Can Carbon Nanotubes Make Chips?
As conventional microelectronics wrestles with the problem of leakage, which has ended one of the benefits of shrinking – the automatic reduction in power – the hope is that new ways of manipulating atoms and molecules will permit the further scaling of electronic devices.
Development has been slow to produce commercial results. At the beginning of this decade, $1 billion worth of VC-backed nanotech companies started, and failed, without bringing a product to market. Nantero of Boston has been struggling for seven years to make a carbon nanotube-based universal memory called NRAM (for Nanotube-based/Non-volatile RAM). Last week came two breakthroughs, one from Korea, one from the US, which could bring carbon nanotube-based manufacturing closer to productisation. 1 Korea: One of the problems with carbon nanotubes is that they don’t lend themselves to manufacture by a planar process but a technique developed by a Korean group led by Kahp Y Suh at the Seoul National University under has announced a technique of aligning nanotubes by including them in a liquid, then flowing them through channels to where they are required, then evaporating the liquid, leaving dense, highly oriented arrays of nanotubes. 2: The US: Another problem with commercial use of carbon nanotubes has been in eliminating the differences in the most promising type of nanotube, the single-walled carbon nanotube (SWNT). A team under Dr Wei Zhao, a professor at the University of Arkansas, is said to have made a major breakthrough in growing ‘near single type purity nanotube production’, making it possible to produce SWNTs with defined electrical properties. If carbon nanotubes become manufacturable, it could have very significant consequences for the semiconductor industry. The prospect of getting back onto the traditional virtual circle of scaling when each shrink delivered lower cost, lower power and higher performance, is a heady one for the chip industry. Resuming a bi-annual shrink path could return the chip industry to the 15 per cent annual growth-rates of the 1960s to 1990s, when each shrink gave the industry’s products such a price-performance advantage over the previous generation that they were irresistible to end-users wanting to remain competitive. TOMORROW MORNING: THE TEN FASTEST GROWING CHIP COMPANIES