Pseudopotentials for correlated electron systems

TL;DR

This paper introduces a new scheme for creating pseudopotentials tailored for correlated-electron calculations, demonstrating improved accuracy over existing Hartree-Fock-based pseudopotentials for light elements.

Contribution

The authors develop correlated-electron pseudopotentials (CEPPs) for light elements using high-level quantum chemical data, enhancing the accuracy of correlated-electron calculations.

Findings

CEPPs outperform Hartree-Fock pseudopotentials in atomic and molecular calculations.

CEPPs provide better dissociation energies and vibrational properties.

Results are validated against CCSD(T) calculations.

Abstract

A scheme is developed for creating pseudopotentials for use in correlated-electron calculations. Pseudopotentials for the light elements H, Li, Be, B, C, N, O, and F, are reported, based on data from high-level quantum chemical calculations. Results obtained with these correlated electron pseudopotentials (CEPPs) are compared with data for atomic energy levels and the dissociation energies, molecular geometries and zero-point vibrational energies of small molecules obtained from coupled cluster single double triple (CCSD(T)) calculations with large basis sets. The CEPPs give better results in correlated-electron calculations than Hartree-Fock-based pseudopotentials available in the literature.