Portland, Ore. - In an experiment at the Ecole Polytechnique Federale de Lausanne, researchers have been able to speed up or slow down light transmitted through an optical fiber with precise control. This capability could have application in telecommunications, optical switching and optical computing.
The experiment, which uses stimulated Brillouin scattering, verifies a prediction of Arnold Sommerfeld and Leon Brillouin, who early in the 20th century theorized that narrowband amplification at a sharp spectral transition could enable the speed of light to become variable.
"We are reporting the first demonstration of a wide optical control of the signal velocity in an optical fiber," said Luc Thevenaz, who leads the group. "The starting and ending points of a pulse carry the information and those still propagate at the normal velocity, but the peak of those same pulses propagates at a variable group velocity that can exceed c [the speed of light in a vacuum]. This is what we experimentally demonstrated and observed."
Thevenaz performed the work with EEs Miguel Gonzalez-Herraez at the University of Alcala (Madrid, Spain) and Kwang-Yong Song at the University of Tokyo (Japan).
The ability to slow light down at a controlled rate would be a powerful tool for designing telecom components like all-optical routers. In current dense wavelength-division multiplexing systems, companies can route light channels optically by using the wavelength in much the same way that a prism will route wavelengths to slightly different positions. But the packet routing required by Internet addressing requires knowledge of the information content of a signal, which is much more difficult to access. The hope is that by slowing light down in the fiber, this type of processing could be performed entirely in the optical domain. Currently, the optical signal has to be converted into electronic form and then processed with electronic circuits.
The experiment enables a single-mode fiber-optic cable to act as if it were of variable length, supplying one of the missing pieces for future all-optical computers and communications. In its program, called Applications of Slow Light in Optical Fibers, the U.S. Defense Advanced Research Projects Agency says that such an achievement is strategically important.
Although there is no known technological application for speeding up light beyond its theoretical limit of approximately 186,000 miles per second, Thevenaz's experiment has generated some controversy, since it appears to violate the basic principle of relativity, which states that nothing in the universe can travel faster than light.
Despite the variable speed of the light in the fiber, Thevenaz pointed out that information still travels at the speed of light. The reasoning behind that is based on the fact that the sharp discontinuities required to toggle a bit of information require a wide bandwidth and thus cannot be entirely contained in Brillouin spectral transitions, which are too narrowband to transmit information faster than c.
"We modified the group velocity so much that it was infinite-that is, the peak of the pulse exits the fiber at the same time it enters it-and even negative velocity was achieved-that is, the peak of the pulse exited the fiber before entering it," said Thevenaz. While this appears to violate basic commonsense notions of cause and effect, the effect was predicted by the Brillouin and Sommerfeld theory and no further theoretical revisions are required.
Other groups have achieved control over the speed of light, but only in special physical systems like ultracold atomic gases. Thevenaz's team used a room-temperature, off-the-shelf two-meter fiber-optic cable in its demonstration, however. The cable appeared to change in length from 3.8 meters down to a paradoxical -3 meters in length, due to changes in the speed at which light was propagated.