[molpro-user] MS-CASPT2 for MOLPRO

[Mitchell Ong]

Dear Molpro users,

I have a few questions about the MS-CASPT2 routine implemented in MOLPRO 2006.1.  

1.)  I would get in the output:

 Recalculated energy wrong
 Eigenvalue(2)=           -155.24783247539050
 D2(t).Heff.D2=           -155.24783247539030
 difference    =    0.1989519660E-12 
 The recalculated energy D2(t).Heff.D2will be stored in MSENERGY

If the energy stored in MSENERGY is incorrect, what would be the correct MS-CASPT2 energy?  

2.)  Also, what is the difference between the Heff-Brandow MS-CASPT2 energy and the Heff Sum Symmetrized MS-CASPT2 energies?  Which one can I trust more?  One of my coworkers is seeing that states can change order from one definition to the other and the energy gaps between states can also be different.

3.)  Do the transition dipole moments at the end of the rs2 calculation correspond to single state CASPT2 or MS-CASPT2 if a MS-CASPT2 calculation is performed?

4.)  Finally, I tried to do a MS-CASPT2 gradient calculation for one of the excited states.  The input for the MS-CASPT2 part is:

 rs2,root=2,mix=4,shift=0.3
 occ,17;closed,13;core,4
 state, 4
 maxiter,100
 forces

I specified the gradient for state 2, but what I see printed out in the output is:

 RSPT2 GRADIENT FOR STATE 1.1

It says it's calculating the gradient for state 1, but I ask for state 2.  Is this gradient for state 2 but mislabeled or is this gradient for state 1?

Any insight to these questions would be appreciated.  Thanks

Mitchell
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20070409/69f18311/attachment.html>