Accurately describing strongly correlated electrons remains a challenge in quantum chemistry, especially for excited states. A promising category of methods that reach near-FCI results often with reduced cost are Selected CI (SCI) methods. Configuration interaction perturbatively selected iteratively (CIPSI) method is an SCI approach that works in the Slater determinant (SD) basis. This SD basis results in fast evaluation but fails to converge quickly enough when strong correlation is present. The result is that the variational calculation quickly becomes intractable. Our recent approach, TPSCI, delays the rapid increase in variational space dimension by working in a basis of Full CI tensor product states, where correlation energy is accounted for locally in the basis vectors. TPSCI thus converges faster than SD methods but at the cost of a more expensive Hamiltonian matrix element evaluation.  Previously, we had limitations on cluster sizes due to our use of FCI as our cluster solver. To overcome this limitation, we have more recently implemented a new Restricted Active Space CI (RASCI) cluster solver. In this talk I will describe the basic formalism, our new implementation, and provide numerical examples on a diverse range of systems including highly-entangled bimetallic complexes and polyaromatic hydrocarbons.

Moving to larger clusters with tensor product selected CI (TPSCI): RASCI cluster state solver

Where thousands of chemistry professionals meet to share ideas and advance scientific and technical knowledge. Spring 2023 abstracts, posters, and presentations will launch at the end of March.

ACS Spring 2023

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