Within the last decade, very significant advances have been made in semiconductor lasers. These advances were direct results of the various specific requirements demanded by optical fiber communications on semiconductor lasers as the light source. In the 1970s, research and development have been concentrated in laser stripe geometries for achieving stable transverse fundamentalmode operation required for efficient and stable coupling into then multimode optical fibers. In the early 1980s, with the advent of low-loss singlemode fiber at 1.57 μm, it became obvious that development of a single-frequency semiconductor laser operating at 1.57 μm would be very important in very high-data-rate long-distance transmission. This fueled the research and development of semiconductor laser schemes for single-longitudinalmode control. Two schemes of major interest are distributed feedback (DFB) and cleaved-coupled-cavity (C3 ). While the wavelength of a DFB laser is fixed at the time of manufacturing by the grating period, that of a C3 laser is electrically tunable. While the side-mode suppression, linewidth, and threshold of a DFB laser depend on the relative position of the end facet with respect to the grating period, those of a C3 laser depend on the gap separation between the two coupled cavities. While the DFB scheme is incorporateable only to certain stripe-geometry laser structures and is technologically feasible for lasers beyond certain lasing wavelengths (due to difficulties in fabricating very short grating periods), the C3 scheme is applicable to all laser structures and wavelengths. Thus the choice depends on the specific applications in mind. Although these lasers were called single-frequency sources, they fell very short of their given name for three major yet unresolved problems.
© 1986 Optical Society of AmericaPDF Article