Completely renormalized EOM-CCSD(T)

[Rulin Feng]

Hi All,

I'm new to the community of EOM-CC and found NWChem useful for doing such calculations.

The CR-EOM-CC module is capable of doing a lot of great things, but I'm having some questions regarding some functionalities.

The CR-EOM-CCSD(T) supposedly only adds the non-iterative triples correction to the exited states but not to the ground state. This is, although not mathematically rigorous, but justified by good agreements with full T calculations.

DOI: 10.1063/1.3132592 introduced a method of balancing the correction to the ground state and to the excited state, however I couldn't find this on the NWChem TCE keyword list. I'm assuming it is not implemented yet?

Last but not the least, correct me if I'm wrong, spin-flipped EOM--CC methods are not implemented, so if I were to do such calculations in NWChem, I'd be doing two EOM-CC calculations with different reference spin states. 

If CR-EOM-CCSD(T) neglects triples correction to the ground states in each calculation, then the ground states would be of CCSD level of theory, and the excited states include triples correction they are triples corrected, so my question is, are the vertical transition energies between these two different calculations comparable, i.e. does it make sense to take one excited state energy from one calculation and subtract that off from the ground or excited state energy from the other calculation?

Furthermore, here are some naive thoughts, since in CR-EOM-CCSD(T), the ground states are not triples-corrected, how much sense does it make to do CR-CCSD(T) for the ground state only and get triples correction from there?

It is a bit long and I apologize but thanks for reading through!

[Kowalski, Karol]

Hi Rulin,

The nested variant of the CR-EOMCC(T) formalism   finds MMCC type corrections starting from the EOMCCSDT problem with the approximate EOMCCSD cluster amplitudes.

As a result, you are getting excited-state corrections due to Triples-H-(Singles+Doubles) & (Singles+Doubles)-H-Triples  blocks of the Hailtonian H  (+ some extra T-H-T terms, which are neglected due to comp. cost).

In fact, they are very similar in total values. If you assume that one of these corrections, for example (S+D)-H-T type,  cancels out the ground-state correction, then

the only surviving correction to the EOMCCSD excitation energy is T-H-(S+D) as in the CR-EOMCCSD(T).

For this reason the CR-EOMCCSD(T) works pretty well for excitation energies.

 

In general, practically in all excited-state corrections due to triples, the ground-state correction is neglected (for whatever reason).

Moreover, in some iterative approaches accounting for the effect of triples it is assumed that ground state is perfectly correlated (which is equivalent to neglecting ground-state corrections),

i.e., T-\bar{H}-reference-function block is assumed to be zero.

 

I hope it helps

Best,

Karol

 

 

 

From: <nwchem...@googlegroups.com> on behalf of Rulin Feng <studen...@gmail.com>
Reply-To: "nwchem...@googlegroups.com" <nwchem...@googlegroups.com>
Date: Saturday, September 4, 2021 at 10:20 AM
To: NWChem Forum <nwchem...@googlegroups.com>
Subject: [nwchem-forum] Completely renormalized EOM-CCSD(T)

 

Check twice before you click! This email originated from outside PNNL.

 

--
You received this message because you are subscribed to the Google Groups "NWChem Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/2515f81e-eb4e-46c9-b6a9-e314b24be451n%40googlegroups.com.