85‐3: Machine Learning Strategy Towards Inverse Design of Blue TADF Emitter: Training Excited State Properties Based on Density Functional Theory Calculations

Abstract

We introduce inverse design strategy utilizing machine learning (ML) models to discover efficient blue thermally activated delayed fluorescence (TADF) organic emitter materials. Here, we leverage graph neural network (GNN) to predict the characteristic intrinsic materials properties of TADF such as excited state energy levels and their transition properties. The GNN model is trained based on density‐functional theory (DFT) calculation results to meet the TADF properties. We discuss consistency between experimental observation and ML predictions, and examined conditions for improving the accuracy of DFT calculations and ML models on top of it.