Exploration of Reduced Scaling Formulation of Equation of Motion Coupled-Cluster Singles and Doubles Based on State-Averaged Pair Natural Orbitals

Abstract

A reduced-complexity variant of equation-of-motion coupled-cluster singles and doubles (EOM-CCSD) method is formulated in terms of state-averaged excited state pair natural orbitals (PNO) designed to describe manifolds of excited states. State-averaged excited state PNOs for the {\em target} manifold are determined by averaging CIS(D) pair densities over the computational manifold. To assess the performance of PNO-EOM-CCSD approach on extended systems the new massively parallel canonical EOM-CCSD program has been developed in the Massively Parallel Quantum Chemistry program that allows treatment of systems with 50+ atoms using realistic basis sets with 1000+ functions. The use of state-averaged PNOs offers several potential advantages relative to the recently proposed state-specific PNOs: our approach is robust with respect to root flipping and state degeneracies, it is more economical when computing large manifolds of states, and it simplifies evaluation of transition-specific observables such as dipole moments. With the PNO truncation threshold of $10^{-7}$, the errors in excitation energies are on average below 0.02 eV for the first six singlet states of 28 organic molecules included in the standard test set of Thiel and co-workers (J. Chem. Phys. 2008, 128, 134110) with 50-70 state-averaged PNOs per pair.

28 pages, 6 figures

References in corpus (4)

• The density-matrix renormalization group
• Tensor factorizations of local second-order Møller Plesset theory
• Scalable Task-Based Algorithm for Multiplication of Block-Rank-Sparse Matrices
• Task-Based Algorithm for Matrix Multiplication: A Step Towards Block-Sparse Tensor Computing

Cited by in corpus (10)

• Rank-reduced coupled-cluster III. Tensor hypercontraction of the doubles amplitudes
• Sub-system quantum dynamics using coupled cluster downfolding techniques
• Optimized pair natural orbitals for the coupled cluster methods
• A complete OSV-MP2 analytical gradient theory for molecular structure and dynamics simulations
• A route to improving RPA excitation energies through its connection to equation-of-motion coupled cluster theory
• A Critical Analysis of Least-Squares Tensor Hypercontraction Applied to MP3
• "Best" iterative coupled-cluster triples model: More evidence for 3CC
• An Efficient Fock Space Multi-reference Coupled Cluster Method based on Natural Orbitals: Theory, Implementation, and Benchmark
• Applications of a perturbation-aware local correlation method to coupled cluster linear response properties
• Towards Balanced Description of Ground and Excited States with Transcorrelated F12 Methods