PNNL

Abstract

In this paper we report on the parallel implementation of the coupled-cluster (CC) Green function formulation (GF-CC) employing single and double excitations in the cluster operator (GF-CCSD). The detailed description of the underlying algorithm is provided, including the structure of ionization-potential- and electron-affinity-type intermediate tensors which enable to formulate GF-CC approach in a computationally feasible form. Several examples including calculations of ionization-potentials and electron a*ffinities for benchmark systems, which are juxtaposed against the experimental values, provide an illustration of the accuracies attainable in the GFCCSD simulations. We also discuss the structure of the CCSD self energies and discuss approximation that are geared to reduce the computational cost while maintaining the pole structure of the full GF-CCSD approach.