The paper deals with the fact that the widely used CASSCF method, if not used carefully, can yield large errors (1-2 eV) in vertical excitation energies. This problem arises for ionic states, as defined within valence bond theory. Within this work we developed a simple diagnostic to identify ionic states. We found a good correlation between the new diagnostic (Qta) and the error, as shown in the figure above.
We hope that the new diagnostic will be useful similar to analogous diagnostics identifying charge transfer states in TDDFT computations. This will give users the possibility to spot potential problems quickly.
On-going work is concerned with going from just diagnosing the problem to developing a numerical correction term to fix the problem.
We just posted a preprint discussing a question I have been wondering about for a while: Why is the lowest excited state of a molecule not always the HOMO/LUMO transition? More generally we show how singlet and triplet state energies are affected in different ways by post-MO energy terms.
The topic of doubly excited states has been discussed quite controversially in the literature over the last couple of years, see for example JACS, 139, 13770 (2017) and JCTC14, 9 (2018), and it is often disputed whether to classify a state as doubly excited at all. To contribute to this discussion we worked on the development of a physically motivated definition of doubly excited character based on operator expectation values and density matrices, which works independently of the underlying orbital representation. We hope that this approach will provide new understanding on these issues.
We were interested in understanding the difference in thermally activated delayed fluorescence (TADF) between two closely related donor-acceptor-donor systems using either an anthraquinone and benzodithiophenedione acceptor units, respectively. The first one was known to be an effective TADF emitter [JACS2014, 136, 18070] whereas the second one had significantly lower quantum yield for TADF [PCCP2019, 21, 10580].
Rather than just presenting energies, it was the purpose of this paper to shed detailed insight into the wavefunctions involved. Notable differences in the wavefunctions and charge-transfer character were found between the two molecules. Even more striking differences existed between different computational methods.
After evaluating electronic structure methods, we presented geometry optimisations in solution, highlighting the importance of symmetry breaking for producing an emissive lowest singlet state. The role of different solvation models was discussed as well.