The Limits of Superluminal Propagation

Last year, L.J. Wang and his colleagues at the NEC Institute reported that a composite wave pulse traveled with little distortion through a medium at a group velocity faster than c, without violating Einstein's theory of relativity, or the notion that cause precedes effect. (Update 495).

Sent into a chamber of specially prepared cesium atoms, the light pulse exited the chamber before the peak of the input pulse entered it. This can happen because the early part of the pulse, made of many component waves, contains all of the information in the wave. Once inside the chamber, the pulse is rearranged such that the peak reappears at a position a little farther ahead in the chamber. This causes the composite pulse to emerge from the chamber earlier than if it had been traveling through the chamber at the speed of c. Potential applications involve the possibility of shuttling along light waves faster in applications such as telecommunications and computers.

How to define and analyze the speed of signal transfer in that setup is a subject of a new paper by the same researchers, along with two other physicists: Peter Milonni of Los Alamos and Raymond Chiao of UC Berkeley (chiao@physics.berkeley.edu). They consider the effect that quantum noise, due in part to random spontaneous emission by the medium, has on the reliability with which a signal can be measured. The more one tries to push along the signal in the medium, the greater the number of noise-producing, signal-obscuring spontaneous emissions that occur, and any attempt to boost the signal's intensity to make it more detectable introduces delays such that the signal velocity always ends up to be less than c. Therefore, the signal velocity is defined operationally as an optical signal-to-noise ratio.

In summary, the researchers extended the special relativity speed limit of c for sharp wavefronts (which act like "on-off" signals), to that of a more realistic smoothly varying signal, based on a speed limit set by quantum fluctuations. (A Kuzmich et al., Phys. Rev. Lett., 30 April 2001; text at Physics News Select.)