Table of Contents
Fetching ...

On the performance of QTP functionals applied to second-order response properties II: Dynamic polarizability and long-range C$_6$ coefficients

Rodrigo A. Mendes, Peter R. Franke, Ajith Perera, Rodney J. Bartlett

Abstract

This work is the second in the series "On the performance of QTP functionals applied to second-order response properties." In the first paper (J. Chem. Phys. 162, 054105, 2025), we demonstrated the good performance of Quantum Theory Project functionals in predicting static perturbed second-order properties, such as static polarizabilities, nuclear magnetic resonance (NMR) spin-spin coupling constants, and NMR chemical shifts. In the present study, we focus on frequency-dependent properties, namely dynamic polarizabilities and C$_6$ dispersion coefficients. For completeness, a total of 25 exchange-correlation (XC) functionals were investigated. Dynamic polarizabilities were evaluated at five different perturbation wavelengths: 632.99 nm, 594.10 nm, 543.52 nm, 514.50 nm, and 325.13 nm. This property was also computed using HF and EOM-CCSD. In general, EOM-CCSD results are very close to those obtained with linear-response CC3, except at the highest frequency. Among Kohn-Sham calculations, TPSS0 and QTP01 showed the best overall performance for dynamic polarizabilities. We also assessed how well QTP functionals reproduce the pole structure of the CO molecule. For the C$_6$ dispersion coefficients, calculations were performed using the Casimir-Polder equation. The best overall performance was obtained with O3LYP; however, the first eleven ranked functionals show very similar accuracy. Within the QTP family, QTP01 and LC-QTP provide the best results for C$_6$ coefficients.

On the performance of QTP functionals applied to second-order response properties II: Dynamic polarizability and long-range C$_6$ coefficients

Abstract

This work is the second in the series "On the performance of QTP functionals applied to second-order response properties." In the first paper (J. Chem. Phys. 162, 054105, 2025), we demonstrated the good performance of Quantum Theory Project functionals in predicting static perturbed second-order properties, such as static polarizabilities, nuclear magnetic resonance (NMR) spin-spin coupling constants, and NMR chemical shifts. In the present study, we focus on frequency-dependent properties, namely dynamic polarizabilities and C dispersion coefficients. For completeness, a total of 25 exchange-correlation (XC) functionals were investigated. Dynamic polarizabilities were evaluated at five different perturbation wavelengths: 632.99 nm, 594.10 nm, 543.52 nm, 514.50 nm, and 325.13 nm. This property was also computed using HF and EOM-CCSD. In general, EOM-CCSD results are very close to those obtained with linear-response CC3, except at the highest frequency. Among Kohn-Sham calculations, TPSS0 and QTP01 showed the best overall performance for dynamic polarizabilities. We also assessed how well QTP functionals reproduce the pole structure of the CO molecule. For the C dispersion coefficients, calculations were performed using the Casimir-Polder equation. The best overall performance was obtained with O3LYP; however, the first eleven ranked functionals show very similar accuracy. Within the QTP family, QTP01 and LC-QTP provide the best results for C coefficients.
Paper Structure (8 sections, 11 equations, 5 figures, 1 table)

This paper contains 8 sections, 11 equations, 5 figures, 1 table.

Figures (5)

  • Figure 1: Heat map illustrating the relative mean absolute deviation (MAD) values in a.u., for the XC functionals compared to the LR-CC3/aug-cc-pVTZ level regarding dynamic polarizability.
  • Figure 2: Mean signed deviation (MSD) values in a.u., for the XC functionals compared to the EOM-CCSD/aug-cc-pVTZ level for the dynamic polarizability.
  • Figure 3: Isotropic dynamic polarizabilities for the CO molecule ($r_e =$ 1.1311),Ranasinghe_2019 computed at the range from $\omega_i = 0.20$ a.u. to $\omega_f = 0.53$ a.u. The basis set used was aug-cc-pVTZ.
  • Figure 4: Percent error (%Error) of isotropic $C_6$ coefficients obtained using different XC functionals w.r.t experimental values.
  • Figure 5: Signed percent error (%Error) of isotropic $C_6$ coefficients obtained using different XC functionals w.r.t experimental values.