Either Accurate Singlet-Triplet Gaps or Excited-State Structures: Testing and Understanding the Performance of TD-DFT...

Area of Science:

• Computational Chemistry
• Materials Science
• Quantum Chemistry

Background:

• The energy gap between singlet and triplet excited states (ΔE_ST) is crucial for thermally activated delayed fluorescence (TADF) emitters.
• Charge-transfer (CT) states, common in TADF emitters, pose significant challenges for computational methods like time-dependent density functional theory (TD-DFT).
• The interaction of polar CT states with their molecular environment further complicates accurate computational modeling.

Purpose of the Study:

• To evaluate the performance of Tamm-Dancoff-approximated TD-DFT (TDA-DFT) for predicting ΔE_ST in CT states.
• To investigate the impact of orbital/structural relaxation and dielectric embedding strategies on TDA-DFT accuracy.
• To compare TDA-DFT performance against more rigorous methods and the ROKS/PCM approach for CT states in dielectric environments.

Main Methods:

• Utilized the STGABS27 benchmark set to assess TDA-DFT performance.
• Explored various strategies for including orbital and structural relaxation.
• Investigated the effects of dielectric embedding (solvent models) on calculated ΔE_ST values.

Main Results:

• The best-performing TDA-DFT strategy involved calculating ΔE_ST at the ground-state structure using functionals with ~10% Fock exchange and no solvent model.
• This seemingly optimal approach relies on error cancellation, leading to poor robustness and systematic deviations in excited state properties.
• More rigorous methods (e.g., state-specific solvation) reduced systematic errors but increased statistical errors in ΔE_ST predictions.

Conclusions:

• No tested TDA-DFT method provides a robust and accurate prediction of ΔE_ST for CT states in dielectric environments.
• Simple TDA-DFT approaches with minimal Fock exchange are not reliable due to reliance on error cancellation.
• The ROKS/PCM approach is currently superior to the evaluated TDA-DFT methods in terms of accuracy, robustness, and computational efficiency for this problem.