Achieving high-power performance at long wavelengths has been a longstanding goal in the field of nitride LEDs, however it poses significant challenges. These include declining efficiency with increasing wavelength, undesired hue shifts with increasing current, and the difficulty of maintaining material quality while incorporating high indium compositions in quantum wells. In this work, I employ predictive modeling techniques to shed light on these hurdles, informing the design of devices that effectively overcome these obstacles. In particular, I will demonstrate how the efficiency-droop and hue-shift problems in the green spectral range are caused by an increase in the operating carrier density rather than material degradation. Then, I will discuss how progress towards longer red wavelengths is hindered by the difficulty in incorporating higher indium concentrations in quantum wells. To alleviate these issues, I explore the high-dimensional configurational landscape of quaternary III-nitride emitters using statistical-learning techniques, and identify promising designs that emit in the red spectral range.
|