| Silicon is incredibly well-studied as an electronic material. Since the out-migration of William Noyce, Gordon Moore, and the rest of the original Fairchild Semiconductor class from Shockley Semiconductor, silicon has only grown in prominence. Untold billions have been expended on research, development, and manufacture, and now silicon is perhaps the most well-controlled commercial material on Earth. For all that, the use of silicon as a mechanical material, though envisioned in the late 1950s, largely became viable only after the advent of large-scale silicon processing for microelectronics. As an optical material, silicon also has a long pedigree, with most of the positive focus on its excellent optical transparency and the enormous potential for improvement residing in its lack of effective emission.;This thesis concerns an alternate route to the generation of light in silicon: the nonlinear route. Resonant elements play a critical role in making this viable. The ability to build up optical intensity in the confined volume of a microresonator reduces the input power required to achieve a given energy density. Silicon also has certain excellent nonlinear properties: a large Raman gain coefficient, for example, and an appreciable Kerr effect. Unfortunately, silicon also exhibits significant two-photon absorption (TPA) in the convenient telecommunications wavelength bands. As such, the build-up of optical energy in silion may also be accompanied by a build-up of TPA-induced free-carrier populations. These populations may serve to limit the efficiency of nonlinear processes or to generate additional nonlinear behavior in their interactions with optical fields.;Thus two important stepping-stones on the way to the low-power, low-footprint use of silicon as an optical material are: the need to reduce optical losses in the optical elements, and to reduce or modify the populations of free carriers generated in the nonlinear optics regime. This thesis will present design considerations of, fabrication techniques developed for, and characterization techniques of high-Q silicon microresonators. In the course of this work, we have created silicon microdisk resonators with quality factors as high as 5 × 106, and high-Q silicon microdisks with free-carrier lifetimes in the deep subnanosecond regime (Q=5–6× 105 and carrier lifetimes ≤ 240 ps). These results, with no indication that higher quality factors or shorter carrier lifetimes are not possible in similar structures, indicate that coherent, CW optical generation in passive silicon microresonators is approaching feasibility. |
