Modeling and analysis of ion-exchanged photonic devices

Photonic devices fabricated by ion exchange in glass have evolved to the point where conventional assumptions of waveguide symmetry and mutual independence are no longer valid. For example, during field-assisted ion exchange processes, the nonhomogeneity of ionic conductivity in the vicinity of the waveguide results in a time-dependent perturbation of the electric field. Previous studies have shown that the depth and vertical symmetry of buried waveguides are noticeably affected by the field perturbation.;This Dissertation describes an advanced modeling tool for guided-wave devices based on ion-exchanged glass waveguides. A genetic algorithm is proposed to determine the physical parameters that drive the ion exchange process. The diffusion equation describing binary ion exchange is solved numerically. The effect of field perturbation, due not only to the conductivity profile, but also to the proximity of adjacent waveguides or partial masking during a field-assisted burial, is accounted for. A semivectorial finite difference method is then employed to determine the modal properties of the waveguide structures. The model is validated by comparison with a fabricated waveguide containing a Bragg grating.;The modeled waveguides are utilized in the design of a multimode interference (MMI) device. A novel genetic algorithm-based design methodology is developed to circumvent issues with the commonly used self-imaging theory that arise when the MMI device operates in the regime of weak guiding. A combination of semivectorial finite difference modeling in two transverse dimensions and mode propagation analysis (MPA) in the propagation direction is used to evaluate the merit of each trial design. Two examples are provided of a 1 x 4 power splitter, which show considerable improvement in power imbalance and polarization dependent loss over that obtained by self-imaging theory.