In high performance electronic systems, with increasing chip speed and larger number of processors, the system performance is being limited by off-chip metal interconnects. In such systems, polymer optical waveguides are being considered to replace electrical interconnects because of their high capacity for bandwidth and less constraints on interconnect length. The optical loss in the polymer optical waveguides is the key criterion used to evaluate their performance, and is significantly affected by thermal curing and aging. The evolution of degree-of-cure is determined from differential scanning calorimetry and compared to optical absorption from spectroscopy. Optical loss due to scattering mechanisms is related to local density fluctuations, which is studied using dielectric analysis. Based on the optical loss trends in uncladded and cladded waveguides, the underlying mechanisms for the optical loss variations are proposed and a cure process schedule to realize the lowest optical loss is recommended. Process-induced thermal stresses can also affect the polymer waveguide by introducing stress birefringence. The stress-optical coefficients of the siloxane polymer are extracted and employed in a numerical modeling method to determine the stress-induced birefringence in an optical waveguide system. The thermal-aging dependent optical loss is determined for waveguide samples at several different accelerated temperature conditions. To get the field-use conditions, the temperature distribution in the vicinity of the embedded laser and the polymer waveguide is determined. Using such thermal experimental data, the analytical reliability models were employed to relate the optical loss with time, and provide a practical way of determining whether the optical waveguides would perform within the optical loss budget during field-use conditions.

Advantage of the single QD active layer is its potential for very low threshold current density, which in turn can produce low internal optical loss. The low threshold current density and low internal loss thus enable a significant increase in laser diode cavity length. Because of the importance of the threshold current density in heatsinking, future technology of broad-area monolithic laser diodes can be implemented. The dissertation describes the development and the unique characteristics of single QD active layer laser with long cavity. The data are presented on single layer QD laser diodes that reach threshold current densities values of 11.7 A/cm2 in a p-up mounted 2 cm long cavity and as low as 10 A/cm2, with CW output power of 2 W in a p-down mounted 1.6 cm long cavity. The 8.8 A/cm2 in a p-down mounted 2 cm long cavity is reported. To our knowledge the value 8.8 A/cm2 is the lowest threshold current density ever reported for a room temperature laser diode. These single layer QD laser diodes reach an internal loss of ~0.25 cm-1, which is also the lowest ever reported for a room temperature laser diode. These unique characteristics of single layer QD and laser diode size are potentially promising for the monolithic mode-locked laser because of relatively high peak power with a low repetition rate that is on the order of a few GHz, which can be the novel device for external clocking in the optical interconnect applications. In this dissertation, the stable optical pulse train in a 40 µm wide stripe with a repetition rate of 3.75 GHz with 1.1 cm cavity length through the passive mode-locked onto the monolithic two-section device fabricated from this single layer QD laser is observed.