[molpro-user] cartesian keyword and core orbitals

Dima Khoroshun dima at mpi-muelheim.mpg.de
Mon Jun 20 10:02:02 BST 2005

Dear Peter!

Thank you very much for the help!

It does indeed seem necessary to specify core explicitly for such heavy
elements, as you suggest; I have not considered a possibility of a Zn
core, I think such a core is much less well defined, and would prefer to
correlate 4s on Br (or whatever other MO(s), one per Br atom, happen to
be chosen as extra's) in my calculations.

On Mon, 20 Jun 2005, Peter REINHARDT wrote:
> in your input. I wonder how one should program orbital freezing
> in cartesian coordinates: internally the program will have to deal
> with spherical d's for the core (for the number of electrons frozen)
> and cartesian ones for the rest, leading to a real mess when counting
> orbitals. Or one has to add a line for adjusting the 'core' card
> automatically, since you rarely want to have a Zn core for your Br-.

I think there is really no problem whatsoever in definitions or
programming. Molecular, not atomic, orbitals are frozen, and I think it
really does not matter which AOs a low-energy MO is composed of, and it
should matter even less how these AOs are represented as a shell. Perhaps
the example of Br, which is an atom, creates an illusion of a connection
between MOs and AOs.

Dmitry Khoroshun

More information about the Molpro-user mailing list