A rchive Date
[ 22-01-2001 ]
Category
[ Science ]
sub-Categoy
[ Physics ]
|
[Dream The Quantum Particle Dream
Matthew Herper
If the computers that you build are quantum, Then spies of all factions will want 'em. Our codes will all fail, And they'll read our e-mail, Till we get crypto that's quantum, and daunt 'em. --Peter Shor, quantum computing guru, with wife Jennifer Shor
NEW YORK - Twenty years ago, quantum computers seemed like an impossible dream. The devices, which use the interactions of tiny particles to do their calculations, existed only in physicists' imaginations. Then, in 1994, AT&T researcher Peter Shor realized something startling: A quantum computer could crack any of the encryption codes used to protect information on computer networks.
That discovery got scientists' attention. But six years later, quantum computers are still no more powerful than pocket calculators. Now, two scientific papers, one published in the Jan. 4 issue of Nature and another that has been distributed over the Internet, have opened up a new avenue for making the arcane devices. The idea is a veritable light bulb: It would build quantum computers that do their calculations using particles of light called photons.
A digital computer is a series of switches, or bits, that can represent "on" or "off" and "0" or "1." Quantum computers are also made up of switches, called qubits, which also represent "0" or "1." But qubits take advantage of one of the more counterintuitive ideas in modern science: that a subatomic particle, like a photon or an atom's nucleus, can be in more than one state at once. In essence, a qubit is a little bit "0" and a little bit "1."
What this means is that a quantum computer can perform a single mathematical function on a whole group of numbers in one step, answering different versions of the same question simultaneously. And as more qubits are added, the computer's power increases exponentially: A 40-qubit computer would be comparable to the ASCI White Supercomputer at Los Alamos National Labs. To run Shor's algorithm, you'd need 1,000 qubits.
The most powerful quantum computer around clocks in at 7 qubits; it was built at Los Alamos National Laboratories by Raymond Laflamme, who is also one of the co-authors on the Jan. 4 Nature paper, along with Los Alamos cohort Emanuel Knill and Gerard J. Milburn of the Centre for Quantum Computer Technology at Australia's University of Queensland.
Part of the problem with quantum computers, like Laflamme's 7-qubit model, is that the methods used to make and run them are so unwieldy. The 7-qubit computer is an organic molecule; the nuclear spins in its component atoms serve as processors. Manipulating and measuring these atoms requires roomfuls of equipment. What makes this bright idea so exciting is that it may allow scientists to build quantum computers out of much less outlandish components.
"The orthodoxy has always been that to get single photons to interact nonlinearly you need to put them in extreme environments," says Andrew White, a researcher at Queensland's Centre for Quantum Computer Technology. "Now, using [relatively cheap] almost 'Wal-Mart' technology, you can get similar interactions."
Milburn, Laflamme and Knill's new approach is to send single photons through a setup that encourages them to interact in a quantum way and detect the single photons coming out the other end. Knill says that this might fail as much as 75% of the time. But if it fails, the researchers can tell and throw it out.
It's what Milburn calls a "gamble." The scientists know the odds, and they can rig the game so they always come out ahead, with a working quantum computer. But how close are we to implementing this brave new idea?
Two vital components are missing: a device that can reliably produce single photons and one that can reliably detect them. Similar devices are available, although they don't work as well as they need to for quantum computing.
This factor--call it imaginability--may be lacking in the other proposal to use light for quantum computing written by John Preskill of Caltech and Daniel Gottesman, a fellow at the University of California, Berkeley. It uses light at room temperature, encoding information in the light's electromagnetic field, but may be more difficult to test. "This is the sort of thing that you can't find out whether it's harder or easier until you've actually done it," Gottesman says.
Queensland's White is already looking at putting the Nature paper into action. He just got a grant of more than $170,000 to start working on that experiment. He thinks he can at least prove that the idea can work within the next year, using photon detectors made by Perkin-Elmer (nyse: PKI). Rockwell Instruments (nyse: ROK) makes detectors that might better fit the job, he says, but those are not available outside the United States for security reasons.
Experiments like White's are exactly what are needed, says AT&T's (nyse: T) Peter Shor as he peruses the paper. "This is the first paper with this idea," he cautions, "and there's going to be a lot more work needed before you figure out whether this idea is feasible."]
|