Sustainable Energy

Manufacturing Method Promises Cheaper Silicon Solar

Ampulse combines low-cost thin-film fabrication with conventional crystalline silicon.

Mar 15, 2012
Amped up: Startup Ampulse thinks its technology could trim the cost of electricity from crystalline-silicon solar panels to less than 50 cents a watt.

Sustainable Energy

Manufacturing Method Promises Cheaper Silicon Solar

Ampulse combines low-cost thin-film fabrication with conventional crystalline silicon.

Mar 15, 2012
Amped up: Startup Ampulse thinks its technology could trim the cost of electricity from crystalline-silicon solar panels to less than 50 cents a watt.

The cost of solar panels has fallen dramatically in recent years, but it’ll need to drop further still if solar power is to compete with electricity from coal or natural gas. The industry needs to find something cheaper than conventional crystalline silicon, which is found in the vast majority of today’s solar cells.

Ampulse, a startup in Golden, Colorado, believes it has the answer. The company says that by combining the high solar-conversion efficiency of crystalline silicon with low-cost, thin-film fabrication, it can slash the cost of producing electricity from crystalline-silicon solar panels to less than 50 cents per watt.

Conventional crystalline-silicon solar fabrication is time consuming, energy intensive, and highly inefficient. Silicon-rich gas is heated to 1,400 °C to make large crystals or ingots that are then cut into thin wafers in a process that takes several days and turns roughly half of the raw silicon material into unusable sawdust.

Ampulse uses a vapor deposition process developed at the National Renewable Energy Laboratory, also in Golden, Colorado, to grow silicon crystals on a flexible metal foil.

The process is known as hot-wire chemical vapor deposition. A tungsten filament similar to the wire found in an incandescent lightbulb is used to heat silicon gas inside a vacuum chamber to 700 °C, causing the gas to decompose and deposit a thin film of silicon directly onto a substrate material in seconds. The resulting silicon layer is five to 10 microns thick, just enough to convert most of the solar energy that hits a panel into electricity. Conventional crystalline-silicon wafers, by comparison, are about 200 microns thick.

Ampulse combines vapor deposition with a novel substrate developed at Oak Ridge National Laboratory. The substrate causes the silicon crystals to grow with uniform alignment and orientation, a key requirement for maintaining high cell efficiency.

Vapor deposition has previously been used with other substrates to produce thin-film amorphous silicon. But this lacks the uniform crystal alignment of crystalline silicon and so has a significantly lower solar-conversion efficiency.

The company is currently installing a pilot-scale production facility at NREL that was built by German photovoltaic equipment manufacturer Roth & Rau Microsystems.

Ken Zweibel, director of the Solar Institute at George Washington University in Washington, D.C., says Ampulse’s approach is interesting, but faces technical challenges and stiff competition. “The history of thin-film silicon is littered with false starts and abandoned approaches,” Zweibel says.“It will be a long, tough climb for Ampulse.”

Zweibel says a glut of conventional crystalline silicon modules from China, and newer thin-film solar cells made from more exotic materials like cadmium telluride and copper indium gallium (di)selenide, makes crystalline silicon a “very dark horse. “First Solar, for example, already manufactures cadmium telluride thin-film modules that can produce electricity for approximately 70 cents per watt, and plans to drive this price down further.

If, however, highly efficient thin-film silicon panels can be manufactured, the material’s low cost and abundant supply may give it an advantage over other, more exotic thin-film materials.

Ampulse officials declined to say what efficiencies they have achieved thus far in laboratory testing, but say their initial product will have an efficiency of 15 percent, and will ship in 12 to 18 months.

Zweibel remains skeptical. “If I’m an investor, I’m not going anywhere with my money until I have some sort of benchmark of what they have achieved in the laboratory.”