'Nanotube' chips set to break silicon barrier

By Anand Parthasarathy

KOCHI, MAY 2. Technology breakthroughs announced in the U.S. last week offer the hope that the ``silicon barrier'' - the limit on the number of transistors that could be squeezed on to a computer chip - would soon be breached. The answer lies in nanotubes - tiny cylinders of carbon atoms, a billionth of a metre in diameter, which could be used to create the millions of interconnections and switches that go into the making of a matchbox-sized microchip.

The latest issue of the American journal Science released over the weekend, reports on the findings of two sets of ``nanotechnology'' scientists - one based at IBM's Watson Research Centre in New York state and the other at Harvard University. Nanotubes - which form spontaneously from six-sided arrays of carbon atoms - were discovered 10 years ago by Dr. Sumio Iijima at Japan's NEC Research Laboratories. They are named after the nanometre, which is one billionth of a metre or approximately 50,000 times thinner than a strand of human hair. In contrast, today's chip technology uses circuits printed on silicon, where the elements are one- tenth of a micron apart, which is about 1,000 times thinner than hair. (A micron is a millionth of a metre). Printing components much closer may hit the physical limits of silicon - and cause parts of the chip to become electrically shorted.

For some years, IBM researchers have been able to work with nanotubes generated in carbon arc furnaces to create the interconnections that go into a simple field effect transistor. They discovered a useful side effect: roll a nanotube of carbon and you have a metal-like conducting wire. Twist it - and the properties became similar to that of a semiconductor switch. Since both switches and interconnections are used in millions on a computer chip - the Pentium 4 has nearly 42 million devices on board - nanotubes are now perceived as ideal material.

Industry commentators were saying this week, that by using a hybrid of silicon and carbon nanotube technology it may be possible to give current chip technology a further lease of life and delay a costly changeover to other radical alternatives like optical computers, which are as yet on the drawing board. While practical nanotube transistors - and hence microchips - may be some years away, IBM feels it has overcome the problems of repetitive production.

One semantic side effect of last week's announcements was the recognition that `Moore's Law lives' - for now. The ``law'' first formulated 25 years ago by Dr. Gordon Moore, a founding father of chip maker Intel, forecast that technology - specifically the number of transistors on a computer chip - would double every 18 months. With fabrication techniques hitting physical limits, it was feared that silicon would finally show the limits of Moore's Law. But nanotechnology appears to have come to the rescue - just in time.

Meanwhile, the question, where will all these nanotube chips go - seems to have been answered. A forecast by market watchers Frost and Sullivan India, quoted by Reuters, says, India's chip market is poised for a compound annual growth rate of 24 per cent, which means a $1.2 billion (Rs. 5,000 crores) value by 2005. Strong demand from the recently untethered telecom sector and a consumer electronics boom, are said to be the reason. While the western and Japanese market for chips is largely driven by the PC and consumer segment, India's more diversified demand from the Information-Communication- Entertainment sector is stated to have an edge.

The U.S. chip majors, Texas Instruments, Motorola and Analog Devices have set up R&D units in India where microprocessors for a global market are designed. These are then sent abroad for manufacture, because prevalent Indian tariffs make local production unviable.

Send this article to Friends by E-Mail

Section  : Science & Tech
Previous : The road to 3G
Next     : Disease resistant variety