Here's Phil Kuekes' recipe for building the computer of the future: Take a glass tumbler, pour in a few jiggers of exotic chemicals, shake it like a Martini, and voila!
"It's computers from a test tube," he says.
Mr. Kuekes is a scientist at Hewlett-Packard Laboratories in Palo Alto, Calif., and one of a growing number of researchers trying to build computers and other machines from the atom up -- a field known as nanotechnology.
If his recipe for a molecular PC pans out, the result could be a machine that crunches numbers a billion times faster than today's computers but is tiny enough to hide behind a mote of dust. It would make possible computerized clothes, furniture, wallpaper, "even fingernail polish," says Mr. Kuekes, who predicts that "a lot of things are going to get `smart."'
The device would revolutionize a computer industry worried about how much longer it can make more powerful machines using silicon.
Until recently, the notion of a high-tech cocktail dusting a Pentium III -- or doing anything at all -- would have given engineers the giggles. But last week Mr. Kuekes and his colleagues announced that they had discovered a molecule that behaves like the silicon transistor on a computer chip. The research, published last week in the journal Science, is the best evidence yet that a molecular computer might not be a crazy idea after all.
"It's a very significant advance," says Tom Theis, a scientist at IBM's T.J. Watson Research Center in Yorktown Heights, N.Y.
And it might be coming just in the nick of time.
Silicon transistors have been the heart of digital electronics for 30 years. Like miniature switches, they toggle on and off, performing the binary computations that allow computer chips to process and store information.
Early on, engineers discovered that cramming more transistors onto the same sliver of silicon boosted the performance of their chips. They've been trying to make smaller transistors -- and faster computers -- ever since.
As a result, chip speed has been doubling roughly every 18 months -- a trend known as "Moore's Law," named after the man who man who first predicted it, Intel Corp. co-founder Gordon Moore.
Moore's Law made possible the personal computer, the Internet, the cell phone -- every gadget that's small and light and has a microchip for brains. You also can blame Moore's Law for making today's top-of-the-line PC tomorrow's paperweight.
But the laws of physics may be closing in on Moore's Law. If the current pace of miniaturization holds up, the smallest component of a silicon transistor will measure just five atoms wide by 2012. Any narrower and transistors go kaput, according to researchers at Lucent Technology's Bell Labs (where the transistor was invented in 1947).
Just as ominously, as transistors get smaller they cost more to make. By 2012, a chip factory could cost more than $50 billion -- far more than even deep-pocketed companies such as Intel can cough up.
"Eventually science and industry will have to find new ways to build faster and larger computers," physicist Max Schulz of the Institute of Applied Physics in Erlangen, Germany, concluded in the British journal Nature.
That's where nanotechnology comes in.
"Even ... in our highest-tech silicon computer chip, the smallest feature is a mountain compared to the size of a single atom," Nobel Prize-winning chemist Richard Smalley testified at a June hearing on Capitol Hill, where Congress is considering doubling the government's $230 million budget for nanotechnology research.
Only in the past two decades have scientists invented devices sensitive enough to see and touch the nano world. In the early 1990s, IBM used these tools and a few hundred atoms to spell out its name.
In their computer research, Mr. Kuekes and his colleagues turned to a chemist at the University of California, Los Angeles.
"Our idea is this: If you really want to build something small, you ask a chemist to do it," says Mr. Kuekes. "These are guys who know how to put atoms together."
Last week, the group announced it had created an exotic class of organic molecules called rotaxanes that act like transistors. When researchers spread rotaxanes onto a bed of silicon, connected the open-faced sandwich to aluminum wires and turned on the electricity, they were able to perform the basic logical operations of a standard computer chip.
Don't chuck your Pentium PCs yet, Mr. Kuekes cautions. Researchers say there's still a long way to go. The rotaxane switches kept sticking in the "on" position. A replacement for the silicon transistor must be able to toggle on and off.