ĎMany Pentiums on a Grain of Sandí
By Kenneth Chang <mailto:email@example.com>
Pat Collier of UCLA holds a chip with the first circuits created from a single layer of organic molecules. (Defense Advanced Research Projects Agency)
Harkening a future where microscopic sensors can be embedded in paint, and medical diagnostic machines are no larger than the bacteria they analyze, researchers have built an electronic device using a specially designed organic molecule instead of the silicon of present-day computer chips.
Writing in todayís issue of the journal Science, the team from the Hewlett-Packard Co. and the University of California at Los Angeles says its work could ultimately lead to ultra-tiny processors 100 billion times faster than the most powerful ones available today.
"Many Pentiums on a grain of sand," says Hewlett-Packard researcher Philip Kuekes, one of the authors. "And thatís not an exaggeration. We could have a prototype of a molecular computer in five years and something you might imagine buying in 10."
The basic component of all digital processors is an on-off switch. The researchers built their switch by simply sandwiching a custom-designed molecule called rotaxane between two criss-crossed wires.
Rotaxanes are a class of organic molecules that could replace the transistors on a microchip.
With the rotaxane, electrons can cross from one wire to the other.
"The molecule is like a stone in a river," says UCLA chemist and lead author James Heath. "And so the electron has only to jump to the molecule and then to the other side, and that means current can flow pretty readily."
This corresponds to the closed position of the switch.
Applying an electric field between the wires breaks the rotaxane, and then the electrons canít hop as easily. This is the open position of the switch. "We essentially remove the stone," Heath says.
Researchers have previously made such molecular switches. The achievement of the Hewlett-Packard and UCLA group was to link together several switches and program them to perform simple logic operations.
"Itís one of these big steps as we try to push to molecular electronics," comments James Tour, a chemist at Rice University in Houston, another researcher in the field.
Not Small Yet
The circuit measures 5 millionths of a meter by 11 millionths of a meter, unimpressively large by computer chip standards. But the researchers say with more work, they should be able to miniaturize the circuit down to a few billionths of a meter on a side.
The other current drawback is that since the rotaxane breaks, the switch canít be reset. That means the switches can be used for static storage of information ó like a CD-ROM or a book ó but not for computer memory or processors.
They hope to create a switchable switch by designing a similar molecule that bends instead of breaks. But Heath says that should be easy. "If we donít get it, someone will get it and itíll be soon. You can bet money on that."
Indeed, at least one other research group says it has built a smaller, reusable switch, but has not yet submitted the work for publication.
A Smaller, Faster Future
Silicon chip technology has been running up against greater and more expensive hurdles as newer computer chips jam more and more transistors into a smaller area. Factories to process the latest chips cost about a billion dollars, and many expect in 10 to 15 years, electrical engineers will have hit the physical limits.
By contrast, the molecular switches are easy and cheap to produce. "We only have about three steps in the fabrication process, and weíre a bunch of dumb chemists," Heath says.
Chemistry doesnít produce perfectly identical molecules all of the time, but the researchers have gotten around that problem as well. Last year, in an earlier paper, the same group showed how to make perfect calculations with imperfect processors, effectively bypassing the faulty parts.
Such powerful and tiny processors could show up in unexpected places. "These things are small," Kuekes says. "Weíre literally talking about computers woven into your clothing or painted on a wall."
Kuekes also imagines embedding such a molecular computer chip into an ultra-tiny medical sensor that literally cozies up to a bacterium to determine what it is. "Itís got very interesting biomedical implications," he says. "We havenít done it, but itís truly a plausible technology."
Used in future storage devices, the molecular switches could lead hard disks that never fill up.
Any maybe even a computer that would be fast enough to run the latest, greatest version of Microsoft Windows? Heath laughs. "I wonít bet on that. So far, theyíve outlived all such predictions. It would be hard. It would be a challenge for them. Yeah, thatís a scary thought."