Tackling Strong Correlation with Natural Determinant Functionals

University Of California-Irvine, Irvine CA

Investigators

Abstract

Filipp Furche of the University of California, Irvine (UCI), is supported by an award from the Chemical Theory, Models and Computational Methods Program in the Division of Chemistry to develop, implement, and apply a new approach to strong electron correlation based on natural determinant functional theory (NDFT). Strong correlation is central to overcoming roadblocks in areas of critical scientific and technological importance, such as organometallic rare-earth and actinide chemistry, metal-metal interactions, reactive intermediates, single-molecular magnets and other quantum materials and devices; however, existing methodology either fails or is computationally prohibitive for real-world applications featuring strong correlation. Furche and his research group will develop, implement, and test NDFT-based approaches addressing these limitations. The computational methods resulting from this project will be widely disseminated through both proprietary and open-source codes. The program supports education and workforce readiness through undergraduate curriculum development at UCI, as well as outreach to high-school students. The project will start from single-reference NDFT developed during the prior funding period, which "exactifies" the popular but empirical generalized Kohn-Sham approaches, including (local) hybrid functionals and fifth-rung random-phase approximation (RPA) methods. NDFT will be extended using a single real parameter s defining an active space through an interval of (interacting) natural occupation numbers around the Fermi occupation number. Strong correlation within the active space will be treated explicitly by configuration interaction, whereas "dynamic" correlation should be accurately captured using appropriately generalized NDFT functionals. This approach does not (i) rely on perturbation theory, and (ii) require problematic active space selections. In parallel, new dynamic correlation functionals based on RPA with local field corrections will be developed. These methods will enable Furche and his group to extend applications to organometallic d and f element compounds with unprecedented electronic structure and properties. Low-power-low-cost single-board computers will be used to bring theoretical and computational chemistry to high-school students nationwide. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.