Time-dependent optimized coupled-cluster method for multielectron dynamics IV: Approximate consideration of the triple excitation amplitudes

TL;DR

This paper introduces a cost-effective, approximate triple excitation method within the time-dependent optimized coupled-cluster framework, enabling accurate multielectron dynamics simulations under intense laser fields with reduced computational cost.

Contribution

It develops the TD-OCCDT(4) method that considers triple excitations up to fourth-order, achieving O(N7) scaling and matching the accuracy of more expensive methods.

Findings

TD-OCCDT(4) accurately reproduces TD-CASSCF results.

The method effectively models nonlinear phenomena like ionization and high-harmonic generation.

Computational efficiency is significantly improved over full triple excitation methods.

Abstract

We present a cost-effective treatment of the triple excitation amplitudes in the time-dependent optimized coupled-cluster (TD-OCC) framework called TD-OCCDT(4) for studying intense laser-driven multielectron dynamics. It considers triple excitation amplitudes correct up to fourth-order in many-body perturbation theory and achieves a computational scaling of O(N7), with N being the number of active orbital functions. This method is applied to the electron dynamics in Ne and Ar atoms exposed to an intense near-infrared laser pulse with various intensities. We benchmark our results against the time-dependent complete-active-space self-consistent field (TD-CASSCF), time-dependent optimized coupled-cluster with double and triple excitations (TD-OCCDT), time-dependent optimized coupled-cluster with double excitations (TD-OCCD), and the time-dependent Hartree-Fock (TDHF) methods to understand how this approximate scheme performs in describing nonperturbatively nonlinear phenomena, such as field-induced ionization and high-harmonic generation. We find that the TD-OCCDT(4) method performs equally well as the TD-OCCDT method, almost perfectly reproducing the results of fully-correlated TD-CASSCF with a more favorable computational scaling.