Cost–accuracy tradeoffs in relativistic many-body theory: recontraction vs. orbital freezing
We present a benchmark study on the impact of basis set recontraction and orbital freezing within the molecular mean-field exact-two-component equation-of-motion coupled-cluster singles and doubles framework (mmfX2C-EOM-CCSD) for computing spin–orbit splittings. The steep computational scaling of EOM-CCSD with respect to virtual and occupied orbital dimensions motivates the use of truncated correlation spaces and recontracted basis functions to reduce cost. Here, we systematically evaluate fine-structure splittings in a set of atomic systems spanning the s-, p-, and d-blocks using basis sets of increasing size. The results demonstrate that orbital freezing, when applied selectively to high-energy virtual orbitals and tightly bound core orbitals, can achieve substantial cost reduction with minimal loss of accuracy. In contrast, basis-set recontraction, while offering significant computational savings through reduced basis dimensions, introduces non-negligible errors, particularly in systems with strong relativistic effects. These errors are most pronounced in p-block elements and at lower basis-set levels, where apparent accuracy may arise from fortuitous error cancellation.
Authors
- Kenneth G. Dyall (ORCID: https://orcid.org/0000-0002-5682-3132)
- A. Eugene DePrince (ORCID: https://orcid.org/0000-0003-1061-2521)
- Xiaosong Li
- Tianyuan Zhang
- Alexandros Peltekis
Publication Details
- Journal
- Figshare
- Published
- 2026-11-17
- DOI
- https://doi.org/10.6084/m9.figshare.35039565
- Primary Topic
- Advanced Chemical Physics Studies
- Type
- article
- Field-Weighted Citation Impact
- 0.00