Molcas Forum

jost

Member

Registered: 2016-03-03

Posts: 2

Calculation of TDMs with RASSI and with (MS)-CAPST2 and/or PCM

Hi all,

I got several questions regarding the use of the RASSI module.

1.) RASSI with MS-CASPT2:
The molcas8.0 manual states that it is usually sufficient to use CASSCF TDMs with CASPT2-energies (Section 10.5, p. 497).
Does this also apply if I want to use the calculated MS-CASPT2 states? As it is stated in the manual they often don't
really correspond to the former unmixed states, so it wouldn't be justified to use the CASSCF-TDMs, would it?

2.) RASSI with CASPT2:
If i want to compute TDMs at CASPT2-level is the "DENSITY"-keyword for the &caspt2-module required or isn't it related
at all to the calculation of the TDMs?

3.) RASSI with PCM:
What is the best strategy to get TDMs within a solvated calculation using PCM? I could think of several strategies:

3.1) RASSI between the single-state CASSCF ground state and every other calculated excited state, where the excited states
      are calculated with "NONEquilibirum" and the corresponding "RFROOT" keywords in &rasscf. I assume this approach would
      respect the fact that different excited states got different "fast"-parts of the reaction field.

3.2) RASSI between the ground state of a SA-CASSCF calculation and every other excited state of this SA-calculation with the reaction field
      of the SA-CASSCF ground state, that is the equilibrium reaction field. With this approach I would neglect the different "fast"
      parts of different excited states. Given the fact that the dipole moments of the ground state and the excited states don't
      differ that much this wouldn't be too much of a problem, would it?
      The input would look like this (only the relevant parts):
     
     

&rasscf
       ciroot
        10 10 1
       rfroot
        1
      
      &rassi
       rfpert
       cipr
       ejob         (assuming i also got a corresponding CASPT2-calculation)
       mees
       prop
        3
       'Mltpl  1'  1   'Mltpl  1'  2   'Mltpl  1'  3

   
       An advantage of this approach would be that i got only one calculation per symemtry in comparison to the next approach (3.3).
 
  3.3) Or the last option. RASSI between the ground state of a SA-CASSCF and every other state of interest within the reaction field
       of the excited state, that is using the "NONEquilibrium"-keyword in &rasscf. A hypothetical input to calculate the TDM for the
       first excited state would look like this:
       
       

&rasscf
        ciroot
         10 10 1
        nonequilibrium
        rfroot
         2
         
       >>LINK $Project.JobIph JOB002
       
       &rassi
        rfpert
        nr of jobiphs
         2 10 10
         1          (JOB001 contains the SA-CASSCF with the equilibrium reaction field)
         2          (JOB002 would contain the SA-CASSCF from the calculation with "noneq" and "rfroot 2")
        cipr
        ejob        (assuming i also got a corresponding CASPT2-calculation)
        mees
        prop
         3
       'Mltpl  1'  1   'Mltpl  1'  2   'Mltpl  1'  3  

 
      These calculations would have to be repeated for every state of interest.
     
Answers/help are greatly appreciated!

Best regards,
Johannes

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,012

Re: Calculation of TDMs with RASSI and with (MS)-CAPST2 and/or PCM

I'm not an expert, but these are my opinions and experiences:

1.) RASSI with MS-CASPT2:
The molcas8.0 manual states that it is usually sufficient to use CASSCF TDMs with CASPT2-energies (Section 10.5, p. 497).
Does this also apply if I want to use the calculated MS-CASPT2 states? As it is stated in the manual they often don't
really correspond to the former unmixed states, so it wouldn't be justified to use the CASSCF-TDMs, would it?

I think you should use the JOBMIX created my MS-CASPT2, which includes the mixed CASSCF states according to the MS-CASPT2 results.

2.) RASSI with CASPT2:
If i want to compute TDMs at CASPT2-level is the "DENSITY"-keyword for the &caspt2-module required or isn't it related
at all to the calculation of the TDMs?

I believe it's unrelated. RASSI operates with CASSCF states (and the CASPT2 energies if you specify so), properties calculated at CASPT2 level probably have no effect.

3.) RASSI with PCM:
What is the best strategy to get TDMs within a solvated calculation using PCM? I could think of several strategies:

Last time I thought about this, I found no strategy was satisfying, because at the end you have a single reaction field in the RASSI calculation, which can't be "correct". But, from a practical point of view, I'd chose an equilibrium SA-CASSCF calculation, especially if, as you say, the dipole moments of the states are similar. A possible improvement would be something like what is done for CASPT2: use equilibrium wavefunctions and nonequilibrium energies.