Duke chemists describe new kind of ’nanotube’ transistor

Duke University researchers exploring ways to build ultrasmall electronic devices out of atom-thick carbon cylinders have incorporated one of these “carbon nanotubes” into a new kind of field effect transistor. The Duke investigators also reported new insights into their previously published technique for growing nanotubes in straight structures as long as half an inch.

Duke assistant chemistry professor Jie Liu will report on these and other nanotube developments during three talks at a national meeting of the American Chemical Society to be held March 28-April 1 in Anaheim, Calif.

Field effect transistors, among the workhorse devices of microelectronics technology, are tiny switches in which the passage of electric current between a “source” and a “drain” is controlled by an electric field in a middle component called a “gate.”

Carbon nanotubes — so named because of their billionths-of-a-meter dimensions (“nano” means billionths) — combine exceptional strength, minuscule size and flexible electronic properties. They can behave either like conducting metals or like semiconductors, depending on how carbon atoms are arranged on their walls. As a result, they offer great promise as components in electronic devices even smaller than those available today.

The Duke research group headed by Liu is among a number that have incorporated a semiconducting nanotube as a component in an experimental field effect transistor. The nanotube is grown on a surface of silicon dioxide with metal electrodes evaporated on the nanotube’s surface serving as the device’s electron source and drain. Meanwhile, a layer of silicon fabricated under the silicon dioxide serves as the transistor’s gate, also called a “back gate.”

However, other groups have found that this back gate of silicon, which is “doped” with other chemicals to fine-tune its electronic properties, is poorly coupled with the rest of the device. The result is excess power demand. “To turn the device from off to on, you need five to ten volts,” Liu said in an interview.

To address this shortcoming, teams at two other universities have found they can reduce the power demand to between 0.3 and 0.5 volts by adding an additional gate made of a tiny droplet of salty water.

“That’s an order of magnitude of difference,” Liu said of what he termed a “water gate.” But “the disadvantage is that water is a liquid. So we looked for a way of replacing this water droplet with something that has similar properties but is a solid.”

In a new paper in the research journal Nanoletters, Liu, graduate students Chenguang Lu and Qiang Fu, and research associate Shaoming Huang describe substituting an electrically conducting polymer that has been developed for dry lithium battery technology.

This substitute compound, called lithium perchlorate/polyethylene oxide (PEO), “can achieve similarly good device performance and avoid the problem of using liquid in the device,” the Duke authors wrote in their paper. This PEO “polymer gate” is placed directly over the carbon nanotube.

Liu’s team found the polymer gate’s electronic properties can also be more easily fine-tuned to control the direction of the electric current by doping the underlying nanotube with other small carbon-containing molecules.

Doping silicon-based semiconductors in that way requires fabricators to precisely incorporate chemicals into those materials’ internal crystal structures. “For a nanotube, you just coat it on the surface, which is a lot easier,” Liu said.

Also at the Anaheim meeting, Liu presented an update on research his group reported in the Journal of the American Chemical Society in April 2003 on growing straight and exceptionally long nanotubes that can be potentially cut into smaller lengths for splicing into electronic nanoarrays.

That 2003 journal report described how quick heating the emerging nanotubes in a continuously flowing feeding gas of carbon monoxide and hydrogen to a temperature hot enough to melt glass made the tubes grow in unusually long and true alignment. “We now have a much better understanding of why this fast heating technology performs differently,” Liu said in an interview before his 2004 presentation.

In previous methods of using this chemical vapor deposition (CVD) process to grow nanotubes, the tubes extend along a surface of silicon dioxide. In the process, they encounter “physical resistance caused by the friction of bumping into other surface features,” he explained. “That stops the growth of the nanotubes.”

But quick-heating in the flowing gas makes the incipient nanotube lift up slightly above the surface as it begins to grow, he said. The growing nanotube follows the direction of the gas and stays slightly suspended, thus avoiding interacting with surface that is rough at molecular dimensions. “It’s like flying a kite,” he added.