An integrated screening approach for designing efficient thermally activated delayed fluorescent materials for OLEDs

Abstract

Thermally activated delayed fluorescence (TADF) materials, as the third generation of organic light-emitting diode (OLED) materials, still lag behind heavy-metal phosphorescent OLEDs in terms of luminescence stability and efficiency, and have not yet been successfully commercialized. Experimental determination is often time-consuming. To screen out efficient and stable TADF molecules, this paper for the first time develops a method for screening organic materials by combining luminescence stability and efficiency. By collecting previous experimental data, we established a candidate dataset of 2533 TADF molecules and an experimental dataset of thermal decomposition temperatures (T_d, equivalent to 5% weight loss) for 365 TADF molecules. Using a machine learning model, we screened 1779 molecules with T_d values greater than 400 °C. From these, 90 molecules were selected for detailed analysis, ultimately identifying 16 candidate molecules with excellent performance. The study also found that molecules with a dihedral angle between the donor and acceptor ranging from 35° to 80° exhibited the best performance. This analysis may provide an effective predictive method. A stepwise screening model with strong theoretical support could be an effective approach for designing TADF molecules.