Size-extensivity correction for the state-specific multireference Brillouin–Wigner coupled-cluster theory

2000 ◽  
Vol 112 (20) ◽  
pp. 8779-8784 ◽  
Author(s):  
Ivan Hubač ◽  
Jiřı́ Pittner ◽  
Petr Čársky
Keyword(s):  
The State ◽  
Coupled Cluster ◽  
Coupled Cluster Theory ◽  
Cluster Theory ◽  
Size Extensivity

The Pole Structure of the Dynamical Polarizability Tensor in Equation-of-Motion Coupled-Cluster Theory

2018 ◽  
Author(s):  
Kaushik Nanda ◽  
Anna Krylov ◽  
Jürgen Gauss
Keyword(s):  
Equation Of Motion ◽  
Reference State ◽  
Polarizability Tensor ◽  
Coupled Cluster ◽  
Amplitude Response ◽  
Major Drawback ◽  
Coupled Cluster Theory ◽  
Cluster Theory ◽  
Size Extensivity ◽  
Pole Structure

<div>In this Letter, we investigate the pole structure of dynamical polarizabilities computed within the equation-of-motion coupled-cluster (EOM-CC) theory. We show, both theoretically and numerically, that approximate EOM-CC schemes such as, for example, the EOM-CC singles and doubles (EOM-CCSD) model exhibit an incorrect pole structure in which the poles that reflect the excitations from the target state (i.e., the EOM-CC state) are supplemented by artificial poles due to excitations from the coupled-cluster (CC) reference state. These artificial poles can be avoided</div><div>by skipping the amplitude response and reverting to a sum-over-states formulation. While numerical results are generally in favor of such a solution, its major drawback</div><div>is that this scheme violates size extensivity.</div>

The Pole Structure of the Dynamical Polarizability Tensor in Equation-of-Motion Coupled-Cluster Theory

2018 ◽  
Author(s):  
Kaushik Nanda ◽  
Anna Krylov ◽  
Jürgen Gauss
Keyword(s):  
Equation Of Motion ◽  
Reference State ◽  
Polarizability Tensor ◽  
Coupled Cluster ◽  
Amplitude Response ◽  
Major Drawback ◽  
Coupled Cluster Theory ◽  
Cluster Theory ◽  
Size Extensivity ◽  
Pole Structure

<div>In this Letter, we investigate the pole structure of dynamical polarizabilities computed within the equation-of-motion coupled-cluster (EOM-CC) theory. We show, both theoretically and numerically, that approximate EOM-CC schemes such as, for example, the EOM-CC singles and doubles (EOM-CCSD) model exhibit an incorrect pole structure in which the poles that reflect the excitations from the target state (i.e., the EOM-CC state) are supplemented by artificial poles due to excitations from the coupled-cluster (CC) reference state. These artificial poles can be avoided</div><div>by skipping the amplitude response and reverting to a sum-over-states formulation. While numerical results are generally in favor of such a solution, its major drawback</div><div>is that this scheme violates size extensivity.</div>

A Case Study of State-Specific and State-Averaged Brueckner Equation-of-Motion Coupled-Cluster Theory: The Ionic-Covalent Avoided Crossing in Lithium Fluoride

2005 ◽  
Vol 70 (8) ◽  
pp. 1082-1108 ◽  
Author(s):  
Marcel Nooijen ◽  
K. R. Shamasundar
Keyword(s):  
Molecular System ◽  
Solution Process ◽  
Iterative Solution ◽  
Equation Of Motion ◽  
The State ◽  
Basis Set ◽  
Coupled Cluster ◽  
Coupled Cluster Theory ◽  
Cluster Theory ◽  
Avoided Crossing

State-specific Brueckner equation-of-motion coupled-cluster theory (SS-B-EOMCC) is summarized, which can be considered an internally contracted version of a state-selective multireference coupled-cluster theory, which, however, is not entirely size-consistent. The method is applicable to general multireference problems, adheres to the space and spin symmetries of the molecular system, is straightforwardly extended to a state-averaged version, and has an associated perturbative variant which yields results close to the full coupled-cluster treatment. A key strength is that Brueckner orbitals are used, such that orbitals are optimized in the presence of dynamic correlation. A number of variations on the theme of SS-EOMCC is applied to study the ionic-covalent avoided crossing in LiF in a 6-311++G(3df,3pd) basis set. While reasonable results are obtained at the state-averaged level, the iterative solution process does not consistently converge for SS-EOMCC, due to the non-Hermiticity of the transformed Hamiltonian which may yield complex eigenvalues upon truncated diagonalization. This leads to an irrevocable breakdown of the state-specific EOMCC approach. We indicate some future directions that can resolve some of the problems with the SS-EOMCC methodology, as revealed by the demanding test case of the LiF potential energy curves.

Sign in / Sign up

Export Citation Format

Share Document