Luring Light From the Straight and Narrow

Like a desert mirage, a small box with a wire and some sodium atoms can make light curve. This new device, described in the current issue of Physical Review Letters, could someday lead to a new kind of switch for optical computer chips.

Researchers at the University of Otago in New Zealand and the University of Zurich in Switzerland passed laser light through a special filter that polarizes it circularly, making its electric field spin like a propeller. Circularly polarized light carries angular momentum, as if the photons were spinning in space. When such a photon collides with a sodium atom, the sodium's own peculiar properties allow it to absorb the photon's angular momentum. But after soaking up the twisting of one photon, an atom can't absorb any more and becomes "bleached"--invisible to the laser light. "If the atom is not able to accept angular momentum from the photon, it doesn't interact," says Otago laser physicist Weston Sandle.

The group figured out how to exploit this bleaching phenomenon to bend light. In a cell filled with sodium vapor, they beamed a laser along a gold-plated tungsten wire carrying a current, which created a magnetic field. The magnetic field drew angular-momentum energy from nearby sodium atoms. As a result, atoms close to the wire were more likely to interact repeatedly with photons than were ones further away. These interactions alter the light's refractive index, essentially slowing it down; the light's wave front bends away from these "roadblocks." The team managed to bend a laser beam 30 micrometers out of true by having it travel along a 3.5-centimeter-long wire.

The ability to bend light on demand could lead to new kinds of switches and routers for optical chips. But such a device is a long way from the market. "It all comes down to cost and system integration," says Richart Slusher, a physicist at Lucent Technologies' Bell Labs in Murray Hill, New Jersey. Sodium vapor is not the easiest medium to work with; a solid-state crystal with the same talent would be easier to integrate into a circuit. And the need to maintain a magnetic field could make the device too expensive to be practical, says Slusher. However, he adds, "Who can say down the line?"

Posted in Math, Physics