Re: CCL:counterpoise with LMP2

[uccatvm <uccatvm (- at -) ucl.ac.uk>]

Hi Gerd,
 Thank you for your answer. However, my question was more specifically how
 to get the best orbital domain match for the dimer and monomer+ghost
 calculations in local MP2, and not so much how to do counterpoise
 calculations in general.
 The SCF-MI methods of Mayer and co-workers are a different approach to
 deal with BSSE, and are not without problems.  For example, with
 SCF-MI/VB method and the aug-cc-pV5Z basis set, the De of He2 is 10.64 K
 at 5.8 bohr - J. Mol. Struct. (Theochem) 549, 77, 2001. The FCI/aug-cc-pV5Z
 estimate is 10.17 K at 5.6 a.u. (J. Chem. Phys. 111, 9248, 1999).  I
 think this shows that SCF-MI overestimates the interaction energy.
 Tanja
 >
 > maybe the work of Istvan Mayer et al.
 >
 >   author = 	 {I. Mayer and \'A. Vib\'ok and G. Hal\'asz and P. Valiron},
 >   title = 	 {A BSSE-Free SCF Algorithm for Intermolecular
 > 		  Interactions. III. Generarlization for Three-Body
 > 		  Systems and for Using Bond Functions},
 >   journal = 	 {Int. J. Quantum Chem.},
 >   year = 	 {1996},
 >   volume =	 {57},
 >   pages =	 {1049}
 >
 > (see also the references there) is useful for you. This approach is
 > especially useful if you are going to use small basis sets.  Here you
 > also get a BSSE free wave function. Furthermore as far as I know you
 > will find also a discussion how the classical Boys Bernardi Method
 > should be applied - this should be sufficient if you just go for BSSE
 > free energies.
 >
 > Gerd
 >
 >    From: uccatvm <uccatvm (- at -) ucl.ac.uk>
 >    Date: Mon, 13 May 2002 21:56:43 +0100 (BST)
 >
 >    Hi all,
 >
 >    I am wondering what the most correct way is to do counterpoise with
 local
 >    MP2. In the local MP2 method originally proposed by Pulay
 (Chem.Phys.Lett.
 >    100, 151, 1983), to each localised MO a subset (orbital domain) of the
 >    virtual orbitals is assigned.  To calculate the interaction energy of a
 >    weakly interacting system, the orbital domains of the subsystems are
 first
 >    determined at large distance, and used in subsequent dimer calculations
 at
 >    smaller intermolecular distances (as recommended in for example Schutz
 et al.,
 >    J. Phys. Chem. 102, 5997, 1998).
 >
 >    Now, I assume that (to keep a true counterpoise) it is best to use the
 >    orbital domains determined at large distance for the monomer+ghost
 >    calculation. For this, one would first have to determine the domains of
 the
 >    monomer+ghost with the ghost at large R, and use the thus obtained
 orbital
 >    domains in the monomer+ghost calculation at the smaller distance. What
 >    do people think? Would this be the correct way of doing it?
 >
 >    Of course, the BSSE is strongly reduced in LMP2, and should in principle
 >    be negligible when using an appropriate basis set. However, when using
 >    small basis sets it may not be negligible, and I would like to know the
 >    best way of doing counterpoise for these cases.
 >
 >    Thanks in advance,
 >
 >    Tanja
 >
 >    --
 >      =====================================================================
 > 	Tanja van Mourik
 > 	Royal Society University Research Fellow
 > 	Chemistry Department
 > 	University College London    phone:    +44 (0)20-7679-4663
 > 	20 Gordon Street             e-mail:   work: T.vanMourik (- at -)
 ucl.ac.uk
 > 	London WC1H 0AJ, UK                    home: tanja (- at -) netcomuk.co.uk
 >
 > 	http://www.chem.ucl.ac.uk/people/vanmourik/index.html
 >      =====================================================================
 >
 >    -= This is automatically added to each message by mailing script =-
 >    CHEMISTRY (- at -) ccl.net -- To Everybody  | CHEMISTRY-REQUEST (- at -)
 ccl.net -- To Admins
 >    MAILSERV (- at -) ccl.net -- HELP CHEMISTRY or HELP SEARCH
 >    CHEMISTRY-SEARCH (- at -) ccl.net -- archive search    |    Gopher:
 gopher.ccl.net 70
 >    Ftp: ftp.ccl.net  |  WWW: http://www.ccl.net/chemistry/   | Jan: jkl (- at -) ccl.net
 >
 >
 >
 >
 >