Slow light on a Silicon Chip - What’s the Limit?

The information bandwidth of lightwave is much higher than today’s electronic information technology. Processing information on lightwave thus has a significant advantage. Temporarily slowing down light on silicon chip allows us to complete the information processing in a small chip before light rushes off the chip. However, significant loss of light intensity occurs as light slows down. This fundamentally limits our capability in optical processing information on a small chip. In the December 15 issue of the journal Physical Review B , Assistant Professor Wei Jiang in Rutgers electrical and computer engineering department elucidates the fundamental mechanism and limit behind such light intensity loss. This could help develop next-generation optical information processing technology on a silicon chip.

Slowing down light is an intriguing topic in optics for decades. In the past, the slow light effect is obtained in bulky, low-temperature apparatuses and/or expensive materials. In the last decade, periodic structures, so-called photonic crystal waveguides, emerged to offer slow light on a compact, inexpensive silicon chip. Photonic crystal waveguides can be extended to a relatively long distance to achieve both a relatively long delay time and a wide bandwidth, while most other slow light approaches are limited to either a narrow bandwidth or a short delay. However, the optical loss due to random scattering from small bumps and dents on a silicon chip fundamentally limits the capability of photonic crystal waveguides.

Prior experimental work on optical loss in slow-light photonic crystal waveguides showed large variation. Prior numerical simulations of optical loss were limited to one or a few instance of structures with specific parameters and were unable to account for the variation. Prof. Jiang has developed an analytic theory that reveals the general characteristics of optical loss in a photonic crystal waveguide over a wide range of parameters. The theory indicates that that spatial phase and polarization variation may hold the key to optical loss reduction. Furthermore, Prof. Jiang and graduate student Weiwei Song have developed numeric code that enables efficient, accurate optical loss simulations over a large parameter range. Currently, Song is using this code to search for a low-loss photonic crystal waveguide design.

Once loss is reduced, slow-light photonic crystal waveguides will find a broad range of applications in optical signal processing and optical delay lines for phased array antennas. The research was conducted in the Center for Silicon Nanomembranes, supported by Air Force Office of Scientific Research through the Multidisciplinary University Research Initiative.