Answers are grounded in this paper and cite their sources.
Extensive calculations on a large set of free radicals containing atoms of the second and third row show that the B3LYP/N07D computational model provides remarkably accurate structural parameters and magnetic tensors at reasonable computational costs. The key of this success is the optimization of core-valence s functions for hyperfine coupling constants, while retaining (and even improving) the good performances of the parent 6-31+G(d,p) basis set for valence properties through reoptimization of polarization and diffuse p functions.
No takes yet. Share an insight, caveat, or question.
Barone et al. (2008) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: