You are here : Home > NPSC team > GaN-based single nanowire photodetectors: comparison of bottom-up and top-down approaches

PhD Defense of Elçin AKAR

GaN-based single nanowire photodetectors: comparison of bottom-up and top-down approaches

​​​​​​​​​
Published on 13 October 2025

Semiconductor nanowires (NWs) offer promising opportunities for ultraviolet photodetectors. Their geometry allows decoupling the optically active area (responsivity) from the electrical cross-section (response time). However, optimizing device architecture and understanding performance—especially disentangling the effects of fabrication and geometry—requires further study. In terms of fabrication, vertically aligned nanowires can be produced via bottom-up growth or top-down fabrication from planar layers. The latter offers finer control over size, doping, composition, and in-plane positioning. In this work, we compare single NW GaN p-n photodetectors made using bottom-up nanowires (BUNWs) and top-down nanowires (TDNWs).
This study presents a detailed investigation of the top-down approach, starting from planar layer growth, through nanosphere lithography, to single NW fabrication. I first focused on Mg-doped GaN planar growth via PAMBE, which suffers from polarity inversion under high Mg flux. I developed an in-situ method using RHEED intensity transients during Ga desorption to detect the inversion threshold. Results show Mg disrupts the Ga wetting layer needed for maintaining planar growth, making high Ga excess at elevated temperature critical for p-type doping.
Next, I optimized nanosphere lithography using polystyrene spheres to pattern a SiO₂ mask on GaN. Then, ICP-RIE etching of GaN with Cl₂ chemistry, followed by a KOH-based crystallographic-selective wet etching results in either vertical NWs with m-plane facets or tapered profiles, depending on sphere size. Finally, I compared BUNW and TDNW single-NW photodetectors in terms of structure, junction quality, and performance. BUNWs featured p-n and p-i-n junctions; TDNWs included p-n, p-n with tunnel junction (TJ), p-i-n with TJ, and p-i-n with InGaN QD and TJ. KPFM confirmed junction integrity and TJ identification. I-V analysis revealed rectifying behavior and space charge–limited transport in all devices. Generally, p-i-n structures showed better linearity and slightly higher responsivity than those containing a p-n junction, in both BUNWs and TDNWs. In general, we demonstrate that both approaches yield comparable device performance despite differing fabrication routes.

Supervision of the thesis:
Eva MONROY- Martien DEN HERTOG