MIT Technology Review

Wednesday, February 22, 2012

Making Solar Power Competitive with Coal

A study identifies early-stage technologies that could be combined to cut the cost of solar panels in half.

By the end of the decade, manufacturers in the United States could make solar panels that are less than half as expensive as the ones they make now. That would be cheap enough for solar power to compete with electricity from fossil fuels, according to a new study in Energy & Environmental Science. The cost reductions will come via technology that's already being demonstrated in research labs at startups, universities, and major solar manufacturers, and could involve silicon, the material most solar panels are made from today.

The report, from researchers at MIT led by Tonio Buonassisi, a professor of mechanical engineering and manufacturing, identifies early-stage technologies that, if employed together, could reduce the cost of making solar panels to 52 cents per watt. Currently, the cost is over a dollar per watt. At 52 cents per watt, assuming similar cost reductions for installation and equipment such as inverters, solar power would cost six cents per kilowatt-hour in sunny areas of the U.S.—less than the average cost of electricity in the U.S. today. Solar power in sunny areas now costs roughly 15 cents per kilowatt-hour, according to the U.S. Department of Energy, although the cost can be sharply higher in small installations or in cloudy areas where solar installations generate less electricity.

The best way to reduce the cost per watt is to make solar cells more efficient—as a result, more power can be produced with a given amount of material and factory equipment. Increasing efficiency also decreases installation costs, since fewer solar panels are needed. But efficiency improvements aren't enough to reach 52 cents a watt. Manufacturers will also need to make solar cells from thinner silicon wafers, make wafers in a way that wastes less silicon, and speed up manufacturing. If a high-efficiency solar cell design slows down manufacturing or requires thick wafers, it likely won't lead to the necessary cost reductions.

One major way to reduce costs involves technologies that offer an alternative to the wasteful process now used to make silicon wafers. Currently, half of the high-quality silicon needed to make wafers ends up as waste. One startup, 1366 Technologies, makes thin wafers directly from a pool of molten silicon. It plans to replace conventional crystallization furnaces, sawing stations, and ingot-handling equipment with a single machine that requires fewer workers to operate. Others startups are replacing sawing with processes that free thin wafers of silicon from a larger piece of silicon using chemical etching, or by peeling them off.  

Once manufacturers have thin wafers, they also need equipment and processes that can handle them without breaking them. It's possible to make silicon solar cells as thin as 25 micrometers while maintaining their performance, but most manufacturers use 180-micrometer wafers that are more durable. One approach to handling thin wafers involves processing wafers on top of a sheet of glass. The glass acts as a support during manufacturing; when the solar panel is complete, it protects the cells from the elements. Magnetic levitation systems that would float the wafers along a production line could also help with the handling of thin wafers.

Some high-efficiency solar cell designs lend themselves to thin wafers. One involves sandwiching a wafer of crystalline silicon between two layers of amorphous silicon, as is done with a type of solar cell now produced by Sanyo. This symmetrical structure reduces stress on the wafer. Such cells can be processed at lower temperatures than conventional solar cells. Other cell designs could also work with thin wafers. One puts all of the electrical contacts on the back of a wafer—a process that could be well-suited to processing the cells on a sheet of glass. The U.S. company Sunpower uses a version of this cell design.

Much of the technology described in the report hasn't been demonstrated at full production scale. The techniques for making wafers without sawing, in particular, face a number of issues, such as producing high enough quality silicon, making wafers in the right shape and size, or producing them reliably and at a high rate.

To make solar power more competitive, installers will also need to reduce costs. Installation and the cost of inverters, wiring, land, and financing account for half—and sometimes as much as 80 percent—of the cost of solar installations. Much of this needed cost reduction could be achieved by improving efficiency, which would reduce the number of panels needed in a project.

Eventually, silicon solar panels could be even cheaper than 50 cents a watt, Buonassisi says. That will require finding ways to manufacture more challenging designs—for instance, including a nanostructured layer that improves light absorption, which would allow silicon cells that are only one micrometer thick to perform as well as conventional solar cells.