Molcas Forum

David

Member

Registered: 2017-05-16

Posts: 85

Oscillator strength job

Hello everyone,according to the molcas 8.0 manual, I computed the oscillator strength in the RASSI program, results seem computed in the CASSCF level, can I compute the oscillator strength in the CASPT2 level? thanks in advance.

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Oscillator strength job

Not exactly. What you can do is use the CASSCF wavefunctions (or a mixture of them as given by a MS-CASPT2 calculation) to compute the transition dipoles, but using CASPT2 energies to compute energy differences. That should give you a good enough approximation. To do this use JOBMIX files for RASSI (instead of JOBIPH) and add the EJOB keyword.

David

Member

Registered: 2017-05-16

Posts: 85

Re: Oscillator strength job

Dear Ignacio, if I understand correctly, is that you mean I need to calculate transition dipole moments (TDM) and CASPT2 energy differences (Delta E)  separately, and then use oscillator strength formula (f=2*(TDM)^2*(Delta E)/3) computes final value manually?

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Oscillator strength job

No need to do it manually, RASSI does is as I said.

David

Member

Registered: 2017-05-16

Posts: 85

Re: Oscillator strength job

Dear Ignacio, this is my example input. I noticed that MS-CASPT2 energies are matched with that in RASSI output, does this mean that oscillator strength was calculated at the MS-CASPT2 level instead of CASPT2 level?

&GATEWAY
        coord   =       se.Opt.xyz
        Basis   =       ano-rcc-vdzp
        Group   =       nosym
        RICD
&SEWARD
&SCF
&RASSCF
        nactel  = 10 0 0 
        ras2    = 8 
        charge  = 0 
        spin    = 1 
        ciroot  = 5 5 1 
        lumorb
&CASPT2
        multi   = 5 1 2 3 4 5 
        imag    = 0.2 
        ipea    = 0.0 

        >>COPY $Project.JobMix JOB001
&RASSI &END
        Nrofjobiphs
        1       5   
        1 2 3 4 5 
        Ejob

David

Member

Registered: 2017-05-16

Posts: 85

Re: Oscillator strength job

~/recent.work/benchmark/osc $ grep '::    CASPT2 Root' sp.log 
::    CASPT2 Root  1     Total energy:  -3288.68896470
::    CASPT2 Root  2     Total energy:  -3288.58062449
::    CASPT2 Root  3     Total energy:  -3288.56993134
::    CASPT2 Root  4     Total energy:  -3288.50069595
::    CASPT2 Root  5     Total energy:  -3288.48470045
~/recent.work/benchmark/osc $ grep MS- sp.log 
       Total MS-CASPT2 energies:
::    MS-CASPT2 Root  1     Total energy:  -3288.69232170
::    MS-CASPT2 Root  2     Total energy:  -3288.58078727
::    MS-CASPT2 Root  3     Total energy:  -3288.56644503
::    MS-CASPT2 Root  4     Total energy:  -3288.50069819
::    MS-CASPT2 Root  5     Total energy:  -3288.48466474
~/recent.work/benchmark/osc $ grep '::    RASSI State' sp.log 
::    RASSI State    1     Total energy:  -3288.69232170
::    RASSI State    2     Total energy:  -3288.58078727
::    RASSI State    3     Total energy:  -3288.56644503
::    RASSI State    4     Total energy:  -3288.50069819
::    RASSI State    5     Total energy:  -3288.48466474

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Oscillator strength job

It is as I said in post #2: they are computed with CASSCF wavefunctions (although mixed in the same way as MS-CASPT2 mixes them), and with MS-CASPT2 energies.