Re: properties with castp2

Posted by Per Ake Malmqvist on June 25, 2001 at 10:01:58:

In Reply to: properties with castp2 posted by Jan Franz on June 25, 2001 at 09:19:21:

: Do molcas calculate the properties for CASPT2-
: wavefunctions on CASSCF-level (if you order only one
: root within the MS-CASPT2) or does is also use
: perturbation-theory to calculate the properties?

For the moment, CASPT2 routinely writes out Mulliken charges and
some properties at the end of each calculation. These are in
principle computed as expectation values for the (normalized)
wave function through 1st order. So it is computed for the
correlated wave function, but it is not exactly the same as
a perturbation-theory property (which is usually taken to mean
the derivative of the energy wrt an added perturbation).
Also, unless one has used the GRADIENT keyword, the active/active
part of ther density matrix is approximated.
Best regards,
Per Ake

Follow Ups:

Post a Followup

Name:
E-Mail:

Subject:

Comments:
: : Do molcas calculate the properties for CASPT2- : : wavefunctions on CASSCF-level (if you order only one : : root within the MS-CASPT2) or does is also use : : perturbation-theory to calculate the properties? : For the moment, CASPT2 routinely writes out Mulliken charges and : some properties at the end of each calculation. These are in : principle computed as expectation values for the (normalized) : wave function through 1st order. So it is computed for the : correlated wave function, but it is not exactly the same as : a perturbation-theory property (which is usually taken to mean : the derivative of the energy wrt an added perturbation). : Also, unless one has used the GRADIENT keyword, the active/active : part of ther density matrix is approximated. : Best regards, : Per Ake