Inquiring about EOM-CCSD calculation of ionization from excited states

Questions
#1

I would like to calculate dyson orbitals for following cases,

EE_state = (HOMO-i)^1…(LUMO+j)^1 -----> IP_state = (HOMO-i)^0…(LUMO+j)^1 or (HOMO-i)^1…(HOMO-k)^1…(LUMO+j)^1
(i,j,k are natural numbers)

These states differ by 1 electron, but the neutral ground state and this IP_state differ by 2 electrons.
If I calculated IP states by reference to a neutral ground state, I could not find most of such IP states.
Are there any general methods of calculation for such cases?

#2

For general information on computing Dyson orbitals from excited states, please refer to 7.10.28 Dyson Orbitals for Ionized or Attached States within the EOM-CCSD Formalism‣ 7.10 Coupled-Cluster Excited-State and Open-Shell Methods ‣ Chapter 7 Open-Shell and Excited-State Methods ‣ Q-Chem 5.4 User’s Manual.

As for the problem with IP states not being single excitations from the reference ground state: One option is to start from the triplet reference and calculate Dyson orbitals for the SF → IP states. In both examples you gave, the SF and IP states would be only one excitation away from the triplet reference.

Another thing you can try is to modify your reference state configuration using the MOM_START option such that your final EE and IP states are only one excitation away from the reference:
https://manual.q-chem.com/5.4/Ch7.S6.SS1.html

1 Like
#3

Yes, this is the way to go – playing with combining different EOM models to find a reference such that both the initial and the final state in Dyson orbital calculations can be described by single-excitation EOM operator (1h, 1h1p, 2h, 2p, etc). This paper illustrates similar issues in the context of XAS calculations: https://aca.scitation.org/doi/full/10.1063/4.0000070 and suggests solutions.

1 Like
#4

Thank you for your kind replies. Starting from the triplet reference seems only applicable to ionization from HOMO-LUMO excited states (i=j=0 in above cases). I understood that an appropriate 1 electron excitation combination and an appropriate reference state should be chosen according to each transition processes. I also refer to papers you provided.