(289c) Repurposing the Hubbard U: A Remedy for Minimal Basis Set Inflexibility and a Path to Large-Scale Calculations At Low Computational Cost

Despite the rapid growth of available computing power, many realistic catalytic systems of interest remain inaccessible to electronic structure methods, in part because large electronic basis sets must often be used to obtain acceptable results. Here we demonstrate that an inexpensive correction improves the accuracy of minimal basis sets, rendering them a viable choice for treating such large systems. The Hubbard U (+U) correction improves density functional theory (DFT) calculations by counteracting the excess electron delocalization that arises from self-interaction error. The correction is tied to orbital occupation numbers, penalizing fractional occupation. We explore here whether this same correction can also improve minimal basis Hartree-Fock (HF) calculations, which improperly describe electron localization due to basis set inflexibility, despite being self-interaction-free (since exact exchange is included). Specifically, we highlight systems in which the inaccurate description of ionization potentials and electron affinities in the minimal basis leads to spurious proton transfers from nitrogen atoms to nearby oxygen atoms. By applying a +U correction to the 2p orbitals of the heavy atoms, we effectively tune these ionization potentials and electron affinities, adjusting the relative energies of the bound states to produce transfer pathways and energetics that closely parallel large basis set results. In the future, we will apply this minimal basis HF+U approach to a diverse range of systems, evaluating its usefulness as a general tool for improving the accuracy of minimal basis set calculations, and thereby opening the door to ab initio treatments of large-scale systems.