Towards a systematic multi-scale method for excitations in molecular materials in the BigDFT code

Understanding excitations is critical in spectroscopy as well as in technological applications, e.g. identifying where excited states lay for molecules in crystals is crucial for guiding experimentalists in locating excitation sources. Another example is thermally activated delayed fluorescence (TADF), a mechanism for designing the next generation of OLEDs (Organic Light Emitting Diodes materials) that are less environmentally harmful than the previous generation. Excitations in TADF materials exhibit an intricate mixture of charge-transfer and local nature and can be strongly influenced by the host material. Modelling TADF (e.g identifying the best performing material) or locating excited states, thus, requires an accurate methodology that explicitly includes environmental effects. We are developing a multi-scale approach in the BigDFT code where high accuracy is combined with the ability of treating big systems to eventually go beyond implicit models. BigDFT runs on parallel architectures and can treat large systems with high, controllable precision. As a first step towards a robust methodology, we assess the performance of a promising new constrained-DFT approach recently developed in our group for various classes of excitations in comparison to standard methods (TDDFT).