A precise understanding of the behavior of electrons in molecules and materials is needed to reliably predict their properties. Unfortunately, the complicated electronic structure of many promising compounds causes standard methods to fail catastrophically. Systems with this behavior are called strongly-correlated, and they are ubiquitous. Here we will discuss how a new framework: the <i>Flexible Ansatz for N-body Configuration Interaction</i> (FANCI), can help us to develop accurate and universally applicable wavefunction methods to treat strongly correlated fermions. In particular, our approach is inspired by the success of Coupled Cluster (CC) methods for weakly correlated systems and the versatility of quasiparticle (QP) wavefunctions to describe complex collective behavior. This motivated us to develop CC methods that can be formulated in terms of general QPs. These wavefunctions capture a large fraction of the static correlation, so to correct for the missing dynamic correlation we are working on Machine Learning-inspired operators that provide extra flexibility without greatly increasing the computational cost.

Quasiparticle, couple cluster, and machine learning-based approaches to strongly correlated systems

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.