An AlGaInP-based red light Micro-LED device design method and system based on artificial intelligence

Abstract

This application discloses an AI-based design method and system for AlGaInP-based red Micro-LED devices. Addressing the core pain points of existing red LEDs—long development cycles, high costs, difficulties in multi-parameter collaborative optimization, and the sharp drop in external quantum efficiency due to chip size effects below 10μm—this invention constructs a dataset corresponding to the entire chain of epitaxial growth, chip fabrication, and performance testing. A multi-branch fused deep neural network is built and trained to obtain a forward predictive proxy model that accurately maps device parameters to chip performance. Combined with a multi-objective optimization algorithm, global optimization in the high-dimensional parameter space is achieved, generating the optimal device parameter set in reverse. Experimental data feedback enables closed-loop iterative optimization of the model. This invention significantly shortens the development cycle, reduces development costs, and improves the luminous efficiency and reliability of small-size red Micro-LEDs. It is suitable for fields such as AR / VR and automotive displays, and has significant industrial application value.

Images