Light-storing chip charted

For years, researchers have been striving to make high-speed, low-power chips that channel light rather than electricity, but finding ways to briefly store light pulses has proved extremely challenging.

Recently, researchers have stored light pulses for fractions of a second in hot gases, extremely cold gases or crystal doped with special metal. But these techniques are challenging to carry out, and would be difficult or impossible to configure into more practical chip form.

Researchers at Stanford University have come up with a scheme to store light pulses under ordinary conditions using photonic crystal -- semiconductor chips that contain regularly spaced holes or rods of a different material. "Our discovery enables quantum coherent storage of light pulses on a microchip about the size of the grain of salt," said Mehmet Fatih Yanik, a research assistant at Stanford University.

The scheme could lead to inexpensive chips that power all-optical communications switches, quantum computers and quantum communications devices. "Operating wavelengths[and] bandwidths... can simply be designed by standard lithographic techniques used in conventional microchip technologies," said Yanik.

The method would allow light pulses to be stored in microchips at room temperature without requiring any special light-matter interactions, Yanik said.

The researchers' findings run counter to the conventional wisdom that devices using optical resonators -- tiny structures that vibrate at light frequencies -- can do no more than slow light by a limited amount. In one type of device, for example, light pulses at the telecommunications wavelength of 1.55 microns and a rate of 10 gigabits per second can be slowed to no less than one hundredth the speed of light in a vacuum, said Yanik.

The key to the researchers' method is a technique that allows them to change -- on-the-fly -- the way portions of the photonic crystal respond to light. "We discovered a practical way to compress light's bandwidth by an unlimited amount... using conventional optoelectronics technologies at speeds sufficient to prevent light pulses [from] passing through our system," said Yanik.

The researchers' simulation shows that light pulses can be slowed to less than 10 centimeters per second, slow enough that the pulses would be held essentially in place for tiny fractions of a second, according to Yanik. This is long enough to make pulses interact to switch light signals for high-speed communications or link photons for quantum computing.

The researchers' light-controlling chip design calls for photonic crystal that contains a series of optical resonators, or cavities. Photonic crystal refracts, or bends, light -- the same effect that produces the familiar bent-drinking-straw illusion. The boundaries made by photonic crystal's holes or rods refract light, and the spacing of these gaps determines the degree to which a given wavelength of light is bent. Photonic crystal can be designed to block or channel specific wavelengths.

In the researchers' design, one series of cavities forms a straight waveguide that allows light pulses to pass through the device. Each cavity in the waveguide is attached to a side cavity that connects to a second side cavity.

The chip would briefly trap a pulse by changing the microcavities' resonant frequencies. Tuning the waveguide to resonate at the same frequency as the light pulse and at the same time keeping the side cavities out of tune would allow the pulse to enter the device. Once the pulse is inside the device, the waveguide would be gradually -- though at very high speed -- detuned while the side cavities were tuned to the pulse frequency. This would shunt the pulse into the side cavities. Reversing the tuning-detuning process would release the pulse into the waveguide, allowing it to continue on its way through the device.

Key to the method is a way to tune the refractive index of the photonic crystal in a way that preserves the shape of the pulse. Light pulses contain multiple wavelengths, and the wavelengths bend to different degrees as pulses travel through matter. This disperses the wavelengths, causing light pulses to spread out, which limits the distance they can travel through a material. Wavelength dispersion also limits the amount light pulses can be slowed, because they can spread only so much before they disappear.

The researchers' technique tunes a device's refractive index in a way that lowers the frequency of all of the pulse's wavelengths consistently, preserving the pulse.

A set of 120 microcavities whose tunings change at a maximum rate of one gigahertz is sufficient to store and release a light pulse, according to Yanik. Multiple light pulses could be stored simultaneously in the device, and specific pulses could be released on demand, he said.

The researchers' scheme could also applied to other systems that involve resonance, said Yanik. It could be used to slow and store microwave signals and ultrasound waves, and possibly detect gravitational waves, he said.

The technique is an advance over previous work on stopped light because it uses microscopic optical cavities rather than atoms, said Raymond Chiao, a professor of physics at the University of California at Berkeley. "This allows much larger bandwidths of light to be stopped."

The work would have been more impressive had the authors demonstrated the stopping of light experimentally, he added.

The researchers are aiming to demonstrate their technique by trapping microwave signals. A demonstration should take place within a year, and a practical prototype that works at optical frequencies could be made in two to five years, said Yanik.

Yanik research colleague was Shanhui Fan. The work is slated for publication in Physical Review Letters. The research was funded by the National Science Foundation (NSF) and Stanford University.

Timeline:   2-5 years
Funding:   Government, University
TRN Categories:  Optical Computing, Optoelectronics and Photonics
Story Type:   News
Related Elements:  Technical paper, "Stopping Light All-Optically," posted at the arXiv physics archive at arxiv.org/abs/quant-ph/0312027