A study of the origin and applications of nonlinear polarization rotation in semiconductor optical amplifiers

In order to fully exploit the capacity of optical networks it is necessary to develop alloptical signal processing techniques. One device which can be used for this purpose is the Semiconductor Optical Amplifier (SOA). A lot of research has been devoted to this device. This work concentrates mainly on the nonlinearities in the device to perform alloptical processes such as switching and wavelength conversion. The main nonlinearities that have been considered are cross-phase modulation, cross-gain modulation and fourwave mixing. However, another type of nonlinearity has recently come to attention, which is due to the intensity dependent rotation in the state of polarization of a signal injected into the SOA. This phenomenon is called Nonlinear Polarization Rotation (NPR). The exact physical origin of this effect is yet to be determined. The physical origins behind the NPR effect are examined using several novel experimental techniques. Polarization dependence of the gain and birefringence are found to be contributing factors. A model is also presented in order to develop an understanding of the physical dynamics occurring in the SOA. This is used to assess the potential of the SOA as a wavelength converter based on NPR. For the first time a detailed investigation of the optimum experimental setup for such an operation is presented. Wavelength conversion based on NPR is shown to have several advantages over XGM. Results are presented to examine the polarization dependence of pulse propagation through the SOA using an accurate and complete measurement technique. A large polarization dependence is present at high intensities. The potential speed of NPR in the SOA is assessed using an original method. This work is based on polarization resolved four-wave mixing. The polarization dependence of several gain dynamics in the SOA are found using this technique. The potential of ultra-fast all-optical signal processing based on NPR in the SOA is demonstrated