The Slow-Light Race Is On

Marie Freebody, Contributing Editor, marie.freebody@photonics.com

Since the notion of slow light first became a reality more than a decade ago, scientists have been exploring its use in fundamental studies of light-atom systems as well as for long-term applications in all-optical data processing, quantum communication and sensors.

When light is slowed down, it is forced to interact more strongly with the confining material. This is key to devices in sensing, quantum information and all-optical signal processing for optical communication. There has been recent progress in exploiting slow light for nonlinear signal processing, such as to demonstrate enhanced four-wave mixing, and in observing striking nonlinearities, including third-harmonic generation in silicon.

This picture illustrates four-wave mixing, a key nonlinear process that is enhanced by slow light. Courtesy of the University of St. Andrews.

“The field has been studied theoretically for about 100 years; however, it wasn’t until 1999, when [Zachary] Dutton, [Stephen E.] Harris and [Lene Vestergaard] Hau demonstrated slow light in a BEC [Bose-Einstein condensate], that it became popular,” said professor John C. Howell of the University of Rochester in New York. “The reason for the popularity was the magnitude of the slowing. When we think about the absurdity of the speed of light – 300 million meters per second – it is difficult to wrap one’s brain around. However, they were able to slow a pulse of light down to about 17 m/s, albeit only for about 100 µm of propagation distance.”

As news of the accomplishment spread, a new field of optics was born. For the first time, light was transformed to a speed in the realm of human comprehension. The race then was on to see which materials provided the best slow-down capability.

Slow light can be achieved by using material resonances; for example, via electromagnetically induced transparency or stimulated scattering effects such as Brillouin or Raman. Recently, some studies have indicated that stimulated Brillouin scattering in silica optical fibers may not provide the desired enhanced light-matter interaction.

A more promising approach, perhaps, is to slow down light using the resonances created by microstructures such as Bragg gratings and photonic crystals. This type of enhancement is similar to that provided by resonant cavities. Importantly, it offers enhancement over a large bandwidth, enabling slow-light waveguides to process high-bandwidth signals.

Schematic of slow-light propagation of pulses in a photonic crystal waveguide. Energy density of pulses in the slow mode is increased and, along with distributed feedback effects, enhances interactions with the material of the photonic crystal waveguide. (Inset) Scanning electron microscopy micrograph of a nanowire coupled to a silicon photonic crystal waveguide. Courtesy of the University of Sydney.

On April 12, 2010, Optics Express published a paper detailing a major advance in photonic crystal engineering that enabled a record sixteenfold increase in processing speed for a silicon device. The authors believe that this improvement has opened the door for slow light to play a key role in ultrahigh-bandwidth telecommunications systems.

Professor Thomas F. Krauss and his team look at slow-light samples on the optical setup. Courtesy of the University of St. Andrews.

Professor Benjamin J. Eggleton, co-author of the paper, is a pioneering scientist in harnessing slow light for telecommunications systems and is director of the Centre for Ultra-high Bandwidth Devices for Optical Systems ARC Centre of Excellence at the University of Sydney in New South Wales, Australia.

“The goal of all photonics researchers is to control the flow of light,” Eggleton said. “Currently, we have tools to control the direction light moves in, and the next big challenge is controlling the speed of light. If we can slow light down, then we have another degree of freedom that we can use to control light.”

Green-light emission as seen by the naked eye. The green light is created from C-band IR light through third-harmonic generation, occurring here in a silicon photonic crystal chip. The red stages are for fiber coupling to the device. Courtesy of the University of Sydney.

The photonic crystal structure used by Eggleton and his colleagues was designed and fabricated by researchers at the University of St. Andrews in the UK, led by professor Thomas F. Krauss. Their special dispersion engineering technique led to broadband slow light, breaking a fundamental barrier in using slow light to effect high-bandwidth signals. The team successfully demonstrated a propagation of 640-Gb/s signals through a photonic crystal chip.

In an earlier example of slow-light-enhanced nonlinearities in photonic crystal structures, the Sydney and St. Andrews researchers reported the observation of green-light emission via third-harmonic generation. As reported online in Nature Photonics on March 22, 2009, green light is ejected out of a slow-light silicon photonic crystal structure, which can be observed easily with the human eye.

Schematic of green third-harmonic emission from a broadband IR pulse in a photonic crystal waveguide. The emission of the green light is mediated by the photonic crystal and directed in certain angles as dictated by phase matching. Courtesy of the University of Sydney.

“As well as being of fundamental significance, this scheme offers a very elegant and compact method for monitoring transmission impairments on optical data streams simply by placing a low-cost, slow detector on the silicon chip,” Eggleton said.

Slow light optical buffers

When slow light first captured the imagination of scientists in 1999, research was geared toward optical buffering. Optical buffers are a key component to overcoming data packet contentions in switches and routers.

Suppose two signals on two different lines must be routed to the same output line. If they arrive at the same time, either one or both of the packets will be rejected. If one of the optical packets could be slowed down until the other packet has been successfully put on the line, then both signals could make it onto the appropriate output line.

“Currently, optical buffers require conversion to electrical signals, computer buffering and then reconversion to optical signals,” Howell explained. “From a bottom line, one wishes to reduce the size, power and cost requirements of the buffers needed to perform this task.”

It turns out, however, that slow light buffering is very hard to achieve, whereas electronics already takes care of this very well. “There have been a number of impressive demonstrations [of slow-light buffering], but they are either incompatible with integration or have other practical limitations,” Eggleton said.

Going one step further and stopping light is a key pursuit for the quantum information processing community. Stopping and storing light is one of the holy grails of photonics and will enable future on-chip, all-optical memories.

For Krauss, professor of optoelectronics at the University of St. Andrews, all-optical memories will not be able to compete with gigabit-size electronic memories. Instead, he envisages future optical buses on multiprocessor chips requiring the storage of a few bits of information to avoid contention between processors.

In more fundamental studies, metamaterials are creating excitement as a new platform for photonics. Metamaterials may have a negative refractive index, which means that they can be arranged to slow down and stop light in its tracks by reversing the direction of its phase.

Group index change, with wavelength for dispersion-engineered photonic crystal waveguides, was produced by the Krauss group at the University of St. Andrews. The ‘s’ values are shifts off the photonic crystal lattice of rows of holes in nanometers. In this case, the lattice period is ~410 nm. The image was reproduced from Monat et al (January 2010), IEEE J. Selected Topics in Quantum Electronics, Vol. 16, No. 1, p. 334.

Scientists such as Ortwin Hess, professor and Leverhulme chair in metamaterials at Imperial College London, are investigating the physics of stopped light and advancing the science that may lead to the much-publicized invisibility cloak.

At this stage, achieving the shielding ability of Harry Potter’s cloak remains a theory, but Hess recently met with Andrea Di Falco at St. Andrews to discuss a recently demonstrated flexible metamaterial that could become a building block for an invisibility cloak.

Much of today’s slow-light research lies in enhancing light-matter interactions. Apart from the potential for all-optical signal processing, enhanced light-matter interactions could boost the efficiency of nanolasers, frequency conversion schemes, quantum information systems and sensors. One issue with using slow light in sensing applications, however, is finding a low-cost solution.

Currently at a loss

The biggest challenge facing today’s slow-light community is overcoming the limitation of loss. Losses go hand in hand with slowing light: As light-matter interactions are enhanced, so, too, are losses. In particular, losses in photonic crystals can increase significantly with reduced velocity, limiting the applications of such waveguides in all-optical devices.

“Most applications are now limited by losses in the slow-light regime,” Krauss said. “Even though our understanding of the nature of losses has improved dramatically and already resulted in substantial improvements that have led to the nonlinear results mentioned above, more needs to be done.”

However, Eggleton points to recent studies in fabrication-induced propagation losses conducted at St. Andrews that hint at methods that may be able to reduce losses associated with slow modes.

A problem that threatens slow-light research in the US concerns another type of loss: funding. For the past five years, DARPA has provided strong support for slow-light research, but Howell believes that this could be coming to an end. “The European community is really the only group that is getting significant funding. I think it will mostly die out in the US, unless some new initiative comes through or unless there is another big ‘breakthrough’ that captures the attention of the broader public again.”

It is always difficult to predict when the next leap in technology will take place. The most tangible applications may be in biosensors, where the enhanced-light-matter interaction provided by slowing down the light may result in exquisitely sensitive devices, which Krauss estimates could become prototypes in three to five years.

Although current state-of-the-art results have been achieved in photonic crystal structures, these have typically been fabricated using electron-beam-lithography processes. The drawback is that these methods do not scale easily to low-cost production. Advanced photolithography or nanoimprint lithography could offer a solution, Eggleton said.

“High-end applications in quantum information might provide a niche without too much advancement in fabrication technology,” he added. “Applications in sensing and communications are likely to require cheap methods of mass production before slow light can make an impact.”