A generalization of the state-specific complete-active-space coupled-cluster method for calculating electronic excited states

Abstract

The complete-active-space coupled-cluster approach with single and double excitations (CASCCSD) based on the ansatz of Oliphant and Adamowicz [J. Chem. Phys. 94, 1229 (1991); 96, 3739 (1992)] is used to derive an approach termed XCASCCSD for calculating potential energy surfaces of ground and excited electronic states with different multiplicities and symmetries. The XCASCCSD approach explicitly includes a procedure for spin and spatial orbital-momentum symmetry adaptation of the wave function that has allowed us to consider states with degenerate formal references. The XCASCCSD method is applied to calculate potential energy surfaces of the ground and some lowest singlet and triplet excited states of the FH and C(2) molecules. Some states of C(2) are known to have a very strong "multireference" character making their description difficult with single-reference methods. The problem of the change of the formal reference determinant along the potential energy surface is discussed. Also, vertical excitation energies of formaldehyde calculated with the XCASCCSD approach are presented.

Block Correlated Coupled Cluster Theory With A Complete Active-Space Self-Consistent-Field Reference Function: The General Formalism And Applications

“…Test applications have showed a very good performance of this approach in describing ground and excited electronic states with strong multireference character [120][121][122]. However, this approach computationally scales as the seventh power of the system size, and thus applications may be limited to quite small systems.…”

Dissociating N2: a multi-reference coupled cluster study on the potential energy surfaces of ground and excited states

“…Both rely on a single-reference Fermi vacuum and are not invariant to the choice of this vacuum. Most recently [81] the same authors tried to address the problem of symmetry breaking by a slight modification of CASCCSD they called XCASCCSD. By employing a symmetry adapted projection and symmetry adapting the t-amplitudes from their corresponding CI coefficients c they were able to improve the symmetry issues with CASCCSD.…”

“…Actually, there is no simple way to apply a reference CSF expansion to SRMRCC since the amplitudes spanning the reference space are 1. not unique for an individual reference and 2. reappear in product terms. It should be noted, that the very recently proposed scheme by Lyakh et al [81] tries to address the symmetry adaption problem at CI level. It applies a symmetry adapted projection and requires an iterative procedure propagating the t amplitudes to c coefficients, symmetry adapting the latter, and back-propagation to t amplitudes.…”

Multi-reference coupled-cluster study of the ionic-neutral curve crossing LiF

“…Although in many cases energetically very accurate the SRMRCC ansatz suffers from the problem of symmetry breaking and missing universality over a full potential surface as it requires a particular Fermi vacuum. Later variants of the SRM-RCC ansatz [48,[51][52][53][54] do not entirely solve this fundamental problem of the SRMRCC approach.…”

“…Recently [47] spin projection and state overlap errors were studied for N 2 . There are also recent attempts to partially cure the symmetry breaking within the SRMRCC ansatz [48]. The MRexpT ansatz is a state specific variant of the state universal approach of Jeziorzki and Monkhorst.…”