Physicists have taken a step toward the ultimate miniature chemical sensors. A single carbon nanotube about two billionths of a meter wide can compete with the best materials of today at detecting ammonia and nitrous oxide, a gas often found in automobile emissions. If these nanotubes could be customized to distinguish other gases, many of them could be placed on a single device, creating a "nose on a chip" that could ferret out pollutants or analyze the atmospheres of other planets.
A nanotube can be imagined as a sheet of carbon atoms that has been rolled up into a cylinder, then twisted to various degrees. Twisted in one way, a nanotube behaves like a metallic wire, freely conducting electricity, while one with a slightly different twist behaves like a semiconductor--it can switch between being an insulator and a conductor. Though nanotubes were first discovered in 1991, and were hailed almost immediately as a technology of the future for sensing devices, it is only in the last couple years that physicists such as Hongjie Dai, of Stanford University, have achieved enough control over the tubes' assembly to make them behave reliably.
Dai noticed right away that nanotubes conduct electricity very differently when exposed to different kinds of gases. "We decided to make use of this, instead of having it as a nuisance, by intentionally putting different kinds of molecules onto the tubes," Dai says. When exposed to nitrous oxide, for example, the tubes became up to 1000 times more conductive. And ammonia, as the team reports in today's Science , decreased conductivity sharply.
The results are "definitely exciting," says Phil Collins, a solid-state physicist at IBM in Yorktown Heights, New York. "But in terms of commercial sensors, there's much work to be done." In particular, he says, the response time of Dai's nanotubes is disappointingly slow--several minutes--and they take up to half a day to recalibrate, during which the tube recovers its original conductivity. Moreover, no one knows yet why the change in conductivity occurs. "The claim that this can be used as a sensor requires a few leaps of faith, and of science too," says physicist Mildred Dresselhaus of the Massachusetts Institute of Technology.