A common strategy to calculate electronic coupling matrix elements for charge or energy transfer is to take the adiabatic states generated by electronic structure computations and rotate them to form localized diabatic states. In this paper, we show that, for intermolecular transfer of singlet electronic excitation, usually we cannot fully localize the electronic excitations in this way. Instead, we calculate putative initial and final states with small excitation tails caused by weak interactions with high energy excited states in the electronic manifold. These tails do not lead to substantial changes in the total diabatic coupling between states, but they do lead to a different partitioning of the total coupling between Coulomb (Förster), exchange (Dexter), and one-electron components. The tails may be reduced by using a multistate diabatic model or eliminated entirely by truncation (denoted as "chopping"). Without more information, we are unable to conclude with certainty whether the observed diabatic tails are a physical reality or a computational artifact. This research suggests that decomposition of the diabatic coupling between chromophores into Coulomb, exchange, and one-electron components may depend strongly on the number of states considered, and such results should be treated with caution.
ASJC Scopus subject areas
- Electronic, Optical and Magnetic Materials
- Physical and Theoretical Chemistry
- Surfaces, Coatings and Films