Cutting the energy of communicating bits.

A century ago, Morse code represented a quick, efficient way of sending information over long distances. A telegraph operator tapped a key that opened and closed a switch to generate a string of electric pulses encoding the news, which was then transmitted over wires to its destination. Nowadays, digital communication involves analogous signals representing 1s and 0s, or bits, to convey information.

All of these forms of communication require a certain amount of energy to produce the electric pulses, radio signals, or photons that transmit an intelligible message.

Information handling invariably involves something physical-whether it's generating an electric pulse or a photon, marking a piece of paper, punching a hole in a card, or magnetizing a spot on a spinning disk-and must therefore conform to the laws of physics.

Rolf Landauer of the IBM Thomas J. Watson Research Center in Yorktown Heights, N.Y., has looked closely at whether the laws of physics mandate that sending a bit requires a certain minimal amount of energy. In the June 28 Science, he concludes that there is no such requirement, and he suggests alternative communication methods that, in theory, offer the possibility of transmitting information without dissipating energy.

"My message is that energy and matter used in sending information need not be thrown away at the receiving end," Landauer says. "They can, in principle, be recycled."

In the usual view of information transfer, researchers typically have argued that the minimum energy needed to send a bit is roughly equal to the energy of motion of an electron or molecule at the given temperature. Otherwise, the signal, whether electric pulses in a wire or photons in an optical fiber, gets drowned out by the background noise of jiggling molecules, atoms, or electrons of any material.

However, these limits don't necessarily apply to all conceivable forms of communication, Landauer insists. For instance, a great amount of information can be stored on a floppy disk and carried from one place to another, demanding considerably less energy expenditure per bit than if those bits travel through a wire.

Landauer has worked out more detailed schemes for moving bits that involve very little consumption of energy. He likens one of his proposed communication links to a ski lift-on a quantum scale. Instead of chairs, it has moving compartments, or potential wells. Adjustable vertical barriers divide these wells in half. A particle loaded into, say, the left half of a well corresponds to 0 and into the right half to 1.

In the downhill loading step, an incoming bit (0 or 1) waits in its own well to meet a well returning from the top of the ski lift. This returning well is initially in the 0 state. If the incoming bit is 1, its well induces the barrier in the returning well to lower temporarily and forces the particle in the returning well to pass from the left to the right half. Then, the original incoming well is reset so the value of its bit is 0.

The freshly loaded well next travels back up the ski lift to the unloading area, where its bit is copied to a receiving well. The unloaded well is reset to 0 and returns downhill to repeat the information transfer cycle.

In principle, each of the operations involved in this scenario and in his other proposals involves no expenditure of energy, Landauer argues. Instead of getting thrown away, the bits are, in effect, recycled. Moreover, the basic setup does not require that the signals overcome background noise.

"These are not practical schemes," Landauer admits. They may, however, stimulate others to invent communication systems that make better use of energy than those now available.

For example, it may be possible to use the timing or angle of polarization of photons to achieve an effect similar to that underlying Landauer's quantum ski lift model of a communication link.

Even if his theoretical models are unachievable in practice, Landauer notes, it is still worthwhile knowing that no physical law prohibits extremely low energy communication.