Wednesday, November 01, 2006

Silicon and Sun

Continued from page 1

By Kevin Bullis

Start from Scratch

The seawater tanks outside Morse's lab are teeming with colorful starfish and corallimorpharians, exotic creatures similar to sea anemones. But Morse and James Weaver, a postdoc in the lab, are more interested in an unremarkable­-looking rust-colored blob: an orange puffball sponge, a type of sponge that ordinarily lives in rock crevices just off the Santa Barbara coast. If the Venus's flower basket is the glass cathedral of sponges, this is the straw hut. The shapeless creature appears not to have a skeleton at all; but once the researchers dissolve away the living material of its exterior, a handful of tiny glass needles remain, each only two millimeters long and thinner than a human hair.

Although Morse ultimately wants to understand sponge skeletons that are more complex, these simple needles are a good place to start. Scientists have long known that at the core of the glass needles are strands of proteins, but no one understood what they did or how they related to the ­needles' construction. So Morse and his colleagues began by isolating the genetic code for one of the proteins--which as a family they came to call "silicateins"--and ran their results through a huge database of known proteins. They weren't expecting a match, but they found one--immediately. The protein was similar to a protease, an enzyme found in the human intestine that is involved in the breakdown and digestion of food.

"It was very bizarre," says Weaver. "Why does the protein that templates the formation of the glassy skeleton of a sponge have anything to do with a protease?" The researchers began to suspect that the silicateins did more than merely serve as a passive template. Indeed, they found that unlike any other enzyme previously studied, a silicatein can do double duty. It actively produces building materials such as silicon oxide--in a sense, by digesting compounds in the seawater--and then causes the materials to line up along its length to form the needle-shaped glass of the sponge skeleton. No such enzyme had been discovered, Morse says, "in all the study of biomineralization, which has gone on for a couple of hundred years."

Morse reasoned that if silicateins were so good at producing silicon oxide, they might also be able to produce the types of metal oxides that make good semiconductors in electronics and in some kinds of solar cells. He was right. "At 16 degrees Celsius, the temperature at which the sponge lives in the cool water right offshore from our lab," Morse says, "this enzyme will catalyze the formation and stabilize the formation of crystal forms of metal-oxide semiconductors that can't be made conventionally except at very high temperatures."

The result suggested a less expensive way to make semiconductors at lower temperatures, but there was a potential problem: contamination. "A biologist is ecstatic when they get a purity of, say, 90 percent. A chemist is ecstatic when they get a purity of 99 percent," says Morley Stone, a biochemist who directs research in biotechnology and materials for the Air Force Research Labs at Wright-Patterson Air Force Base, near Dayton, OH. "But an electronics engineer or someone else who needs to make devices--they want to see materials that have five nines of purity behind them, at least." He adds, "Oftentimes, when you take these biological approaches, you can grow some interesting things and get some interesting morphologies, but they're nowhere close to having the end-state purity that you would need in a final device."

Morse and his colleagues knew that if they hoped to make semiconductor materials for cheap but efficient solar cells, they would probably need a chemical synthesis technique that took its cue from the sponges but avoided the messy biology. The sponge's secret, they discovered, was that amine and hydroxyl chemical groups in the enzyme produce the silicon oxide and assemble it in the required way. That meant that all the chemicals a new synthesis technique would require could be found in ammonia and water. The researchers found that by mixing molecules containing the metal oxides' precursors into water, and then exposing the mixture to ammonia gas, they could create thin films of highly crystalline semiconductors--materials useful for electronics. "This is the breakthrough that gets us into the domain of practical usefulness," Morse says.

Moreover, the crystals have a complex nanostructure that could improve the performance of photovoltaic devices. Near the surface of the water, the concentration of ammonia gas is relatively strong, so this is where the semiconductor crystal starts to form. As the ammonia slowly diffuses deeper into the water, however, it causes crystals to grow down into the mixture, producing a thin film that is not uniform but rather comprises a network of needles or flat plates each merely a few billionths of a meter thick. That network could be the basis for a more efficient solar cell.