Title:	High-accuracy vdW-DFs: Nonempirical theory for sustainable-energy solutions and DNA-marker spectra
DNr:	NAISS 2025/3-25
Project Type:	NAISS Large
Principal Investigator:	Per Hyldgaard <hyldgaar@chalmers.se>
Affiliation:	Chalmers tekniska högskola
Duration:	2026-01-01 – 2027-01-01
Classification:	10304 10407 10402
Homepage:	http://fy.chalmers.se/~hyldgaar/SNIC
Keywords:

Abstract

We apply for a 12 month NAISS large-computing allocation during 2026 (as well as a NAISS allocation for active storage at NSC and PDC). The computing resources that we seek are to enable the research programs in the groups of two of the original architects of the van der Waals (vdW) density functional (vdW-DF) method for truly-nonlocal density functional theory (DFT). We apply to continue our present large computing NAISS2024-3-16 allocation, extending our work in the Chalmers Spectra-Design Initiative with databases for combined atomic- and electronic structure characterizations. Our proposal is motivated by 2022-2025 progress suggesting exciting opportunities for our set of new range-separated-hybrid- and Koopmans-vdW-DFs to facilitate highly accurate and robust Kohn-Sham DFT predictions of structure and energies and at the same time fast, direct-from-generalized-DFT but accurate, predictions of quasiparticles. Those quasiparticles are the main computation target and they will in turn be use now to also seek fast predictions of excitons (as we are set to bypass the costly GW step), in international collaborations. With these new generalized-DFT tools we intend to help the development of optical DNA markers as well as better use of MOFs. To accelerate those developments we propose the formation of a new class of databases that contains both key atomic-structure and (binding) energy results but also electronic-structure characterizations.