Nano Technology :: Mega Steps Toward the Nanochip

The quest for nanometer-scale computing is now a quantum leap closer to reality.
In Friday's issue of the journal Science, physicists from IBM's Thomas J. Watson Research Center announce their fabrication of the world's first array of transistors made from carbon nanotubes.

The announcement recalls the breakthroughs of the late 1940s, when scientists first began developing the bipolar transistor, the device that spawned the microchip age.

Like their postwar predecessors, the IBM scientists -- Philip G. Collins, Michael S. Arnold and Phaedon Avouris -- have established an important proof of principle.


"It's a beautiful paper and a major step forward," said Richard Smalley of Rice University, who won the Nobel Prize, in part, for his pioneering work with nanotubes. "But we're still far away from a practical technology to build integrated circuits that would compete with silicon."

The innovation centers around one fundamental problem in molecular-scale electronics: The most sensible medium for computing typically exists in a "fluid" of near-lookalikes.

Although silicon will undoubtedly remain the basis for computing hardware for years to come, silicon technology is also approaching a dead end. The kinds of transistors that can be made out of nanotubes may only consist of a hundred or a thousand atoms, while present semiconductor materials can't even approach this level of miniaturization.

"I can't imagine a silicon transistor that doesn't contain a few million atoms, even in the far future," said IBM's Tom Theis, director of physical science research. "So we're talking about devices that are drastically smaller, and because their key components are fabricated by chemical synthesis, they can be drastically cheaper than silicon transistors."
Smalley added that even the nanotube wires will be important in nano-circuits of the future.
"These things look like great answers to the question of how to conduct electricity in practical circuits on the nanometer scale with air and water around in the real world."

Since they're a thousand times stronger than steel and can serve as both transistors and wires, nanotubes may indeed be the ultimate last step in conventional computing technology before the realm of the quantum computer.

"What certainly is going to happen is that these nanotubes are going to be used in damned near any area you can think of where electrons move from here to there," Smalley said. "And like all other new things, it's going to have to find niches in which it's competitive to displace the existing answers."
The nanotube is a long hollow cylindrical molecule composed of carbon and with a typical width of only 10 times the size of an individual atom. It was discovered in 1991, and in 1998 several teams of researchers began investigating the nanotube's ability to serve as a nanoscale transistor, the basic element of any conventional computer.

The trouble is when nanotubes are fabricated -- typically involving laser heating of carbon soot -- only some of the end product is the desired semiconductor nanotubes. The same recipe also produces a cohort of metallic nanotubes, which can't be used to make a transistor.

Previously, any attempt to build a nanotube circuit has involved the painstaking process of cherry picking the desired semiconductors one by one using atomic force microscopes. (No technique has yet been devised to create semiconductor nanotubes only.)

What Collins and company did, though, was to gather both types of nanotubes in a circuit and then exploit the fact that metallic nanotubes ultimately shatter if enough current is run through them.
"They've come up with a recipe that anyone can follow that will allow you to make many thousands of these transistors simultaneously on a silicon substrate," IBM's Theis said.

"With the proper sequence of electrical pulses, we're able to fuse out the tubes that are wires -- the ones we don't want -- and pick out the ones that are semiconducting."