CCL:G: AIMall energy of water dimer gives huge error with CP/ BSSE correction

 Sent to CCL by: "Laura  Albrecht" [lauraalbrecht[*]dal.ca]
 Hi Corey!
 It turns out that the .fchk file written by Gaussian does not include the MP2
 density matrix when counterpoise option is requested.  I'm trying to run with
 the old massage option to set the other fragment's nuclei to charge zero and no
 electrons (I would need to evaluate the wavefunction for each system
 individually as if I were doing the counterpoise-type correction manually - TAKs
 suggestion) but now formchk utility errors when I try to write the formatted
 checkpoint file (for the massage option, or also if I set the other fragment up
 as ghost atoms with A-Bq, which is essentially the same thing). Am right now
 trying to run with output=wfn in the route section.
 If it's the dimension causing this, what does that mean for the AIM
 calculations? (ie, can I fix this?)
 cheers, and thanks!
 Laura
 > "Cory Pye cpye[a]ap.smu.ca"  wrote:
 >
 > Sent to CCL by: Cory Pye [cpye ~~ ap.smu.ca]
 > Hello Laura (from down the street),
 > sounds like it might be a problem with dimension. When a BSSE calculation
 is
 > done using the counterpoise correction, the wavefunction of the monomer is
 > expanded to include the basis functions of the other monomer. So the
 effective
 > basis set size is different. The way Gaussian does this is to add the basis
 functions of the other monomer at the same geometry, but setting the nuclear
 charge to zero.
 > -Cory
 >
 >
 > On Tue, 11 Oct 2011, Laura Albrecht lauraalbrecht/./dal.ca wrote:
 >
 > >
 > > Sent to CCL by: "Laura  Albrecht" [lauraalbrecht..dal.ca]
 > > Hello CCL subscribers.
 > >
 > >
 > > I am attempting calculation of water dimers at MP2/aug-cc-pVTZ level
 theory corrected for BSSE with the Boys and Bernardi counterpoise approach
 (implemented with the counterpoise=2 option in G09 vs A.02). After a single
 point calculation the formatted checkpoint file is analyzed using the AIMall
 programs. I use AimQB to write the .wfn file.
 > >
 > > The resulting total energies as given by AIMall are ~318 kcal/mol
 greater than the G09 EUPM2 final energies. If I neglect the CP correction in G09
 calculation, the AIM calculated total energies are only ~0.01 kcal/mol greater
 than the (more accurate) G09 energies. AIM energies are taken from the .sumviz
 file, no modifications.
 > >
 > > Optimizing the geometry to include CP correction does not solve the
 issue, and neither does improving the basis set. The virial ratio is 2.00173 for
 single point and 2.000978 for optimized (no CP) and 2.001699 (with CP)
 > >
 > > Is there a known issue with AIMall calculations on G09/Counterpoise
 output? Is it possible that G09 writes the (formatted)checkpoint file in a
 different way when it applies CP correction, and thus the aimQB program is
 misprinting the .wfn file? Finally, is it even necessary to use the BSSE
 corrected .fchk file for the AIM calculation since the final wavefunction should
 be essentially the same, since CP correction on a SP calculation just yields an
 energy correction factor and does not change the dimer electron density topology
 (?).
 > >
 > > Any hints or literature direction would be immensely appreciated.
 > >
 > >
 > > Sincerely,
 > >
 > >
 > > Laura Albrecht
 > > Dalhousie University,
 > > Halifax, NS, Canada>
 > >
 >
 >    *************    !  Dr. Cory C. Pye
 >  *****************  !  Associate Professor
 > ***   **    **  **  !  Theoretical and Computational Chemistry
 > **   *  ****        !  Department of Chemistry, Saint Mary's University
 > **      *  *        !  923 Robie Street, Halifax, NS B3H 3C3
 > **      *  *        !  cpye{:}crux.stmarys.ca   http://apwww.stmarys.ca/~cpye
 > ***     *  *    **  !  Ph: (902)-420-5654  FAX:(902)-496-8104
 >  *****************  !
 >    *************    !  Les Hartree-Focks (Apologies to Montreal Canadien
 Fans)
 >
 >