[molpro-user] State-Specific CASSCF?

[Alan Chien]

I think I have a resolution.

It seems that the problem may be due to the order of the states. The two
excited states are rather close in energy, and so targeting the state that
I want may be chancy at best. Especially considering the fact that the
geometry is taken from force field energy minimizations.

State-averaging over three states with SymmOff gives CI vectors that
qualitatively match up with those of the state-specific SymmOn calcs.

Extra opinions always welcome though.
~Alan Chien


On Wed, May 21, 2014 at 8:34 AM, Alan Chien <alandc at umich.edu> wrote:

> Dear Molpro users,
>
> I'm trying to do state-specific CASSCF and wanted to check whether the
> same states are found with symm on or off. Im checking if the states are
> the same based on the CI vectors that I get out
>
> I get the exact same coefficients for the ground state. But I'm having
> trouble getting the excited states to match. My input for the casscf
> portion is as below
>
> *Symm Off*
> basis=6-31G
>                                                           gprint,civector
> hf
>
> {casscf
>
> closed,72     !core orbs
>
> occ,76        !active orbs
>
> wf,148,1,0    !nelectrons,symm,singlet
>
> state,3       !nstates
>
> weight,0,0,1
>
> pspace,100.0
>
> }
>
> *Symm On (C2h symmetry)*
> basis=6-31G
> gprint,civector
> hf
> {casscf
> closed,31,6,30,5 !core orbs      - 72 total
> occ,31,8,30,7     !active orbs    - chosen for HOMO-1,HOMO,LUMO,LUMO+1, 76
> total
> wf,148,1,0 !nelectrons,Ag symm,singlet
> state,2
> weight,0,1
> pspace,100.0  !cause small active space, so include everything in
> }
>
> Assuming the third excited state with symm off is indeed the same as the
> 2nd Ag state, the CI vectors for the third SymmOff state and the 2nd SymmOn
> state should be identical. But they are not.
>
> My active space is identical, I double checked and the fact that the
> ground state turns up the same exact CI vec also supports this. Any
> suggestions? Am I using weight wrong?
>
> ~Alan Chien
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20140521/2bde0cdd/attachment.html>