Enough sunlight bathes Earth's daytime half in an hour to meet all human energy needs for a year. Sadly, there are several problems with meeting human energy demands by tapping such abundant, free solar power—not least of which is the cost of making semiconducting material that can cheaply harvest the power in sunlight. But material improvements from the California Institute of Technology and IBM might just lower the cost of solar power.

Graduate student Michael Kelzenberg and other materials scientists at Caltech employed vertical crystals of silicon—microwires, like "blades of grass," Kelzenberg says—to capture as much as 85 percent of the full spectrum of incoming sunlight, the researchers report in the February 14 Nature Materials. (Scientific American is part of Nature Publishing Group.) Their efficiency is almost as good as that of traditional silicon wafers, yet they require just one percent of the silicon in such wafers.

"With one one-hundredth of the material, we've gotten it to absorb 96 percent of the peak visible light," Kelzenberg says. "There's lots of reasons to believe this could be scaled to make thin-film solar cells."

The researchers embedded the silicon blades in a bed of "fish-tank material," flexible, clear silicone plastic known as polydimethylsiloxane. That alone was enough for the rods to efficiently capture sunlight coming in at an angle. But the direct sunlight of high noon failed to hit enough of the blades to efficiently initiate the flow of electrons. So the team added nanoparticles of aluminum oxide to reflect and scatter the incoming sunlight, enabling it to hit more of the microwires. "We have access to all the side walls of the wire," Kelzenberg explains. "It creates interesting junction geometries that absorb light and collect electricity."

No actual solar cells have been produced from the new microwires, yet. But "if you can't absorb light efficiently, then you certainly can't convert it to electricity efficiently," notes Caltech chemist Nate Lewis, who was also involved in the research. The silicon blades show enough light absorption to make them "interesting candidates from which to make solar cells."

Of course, thin-film silicon solar cells already exist, but they have struggled to match the efficiency of traditional silicon photovoltaics at absorbing light or turning it into electricity. The new microwires of silicon achieve similar efficiency at a fraction of the material cost. "Our goal is to make a thin-film [solar] cell that gets you the efficiency of a regular wafer-based solar cell," Kelzenberg says. "I certainly hope to see that come to fruition within the next few years."

There are also other semiconducting materials that might prove as cheap as the silicon microwires to produce. Already, thin-film photovoltaic cells made from copper indium gallium selenide (CIGS) or cadmium telluride are available with efficiencies in the field of roughly 11 percent and prices as low as $1 per watt, according to manufacturer First Solar. But both of these types of cells are made from rare or expensive materials: tellurium is rare, and indium is expensive because it is also employed in flat-panel televisions.

Researchers at IBM, however, have produced a similar thin-film solar cell from a combination of copper, zinc, tin, selenium and sulfur—all relatively abundant, inexpensive materials. The new cell can turn 9.6 percent of incoming sunlight into electricity, according to testing from the U.S. Department of Energy's National Renewable Energy Laboratory in Golden, Colo.

Materials scientist David Mitzi of IBM and his team report in the February 8 Advanced Materials that they replaced the indium in a CIGS cell with zinc and tin. Previous work in Japan had shown that such "kesterite" cells could achieve efficiencies of nearly 7 percent—not enough for deployment in the field, where lab efficiencies are usually cut at least in half—but promising. By varying the ratio of sulfur and selenium, Mitzi and his colleagues were able to boost the overall efficiency of the kesterite solar cell by 40 percent. "Increasing the sulfur does increase the open-circuit voltage of the device," Mitzi says. And "it's not obvious that this is limited to low efficiency," meaning there is more room to improve.

IBM itself has no plans to manufacture such kesterite cells, preferring to license the technology. But the group has also made a breakthrough in the manufacturing process. Typical thin-film solar cells are made from depositing a layer of the applicable materials as a vapor in a vacuum, which necessitates high-energy expenditures to both vaporize the materials and maintain the vacuum as well as intensive quality control to minimize introduced defects. But the IBM team deposited most of its material as a mixture of particles and solution, like printing with kesterite "ink" made from copper and tin in solution with particles of zinc. The sulfur and/or selenium was then added as a vapor. Similarly, companies such as XsunX are working on importing some of the vapor-based deposition improvements from industries such as computer hard-drive manufacturers to the solar cell business.

The ultimate benefit of both of these technologies may be flexibility as well as cost—both the silicon microwires and kesterite cells could conceivably be made into a variety of flexible devices, including some woven into fabrics, Mitzi says. Already, Dow Chemical and Global Solar have created a solar roof shingle that employs CIGS thin-film cells to generate electricity.

But, as it stands—despite an ongoing boom in rooftop and ground-based photovoltaic panels—solar power contributes roughly 0.1 percent of U.S. electricity needs. New thin-film technologies aim to boost that, cheaply. "Unless solar can get to a cost point that is competitive with conventional carbon-based energy sources, it's not going to take off," Mitzi says. Solar "production is approaching gigawatt scale. But if one wants to really have solar take off and address large-scale energy issues in the world, one needs to increase that by orders of magnitude."