Quantifying Strong Correlation with AFQMC

ORAL

Abstract

Strong correlation is challenging to describe with most approximate quantum chemical models, including single-reference methods such as Density Functional Theory (DFT), low-order perturbation and coupled cluster theories. Meanwhile, exact methods such as Full CI (FCI) are only feasible for very small systems (generally less than 18 electrons). We aim to show that auxiliary-field quantum Monte Carlo (AFQMC) is a promising candidate for describing correlated electronic systems, given its low-polynomial scaling with system size and its demonstrated ability to capture strong correlation when using suitable trial wavefunctions such as symmetry-broken mean-field states. Our group has previously argued(*) that the trace of the two-body cumulant is a justifiable metric for describing the amount of strong correlation in an electronic system; here, we show that AFQMC produces results that closely resemble those of FCI in a set of molecules with challenging bond dissociation coordinates. We then turn to the investigation of cyclic conjugated systems and explore the relationship between antiaromaticity and strong correlation.

(*) Ganoe, B., & Shee, J. (2024). On the notion of strong correlation in electronic structure theory. Faraday Discussions, 254, 53-75. https://doi.org/10.1039/d4fd00066h

Presenters

  • Saisrinivas Gudivada

    • Rice University

Authors

  • Saisrinivas Gudivada

    • Rice University
  • Brad Ganoe

    • Rice University
  • James Shee

    • Rice University