Drawing on somewhat of an “old-school” crystal-growth method, Milton A. and J. Maude Shoemaker Chemical and Biological Engineering Professor Tom Kuech and graduate student Kevin Schulte are studying how rapid synthesis affects the chemistry and electrical properties of gallium arsenide. Their results could point the way to faster, less costly manufacturing processes for solar cell materials.
Most solar cells are made from silicon; however, gallium arsenide is a promising alternative as one component of multijunction, or “stacked,” solar cells, says Kuech. “The highest-efficiency solar cells consist of a solar cell of gallium arsenide, and then what you do is put another solar cell on top of that, and another one, and stack them—each one grabbing its own little bit of the solar spectrum,” he says.
The process for making gallium arsenide for solar cells is much like making transistors for cell phones and lasers for communications. Yet, particularly for solar cells, the process is costly and slow.
Working with researchers at the National Renewable Energy Lab, Kuech and Schulte are studying ways to synthesize materials for solar cells at rates that are orders of magnitude faster than the current methods, molecular beam epitaxy and metal-organic vapor phase epitaxy. Those processes require expensive equipment, within which researchers very slowly deposit atoms onto a wafer, and the atoms arrange themselves and continue to grow in the crystal form. Materials typically grow at the slow rate of a micron or two an hour.
Using a unique reactor they are building, Kuech and Schulte are investigating whether entirely different growth technology and material chemistry could accelerate the manufacturing process without sacrificing solar cell performance or introducing material defects. Their system takes advantage of an early-’80s semiconductor thin-film growth technique called hydride vapor phase epitaxy. While they are using gallium chloride and arsine for this research, Kuech and Schulte have shown they can grow gallium nitride materials at a rate of hundreds of microns an hour. It’s a system that could scale up from a batch process to a continuous manufacturing process, says Kuech. “So now you’re talking about time-on-tool that went from hours down to minutes,” he says.