2020: nanotubes based chips?

Each chip on this wafer has 10,000 nanotube transistors on it. IBM hopes to be able to put billions of the devices on a single chip soon after 2020. Credit: IBM

It is now a few years that carbon nanotubes have been demonstrated to have a Schottky junction, that is they can operate as a transistor. 

Having a chip based on carbon nanotube is quite interesting since that would lead to higher density (more capacity and performance) with less use of energy. In other words moving from silicon to carbon would sustain the Moore's law for another decade.

There have been a number of trials to create circuits out of nanotubes and it has been shown that this is possibile. The point is to create an "industrial process" that can lead to the manufacturing of carbon based nanotubes in the billions at a sustainable cost (comparable to the one of today's chip manufacturing).

Hence the interest in this news for IBM, the first to my knowledge, promising an industrial process leading to carbon tube chip manufacturing bu the end of this decade.

They have already been able to produce, using industrial processes, a wafer for chips with up to 10,000 carbon nanotube. 10,000 is nothing, of course, chips have hundred millions of transistors, in some cases over a billion and the challenge will be to be able to step up in density. And here comes what really fascinated me in the news.

IBM researchers have chosen to adopt a process for chip production that does not exist today but that they believe will exist by 2019. Today manufacturing process is based on 14nm etching. This would not work for nanotubes. However, in order to keep up with the Moore's law industry will have to step up to 5nm etching by 2019 and with the tools supporting that technology they will be able to manufacture carbon tube based chips. Amazing how evolution and progress is taken for granted!

Author - Roberto Saracco

© 2010-2018 EIT Digital IVZW. All rights reserved. Legal notice