This thesis explores the realization of low facet reflectivity using self-aligned stripe buried waveguide configuration and its implementation in optoelectronic devices such as superluminescent diodes (SLDs) and semiconductor optical amplifiers (SOAs). I explored the development of the buried waveguide in AlGaAs/GaAs material system ,since its first presentation in 1974 by Tsukada, in order to identify the problems associated with this technology. A novel window-faceted structure is demonstrated. The experimental measurements demonstrated effective reflectivity <10-14 as a result of both divergence and absorption within these window-like regions (i.e. not transparent). Its implementation to suppress lasing in tilted and normal-to-facet waveguide SLDs was thoroughly investigated in chapters 3 and 4. In the tilted devices, ~40mW output power with spectral modulation depth < 2% is demonstrated. In the latter types of SLDs, up to 16mW output power with <5% spectral modulation depth was recorded, which is the highest power demonstrated for such configurations. The performance of the two types of devices was measured without the application of anti-reflective coatings on the rear facet, which makes them inherently broadband. By incorporating a windowed facet at each end of a waveguide I could realize an SOA with window structured facet. Promising results were demonstrated in this configuration including 33dB gain and <6dB noise figure, which are comparable to the state-of-the-art. A trial was held to extend the concept of absorptive rear window to visible wavelengths available in the GaInP/AlGaInP material system. Problems associated with such devices were explored briefly and two solutions are suggested. Simulations were performed to realize design of an optimized device. Unfortunately, the experimental implementation of the design was not successful but suggestions for strategies to overcome these problems are discussed.