Physicists hope to use subatomic particles' imprecise nature to answer questions beyond the reach of today's computers

At IBM's research labs, Isaac Chuang uses an NMR machine to turn alanine molecules into tiny quantum computers, with each molecule's three carbon atoms serving as its working memory.
Electrons would make terrible golf balls. They're just too ill-behaved. When an ordinary golf ball rolls across the green and comes to a stop, it's either in the hole or it's not. An electron, on the other hand, can be in many places at once--in the hole, beside it, and at the edge of the green. Like all submicroscopic particles, an electron tends to spread itself out in a sort of hazy ''cloud'' of probability. It's impossible to keep track of where it is at every moment. With quantum mechanics, we can work out the probability that an electron is in a given spot, but the electron won't settle on a single location until something forces it to. This unruly mix of chance and imprecision would ruin a golf game. But physicists and computer scientists are finding that they can harness it to crack problems that were long thought unsolvable. The resulting quantum computers--which may be available in some 15 to 20 years--will speed drug discovery, let forecasters nail the weather with precision, and help chipmakers design circuits that are now impossibly complex. (Unfortunately, they will also allow hackers to break codes protecting secure traffic on the Internet.)

Today's computers still solve problems the same way their ancestor, the Eniac, did back in 1945. They follow instructions, step by step. Their painstaking obedience has freed humans from tedious mathematical calculations, bringing changes in communication, entertainment, and scientific research that the Eniac's inventors never dreamed of.

But their dutifulness is also their undoing. Many important problems cannot be solved efficiently by following rules. For example, to find a name in an unalphabetized phone book, an ordinary computer would have to check each listing one at a time to see if it matches. With a large enough ''phone book''--say, a database of all possible combinations of human genes--this trial and error process could take centuries, even on the fastest supercomputer. That's because any ordinary computer has to represent each listing with a series of switches, each on or off, representing 1 or 0.

A quantum computer, on the other hand, could check all the listings at once, using quantum ''golf balls'' in place of switches. How? Let's call a ball a 1 if it's in the hole and a 0 if it's out. Because of the everywhere-at-once nature of quantum particles, we needn't make each ball either a 1 or a 0. Instead, we can assign it a set of probabilities: say, a 50% chance that it is in the hole and a 50% chance that it's not. With enough of these 50-50 quantum balls, we can represent all the names in our phone book simultaneously.

This is the input to our quantum computer--a billion or more names, each represented with equal probability. The computer's program nudges and shapes this ''cloud'' of probabilities, checking all the listings against the desired name at once. The quantum programmer's job is to manipulate the odds, loading the dice in a way that leads quickly to the correct answer.

Nabil Amer, manager of IBM's quantum computing effort, predicts that this kind of subatomic gambling will have huge payoffs for complex simulations. Engineers and drug designers will be able to ''just shut down the lab and do it on the computer,'' he says. Picture a database containing all the known rules for how chemicals interact. A quantum computer could sift through in an instant to find a molecule to fit any wish list of drug properties.

Actually building one of these devices, however, is a major challenge. And around the world, radically different designs are emerging. At the National Institute of Standards & Technology, scientists run quantum calculations by flashing brief pulses of laser light onto beryllium atoms that have been chilled to near absolute zero. Light particles bouncing between two mirrors act as the quantum golf balls in a California Institute of Technology experiment. And at NEC Corp.'s Fundamental Research Labs in Tsukuba, Japan, a tiny superconductor attached to an ordinary silicon chip generates 1-0 mixtures.

So far, the most successful design sits in Isaac Chuang's lab at IBM's Almaden Research Center in San Jose. A pencil-size glass tube filled with a yellow liquid containing millions of alanine molecules rests inside an NMR machine, a small version of the imaging machines found in hospitals. Each molecule is a tiny quantum computer, with its three carbon atoms serving as its working memory. To do a calculation, Chuang uses a series of NMR pulses to flip and jiggle the atoms' probabilities. A tenth of a second later, a final pulse forces the atoms to make up their minds and cough up an answer.

Chuang likes to joke that he has the world's largest functioning quantum computer, but he readily admits that it's still far from a useful size. With only three manipulable atoms per molecule, it can search a phone book containing only eight names. To search a bigger list, he needs a bigger molecule.

Researchers at Lucent Technologies' Bell Labs and Michigan State University have proposed a different way to solve that problem. They believe they can float any number of electrons on a layer of liquid helium. An orchestrated series of microwaves would run the electrons through their calculations.

Whether we want quantum computers or not, we may find that we need them. As ordinary computer circuits continue to shrink, eventually they'll reach quantum mechanical size, and their behavior will change radically. Many in the computer industry see this size limit--expected sometime around 2012--as a roadblock, but Chuang calls it the quantum mechanical pot of gold at the end of the rainbow. ''We're starting at the end,'' he muses, ''and walking backwards toward civilization.'' You can't get more quantum than that.


_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

21: The Toughest Problems Will Be Solved With a Roll of the Dice

PHOTO: Research Scientist Isaac Chang Holding Quantum Computer Liquid

ONLINE ORIGINAL: Relying on Mother Nature to Make a Quantum Computer

Quantum Computing Information

NMR Quantum Computing Home Page

E-Mail to Business Week Online

Copyright 1999, Bloomberg L.P.
Terms of Use   Privacy Policy