Comparative analysis of electric potential in p-GaN/InGaN/n-GaN nanowire LEDs

Recent advances in the development of deep ultraviolet, light emitting diodes (LEDs) have been reported for nanowire (NW) device geometries composed of nitride semiconductors. Typically, these involve arrays of NWs where the level of variation in the electrical and optical properties of individual NW LEDs is unknown. In this work, the electric potential distributions in axial p-GaN/InGaN/n-GaN NW LEDs grown by plasma-assisted molecular beam epitaxy are investigated using electron holography (EH). Two kinds of NWs are observed to grow simultaneously on the same substrate. One type exhibits a long, thin morphology and a varying diameter, while the other has a short, wide morphology with a uniform diameter. Although the bottom p-GaN and InGaN regions have similar lengths in both types, the top n-GaN region are five times longer in the first type. Photoluminescence spectra from arrays, show an InGaN emission peak ranging from 2.55 eV to 2.65 eV, which indicates an average In composition of 20 ± 3 percent. This is consistent with energy dispersive x-ray maps from individual NWs of both types, which reveal a core/shell InGaN/GaN structure with similar composition. However, the EH potential maps reveal a built-in junction voltage of approximately 3 V in the long, thin NWs, while the short NWs exhibit a drastic reduction, with a junction voltage of only 0.6 V. The difference is primarily attributed to the length of the short wire n-doped segment being too short to reach the flat potential of a complete p-n junction.