Catastrophic optical damage in semiconductor lasers
The catastrophic optical-damage (COD) mode in semiconductor lasers is a well-known degradation effect, resulting in more or less pronounced damage signatures at the front facet of diode light sources. By employing increasingly advanced facet treatments, COD thresholds of more than 200mW per micrometer aperture width have been achieved even for broad-area devices and bars.1 This ensures that—for regular operating conditions—COD is not necessarily the dominant degradation mechanism limiting the overall lifetime of current devices. Nevertheless, the microscopic details and kinetics of the actual COD process are not yet well understood because they have thus far not been accessible by direct measurements.
Here, we focus on progress in instrumentation development for such direct experiments. First, we address direct measurements of surface-recombination velocities at diode-laser-front facets.2 Knowing these kinetics is crucial since they describe the status of the surface as affected by defects. Although surface recombination itself unlikely causes COD directly, it is a primary source of initial facet heating that may trigger follow-up responses, which eventually lead to degradation.
We report on real-time COD monitoring in red-emitting aluminum-gallium-indium-phosphide (AlGaInP) diode lasers during continuous-wave operation.3 This is achieved by simultaneous optical near-field and thermal imaging of the emitting aperture (see Figure 1). Using two cameras simultaneously allows for monitoring the emitter's evolution in the near field and its temperature distribution with a temporal resolution of 2.3ms. This is accomplished by operating both cameras in synchronized mode while sweeping the laser current. The operating current was increased at a rate of 20mA/2s until COD occurred (see the inset of Figure 1). Figure 2 shows the intensity evolution of optical near-field and selected thermal images, taken shortly before and after COD, which for this particular device occurred at a current I=2.04A.
Figure 3 shows the thermal transients compiled from the thermal-image sequences. The data in Figures 2 and 3 suggests that COD is initiated in a spatially highly confined section of the chip. The COD mode—as monitored by the thermocamera—manifests itself as a single-pixel, single-frame event, from which we derive an upper limit of Δx × Δy × Δt = 8.8 × 8.8 × 2.3μm2ms for the spatio-temporal dimensions of the COD seed. Thus, the camera still imposes resolution limits in space and time. The measured ‘thermal flash’ of ~4K represents an average over these dimensions (i.e., a lower limit), with the potential presence of much higher temperatures.
Thus, we propose a method to dynamically image COD in semiconductor lasers with a thermal-infrared camera. We observed a pronounced temperature spike of 2.3ms duration at the location of the COD seed and a subsequent jump in temperature of the entire device. Our study strongly suggests that the COD mode of continuous-wave-operating red-emitting high-power AlGaInP lasers is due to a thermal-runaway process. The observed correlation with the optical near-field temperature provides further evidence of the critical nature of the COD process driven by both the optical load at the facet and the thermal load due to bulk heating. We think that using a thermocamera as ‘COD monitor’ will enable preparation of devices into the very early stages of COD. This should allow further detailed studies of the relevant kinetics.
Jens W. Tomm received his diploma and PhD in physics from the Humboldt University in Berlin (Germany). From 1993 to 1995, he held an appointment at the Georgia Institute of Technology in Atlanta, followed in 1999 by a post at RIKEN in Sendai (Japan). His research interests include photonics, optoelectronics, and device physics. He has authored more than 200 peer-reviewed papers, reviews, and a book.
