[molpro-user] non-consecutive frozen-core orbitals

[Kirk Peterson]

Dear Adriana,

one just needs to use the rotate option to change the orbital ordering, 
e.g.,

ccsd(t);core,3,1,1,0;orbit,2100.2;rotate,3.1,4.1,90;

would swap the order of orbitals 3 and 4 in symmetry 1.

regards,

Kirk

On Jun 15, 2005, at 5:40 AM, Adriana Olleta wrote:

> Hello,
>
> I would like to get some help related to how to specify the orbitals 
> that
> should be frozen, if they are not in ascending order of energy.
>
> If I am running a CCSD calculation for the diatomic NaBr molecule with
> cc-pVDZ basis set for the Na atom and CRENBL ECP + corresponding basis 
> for
> the Br atom, by using G03 and MOLPRO the following problem occurs:
>
> When running these calculations with G03, five orbitals are frozen 
> which
> correspond to the (1s2s2p) set of the Na atom. In the case of MOLPRO, 
> the
> frozen orbitals additionally include the five 3d orbitals of bromine,
> totalling to a number of ten core orbitals.
>
> Although the approach of Molpro seems more consistent, I would like to 
> be
> able to reproduce the G03 calculation by performing the same type of
> calculation (with the five core orbitals 1s2s2p(Na)) using Molpro, and 
> at
> the same time correlate the 3d orbitals of Br.
>
> Since these 3d orbitals lie energetically in between the 1s(Na) and the
> 2s(Na) orbitals, it is not possible to freeze 1s2s2p(Na) only by
> specifying "core,5" (i.e., "core,3,1,1,0" in C2v symm).
>
> Rather, 1s(Na) and four of the five 3d orbitals of Br are frozen, while
> one 3d orbital and 2s2p(Na) are correlated.
>
> Is there a possibility to freeze orbitals, which are not in consecutive
> energetic ordering?
>
> Thank you
>
> Adriana