Both broad-area and single-mode strained InGaAs-AlGaAs single quantum well (QW) lasers are indispensable components for both terrestrial and space satellite communications systems due to their excellent power and efficiency characteristics. However, their degradation mode (catastrophic and sudden degradation) due to catastrophic optical damage (COD) is a major concern especially for space applications, since COD-prone lasers typically show no obvious precursor signature of failure. Furthermore, as our group first reported in 2009, these lasers predominantly degrade by a new failure mode (bulk failure) due to catastrophic optical bulk damage (COBD) unlike AlGaAs QW lasers that degrade by a well-known failure mode (facet failure) due to catastrophic optical mirror damage (COMD). Unlike COMD, there have been limited reports on root causes of COBD. In addition, none of decades-long studies of reliability and degradation processes in (Al)GaAs or InGaAs QW lasers by many groups have yielded a reliability model based on the physics of failure. As part of our efforts to develop a physics of failure-based reliability model of InGaAs-AlGaAs strained QW lasers, we continued our investigation by performing short-term and long-term lifetests, failure mode analyses, and root causes investigations using various destructive and non-destructive techniques. All of broad-area and single-mode lasers that we tested degraded by COBD. We employed electron beam induced current (EBIC) techniques to study formation of dark line defects (DLDs) of lasers stressed under different test conditions and time-resolved electroluminescence (TR-EL) techniques to study the dependence of DLD propagation on electrical-thermal stresses via recombination enhanced defect reaction. Also, we employed high-resolution TEM and deep level transient spectroscopy (DLTS) techniques to study extended defects and point defects (and electron traps), respectively. Finally, we report on reliability model parameters obtained from our physics of failure investigation and compare them with those extracted using an empirical model.
|