III-N blue-emitting epi-structures with high densities of dislocations: Fundamental mechanisms for efficiency improvement and applications

Abstract

Since the dawn of humanity, human beings seeked to light their surroundings for their well-being, security and development. The efficiency of ancient lighting devices, e.g. oil lamps or candles, was in the order of 0.03-0.04% and jumped to 0.4-0.6% with the use of gas during the first industrial revolution. Electricity allowed to reach 3-4% efficiency with incandescent bulbs, which contextualises the importance taken by III-nitrides (III-Ns) when > 10% external quantum efficiency (EQE) light-emitting diodes (LEDs) were first demonstrated in the 1990s. Nowadays, this technology is mature and state-of-the-art devices show > 80% EQE. This was only made possible through multidisciplinary research in the fields of semiconductor materials science, optics and photonics, and electronics. Of high interest is that such high efficiencies were obtained after numerous trial-and-error growth iterations, and we are still on the way to understand the intricacies of the underlying physics. For the present work, we choose the Si(111) platform to grow thin III-N epilayers containing blue-emitting InGaN/GaN quantum wells (QWs). These systems present high threading dislocation (TD) densities > 10^10 cm^-2 which make them an ideal testbed for investigating the impact of point defects (PDs) versus TDs on the internal quantum efficiency (IQE). By using the common In-containing underlayer (UL) mitigation strategy of burying the deleterious PDs below the QW, we have control on their density in the active region. We first perform various mesoscopic photoluminescence (PL) measurements giving quantitative and comparative efficiency assessments for samples with and without a consequent density of ~ 10^9 cm^-2 PDs. Our findings emphasize the importance of these defects, even when the TD density outscales the PD density by one order of magnitude. We then employ high resolution (~ 60 nm) cathodoluminescence (CL) tools to investigate the detrimental impact of PDs at the nanoscale. As suggested by scanning electron and atomic force microscopy, we discover that instead of being infinite two-dimensional landscapes featuring random energy variations, QWs are concatenated growth grain domains. This new, nuanced view was confirmed by time-resolved CL measurements, offering a deeper insight into the physical processes at play in InGaN/GaN QWs. We finally design and fabricate applicative demonstrator devices such as suspended membrane photonic crystal nanobeam lasers and microcavity LEDs, paving the way toward efficient and versatile integrated silicon photonics. The outcomes of this work enlighten the impact of point defects in the regime of high dislocation density. By delving deeper into this unexpected finding, nanoscale studies and experiments reshape our understanding of QWs properties.