By Kurt Kleiner
A new technique that places carbon nanotubes exactly where they are needed could help overcome one of the biggest obstacles blocking the development of nanotube-based electronic devices.
The method uses a specially constructed molecule that attaches one end to a carbon nanotube and the other end to a strip of metal oxide that has been placed on piece of silicon. The nanotubes are just a few nanometres in diameter, and knowing exactly where a tube is means researchers can use it to make a transistor.
"We can use this approach to make lots of devices," says team member James B Hannon, at IBM's T J Watson Research Center in Yorktown Heights, New York, US.
Researchers have previously created nanotube-based transistors, which could eventually be used to make smaller, faster computers. But there was no efficient way of exactly positioning nanotubes. Some researchers have manipulated tubes one at a time with atomic force microscopes. Others have laid down thousands at random, and then created contacts and repeatedly tested them until they found a working circuit.
To line up the nanotubes more reliably, the IBM researchers created a molecule that bonds to nanotubes at one end and metal oxides at the other, and mixed this molecule into a solution containing nanotubes. This left each tube attached to a molecule.
They then took a piece of silicon and dipped it into the nanotube solution – the silicon had previously had aluminium oxide stripes laid onto it using electron beam lithography. The free end of the molecule stuck to the aluminium oxide, holding the nanotubes in place.
The next step was to heat the silicon to 600°C, which removed the sticky molecules and left the nanotubes in place, held there by Van der Waals forces (weak intermolecular attraction). Finally, the team used lithography to attach palladium leads to the nanotubes, completing the transistor.
"What needs to be examined further is whether it can be used for large-scale integration. It's a step in that direction," says Chris Papadopoulos, an electrical and computer engineer at the University of Victoria in British Columbia, Canada.
He says that as the researchers try to pack more devices closer together, nanotubes might start crossing each other, ruining many of the transistors.
And the technique is still far from perfect. From a total of 49 measurements made by the researchers, 28 working transistors were detected. But Hannon believes that refinements will eventually allow the technique to produce hundreds or thousands of working devices in precise locations.
Journal reference: Nano Letters (vol 6, p 906)