Reproduce “Benchmarking DFT-based excited-state methods for intermolecular charge-transfer excitations” on OpenResearch

About this paper

Nicola Bogo, Christopher J. Stein

Bogo and Stein benchmarked various density functional theory approaches for modeling intermolecular charge transfer excitations against high-accuracy coupled-cluster data. Their study found that orbital-optimized DFT methods, particularly IMOM, offer the best balance of accuracy and computational efficiency, outperforming standard and optimally-tuned TD-DFT for predicting excitation energies and their distance dependence.

Key insights

• Orbital-optimized DFT methods, particularly IMOM, demonstrate superior accuracy and stability for intermolecular charge transfer (ICT) excitation energies compared to TD-DFT.
• Accurate characterization of charge transfer (CT) states requires density-based descriptors computed from relaxed excited-state densities, with D_CT identified as the most reliable metric.
• Optimal tuning of range-separated hybrid functionals significantly improves TD-DFT performance for ICT, and smaller basis sets can provide reasonable accuracy when combined with optimally-tuned functionals or IMOM, which has implications for large systems.
• OO-DFT methods, especially IMOM, accurately reproduce the correct distance dependence of ICT excitation energies, capturing the asymptotic behavior at large donor-acceptor separations better than TD-DFT.

Summary by alphaXiv.