[molpro-user] CASSCF, orbital reordering

[Dumitru SERGENTU]

Hello,

I am a user of Molpro, version 2009.1.
1) I am performing some CASSCF calculations on the H2O--AtO+ system and I
try to get in the active space the 6 orbitals of AtO+
(sigma,pi,pi,pi*,pi*,sigma*), knowing that the water molecule doesn't
perturb much the electronic strocture of AtO+ moiety.
2) Starting from correctly ordered ROHF orbitals, I always end up in the
active space with a dyz orbital of At.  If I try further to rotate the
orbitals it gives me the same result.
3) Starting from correctly ordered ROHF orbitals, using FROZEN keyword in
MULTI, I get (of course) the correct active space and the energies around
the expectations, if I take these orbitals and replace FROZEN with CLOSED,
I get again the dyz orbital in the cative space (energies getting higher
than in previous case when I used FROZEN).
4) Starting from atomic densities guess, I get the correct active space
Furrthermore I would like to include some lone pairs of the oxygen in the
water molecule, this task is so difficult since after reordering, FROZEN
etc, they switch back to their positions. Somehow, I found natural that
these orbitals switch back since I basically don't apply any shift after
the ordering, to help convergence.  I did not found in the Manual anything
related to this, I wonder how this all work?

Best,
Claudiu.