Small basis set density functional theory method for cost-efficient, large-scale condensed matter simulations
We present an efficient first-principles based method geared toward reliably predicting the structures of solid materials across the Periodic Table. To this end, we use a density functional theory baseline with a compact, near-minimal min+s basis set, yielding low computational costs and memory demands. Since the use of such a small basis set leads to systematic errors in chemical bond lengths, we develop a linear pairwise correction, available for elements Z = 1-86 (excluding the lanthanide series), parameterized for use with the Perdew-Burke-Ernzerhof exchange-correlation functional. We demonstrate the reliability of this corrected approach for equilibrium volumes across the Periodic Table and the transferability to differently coordinated environments and multi-elemental crystals. We examine relative energies, forces, and stresses in geometry optimizations and molecular dynamics simulations.
Authors
- Karsten Reuter (ORCID: https://orcid.org/0000-0001-8473-8659)
- Johannes T. Margraf (ORCID: https://orcid.org/0000-0002-0862-5289)
- Elisabeth Keller (ORCID: https://orcid.org/0000-0001-8946-3308)
- Jack Morgenstein (ORCID: https://orcid.org/0009-0005-0620-6485)
Institutions
- Duke University (US)
- Fritz Haber Institute of the Max Planck Society (DE)
- University of Bayreuth (DE)
Publication Details
- Journal
- The Journal of Chemical Physics
- Published
- 2024-08-15
- DOI
- https://doi.org/10.1063/5.0222649
- Citations
- 2
- Primary Topic
- Machine Learning in Materials Science
- Type
- article
- Field-Weighted Citation Impact
- 0.23
Funders
- International Max Planck Research School for Advanced Methods in Process and Systems Engineering