- 7,916 visits since January 2015
Connecting Theory and Practice in Optoelectronics
I just came across a new paper from Soraa Inc. on highly efficient violet light-emitting diodes (LEDs) grown on bulk GaN substrate . The authors achieve a power conversion efficiency (PCE) of 70% at 100 A/cm2 current density by using a novel flip-chip architecture. These are certainly impressive results but the part that puzzles me is their explanation for measuring an electrical efficiency above 100%, i.e., the emitted photons have a higher energy than the injected electrons. This phenomenon is commonly observed in high-performance LEDs at low current density but it is seldom discussed. The authors attribute it to the absorption of thermal energy by injected electrons. In other words, the photons are emitted by electrons located in the tail region of the Fermi distribution, while the applied bias (times charge) is still below the photon energy. In my view, this explanation is acceptable at very low current density, but the authors observe the effect up to high current densities beyond the PCE maximum, i.e., at carrier densities that are high enough to trigger strong Auger recombination or other efficiency droop mechanisms. Does anybody have an alternative explanation for such low bias at high current ?