Nanotube transistors shrink smaller than silicon sizeIS THERE anything carbon nanotubes can't do? Using them to make transistors about half the size of the silicon ones available today suggests they might help maintain the continual growth of computing power that we have come to rely on.
There are fears that this growth is threatened as engineers run out of ways to shrink silicon transistors and cram more power into chips. Finding new ways to make smaller transistors has become a priority (New Scientist, 6 December 2008, p 35).
But while carbon nanotubes had been considered a potential saviour, making transistors with them has proved to be difficult.
Performance in such transistors is limited by an effect that creates an electrical barrier at each point a nanotube joins any metal, impeding current flow. This seemed a deal-breaker because a nanotube in a transistor must connect to two metal electrodes, with a third "gate" electrode placed nearby. Using fatter nanotubes reduced the size of the electrical barriers, but goes against the computer scientists' goal of constantly making things smaller.
Now Aaron Franklin and colleagues at IBM's Watson Research Center in Yorktown Heights, New York state, have found a way to use thinner tubes to build a competitive nanotube transistor.
"That success is largely due to its geometry," says Franklin. His team placed the gate electrode, which controls the transistor, below the nanotube instead of in its usual position above it. This makes it possible to position the two closer together, and increases the gate's influence on electrons inside the nanotube, enabling them to punch through the electronic barriers.
Moving the nanotube and gate electrode closer together also makes it possible to shrink the device's length down to 15 nanometres. "That's about half the length of the best silicon technology on the market today," says Franklin.
The carbon nanotube transistor is half the length of the silicon alternatives on the market today
The work was presented at the International Electron Devices Meeting in Baltimore, Maryland, last month.
"This is great work that helps shed light on the scaling of carbon nanotube transistors," says Yu Cao at Arizona State University in Tempe, who was not involved with the study.
But Cao adds that nanotubes are still not ready for commercialisation. For example, physically manipulating them to build devices is tricky.
Franklin agrees. It's difficult to predict whether nanotubes can yet compete with silicon, he says.