TADF emission is based on inverse inter-system crossing from triplet state to singlet states. These excitations have shown to exhibit an intricate mixture of charge-transfer and local nature. Modelling TADF (e.g to identify the best performing material) as well as locating excited states, thus, requires a methodology able to provide high accuracy while explicitly including environmental effects. However, developing a versatile theoretical approach for the characterisation of excitations can be challenging due to the complexity of the methods available and the variety of sources of error associated with them.

We are developing a multi-scale approach within the BigDFT code where we combine the needed accuracy with the ability of treating big systems, which would allow one to go beyond implicit models. BigDFT is designed to run on parallel architectures and can treat large systems while ensuring high, controllable precision. 

As a first step towards a robust methodology, we assess the performance of a novel promising constrained-DFT (T-CDFT) approach we recently developed while I was at Imperial College London and compare the results with standard methods (e.g. TDDFT) and simulation conditions. Such investigation is conducted on a diverse set of molecules (e.g. TADFs, acenes) in order to cover various classes of excitations. This phase is carried out by also developing portable jupyter-notebooks for the analysis of excitations.