Default PCM during a TDDFT/PCM calcualtion

While there are ground state equilibrium and excited state non-equilibrium solvent approaches, it is confusing for these reasons:

  1. a default tddft/pcm calculations seems to lead to a LR-response to solvation, no state specific correction or fast solvent response
  2. it seems that state specific approach is only possible during non-equilibrium solvation which results in 0th order and perturbative LR and SS energies.
  3. So with a simple pt-ss approach, with simple default ground state equilibrium is not possible (as is possible for LR as mentioned in point 1)

It will be of great help if clarified, thanks!!

The easiest thing is to use an input file like this:

$rem
method hf ! or whatever
basis sto-3g ! for example
cis_n_roots 5
cis_triplets false
solvent_method pcm
$end

$pcm
NonEquilibrium
StateSpecific Perturb
$end

The key parts are in the $pcm section. This job will give you 0th order, LR, and SS corrections at the end. The cost to do all of them is fairly inconsequential overhead on top of normal TDDFT. I’ve omitted the $solvent section here, using defaults for water, equivalent to:

$solvent
dielectric 78.4
OpticalDielectric 1.78
$end

Thank you for the explanation @jherbert!
I understand this, so now my confusion is when we do a simple tddft without “Nonequilibrium” which should be the “no separation of charges” and so “no fast response from solvent” scheme and without mentioning “statespecific = perturb”, what kind of treatment are the excited states getting?

I believe we are performing an LR approach with no solvent’s fast electronic response?
Am I right? Which is not possible with SS? (SS seems to be possible with Nonequilibrium solvation only)

Thanks!

It appears that the LR-PCM is turned on by default for TDDFT with SOLVENT_METHOD=PCM, even when no $pcm section is included. That would be nonequilibrium solvation with the LR formalism. You should be able to check this, because with the input file that I suggested above, everything comes out nicely formatted, LR and SS (and 0th-order). Let me know if it looks different or if you can’t figure out the numbers. I’m not sure that I like the idea of LR being turned on by default (as compared to 0th-order only, which would be fine) and might disable this in a future release.

Yes, it appears to me too that LR is the default (not mentioned anywhere in the manual, as far I have searched and that is everywhere haha) If it was so written, I think there would be less confusion.

The difference, in my observation so far, when we specify “Nonequilibrium” is that we get pt-LR values to 0th order and not full LR values. My reasons:

  1. Qchem manual says something in this direction:
    “In the perturbation scheme, we can calculate the LR contribution using the zeroth-order transition density, in what we have called the “ptLR” approach.”

  2. Numbers for LR-PCM in default and with Nonequilibrium are different although close

Thanks!

With “NonEquilibrium” in the $pcm section you are definitely getting ptLR, as you suggest.

I was unaware until recently that LR is turned on by default, need to look into it more carefully. Thanks for investigating.

It appears that you can set TDDFT_PCM = FALSE in $rem and this will turn off LR-PCM, which would then just give you the “0th-order” correction, meaning normal TDDFT applied to MOs that are solvent-polarized in the ground state. This keyword does not seem to be documented in the manual. I will fix that for future versions of the manual.