[molpro-user] some general questions about Multi code

Mehdi D. Davari mehdi.davari at gmail.com
Wed Feb 22 20:07:01 GMT 2012

I apologize for sending this message again.

Hello Molproers,

I was brought up with Gaussian as my main computational chemistry
package during the last years. I moved to Molpro recently.
I am wondered to ask some comparative questions about MCSCF code
inside Molpro as following.

First, in CASSCF calculations by Gaussian usually if we correlate
some, let,s say unimportant orbitals in the active space we got
warnings of large rotation of orbitals. What is the equivalent
mechanism in Molpro?

Second, if we obtain a set of orbitals for a geometry (e.g. planar
structure), and then if we want to dump this set of orbitals to a
neighboring twisted geometry by changing the structure abruptly we
will have many warnings due to the large rotation of orbitals which
indicates that there could be problems. It is useful though to check
that the active space still contains the correct orbitals. The way to
solve that is to interpolate some structures between these two states
(changing the structure smoothly)  and project the orbitals in some
steps through a multi-step job. Does the same hold in Molpro? In other
words, How we can make sure that we dump the same orbitals obtain from
a structure to a connecting structure  with different geometry (other
than visualizations and checking the occupancy numbers)? And how we
could run calculation on different geometries in  a single input file?

Many thanks for your kind answers,

More information about the Molpro-user mailing list