A revised version of the well-established B97-D density functional approximation with general applicability for chemical properties of large systems is proposed. Like B97-D, it is based on Becke’s power-series ansatz from 1997 and is explicitly parametrized by including the standard D3 semi-classical dispersion correction. The orbitals are expanded in a modified valence triple-zeta Gaussian basis set, which is available for all elements up to Rn. Remaining basis set errors are mostly absorbed in the modified B97 parametrization, while an established atom-pairwise short-range potential is applied to correct for the systematically too long bonds of main group elements which are typical for most semi-local density functionals. The new composite scheme (termed B97-3c) completes the hierarchy of “low-cost” electronic structure methods, which are all mainly free of basis set superposition error and account for most interactions in a physically sound and asymptotically correct manner. B97-3c yields excellent molecular and condensed phase geometries, similar to most hybrid functionals evaluated in a larger basis set expansion. Results on the comprehensive GMTKN55 energy database demonstrate its good performance for main group thermochemistry, kinetics, and non-covalent interactions, when compared to functionals of the same class. This also transfers to metal-organic reactions, which is a major area of applicability for semi-local functionals. B97-3c can be routinely applied to hundreds of atoms on a single processor and we suggest it as a robust computational tool, in particular, for more strongly correlated systems where our previously published “3c” schemes might be problematic.
REFERENCES
Note, however, that SCAN-D3/def2-QZVP4,62 has an even smaller MAD for the CCse21 bond lengths of 0.37 pm and yields a small average structure RMSD of 0.42 pm and is thus the most accurate standard (non-double-hybrid) functional for geometries of small semi-rigid molecules.
On the other hand, both M06L and SCAN converge slow with respect to the single particle basis set and require large radial grid sizes.62 In practice, B97-3c is about 2 orders of magnitude faster as confirmed for a standard single-point energy of phenanthrene and [7]helicene. B97-3c is used in default settings (see Sec. III B), while M06L is evaluated with RI-J in a def2-QZVP basis set and increased gridsize (grid7, IntAcc = 33.5), i.e., with ≥400 radial grid points per atom in following the recommendation in Ref. 62. With these settings, a single-point calculation with M06L of phenanthrene and [7]helicene takes 01:45:46 and 09:42:35, respectively (hh:mm:ss).
