Dave,

> I am trying to correct for BSSE in some calculations using
>ghost orbitals. The calculations use relativistic core potentials
>as calculated with the DIRAC utility. How do I use these in
>the ghost orbital calculations?

Ghosts have no electrons, hence no core and/or total-atomic potential or
density, so there's nothing to replace with Dirac stuff. Answer: NO as far
as the Ghost-create runs are concerned, but YES, you do include them for
the 'normal' atoms, in the standard molecular calculation as well as in the
calculation with the ghost atoms added.

Hmm, come to think of it: it might be that the program starts complaining
about missing core-potentials info in the second run, with Ghosts added.
I'll check it out later and repair it if there is an 'impossible' check
inside the program.

>If I neglect the relativistic corrections in the ghost orbital part
>then the overall interaction energies come out positive (i.e.
>the complexes that I am calculating are un-bound).
> I have played around with the input files for DIRAC, but
>get error messages all the time.
> Anyone managed this?

Dirac will fail if there is no electronic density at all. About the
positive (un-bound) results: I've no idea whether that indicates an error
of some kind.

Hope this helps a little to clarify things.

Take care
Bert

-------------------------------------------------------------------
Bert te Velde SCIENTIFIC COMPUTING & MODELLING NV
Phone: +31-20-44 47 626 Vrije Universiteit, Theoretical Chemistry
Fax: +31-20-44 47 629 De Boelelaan 1083
E-mail:tevelde_at_scm.com 1081 HV Amsterdam; The Netherlands
                          http://www.scm.com
-------------------------------------------------------------------
Received on 1999-11-01 12:58:19