Reducing the Cost of Solar Cells
Solar-powered electricity prices could soon approach those of power from coal or natural gas thanks to collaborative research with solar startup Ampulse Corporation at the National Renewable Energy Laboratory. Silicon wafers account for almost half the cost of today's solar photovoltaic panels, so reducing or eliminating wafer costs is essential to bringing prices down. Current crystalline silicon technology converts energy in a highly efficient manner; however, that technology is manufactured with processes that could stand some improvement. The industry needs a method that is less complex, creates less waste and uses less energy. First, half the refined silicon is lost as dust in the wafer-sawing process, driving module costs higher. Wafers are sawn off of large cylindrical ingots, or boules, of silicon. A typical 2-meter boule loses as many as 6,000 potential wafers during sawing. Second, the wafers produced are much thicker than necessary. To efficiently convert sunlight into electricity, the wafers need be only one-tenth the typical thickness. NREL, the Oak Ridge National Laboratory and Ampulse have partnered on an approach to eliminate this waste and dramatically lower the cost of the finished solar panels. By using a chemical vapor deposition process to grow the silicon on inexpensive foil, Ampulse is able to make the solar cells just thick enough to convert most of the solar energy into electricity. No more sawdust - and no more wasting refined silicon materials. NREL developed the technology to grow high-quality silicon and ORNL developed the metal foil that has the correct crystal structure to support that growth. Ampulse is installing a pilot manufacturing line in NREL's Process Development Integration Laboratory, where solar companies can work closely with lab scientists on integrated equipment to answer pressing questions related to their technology development, as well as rapidly overcoming R and D challenges and risk. NREL's program is focused on transformative innovation in the domestic PV industry. With knowledge and expertise acquired from the PDIL pilot production line tools, Ampulse plans to design a full-scale production line to accommodate long rolls of metal foil. The Ampulse process 'goes straight from pure silicon-containing gas to high-quality crystal silicon film,' said Brent Nelson, the operational manager for the Process Development Integration Laboratory. 'The advantage is you can make the wafer just as thin as you need it - 10 microns or less.' Most of today's solar cells are made out of wafer crystalline silicon, though thin-film cells made of more exotic elements such as copper, indium, gallium, arsenic, cadmium, tellurium and others are making a strong push into the market. The advantage of silicon is its abundance, because it is derived from sand. Silicon's disadvantage is that purifying it into wafers suitable for solar cells can be expensive and energy intensive. Manufacturers add carbon and heat to sand to produce metallurgical-grade silicon, which is useful in other industries, but not yet suitable for making solar cells. So this metallurgical-grade silicon is then converted to pure trichlorosilane (SiCl3) or silane (SiH4) gas. Typically, the purified gas is then converted to create a silicon feedstock at 1,000 degrees Celsius. This feedstock is melted at 1,414 C and recrystallized into crystal ingots that are finally sawed into wafers. The Ampulse method differs in that it eliminates the last two steps in the traditional process and works directly with the silane gas growing only the needed silicon right onto a foil substrate. A team of NREL scientists had developed a way to use a process called hot-wire chemical vapor deposition to thicken silicon wafers with near perfect crystal structure. Using a hot tungsten filament much like the one found in an incandescent light bulb, the silane gas molecules are broken apart and deposited onto the wafer using the chemical vapor deposition technique at about 700 C - a much lower temperature than needed to make the wafer. The hot filament decomposes the gas, allowing silicon layers to deposit directly onto the substrate. Armed with this new technique, Branz and Teplin searched for ways to grow the silicon on cheaper materials and still use it for solar cells. They found the ideal synergy when visiting venture capitalists from Battelle Ventures asked them whether they could do anything useful with a breakthrough from Oak Ridge's superconducting wire development group. The new development, called the rolling assisted biaxially textured substrate (RABiTS), was just the opportunity the two scientists had been seeking. If metal foil is to work as a substrate, it must be able to act as a seed crystal so the silicon can grow on it with the correct structure. The RABiTS process forms crystals in the foil that are correctly oriented to receive the silicon atoms and lock them into just the right positions.