Quantum-Chemical Simulation of Multiresonance Thermally Activated Delayed Fluorescence Materials Based on B,N-Heteroarenes Using Graph Neural Networks

Multiresonance thermally activated delayed fluorescence (MR-TADF) emitters are crucial for the next generation of electroluminescent devices due to their high efficiency and narrowband emission. In this study, we developed a simple molecular design for MR-TADF materials based on a π-extended DABNA core decorated with four different framework types (carbazole (X = none), acridine (X = C(Me)2), phenoxazine (X = O), and phenothiazine (X = S)) and further modified with 18 different annulated systems. The optoelectronic properties of these compounds were modeled using density functional theory. Based on quantum chemical calculations, an accelerated search tool for MR-TADF emitters was developed using deep learning methods, enabling the prediction of energy values approximating experimental results.