Synergistic enhancement of light extraction in deep ultraviolet LEDs via holistic co-optimization of dual-layer nanostructures

AlGaN-based deep-ultraviolet light-emitting diodes (DUV-LEDs) suffer from low light extraction efficiency (LEE), primarily constrained by strong optical absorption in the p-GaN layer and total internal reflection (TIR), particularly for the dominant transverse-magnetic (TM) polarization. To address this, we introduce a dual-layer nanostructure design combining surface parabolic cones with embedded p-GaN nanopillars. We demonstrate that conventional layer-by-layer optimization (Strategy A) is limited by inefficient inter-layer mode coupling, whereas our proposed holistic co-optimization (Strategy B) utilizing particle swarm optimization (PSO)-based finite-difference time-domain (FDTD) simulations effectively harnesses the inter-layer synergistic effect. Results reveal that Strategy B boosts the total LEE to 12.98% at 280 nm. Crucially, the TM-mode extraction is enhanced by ∼36-fold (from 0.29% to 10.70%), effectively converting trapped lateral guided modes into upward extracted radiation. This work proves that holistic co-optimization is a necessity for maximizing the performance of coupled multi-parameter photonic devices.