UC San Diego

Initially motivated by the vast data transmission capacity, photonics has also been recently recognized for its potential capability in the so-called parallel processing. The significance of photonics in either data transmission, or processing, in conjunction with the fact that any domain conversion (i.e., optical to electrical or vice versa) is a cost inefficient operation, suggests for a necessity of a device/mechanism capable of direct manipulation/processing of the optical signals, independent of any electrical processing. The nonlinear optics and specifically the optical parametric process has been recognized as one of the few vastly under- investigated apparatuses proficient at handling a number of important all-optical functionalities, essentially serving as a pre-processor. As an optical pre-processing unit, the parametric process is susceptible to multiple impairment mechanisms. In this dissertation, the main sources of the performance degradation in a parametric mixer device, which can be effectively classified with respect to their origin into the random field fluctuations (i.e., noise) and deterministic distortions, are introduced and the corresponding compensation/equalization techniques are proposed and verified through a set of experimental demonstrations. In particular, the investigated equalization methods either rely on the digital post processing technique or benefit from the specific physical configurations leading to the superior mixer performances. As demonstrated in the dissertation, the incorporation of these mitigation techniques reinforces the utilization of fiber-based parametric devices as by far the best performing and the most versatile all-optical pre-processors to date