Abstract

The tailored approach is applied to the distinguishable cluster method together with a stochastic FCI solver (FCIQMC). It is demonstrated that the new method is more accurate than the corresponding tailored coupled cluster and the pure distinguishable cluster methods. An F12 correction for tailored methods and FCIQMC is introduced, which drastically improves the basis set convergence. A new black-box approach to define the active space using the natural orbitals from the distinguishable cluster is evaluated and found to be a convenient alternative to the usual CASSCF approach.

Highlights

  • Coupled cluster (CC) theory[1−3] is a reliable and standard tool in quantum chemistry to treat many-body electron correlations at high accuracy level.[4−7] the standard CC methods yield qualitatively erroneous results when applied to strongly correlated systems, e.g., multicenter transition metal complexes and other multiradicals, or in other highly degenerate situations encountered, e.g., in systems away from the equilibrium geometry

  • The high accuracy of the DC with singles and doubles (DCSD) method and improved stability to the onset of the static correlation suggest its superiority over CC with singles and doubles (CCSD) in the tailored formalism, and in this work we investigate the applicability of the tailored CC methodology to the DCSD method, denoted in the following as TDC

  • For all these systems the corresponding full configuration interaction quantum Monte Carlo (FCIQMC)-tailored CCSD (TCCSD) potential energy curves (PECs) have been calculated for comparison

Read more

Summary

INTRODUCTION

Coupled cluster (CC) theory[1−3] is a reliable and standard tool in quantum chemistry to treat many-body electron correlations at high accuracy level.[4−7] the standard CC methods yield qualitatively erroneous results when applied to strongly correlated systems, e.g., multicenter transition metal complexes and other multiradicals, or in other highly degenerate situations encountered, e.g., in systems away from the equilibrium geometry. Calculation of the external correction can become a bottleneck in this method, since the tailored methods generally require large active spaces to accurately describe multiradicals In this respect different ideas have been investigated, and especially the density-matrix-renormalization-group (DMRG) based TCC38−40 demonstrated very promising results. Apart from the usual (but expensive) CASSCF route, we employ natural orbitals from a preceding DCSD calculation, with occupation numbers significantly differing from zero and two, as the active orbital subspace for FCI or FCIQMC calculations This route is especially appealing as a basis set correction for the expensive FCIQMC calculations, and in this respect we introduce a simple F12 explicit correlation correction on top of the TCC/ TDC methods based on the Valeev’s perturbative F12 approach.[45]. The embedded FCI calculation is performed in the space of most active DCSD natural orbitals, and the correction to the total DCSD correlation energy is calculated as the difference between the embedded FCI energy and the embedded DCSD energy

OVERVIEW OF COMPUTATIONAL APPROACHES
RESULTS
CONCLUSIONS
■ REFERENCES
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call