Input-wavelength-insensitive tunable wavelength conversion was achieved in the range of 1530 to 1560 nm using cascaded semiconductor laser wavelength converters (a DFB laser and an SSG-DBR laser). The power penalty in the wavelength conversion of input signal between 1530 and 1555 nm, where the wavelength ranged between 1537 and 1557 nm, is less than 1 dB for 5 Gbit/s signals.

We report a low loss strip loaded single mode waveguide with a GaAs/Al0.3Ga0.7As double heterostructure. The propagation loss is derived to be 1.3 dB/cm for 1.3µm single mode propagation from Fabry-Perot resonances of a waveguide with good cleaved facets using DFB laser.

As one of the approaches to developing a monolithic integrated device, optical coupling efficiency between a monolithically integrated DFB laser and a passive waveguide is experimentally evaluated. A butt-joint coupling is selected in this study, because it is expected to produce a higher coupling efficiency than any other coupling techniques. The DFB laser is butt-jointed with the waveguide by a low-pressure MOVPE selective growth, which produces no overgrowth at the butt-joint region. The average efficiency is evaluated to be 60%. The distribution of the evaluated efficiency indicates that the coupling efficiency of more than 50% can be obtained with good reproducibility.

The paper discusses the possibility of building semiconductor lasers whose wavelength stays nearly constant with ambient temperature variation. Several factors affecting the lasing wavelength change with temperature variation in both distributed feedback lasers and Fabry-Perot lasers are addressed and the optimum design of bandgap temperature dependence for the active layer material is discussed. It is pointed out that the most important challenge we face in building temperature-insensitive wavelength lasers is the development of a temperature-insensitive bandgap material for the active layer. Based on published data, it is speculated that such a laser could be developed using a Hg1-xCdxTe/CdTe double heterostructure. Although no data is available yet, we expect a Ga1-xInxAs1-yBiy III-V alloy semiconductor can be used for this purpose. Recently reported T1xIn1-x-yGayP III-V alloy semiconductor might be another promising candidate. Such lasers will greatly advance applications of WDM (Wavelength-Division-Multiplexing) technology to optical fiber communication systems and contribute to network innovations.