IBM's research division says it's found a way to overcome a key roadblock on the way to practical carbon-nanotube transistors.
The company notes that scaling carbon nanotube-based transistors to smaller sizes is limited by two things: the size of the device's channel and the size of its contacts. IBM says making a transistor with a carbon-nanotube channel is relatively practical. Connecting such a channel to its accompanying wires has been a major challenge, however, because of an increase in electrical resistance with traditional contacts (the part of the transistor that connects the channel to its wires) when they're used at such a small scale.
The firm announced an innovation today that may be a breakthrough on this front. It involves bonding contacts made of molybdenum to the ends of the nanotube channel. Those contacts are then connected directly to the sides of the transistor's wires. By making contacts in this fashion, the scale of the transistor can be reduced without also increasing resistance at the contacts.
IBM's researchers say they've successfully created a 9-nm transistor using this method, and the company expects the technology could scale down as far as the 1.8-nm node.
IBM researcher Shu-Jen Han outlined the next challenges for carbon-nanotube transistors in a blog post. The company now has to orient and position the nanotubes from a solution. It also has to find a way to remove the roughly 33% of nanotubes that are metallic and therefore not useful for building transistors.
Right now, Han says his team has developed a way to self-assemble carbon nanotubes and position them on a wafer. The next challenges he anticipates are to increase the density of the nanotubes so that they're only placed 10 nm apart and to scale that process over an entire wafer.