Catching errors in scrambled quantum bits.

When the noise of a party threatens to drown out a conversation, the speakers typically respond by talking more loudly or slowly or by repeating key phrases. Engineers have developed roughly analogous strategies to maintain the integrity of digital information exposed to static and other types of environmental noise, whether these bits are transmitted as a radio signal or stored in a computer or on a compact disc.

Now, researchers have developed techniques for correcting or avoiding errors when transmitting or storing quantum states such as electron spins or photon polarization angles. As a step toward superfast computation, these methods can evade the usual limitation that observations inevitably disturb the states of quantum systems.

"It's now known that you can send quantum information reliably through a noisy channel," says Charles H. Bennett of the IBM Thomas J. Watson Research Center in Yorktown Heights, N.Y. "In fact, there are two, somewhat distinct approaches for solving this problem."

The advance represents an important step toward the development of quantum computers, in which the binary logic of 0s and 1s used in conventional computers is replaced by computing elements that follow the laws of quantum mechanics (SN: 1/14/95, p. 30). Called a quantum bit, or qubit, such an element has not only a value but also a phase, which reflects the wavelike aspect of a quantum system.

In principle, quantum computers can solve certain mathematical problems, such as factoring whole numbers (SN: 5/14/94, p. 308), in dramatically fewer steps than conventional computers.

The development of quantum computers is stymied at present by the fact that no quantum system can be perfectly isolated from the rest of the world. Individual components, therefore, can't be trusted to maintain their states over long periods of time.

Moreover, because any measurement of a quantum state disturbs it, determining the values of the bits in a quantum computer would preclude knowing the phases of those bits and vice versa. Thus, engineers can't use conventional schemes, which rely on some sort of redundancy-for example, making and comparing copies of digital data-to detect and correct errors (SN: 3/12/94, p. 170).

To avoid this problem, Peter W. Shor and his colleagues at AT&T Bell Laboratories in Murray Hill, N.J., have invented an error correction method in which each original quantum bit of data is encoded in a particular way as nine quantum bits. If one of these bits is somehow spoiled during transmission, it's possible to recover the original quantum bit from the information that does get through.

That encoding operation can be imagined as shifting the original quantum bit into a protected position so that it leaks only a minimal amount of information to the environment, Bennett comments.

Bennett and his collaborators have developed an alternative strategy based on quantum teleportation (SN: 4/10/93, p. 229). This technique exploits a quantum phenomenon in which a single process within an atom simultaneously generates two photons that travel in opposite directions. According to quantum theory, neither photon has a particular electric field orientation, or polarization, until it's detected.

This measurement transforms a photon's polarization from a range of possibilities into a specific, random value. Astonishingly, measuring one photon's polarization causes the other photon to acquire the opposite polarization at the same instant-even if the second photon is at the other end of a room or across the galaxy.

Such special pairs of photons can be used to convey quantum information, such as a photon's quantum state.

Bennett and his coworkers have now shown that it's possible to send these photons through a noisy channel, which corrupts their quantum states, and still recover useful information. Applying quantum operations, one can "purify" the many corrupted photons to obtain a few "good" ones, which can be used as information carriers, Bennett says. The researchers describe this approach in a paper accepted for publication in Physical Review Letters.

"This doesn't solve the full problem," Bennett notes. "We can now protect [quantum information] while it's being stored and while it's being transmitted, but we don't know how to protect it while it's being processed."

Quantum computation also awaits the development of components for implementing its peculiar logic.