Abstract

The reliability of semiconductor laser diodes and related devices is a significant issue for their deployment in many applications [1, 2], creating demand for device diagnostics applicable to packaged devices. Measurements of the transparency current density (J₀) in laser diodes and traveling-wave semiconductor optical amplifiers (SOAs) can provide such a diagnostic. It is essential, however, to measure J₀ on packaged devices, so that they can be characterized after aging or degradation. This precludes techniques requiring data from multiple devices (e.g., an ensemble with different lengths). J₀ is conventionally measured using a junction-voltage technique [3,4], in which an input optical signal induces a change in carrier density in the active region due to stimulated absorption or emission. The result is a voltage drop across the diode. At material transparency, the stimulated absorption is exactly balanced by the stimulated emission and the voltage drop goes to zero. Since the polarity of the voltage drop changes sign at current densities above J₀, the optical input beam is typically modulated and lock-in amplification is employed to sensitively detect the polarity sign flip. Here we show that this technique is not reliable for certain types of laser structures, because the deduced J₀ is strongly dependent on device packaging—that is, the measured Jo varies with the manner in which light is coupled into the diode waveguide. For packaged SOAs, we propose and demonstrate an alternative all-optical technique to overcome this problem. One important advantage of the optical method over the junction voltage method is that the actual device optically-guided mode is directly sampled. In the case of packaged semiconductor lasers, our all-optical method can complement the junction-voltage method to resolve device versus packaging degradation.

© 1995 Optical Society of America