By Jeff Karoub
Small Times Staff Writer
IBM researchers said they have taken a major step in the race to replace silicon with carbon nanotubes in microchips.
Researchers said they created the first logic-performing computer circuit from a single molecule. The
 |
| IBM researchers created a protected polymer coating (darkened areas) and dropped potassium atoms through the open area into the carbon nanotube (blue line). The atoms converted unprotected positive, or p-type transistors, into negative, or n-type, both of which are needed to build a logic circuit. |
The findings were published in Sunday’s edition of Nano Letters, a journal of the American Chemical Society, and presented the same day at the society’s national meeting in Chicago.
IBM and other researchers are working on molecular-based microchips because they are expected to lead to smaller and faster computers that operate at much lower power. But the new technology also could replace silicon if it reaches its physical limits in 10-15 years, as many scientists believe.
It’s the latest but by no means the only research into new ways to maintain Moore’s Law, the oft-quoted prediction of Intel Corp. co-founder Gordon Moore. He said the processing power of a chip will double every 18 months.
IBM’s work builds its progress reported in April, when researchers announced they had built arrays of carbon nanotube transistors.
Scientists found a way to separate metallic nanotubes from the semiconducting nanotubes — the only kind that can be used as transistors — by destroying the metallic nanotubes with an electric shockwave.
The new technique bypassed the need to separate nanotubes with a scanning tunneling microscope, turning a tedious task that took a half-day down to about a minute.
The most recent discovery solved yet another problem of carbon nanotube transistors. All of them are positive, or p-type, which lack electrons found in n-type, or negative transistors. Both are necessary to build any type of logic circuit.
Researchers discovered they could selectively convert part of a single nanotube to n-type by pouring a polymer coating over the whole thing, etching out part of it to create an opening and dropping in potassium atoms, which change unprotected p-type transistors into n-type.
The discovery eliminates wiring two different types of transistors together. The new process allows scientists to make as many transistors on a single nanotube as needed.
“We’re making some good progress, but there is still a lot of work to be done,” said Matt McMahon, a spokesman for IBM’s research division.
He said the researchers’ next step will be developing more complex circuits and then creating a working microprocessor. The work could move into the engineering and design stage in about two years, and full production could be at least a decade away.
“It’s an exciting development,” said Mark Reed, a Yale University professor who specializes in molecular and nanoscale electronics.
“(IBM) is a real leader in doing nanotube transport. This is an additional accomplishment in their laboratory. It keeps them established as a leader in the field.”
Still, IBM is by no means the only company or institution seeking a nanoscale solution to silicon, or stretching the limits of silicon’s useful life:
- Hewlett-Packard in July received another in a series of patents in a process to create a molecular chip. The latest discovery is a system that allows information to be stored on and retrieved from molecules found at the intersection of each tiny, silicon wire.
- Intel Corp. announced in June it had built what it called the fastest silicon transistors. These transistors, considered nanoelectronic devices, are made of about 80 atoms in width and three atoms in thickness and are 1,000 times faster than today’s existing microprocessors.
- Researchers at Osaka University in Japan said last week they have produced long crystalline nanochains and silicon dioxide stems using little more than a furnace, vacuum chamber and a quartz container, the report said. The researchers said the process is much simpler and cheaper than nanotube techniques, and could be commercially produced within 10 years if they can get the money to move forward.
Reed said the efforts at Intel — as well as IBM and other labs — suggests nobody should be so quick to dismiss the durability of silicon.
“Silicon (makers) are working very hard, making incredibly impressive advances,” he said. “At some stage, out in something like 2015, they kind of reach their limits. Of course, those projections can be extended, depending on technological advances.”
McMahon agreed, but said IBM holds out the highest hope for nanotubes.
“We’re not saying this is the only possible solution, but it’s the one we see as the front-runner and the one we’re focusing on.”
One nanotechnology expert said whatever technology emerges, it’s still far away from the marketplace. Until then, it’s good to have time to work on as many solutions as possible.
“Some are far-fetched, some are straight-forward,” said Jim Murday, executive secretary of the U.S. government’s Nanoscale Science and Engineering Technology group, which oversees the National Nanotechnology Initiative, created last year to coordinate research and development among several federal agencies.
“With (so many projects), there is a much higher probability you’re going to be successful at it.”
“There’s not just one approach, and not just one end application,” said Reed, who started Molecular Electronics Corp. last year with the aim of pursuing commercial application for microchips.
“IBM’s lab is one of hundreds. At Yale, we have 10 people looking at researching technologies beyond (silicon).
“We don’t know what’s going to win. No one knows.”
Related links
Carbon may be key to computing in follow-up to IBM’s research
Full report from IBM
null
CONTACT THE AUTHOR:
Jeff Karoub at [email protected] or call 734-528-6291.