First principles calculation of inhomogeneous broadening in solid-state cw-EPR spectroscopy

Excited-State Forces with the Gaussian and Augmented Plane Wave Method for the Tamm-Dancoff Approximation of Time-Dependent Density Functional Theory

Augmented plane wave methods enable an efficient description of atom-centered or localized features of the electronic density, circumventing high energy cutoffs and thus prohibitive computational costs of pure plane wave formulations. To complement existing implementations for ground-state properties and excitation energies, we present the extension of the Gaussian and augmented plane wave method to excited-state nuclear gradients within the Tamm-Dancoff approximation of time-dependent density functional theory and its implementation in the CP2K program package. Benchmarks for a test set of 35 small molecules demonstrate that maximum errors in the nuclear forces for excited states of singlet and triplet spin multiplicity are smaller than 0.1 eV/Å. The method is furthermore applied to the calculation of the zero-phonon line of defective hexagonal boron nitride. This spectral feature is reproduced with an error of 0.6 eV in comparison to GW-Bethe-Salpeter reference computations and 0.4 eV in comparison to experimental measurements. Accuracy assessments and applications thus demonstrate the potential use of the outlined developments for large-scale applications on excited-state properties of extended systems.

Performance of DFT methods in the calculation of isotropic and dipolar contributions to 14N hyperfine coupling constants of nitroxide radicals

In the present study, we tested the widely used density functionals BP86, PBE, OLYP, TPSS, M06-L, B3LYP, PBE0, mPW1PW, B97, BHandHLYP, TPSS0, M06, M06-2X, CAM-B3LYP, ωB97x, and B2PLYP with the cc-pCVQZ basis set in calculations on a set of 23 nitroxide radicals with well-resolved ¹⁴N anisotropic hyperfine coupling (HFC) constants. The results were compared with those obtained using the B3LYP/N07D and PBE/N07D methods. The convergence of the HFC values to the complete basis set limit is briefly discussed. The best results were obtained using the M06/COSMO method, with a mean absolute deviation (MAD) of 0.4 G for the dipole-dipole contribution and MAD = 0.6 G for the contact coupling contribution (as compared to 1.1 G and 1.0 G, respectively, for the B3LYP/N07D/COSMO method and 1.7 G and 0.5 G, respectively, for the B3LYP/N07D method). The majority of the functionals yielded satisfactory results for the dipole-dipole contribution, but only the M06 functional yielded similar errors for both the dipole-dipole and isotropic contributions. The RIJCOSX and RI approximations introduced errors equal to or smaller than 0.01 G.