Semiconductor slow-light device

Slow-light, i.e. light with an ultra-low group velocity, has experienced unprecedented progress in the last decade since the discovery of electromagnetically induced transparency (EIT) in 1991. Artificial manipulation of a highly dispersive medium while keeping reasonably low optical absorption allows the propagation speed of an optical signal being considerably slowed down without much distortion. Despite its popularity, slow-light has so far been merely of scientific interest mainly because it can only be observed in atomic vapor or solid crystalline materials. It is thus the purpose of this dissertation to study the aspect of slow-light in engineering. Our goal is to answer the following three questions: (1) "Why" is slow-light useful? (2) "How" can slow-light be useful? and (3) "Can" slow-light be useful?; Historically, semiconductor devices are the key to practical applications. A realization of a compact and fast semiconductor slow-light device will give positive answers to all three questions above. We will show that an efficient slow-light device can indeed be implemented in semiconductor materials by two different approaches, namely the waveguide dispersion engineering and the material dispersion engineering. In the case of waveguide dispersion, we propose and fabricate a horizontal cavity buried aluminum oxide grating. The maximally achievable slow-down factor is on the order of 20. We also propose the implementation of an optical buffer in a 40 Gbit/s optical network using ring resonator structures to store an optical signal up to 3.4 ns. In the case of material dispersion, we demonstrate EIT and population pulsation in a semiconductor quantum well. EIT signature via spin-coherence between two doubly-generate conduction bands is demonstrated in a GaAs/AlGaAs waveguide at room temperature. A group velocity as small as 9600 m/s which corresponds to a slowdown factor of 31200 and an operating bandwidth as large as 2 GHz are measured in the population pulsation experiment at 10 K.; We also study the theoretical aspects of a semiconductor slow-light device. We adopt a semiclassical approach in which quantized material and classical light are used. For the case of waveguide dispersion, we develop optical, growth, and oxidation models for the buried oxide grating. For the case of material dispersion, we use the density matrix framework to prove the feasibility of EIT and population pulsation in semiconductors. We also obtain good theory-experiment agreement in all cases.