Breaking the light extraction limit in AlGaN-based deep-UV LEDs via AI-optimized mesa-sidewall nanoscale reflective...

Area of Science:

• Optoelectronics
• Materials Science
• Photonics

Background:

• Deep-ultraviolet light-emitting diodes (DUV-LEDs) are crucial for applications requiring short wavelengths.
• Enhancing light extraction efficiency (LEE) is a key challenge in DUV-LED development.
• Conventional planar reflectors often suffer from absorption in the p-GaN layer, limiting performance.

Purpose of the Study:

• To introduce and optimize a mesa-sidewall nanoscale reflective structure for DUV-LEDs.
• To minimize ultraviolet absorption and create a novel reflective pathway for photons.
• To significantly enhance the light extraction efficiency (LEE) of DUV-LEDs.

Main Methods:

• An artificial intelligence (AI)-assisted inverse design approach using the Jaya algorithm was employed for optimization.
• The mesa-sidewall nanoscale reflective structure was engineered on the p-GaN surface.
• Simulations and experimental validations were conducted to assess performance.

Main Results:

• The optimized mesa-sidewall structure demonstrated a novel reflective pathway for photon propagation.
• Light extraction efficiency (LEE) was enhanced by over 200% compared to conventional aluminum (Al) reflective layers.
• The structural parameters' effects on light propagation and LEE were comprehensively investigated.

Conclusions:

• The mesa-sidewall nanoscale reflective structure offers a viable and efficient strategy for high-power DUV-LED development.
• AI-driven design shows significant potential for advancing photonic devices.
• This approach paves the way for next-generation high-performance DUV optoelectronic applications.