Abstract
Charge-transfer excited states are crucial to modern electronics, particularly organic light-emitting diodes (OLEDs) based on thermally-activated delayed fluorescence (TADF). However, accurately modeling CT states remains challenging, even with modern implementations of (time-dependent) density functional theory [(TD-)DFT], especially in a dielectric environment. To identify short-comings and improve the methodology, we previously established the STGABS27 benchmark set with highly accurate experimental references for the adiabatic energy gap between the lowest singlet and triplet excited states (∆EST). Here, we diversify this set to the STGABS27-EMS benchmark by including experimental emission energies (Eem) and use this new set to (re)-evaluate various DFT-based approaches. Surprisingly, these tests demonstrate that a state-specific (un)restricted open-shell Kohn-Sham (U/ROKS) DFT coupled with a polarizable continuum model for perturbative state-specific non-equilibrium solvation (ptSS-PCM) provides exceptional accuracy for predicting Eem over a wide range of density functionals. In contrast, the main workhorse of the field, Tamm-Dancoff-approximated TD-DFT (TDA-DFT) paired with the same ptSS-PCM, is distinctly less accurate and strongly functional dependent. More importantly, while TDA-DFT requires the choice of two very different density functionals for good performance on either ∆EST or Eem, the time-independent U/ROKS/PCM approaches deliver excellent accuracy for both quantities with a wide variety of functionals.
Supplementary materials
Title
Supporting Information
Description
Detailed description of the computational workflow, methods, and used programs; complete plots of the emission energies for all used methods; investigation of basis set effects; definition of the used statistical measures
Actions