Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A reduced-cost two-component relativistic equation-of-motion coupled cluster method for the double electron attachment problem

Sujan Mandal, Tamoghna Mukhopadhyay, Achintya Kumar Dutta

Abstract

We present a computationally efficient relativistic formulation of the equation-of-motion coupled-cluster method for the double electron attachment problem. In this work, the exact two-component Hamiltonian within the atomic mean-field approximation is employed, yielding results that are in close agreement with the corresponding four-component calculations. However, canonical DEA-EOM-CCSD calculations become prohibitively expensive for heavy elements and large basis sets due to the substantial memory requirements associated with complex 3p1h excitation manifold. To address this limitation, we introduce a state-specific frozen natural spinor basis that significantly reduces the virtual space through two controllable truncation thresholds. Furthermore, the use of Cholesky decomposition for the two-electron integrals provides an additional reduction in computational cost and memory. The performance of the proposed approach is demonstrated through calculations of double ionization potentials and excitation energies for group-12 and group-14 heavy elements. Vertical excitation energies for heavy chalcogen dimers are also presented. In addition, a range of diatomic spectroscopic constants is evaluated for group-13 halides.

A reduced-cost two-component relativistic equation-of-motion coupled cluster method for the double electron attachment problem

Abstract

We present a computationally efficient relativistic formulation of the equation-of-motion coupled-cluster method for the double electron attachment problem. In this work, the exact two-component Hamiltonian within the atomic mean-field approximation is employed, yielding results that are in close agreement with the corresponding four-component calculations. However, canonical DEA-EOM-CCSD calculations become prohibitively expensive for heavy elements and large basis sets due to the substantial memory requirements associated with complex 3p1h excitation manifold. To address this limitation, we introduce a state-specific frozen natural spinor basis that significantly reduces the virtual space through two controllable truncation thresholds. Furthermore, the use of Cholesky decomposition for the two-electron integrals provides an additional reduction in computational cost and memory. The performance of the proposed approach is demonstrated through calculations of double ionization potentials and excitation energies for group-12 and group-14 heavy elements. Vertical excitation energies for heavy chalcogen dimers are also presented. In addition, a range of diatomic spectroscopic constants is evaluated for group-13 halides.

Paper Structure

This paper contains 14 sections, 37 equations, 1 figure, 6 tables.

Table of Contents

  1. Introduction
  2. Theory
  3. The X2C Hamiltonian with AMF approximation
  4. EOM-CC theory for double electron attachment
  5. State-specific frozen natural spinors
  6. Cholesky Decomposition
  7. Computational details
  8. Results and discussions
  9. Selection of optimal truncation threshold
  10. Validation of CD-X2CAMF results against 4c SS-FNS-DEA-EOM-CCSD
  11. Basis set convergence
  12. Excitation energies of atoms
  13. Heavy diatomic molecules
  14. Conclusion

Figures (1)

  • Figure 1: Convergence of the absolute error in the DEA value (in eV) obtained using the SS-FNS-CD-X2CAMF-DEA-EOM-CCSD method for the InH molecule with the s-aug-dyall.v2z basis set. (a) Variation of the error with respect to the excited-state FNS threshold, where the untruncated canonical result is used as the reference. (b) Variation with respect to the ground-state FNS threshold, with the excited-state threshold fixed at $10^{-4.5}$; the corresponding value at this threshold is taken as the reference.