Hyperfine coupling constants from internally contracted multireference perturbation theory

TL;DR

This paper introduces a precise computational method for hyperfine coupling constants using CASPT2 with full internal contraction, accounting for electron correlation effects and relaxation contributions, validated on various radicals and complexes.

Contribution

The authors develop an accurate, efficient approach for calculating hyperfine coupling constants with CASPT2, incorporating orbital and configurational relaxation effects through an extended automatic code generator.

Findings

CASPT2 results are comparable or superior to coupled-cluster and DMRG methods.

The method accurately computes HFCCs for radicals like CN and AlO.

The approach is effective for transition metal complexes such as hexaaqua complexes.

Abstract

We present an accurate method for calculating hyperfine coupling constants (HFCCs) based on the complete active space second-order perturbation theory (CASPT2) with full internal contraction. The HFCCs are computed as a first-order property using the relaxed CASPT2 spin-density matrix that takes into account orbital and configurational relaxation due to dynamical electron correlation. The first-order unrelaxed spin-density matrix is calculated from one- and two-body spin-free counterparts that are readily available in the CASPT2 nuclear gradient program [M. K. MacLeod and T. Shiozaki, J. Chem. Phys. 142, 051103 (2015)], whereas the second-order part is computed directly using the newly extended automatic code generator. The relaxation contribution is then calculated from the so-called Z-vectors that are available in the CASPT2 nuclear gradient program. Numerical results are presented for the CN and AlO radicals, for which the CASPT2 values are comparable (or, even superior in some cases) to the ones computed by the coupled-cluster and density matrix renormalization group methods. The HFCCs for the hexaaqua complexes with VII, CrIII, and MnII are also presented to demonstrate the accuracy and efficiency of our code.