Once again, physicists have done the seemingly impossible--broken the light-speed barrier. Stranger still, the light pulse exits a box before entering it. But the trick, although baffling, doesn't upset any physics dogma.
For several years, scientists have been making light travel faster than light speed, dubbed c. This hasn't broken the laws of relativity; Einstein posited that information can't be transmitted faster than c. These studies don't contradict him: Previous faster-than-light demonstrations aren't able to send messages at "superluminal" speeds (ScienceNOW, 1 June). And this experiment, described in the 20 July issue of Nature, is no exception.
To boost light beyond its apparent limits, physicist Lijun Wang of NEC Research Institute in Princeton and colleagues set up a 6-centimeter chamber of cesium gas that has a peculiar property: slightly different wavelengths of light go through at very different speeds. This property is important because a pulse of light can be thought of as many superimposed light waves of different frequencies that extend throughout space. The pulse is merely the region where all of the waves reinforce each other, rather than canceling each other out. That region travels along as the waves propagate--and scientists observe the pulse moving at the speed of light. When the light is shined through the cesium cell, however--because of the strange dispersion of light waves within the cell--the cell changes the relative speeds of the component waves. This, in turn, messes up the established wave cancellation. Instead of having just a single region where the waves reinforce--the pulse itself--the waves get reset by the cell and create a second region beyond the cell where the waves reinforce again: It is a duplicate of the original pulse. So, to the observer, a copy of the original pulse is coming out of the cell 62 nanoseconds before the pulse even enters.
According to Wang, this effect, though bizarre, can't transmit information faster than the speed of light, in part because the duplication of a finite pulse is imperfect. Physicist Aephraim Steinberg at the University of Toronto hopes that superluminal tricks might speed up computer circuits, although he admits such applications are currently just a dream.