A cascade model for the defect-driven etching of porous GaN distributed Bragg reflectors
Ben Thornley, Maruf Sarkar, Saptarsi Ghosh, Martin Frentrup, Menno J. Kappers, Thom R. Harris-Lee, Rachel A. Oliver
TL;DR
This study addresses the mechanism of defect-driven porosification in GaN-based distributed Bragg reflectors by leveraging a lithography-free electrochemical etching process. Using serial-section FIB-SEM tomography, the authors reconstruct 3D pore networks across a 5-pair GaN-on-Si stack etched at 5, 8, and 10 V, and demonstrate that etching propagates as a cascade involving multiple dislocations rather than a single kebab pathway. Quantitative dislocation statistics reveal voltage-dependent increases in active dislocation density and a shift toward kebab-like networks at higher voltages, while many dislocations exhibit interrupted or activation-then-inactivation behavior. The proposed cascade model provides a mechanistic framework for understanding and optimizing porosification, with broad implications for designing high-reflectance porous GaN DBRs and for characterizing sub-surface porosity in complex multilayer materials.
Abstract
Fabrication of porous GaN distributed Bragg reflectors (DBRs) via the selective electrochemical etching (ECE) of conductive Si-doped layers, separated by non-intentionally doped (NID) layers, provides a straightforward methodology for producing highly reflective DBRs suitable for device overgrowth and integration, which has otherwise proven difficult in the III-nitride epitaxial system via conventional alloying. Such photonic materials can be fabricated by a lithography-free defect-driven etching process, where threading dislocations intrinsic to heteroepitaxy form nanoscale channels that facilitate etchant transport through NID layers. Here, we report the first three-dimensional characterisation of porous GaN-on-Si DBRs fabricated in this methodology with different ECE voltages, using serial-section tomography in a focused ion beam scanning electron microscope (FIB-SEM). These datasets reconstruct the pore morphology as etching proliferates through the alternating Si-doped/NID layer stack. Volumetric reconstruction enabled us to enhance the established `kebab' model for defect-driven etching by proposing a `cascade' model where etchant cascades through the material via vertical etching down nanopipes and horizontal etching across pores, forming complex networks directly related to the pathways taken. This accounts for premature nanopipe termination and discontinuities in nanopipe formation, where dislocations are observed to activate and deactivate individually. Statistical analysis of individual etching behaviour, across all dislocations for each tomograph, revealed a greater tendency to form continuous structures that follow conventional kebab behaviour at higher ECE voltages. We propose that higher ECE voltages alter the probability of dislocation etching relative to doped layer etching, thereby empowering morphological optimization through improved mechanistic understanding of ECE.
