Predicting Inverted Singlet-Triplet Gaps with XYG3-Type Doubly Hybrid Functionals

Molecules exhibiting an inverted energetic ordering of the lowest singlet and triplet excited states, commonly referred to as having an inverted singlet-triplet gap (ΔE _ST), have attracted significant interest due to their potential in enabling next-generation organic light-emitting diode (OLED) materials. Conventional time-dependent density functional theory (TD-DFT) methods, however, often fail to predict such inversions because they cannot adequately account for double excitations. Doubly hybrid functionals (DHs), which incorporate double excitations in a perturbative way, offer a promising alternative. In this work, we demonstrate that XYG3-type doubly hybrid functionals (xDHs) are able to correctly predict the inverted ΔE _ST. These functionals provide a cost-effective approach compared to highly accurate but prohibitively expensive coupled-cluster-based methods. Among the xDHs benchmarked, the nonempirical XYG2 functional exhibits the best performance and is strongly recommended for predicting the inverted ΔE _ST. It is anticipated that xDHs hold considerable promise for future use in the design and discovery of advanced OLED materials.