[molpro-user] CI coefficients of a MCSCF calculation in the canonical basis

[Snake]

Dear Molpro users

I have a question about the CI expansion coefficients in a CASSCF
calculation. If I include in the input file
  gthresh, printci = 0
  GPRINT, civector
in the output is printed the CI vector in terms of determinants (In the
Natural MO basis). It's supposed that if is included "NATORB, CI" or
"CANORB, CI" in the MCSCF module, the Hamiltonian is diagonalized with
that's orbitals and the CI vector is printed in that basis.

However, I get the same values ​​for the CI coefficients using natural or
canonical orbitals (the difference betwen is less than 8.d-7 for the same
configuration). Also, if I compare the canonical and natural orbitals of
one particular calculation, the difference between the basis functions
coefficients is as well very small (the molecular orbitals are practically
the same).

I know that the transformation between natural to canonical orbitals must
change the CI expantion coefficients, but i don't know in that magnitude.
Seems that the transformation between both bases leaves the MO and the CI
coefficients invariant.

I would greatly appreciate any help that could provide.


Regards



José Jara
Universidad Nacional
Autonoma de México
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20150423/24fbe4fb/attachment.html>