where we can very clearly see how to halve the cost from where we are today. And that is sufficient to fuel explosive industry growth.”

Turning the Corner
The solar industry might not need a breakthrough to continue healthy growth rates. But many scientists say that without dramatic advances, solar power will never supply the vast amount of power needed to eventually displace fossil fuels.

Of the 46 new energy research centers announced by the secretary of energy in late April, 24 are doing work related to solar power, and each is receiving $2 million to $5 million annually over the next five years. Likewise, two of the eight new DOE innovation hubs will focus on solar technologies: one on electricity and the other on techniques for storing the energy from sunlight in the form of fuels. And the proposed 2010 DOE budget, which (coming just a few months after the stimulus bill) contained relatively modest increases for most new energy technologies, nearly doubled the research budget for solar power.

Much of the research focuses on overcoming the fundamental dilemma of photovoltaic technology: the trade-off between cost and efficiency. Conventional solar cells are efficient because the silicon from which they’re made is grown as a single crystal, yielding a perfectly ordered molecular structure; when the semiconductor absorbs sunlight, the light’s energy excites electrons that can travel through this crystal structure unimpeded, escaping to create an electrical current. But making devices out of single-crystal silicon is relatively difficult and expensive. Newer photo­voltaic technologies use materials that have a less ordered structure and can be deposited as thin films; they are potentially easier and cheaper to make, but they’re also less efficient.

“With photovoltaics you have either high efficiencies or low cost, but what we urgently need are [photovoltaics] with both attributes,” says Harry Atwater, a professor of physics and materials science at Caltech. “One of the challenges of solar power is how to get hundreds of gigawatts to a terawatt of power in a way that is cost effective.” Achieving that, he says, may take technology “very different than what we use today.”

Researchers elsewhere are hoping to overcome the challenges inherent in using disordered materials for photovoltaic cells. When light strikes the jumble of molecules in such materials, the excited electrons and the electron “holes” left behind when they’re knocked free form particle-like pairs called excitons. Excitons play a role in the process that plants use to capture energy through photosynthesis, says Marc Baldo, a professor of electrical engineering at MIT; in addition, organic light-emitting diodes use them to generate light. And, he says, it might be possible to manipulate these excitons on the nanoscale to improve the photovoltaic properties of disordered materials. Baldo heads a DOE-funded energy research center for excitonics, which includes researchers from MIT, Harvard University, and Brookhaven National Laboratory.

Ultimately, however, using sunlight to produce electricity will never supply enough of the energy we need: existing solar technologies, after all, produce power only during the day, and electricity can’t easily be stored. Instead, we must find a way to use sunlight to make fuels such as hydrogen, which can readily and cheaply be stored until they’re needed.

Learning how to efficiently make such fuels directly from the sun–a process called artificial photosynthesis, because the aim is to essentially mimic the natural process used by green plants–is “still 20 to 30 years down the road,” says Harry Gray, a chemist at Caltech and director of a solar-research collaboration that includes scientists from a number of universities. Although researchers, including some in his group, are getting “nice results” on certain aspects of artificial photosynthesis, lots of difficult problems remain to be solved. “It’s going to take a long time to get it together,” he says.

Silicon photovoltaics will be the dominant solar technology “for quite a while,” says Gray. “If all goes well, we will move into cheaper solar cells that are not single-crystal silicon, such as organic photovoltaics. But the transition [to cheaper photovoltaics] is not going to come all that fast.”

Flying Tomatoes
Will the stimulus bill facilitate that much-needed transition to more efficient technologies? Severin Borenstein, for one, is doubtful. Borenstein, the director of the University of California Energy Institute, says the problem with the stimulus funding is that when it comes to existing technologies, the DOE will need to pick which projects to support. “The worry is that the government will invest in the wrong technologies,” he says; picking technology winners is something that “historically it has not been very good at.”A far more effective way to promote the growth of renewable energy, he believes, is to put a price on carbon dioxide emissions through a carbon tax or a cap-and-trade scheme (see “Carbon Trading on the Cheap”). Either approach would provide market-based incentives for deploying renewables and would represent a more efficient and “technology-­neutral” government policy. At the same time, he says it is important for the government to fund research into new renewable technologies.

From an economist’s perspective, Borenstein says, government subsidies are justified to address “market failures”: cases in which the market doesn’t allocate enough resources to the pursuit of socially desirable goals, such as reducing greenhouse-gas emissions. The government incentives then support efforts that are financially risky but are likely to provide a common benefit. In such a context, he says, the argument for public spending on research into new solar technologies is strong–but the case for subsidizing the current commercial technologies, particularly photovoltaics, is “really weak.” Existing photovoltaics are expensive even compared with other renewables such as wind and solar thermal, he says, and they won’t necessarily lead to cheaper technologies, either. “You’re obviously going to get [solar] panels put in, but is that going to generate something that will have a lasting benefit? Will it help you build a solar industry? I think the answer is probably not.”

In a recent paper, researchers at Carnegie Mellon University’s Department of Engineering and Public Policy surveyed leading solar-power experts on the future of photovoltaics and concluded not only that the technology is far more expensive than other renewable sources of energy, such as wind and even solar thermal power, but that it “may have difficulty becoming economically competitive” in the next 40 years. The results are “dismaying,” says Granger Morgan, a Carnegie Mellon engineering professor and the department’s head.

Short-term subsidies for wind and solar thermal power could help them become cheap enough to compete with conventional sources of electricity, Morgan says. “But silicon photovoltaics are really a different matter. With existing technology, I just don’t see it happening.” He doubts that even doubling or tripling the use of current photovoltaic technologies will dramatically bring down the prices. “Of course,” he says, “if you get up in a room and say this, the tomatoes start flying.”

Indeed, plenty of experts believe that existing photovoltaics have an important role to play in promoting new forms of solar power. Deploying them on a larger scale will “pave the way for the next technologies,” insists Lawrence Berkeley’s Alivisatos. For that reason, he says, it is important to establish photovoltaics in the market. “It makes sense to have an industry that can gear up now,” he says. “Hopefully, that industry will absorb the new developments and bring out newer products over the next couple of decades.”

Caltech’s Harry Gray agrees: the urgency in his voice is palpable when he argues that we need to install as much solar power as we can, as soon as possible. “We need to make investments now in the technology we have, which is silicon photovoltaics,” says Gray. “We should be putting in [solar power] everywhere we can so people can see that it can make a difference. We can’t sit back and wait for breakthroughs. We need to show people that solar can work.”

The disagreement on the role of solar photovoltaics illustrates the larger debate over the best way for government policy to encourage a national shift to cleaner energy. And it is about to be played out around the country.

The DOE will probably decide soon on Exelon’s loan application to build the Chicago solar plant. If it gets built, the facility will represent only a tiny fraction of the nation’s total solar capacity, or even of Exelon’s electricity portfolio. But Gray is surely right on one point: such a facility, located in one of the nation’s largest cities, would be the face of solar power for many. Its fate will matter.

According to accounts by several companies, the DOE is thorough in reviewing the hundreds of loan applications it’s received, rigorously evaluating the financial health of the applicants and the market potential of proposed facilities. Nevertheless, it is hard to ignore the role of politics in deciding whether Exelon will build its facility on the South Side of Chicago. After all, President Obama’s home is just 13 miles away, and local politicians have been trying for years to reinvigorate the neighborhood that surrounds the site.

But then, politics have always played a major role in deciding the nation’s energy future, especially when it comes to solar power. Just ask Arnold Goldman.

David Rotman is the editor of Technology Review.