Ever since Sir Isaac Newton sent sunlight streaming through a prism, showing there was more to light than meets the eye, scientists have bent it, focused it, amplified it, explained it, even slowed it from a torrid 670 million miles an hour to bicycle speeds.
But never have they been able to stop it and bottle it for later use.
In separate experiments, two teams of scientists say they have caught light in high-tech traps, then released it again unchanged. During its pause, its information was imprinted on atoms in a target gas.
It's a bit like watching a bullet train vanish into a paper-thin sheet of gauze for a period, only to reappear out the other side at full speed with cars in order and people in their original seats, notes Eric Cornell, a physicist at the University of Colorado in Boulder.
The results, researchers say, represent a key step toward building "quantum" computers that use molecules or atoms to store information and carry out calculations that would choke a conventional supercomputer.
This newfound ability to make light pause then release it "is one of the decisive points to make quantum computing possible," says Lene Hau, a Harvard University physicist and leader of one of the teams formally reporting results this week and early next.
And if the gravitas of quantum computing fails to capture some imaginations, there is the gee-whiz aspect of stopping the fastest known phenomenon dead in its tracks.
"That's not a thing to discount," acknowledges David Phillips, a member of the second team, "these are fun experiments to do."
Dr. Phillips, a physicist at the Harvard-Smithsonian Center for Astrophysics (CFA) in Cambridge, Mass., says his team's work came from research into making more-accurate atomic clocks. The clocks help astronomers synchronize and process signals coming in from radio-telescopes around the world observing the same object at the same time.
"One of our latest projects was to design a new set of rubidium clocks as small and as stable as existing clocks. The equipment you need to study the clocks is about equal to what you'd need to stop light," he says.
Last May, CFA physicist Mikhail Lukin proposed a way that the team could use the equipment to stop light.
The team filled a 3-in. by 3/4-in. glass tube with rubidium vapor and helium, then aimed a specially tuned laser at the target. The team first hit the target with a "control" burst that altered the rubidium atoms so they would not absorb the light in the manner they typically do. Then they fired the "signal" pulse that contained information they wanted to store.
The signal pulse lasted about 10 to 30 microseconds - long enough for the beam to stretch 2 kilometers if it had been shining in space. When the beam hit the rubidium, it was as though it hit light's equivalent of molasses. Its speed dropped to about 2,000 miles an hour. Meanwhile, the tail of the beam kept coming, compressing the signal beam within the rubidium cloud.
Then the team smoothly dimmed the control beam, shutting it down in about 3 microseconds. While the control beam's tail kept coming, the beam inside the rubidium effectively vanished - its information stored by uniform changes it made in a property of the atoms known as spin. When the team blasted the rubidium again with the control beam, the atoms released the information as a pulse of light identical to the pulse that entered the rubidium trap.
The portion of the beam trapped in the rubidium was squeezed into a length a hundred thousand times shorter than its free-space length, the team estimates.
"We were able to confine the beam for hundreds of microseconds," Phillips says.
Meanwhile, across town, Dr. Hau and colleagues at Harvard and the Rowland Institute for Science took a similar tack, but with important differences. Where the CFA experiment used a gas at temperatures near the boiling point of water, Hau's team used lasers and magnetic fields to chill sodium vapor to within 1 millionth of a degree above absolute zero.
At that temperature, the atoms virtually freeze in place, eliminating the "noise" their motions can generate, she says. Holding the sodium at such frigid temperatures also slowed the incoming beam to bicycle speeds, instead of Concorde airliner speeds. When Hau's team turned off its control laser, the entire 2-kilometer signal beam found itself mired in a layer of atomic molasses only a thousandth of an inch thick.
Hau's group was able to keep the light - and more important, its information - bottled up five times longer than the CFA team. Moreover, they also demonstrated that they could release partial chunks of information from its "atomic memory" by sending pulsed "release" commands from the control laser.
"This really opens so many doors," Hau says.
In a quantum computer, "you'd like to have some sort of central processing unit," she notes. Light itself doesn't look like a promising candidate for a CPU because light fields don't interact with each other readily. Atoms, on the other hand, do readily interact, allowing for arrays of "gates" that would steer data and conduct calculations. The two Cambridge experiments point to a way to send information to the atomic CPU and retrieve it again. The approach also may find uses as optical switches for a quantum-computing Internet.
Hau says her team's goal is to combine what they have learned with nanofabrication techniques to reduce tabletop atomic memory to a single computer chip. Her work, which appears in today's issue of the journal Nature, and the CFA team's efforts, which will be published on Monday in Physical Review Letters, represent the latest in a wide range of experiments that could open the door to quantum computing. (Quantum computing would rely on the rules of quantum mechanics to process information, rather than on classical electronics, which undergirds today's computers.)
The interest in quantum computing lies in part in the drive to pack more computing power into smaller packages. At some point, experts say, individual logic devices will be built out of a few atoms. Beyond the notions of smaller and faster, however, lies the belief that working in the quantum world would allow for a new type of computing, complete with new types of programming approaches that, for some problems, would leave classical computers spinning their hard drives in futility.
For example, trying to establish all the possible factors for a 400-digit number would take today's fastest supercomputer about a billion years, according to Neil Gershenfeld, a physicist with the Media Lab at the Massachusetts Institute of Technology, in Cambridge. Yet, he estimates, a quantum computer could solve the problem in about a year.
Moreover, quantum computers are thought to be uniquely suited to solving problems in quantum mechanics - problems that could stymie efforts to build ever smaller chips using classical microchip-making techniques.
"What we've demonstrated is the storage of classical information," Phillips says. "Next, we want to demonstrate the storage of quantum information."
The theory that propelled his team into trying their experiment, he says, indicates that the same approach should be capable of reaching this next rung on the quantum-computing ladder.
(c) Copyright 2001. The Christian Science Publishing Society