AI-Enhanced High-Resolution Functional Imaging Reveals Trap States and Charge Carrier Recombination Pathways in Perovskite

Understanding and managing charge carrier recombination dynamics is crucial for optimizing the performance of metal halide perovskite optoelectronic devices. In this work, we introduce a machine learning-assisted intensity-modulated two-photon photoluminescence microscopy approach for quantitatively mapping recombination processes in MAPbBr₃ perovskite microcrystalline films at micrometer-scale resolution. To enhance model accuracy, a balanced classification sampling strategy was applied during the machine learning optimization stage. The trained regression chain model accurately predicts key physical parameters—exciton generation rate (G), initial trap concentration (N_TR), and trap energy barrier (E_a)—across a 576-pixel spatial mapping. These parameters were then used to solve a system of coupled ordinary differential equations, yielding spatially resolved simulations of carrier populations and recombination behaviors at steady-state photoexcitation. The resulting maps reveal pronounced local variations in exciton, electron, hole, and trap populations, as well as photoluminescence and nonradiative losses. Correlation analysis identifies three distinct recombination regimes: 1) a trap-filling regime predominated by nonradiative recombination, 2) a crossover regime, and 3) a band-filling regime with significantly enhanced radiative efficiency. A critical trap density threshold (~10¹⁷ cm^-3) marks the transition between these regimes. This work demonstrates machine learning-assisted intensity-modulated two-photon photoluminescence microscopy as a powerful framework for diagnosing carrier dynamics and guiding defect passivation strategies in perovskite materials.