Basis-set correction for coupled-cluster estimation of dipole moments

TL;DR

This paper introduces a DFT-based basis-set correction method for coupled-cluster calculations of dipole moments, significantly improving basis convergence with minimal additional computational cost.

Contribution

It generalizes previous basis-set correction methods to non-variational properties like dipole moments in CCSD(T) calculations, using only Hartree-Fock densities.

Findings

Enhanced basis convergence of dipole moments in CCSD(T)

Minimal additional computational cost

Validated on fourteen molecules with diverse dipole moments

Abstract

The present work proposes an approach to obtain a basis-set correction based on density-functional theory (DFT) for the computation of molecular properties in wave-function theory (WFT). This approach allows one to accelerate the basis-set convergence of any energy derivative of a non-variational WFT method, generalizing previous works on the DFT-based basis-set correction where either only ground-state energies could be computed with non-variational wave functions [J. Phys. Chem. Lett. 10, 2931 (2019)] or properties where computed as expectation values over variational wave functions [J. Chem. Phys. 155, 044109 (2021)]. This work focuses on the basis-set correction of dipole moments in coupled-cluster with single, double, and perturbative triple excitations (CCSD(T)), which is numerically tested on a set of fourteen molecules with dipole moments covering two orders of magnitude. As the basis-set correction relies only on Hartree-Fock densities, its computational cost is marginal with respect to the one of the CCSD(T) calculations. Statistical analysis of the numerical results shows a clear improvement of the basis convergence of the dipole moment with respect to the usual CCSD(T) calculations.