Sun and Sand

An energy source that requires no burning, leaves no waste, and will never run out, harnessed to a technology built from sand. What could be better?

solar panels on roof of building

Solar power is clean, renewable, and pretty much evenly distributed around the globe. "It's more or less perfect," says Joshua Pearce, a Penn State graduate student in electrical engineering. The hitch has always been cost. Most solar cells are made of silicon, which starts out as sand. The sand itself is cheap. It's the making into silicon that isn't. Not yet, anyway.

Solar cells work by converting sunlight into electricity. Incoming light is absorbed in a flat panel of silicon, a semiconductor material. The light knocks loose electrons from the silicon atoms, allowing them to roam free within the solid's molecular matrix. Because the solar cell has a built-in electric field, however, all of those free electrons move in the same direction: They create a current, which is drawn from the cell by metal contacts.

Crystalline silicon was the early standard. Its uniform molecular structure—neatly ordered rows of atoms—makes a smooth road for carrying current. But crystalline silicon is expensive to grow, requiring a complex process that starts with dipping a single "seed" crystal into a crucible of molten material.

Back in the early 1970s, Christopher Wronski, then working at RCA's Sarnoff Research Laboratory and now a professor of electrical engineering at Penn State, devised a cheaper alternative. With David Carlson, now chief scientist for British Petroleum's solar division, Wronski invented a solar cell made from amorphous silicon.

Amorphous silicon, Pearce notes, is "the stuff used for thin-film transistors, for flat-panel laptop screens—because it's cheap." It's made by a process called chemical vapor deposition: In a closed chamber a source gas containing silicon is blasted with electrons until it breaks down into a mixture of charged particles which settle onto a substrate as a solid film. One important advantage of this method is that it allows for depositing material over a large surface, all at once. Another, says Pearce, is that "a thin film can be put on anything. And it's flexible. You can coat it onto shingles, or on steel, or on a sheet of plastic that you roll out."

The trade-off is that amorphous silicon, true to its name, is amorphous: It lacks that neat crystalline structure. At the molecular level it's an irregular jumble. The chemical bonds between the atoms, Pearce says, "are imperfect. There are loose ends, where bonds are left dangling." As a result of these defects, amorphous silicon is a poor conductor, with an efficiency below five percent. (Efficiency, Pearce explains, is the percentage of incoming energy that is converted to electricity.) Worse, as Wronski discovered soon after his invention, amorphous silicon degrades with exposure to light. (Pearce: "Not great for a solar panel, right?")

But Wronski and others, after years of experiment, found their way past these obstacles. By depositing the silicon with hydrogen—adding a little bit of hydrogen to the vapor-deposition process—they discovered that they could bump amorphous silicon's efficiency to near 15 percent. "The hydrogen atom binds with those extra electrons, and ties off those dangling bonds," Pearce explains. And fifteen percent efficiency, while not as good as crystalline silicon's 20-25 percent, is good enough, given the abundance of available sunlight.

Adding hydrogen improves cell stability, too. Until recently, however, when Wronski and Pearce began collaborating with Penn State professor of physics Robert Collins, it wasn't clear exactly how that happened.

Collins, a condensed-matter physicist who studies the optical properties of thin films, has engineered a probe that shoots polarized light into the vapor-deposition chamber, bouncing it off the film/substrate "growing" surface. By analyzing the polarization of the reflected light—the process is called spectroscopic ellipsometry—he can in effect watch individual thin-film layers forming in real time. What this method has shown about Wronski's material has been somewhat surprising, Pearce says.

"Everyone thought that as you grew amorphous silicon, it was just amorphous silicon. But it changes as you grow it. With thickness, the properties of the film change. You get little crystallites, first nano-, then micro-sized. They have different properties."

When Wronski and Pearce started experimenting with adding different amounts of hydrogen, things got even more interesting. "What we found is that at low dilutions, at a certain thickness you get an unstable material," Pearce recounts. "At high-level dilution, you quickly go into phase transitions and micro-crystal. But at mid-level dilution, you get a stable material all the way-no large micro-crystals are formed."

By adding just the right amount of hydrogen, in other words, "we have been able to produce a solar cell that is both efficient and extremely stable," Pearce says. "We call it protocrystalline." The next challenge, he says, is to figure out how to grow this new material as rapidly as possible without sacrificing the gains in stability. "We're constantly balancing optimal dilution against optimal growth rate—trying to find that sweet spot."

"Rate is especially important," Pearce adds, "because if you can grow the stuff at twice the rate, you double the capacity of a solar-cell-manufacturing plant without changing anything else." That increased capacity is crucial to reducing production costs. Which is key, Pearce says, to making solar cells ubiquitous.

"It's a matter of scale-up. At the moment when we get a 100 megawatt-peak plant" (that is, a manufacturing plant large enough to make enough solar cells in a year to produce 100 megawatts of electricity), "solar power becomes cheaper than coal power." Japan will build such a plant within three to five years, he predicts.

In the U.S., where coal reserves are more plentiful, "We seem destined to be behind on solar power," Pearce says. Still, United Solar Systems Corporation, "one of the big U.S. players," is building a 25-megawatt plant in Auburn Hills, Michigan. "If that works, they'll go for 100 megawatts," he says. And once there's one plant in operation that can produce that many solar cells, he predicts, the fall in prices will quickly create demand for many more.

"It's going to happen," Pearce says. "It's just a matter of time."

Joshua M. Pearce is a doctoral candidate in the Intercollege Graduate Degree Program in Materials Engineering, 121 Electrical Engineering West, University Park, PA 16802; 814-865-2063; solar@psu.edu. Christopher Wronski, Ph.D., is the Leonhard Professor of Microelectronic Materials and Devices in the College of Engineering, 215 EEW; 865-0930; crwece@engr.psu.edu. Robert W. Collins, Ph.D., is professor of physics in the Eberly College of Science, 275 Materials Research Laboratory, 865-3059; rwc6@psu.edu. The work described above is funded by the National Renewable Energy Laboratory.

Last Updated January 01, 2002