IBM’s physical sciences director Supratik Guha is pursuing a new type of chip that has the potential to be three times as fast as those made of silicon while requiring only a third as much energy.
A transistor has been assembled from 10,000 carbon nanotubes by Guha’s team at IBM’s TJ Watson Research Center in New York. It’s only a rough prototype that will have to refined over the next eight to ten years. But it promises to take over as silicon chips are expected to hit the practical production limits at about the 11-nanometer level. When that happens the IT industry’s ability to double the number of transistors on a single processor chip every two years — known as Moore’s Law — may come to a screeching halt in the absence of a fundamentally new way of creating semiconductors.
That’s where Guha hopes his nanotube chip will be able to step in to continue improvements in the speed and efficiency of processors for another couple of decades — or longer if the nanotubes can ultimately be linked in three dimensions.
That’s because a nanotube is just two nanometers in diameter. The problem is in finding a cost-effective way to produce high-quality nanotubes and controlling their placement into arrays of billions to form processors.
Guha’s team is etching tiny trenches on silicon and using a multistep process to precisely align nanotubes in them. They then have to add metal contacts to test the nanotubes’ performance. He hopes that the use of a silicon substrate will allow the nanotubes to be inserted efficiently using equipment in existing fabrication plants.
Guhe’s current prototype has the nanotubes arranged 150 nanometers apart, about 30-40 times farther apart than they must be if they are to compete against current silicon transistors.
“We need to lay down a single layer of carbon nanotubes spaced a few nanometers apart,” says Guha.
There are other major hurdles before nanotube chips become reality. A technique must be devised to add atomic-scale vertical posts to each of potentially billions of nanotube transistors to serve as electrical contacts. Production of ultrapure supplies of carbon nanotubes must become far more cost-effective and reliable.
“Nanotubes are an excellent candidate to keep the scaling of microelectronics technology going,” insists Guha, reckoning the time span to overcome these obstacles at five to ten years.
Guha joined IBM Research’s physical science department in 1995. Since then he has worked on new materials for silicon microelectronics, specifically, high-k oxide-metal gate technology. More recently, he has established IBM’s research programs in the areas of photovoltaics and semiconductor nanowires. Guha also manages the firm’s programs in the thermal physics of datacenters, and silicon based nanophotonics. His research focus is in new semiconductors and oxides for logic, and energy conversion applications.
Guha is also serves as an adjunct professor of materials science at Columbia University. He is a Fellow of the American Physical Society.
Supratik Guha received his PhD in materials science in 1991 from USC and a bachelors in technology in 1985 from the Indian Institute of Technology in Kharagpur.