2008 | Michael J. G. Peach, Peter Benfield, Trygve Helgaker, David J. Tozer
This study evaluates the performance of three density functional theory (DFT) functionals—PBE, B3LYP, and CAM-B3LYP—in calculating electronic excitation energies for 59 excitations in 18 main-group molecules. The functionals are tested for their ability to accurately predict local, Rydberg, and intramolecular charge-transfer (CT) excitations. The study finds that CAM-B3LYP provides the best overall performance, with no significant correlation between excitation energy errors and spatial orbital overlap, as measured by the quantity Λ. In contrast, PBE and B3LYP show a clear correlation between errors and Λ, indicating that these functionals are less reliable for certain types of excitations. The study also highlights the ambiguous nature of the term "charge transfer," showing that while some CT excitations are poorly described by GGA and hybrid functionals, others are accurately reproduced. The results suggest that CAM-B3LYP is the most reliable functional for general excitation energy calculations. The study proposes a diagnostic test for evaluating the reliability of excitation energies from PBE and B3LYP: if Λ is below a certain threshold, the excitation is likely to be in significant error. The study also discusses the importance of spatial orbital overlap in determining the accuracy of excitation energies and provides insights into the behavior of different types of excitations in various molecules. The results are supported by comparisons with reference values and experimental data.This study evaluates the performance of three density functional theory (DFT) functionals—PBE, B3LYP, and CAM-B3LYP—in calculating electronic excitation energies for 59 excitations in 18 main-group molecules. The functionals are tested for their ability to accurately predict local, Rydberg, and intramolecular charge-transfer (CT) excitations. The study finds that CAM-B3LYP provides the best overall performance, with no significant correlation between excitation energy errors and spatial orbital overlap, as measured by the quantity Λ. In contrast, PBE and B3LYP show a clear correlation between errors and Λ, indicating that these functionals are less reliable for certain types of excitations. The study also highlights the ambiguous nature of the term "charge transfer," showing that while some CT excitations are poorly described by GGA and hybrid functionals, others are accurately reproduced. The results suggest that CAM-B3LYP is the most reliable functional for general excitation energy calculations. The study proposes a diagnostic test for evaluating the reliability of excitation energies from PBE and B3LYP: if Λ is below a certain threshold, the excitation is likely to be in significant error. The study also discusses the importance of spatial orbital overlap in determining the accuracy of excitation energies and provides insights into the behavior of different types of excitations in various molecules. The results are supported by comparisons with reference values and experimental data.