A Combined Selected Configuration Interaction and Many-Body Treatment of Static and Dynamical Correlation in Oligoacenes

TL;DR

This paper introduces a novel computational method combining adaptive configuration interaction with a density-fitted multireference perturbation approach to accurately predict electronic properties of oligoacenes, capturing both static and dynamical correlation effects.

Contribution

It develops ACI-DSRG-MRPT2, a new method that handles large active spaces and provides improved predictions of electronic properties in oligoacenes.

Findings

Accurately predicts singlet-triplet gaps in oligoacenes.

Reduces radical character estimates compared to previous methods.

Achieves good agreement with experimental data.

Abstract

We have combined our adaptive configuration interaction (ACI) [J.B. Schriber and F.A. Evangelista, J. Chem. Phys. 144, 161106 (2016)] with a density-fitted implementation of the second-order perturbative multireference driven similarity renormalization group (DSRG-MRPT2) [K.P. Hannon, C. Li and F.A. Evangelista J. Chem. Phys. 144, 204111 (2016)]. We use ACI reference wave functions to recover static correlation for active spaces larger than the conventional limit of 18 orbitals. The dynamical correlation is computed using the DSRG-MRPT2 to yield a complete treatment of electron correlation. We apply the resulting method, ACI-DSRG-MRPT2, to predict singlet-triplet gaps, metrics of open-shell character, and spin-spin correlation functions for the oligoacene series (2-7 rings). Our computations employ active spaces with as many as 30 electrons in 30 orbitals and up to 1350 basis functions, yielding gaps that are in good agreement with available experimental results. Large bases and reference relaxation lead to a significant reduction in the estimated radical character of the oligoacenes with respect to previous valence-only treatments of correlation effects.