Perturbative triples correction for the equation-of-motion coupled-cluster wave functions with single and double substitutions for ionized states: Theory, implementation, and examples

Abstract

A noniterative N(6) triples energy correction is presented for the equation-of-motion coupled-cluster method with single and double substitutions for ionized states (EOM-IP-CCSD). The correction, which is size intensive, is derived using a second-order Rayleigh-Schrodinger perturbative treatment and is similar to the approach of Stanton and Gauss [Theor. Chim. Acta 93, 303 (1996)]. In the present implementation, only the target EOM-IP states are corrected, and the reference state is described by CCSD; the method is therefore more useful for the study of the target states themselves than ionization potentials. The performance of the correction, which demonstrates the caveat above, is demonstrated by applications to singlet methylene, BNB(-), nitrogen, carbon monoxide, acetylene, benzene, thymine, and adenine.

Non-iterative triple excitations in equation-of-motion coupled-cluster theory for electron attachment with applications to bound and temporary anions

“…75 have to be used instead of aug-cc-pVX Z to obtain converged energies for the parent Rydberg states and moreover, the CAP onset has to be chosen much larger than the spatial extent of the neutral ground state and also that of the parent states in order to limit the perturbation of the latter states to an acceptable level. EOM-EA-CCSD* and EOM-EA-CCSD(fT), which correct only target states, do not yield accurate attachment energies, although significant improvements over EOM-CCSD have been reported for vertical ionization potentials 49 and energy differences between EOM-SF states 48 . In contrast, the EOM-EA-CCSD(T)(a)* method, in which the CCSD reference state is corrected for the effect of triples excitations before constructing the similaritytransformed Hamiltonian, reliably improves upon EOM-EA-CCSD attachment energies and is comparable in accuracy to EOM-EA-CC3.…”

“…There have also been many efforts [41][42][43][44][45][46][47][48][49][50] to devise EOM-CC methods that account for the effect of triples excitations in a non-iterative fashion similar to CCSD(T) 51,52 for groundstate energies and properties. It has become apparent, however, that a balanced correction of several states is more difficult to achieve than that of a single state.…”

“…As for the first objective, EOM-CCSD* 44,45 and EOM-CCSD(fT) 48,49 are investigated as examples of methods that correct only the target state and the novel EOM-CCSD(T)(a)* method 50 is considered because it includes a correction to the CCSD reference state as well. As for the second objective, different effects can be anticipated for shape and Feshbach resonances.…”

The effect of microhydration on ionization energies of thymine

“…Typical errors in excitation energies associated with "single excitation processes" for EOM-CC with single and double substitutions are in the 0.10.3 eV range [37,38]. However, energy differences between different ionized states are described much better (errors below 0.05 eV) [23].…”

Perturbative order analysis of the similarity transformed Hamiltonian in Fock-space coupled cluster theory: difference energy and electric response properties

“…However, inclusion of triples in the calculation gives enhanced accuracy at the expense of added computational cost [18][19][20][21][22][23][24][25][26]. The inclusion of partial as well as full triples has been tried within the framework of both EOMCC [27][28][29][30][31][32][33] and FSMRCC [34][35][36][37][38][39]. Significant development has also taken place in the context of analytic derivative calculation within the framework of EOMCC [40][41][42][43][44].…”