[molpro-user] Questions about frozen core in SAPT calculation

[cong.wang]

Dear Dr. Hesselmann,

  Many thanks your reply. I was considering \delta HF term from HF-SAPT with
frozen core.  I thought density fitting (which can affect freeze core) is
only for non-hybrid DFT. Actually as mentioned in the manual, it is already
implemented for HF at level 1 and 2, which is sufficient for \delta HF
calculation. Now everything is clear.

Thanks once more for your help
Best regards
Cong Wang

Ph. D. Student
Department of Chemistry
Laboratory for Instruction in Swedish
University of Helsinki
A.I. Virtanens plats 1
P.O. Box 55
FI-00014 University of Helsinki
FINLAND

On Mon, Feb 22, 2010 at 5:40 PM, Andreas Hesselmann <
andreas.hesselmann at chemie.uni-erlangen.de> wrote:

> Dear Cong,
>
> yes, the core orbitals are usually defined as the
> orbitals with the lowest orbital energies.
>
> I am a bit uncertain about your 2nd question. If you
> are interested in accurate interaction energies you
> shouldn't use either HF or LHF for the monomer calculations
> as you wouln't account for intramonomer correlation effects.
> As described in the manual, it will be much better
> to use density functionals in combination with an
> asymptotic correction for the xc-potential (see manual).
>
> Best wishes,
> Andreas
>
>
>
>
>
>
> On Monday 22 February 2010 13:32, you wrote:
> > Dear Dr. Hesselmann,
> >
> >   Many thanks for your help!
> >
> >   Excuse me, what does "excluding the valence orbitals" mean? (will the
> > frozen orbital automatically be chosen from low to high orbital energy?)
> > Also, as you mentioned, the frozen core only affects the density-fitting
> > calculation, does it imply if I use Hartree-Fock monomer, it is
> impossible
> > to freeze any orbital unless circumvent the problem by lhf. If so, is
> there
> > any comparison available about the accuracy HF-SAPT vs LHF-SAPT for some
> > given set of molecules?
> >
> > Thank you very much
> > Best regards
> > Cong Wang
> > Ph. D. Student
> >
> > Department of Chemistry
> > Laboratory for Instruction in Swedish
> > University of Helsinki
> > A.I. Virtanens plats 1
> > P.O. Box 55
> > FI-00014 University of Helsinki
> > FINLAND
> >
> >
> >
> > On Mon, Feb 22, 2010 at 12:44 PM, Andreas Hesselmann <
> >
> > andreas.hesselmann at chemie.uni-erlangen.de> wrote:
> > > Dear Cong,
> > >
> > > The number to be specified in the frozena/b
> > > option is the number of occupied orbitals
> > > excluding the valence orbitals.
> > >
> > > The frozen core option will only effect the 2nd
> > > order terms in the density-fitting part of the
> > > SAPT program (and indunction only if SAPT_ICPKS=0).
> > > For the 1st order terms it is important to include
> > > all occupied orbitals, otherwise the interaction energy
> > > will not be very accurate for small distances of the
> > > monomers.
> > >
> > > Best wishes,
> > > Andreas
> > >
> > > On Saturday 20 February 2010 17:46, cong.wang wrote:
> > > > Dear everyone,
> > > >
> > > >   Hello,
> > > >
> > > >   Excuse me, I have two questions about the frozen core in the SAPT
> > > > calculation in molpro 2009.
> > > >
> > > > (i) in the manual,
> > > > http://www.molpro.net/info/current/doc/manual/node417.html
> > > >   it is said *
> > > >
> > > > *********
> > > >    SAPT_FROZENA*Number of frozen electrons in the response
> calculations
> > >
> > > for
> > >
> > > > monomer A (default 0) ***************
> > > >   If I understood correctly, it is the number of electrons (number of
> > > > orbital *2 in closed-shell case). But when I do a SAPT for neon
> dimer,
> > > > setting SAPT_FROZENA=6
> > > >
> > > >   the output file said
> > > >
> > > >
> > > >  Occupied space for monomer A:    5
> > > >  Virtual  space for monomer A:   87
> > > >
> > > >  Occupied space for monomer B:    5
> > > >  Virtual  space for monomer B:   87
> > > >  ? Error
> > > >  ? frozena>noa!
> > > >  ? The problem occurs in sapt_interface
> > > >
> > > >   does it imply the FROZENA actually is the number of orbital to be
> > >
> > > frozen?
> > >
> > > >  (ii) I set FROZENA=0,1,2; FROZENB=1, the  E1tot+E2tot  is invariant
> > > > respect with the frozen core set up. Why the frozen core does not
> > > > change the interaction energy?
> > > >
> > > > Here is an input file, if you need any other information, please let
> me
> > > > known
> > > >
> > > > *************************
> > > > memory,80,m
> > > > symmetry,nosym
> > > > geomtyp=xyz
> > > > geometry={
> > > > 2
> > > > title
> > > > Ne 0. 0. 0.
> > > > Ne 0. 0. 3.0}
> > > >
> > > > basis=avtz
> > > >
> > > > !wf records
> > > > ca=2101.2
> > > > cb=2102.2
> > > >
> > > >
> > > > !monomer A
> > > > dummy,2
> > > > {hf; save,$ca}
> > > > sapt;monomerA
> > > >
> > > > !monomer B
> > > > dummy,1
> > > > {hf; start,atdens; save,$cb}
> > > > sapt;monomerB
> > > >
> > > > !interaction contributions
> > > > {sapt,SAPT_FROZENA=6;intermol,ca=$ca,cb=$cb}
> > > >
> > > > *************************
> > > >
> > > >
> > > > Thank you very much in advance
> > > > Best regards
> > > > Cong Wang
> > > > Ph. D. Student
> > > >
> > > > Department of Chemistry
> > > > Laboratory for Instruction in Swedish
> > > > University of Helsinki
> > > > A.I. Virtanens plats 1
> > > > P.O. Box 55
> > > > FI-00014 University of Helsinki
> > > > FINLAND
> > >
> > > --
> > > --------------------------------------------------
> > > Andreas Hesselmann
> > > Institut für Physikalische und Theoretische Chemie
> > > Universität Erlangen
> > > Egerlandstraße 3
> > > 91058 Erlangen / Germany
> > > Phone:  +49 9131/85-25021
> > > E-Mail: andreas.hesselmann at chemie.uni-erlangen.de
> > > -------------------------------------------------
> > > _______________________________________________
> > > Molpro-user mailing list
> > > Molpro-user at molpro.net
> > > http://www.molpro.net/mailman/listinfo/molpro-user
>
> --
> --------------------------------------------------
> Andreas Hesselmann
> Institut für Physikalische und Theoretische Chemie
> Universität Erlangen
> Egerlandstraße 3
> 91058 Erlangen / Germany
> Phone:  +49 9131/85-25021
> E-Mail: andreas.hesselmann at chemie.uni-erlangen.de
> -------------------------------------------------
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20100222/d2c85cc4/attachment.html>