From chemistry-request@www.ccl.net  Fri Feb 12 19:47:48 1999
Received: from bioorganic.mechanisms.ucsb.edu. ([128.111.114.116])
        by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id TAA11488
        Fri, 12 Feb 1999 19:47:47 -0500 (EST)
Received: (from kalju@localhost)
	by bioorganic.mechanisms.ucsb.edu. (8.9.1a/8.9.1) id QAA17983
	for chemistry@www.ccl.net; Fri, 12 Feb 1999 16:44:26 -0800 (PST)
From: "Kalju Kahn" <kalju@bioorganic.ucsb.edu>
Message-Id: <9902121644.ZM17981@bioorganic>
Date: Fri, 12 Feb 1999 16:44:26 -0800
X-Mailer: Z-Mail (3.2.3 08feb96 MediaMail)
To: chemistry@www.ccl.net
Subject: BSSE in calc. of activation energies?
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii



Dear CCLers,

I realize there has been some controversy about basis set
superposition error (BSSE), and that the current consensus
favors the view that full counterpoise procedure (CP) of
Boys and Bernardini is a good way to correct for BSSE.
(e.g. van Duijneveldt et. al. in Chem. Rev, 1994, 94, 1873-1885)

My question concerns application of BSSE correction via CP
to bimolecular reactions. Let's take gas phase SN2
displacement as an example. Here we have a two reactants, A and B,
forming first a bimolecular "reactants complex" (RC), followed by
transition state (TS):
          A + B -> (A...B)[RC] -> A-B[TS]
The activation barrier can be defined as an energy difference
between TS and RC. What is the correct way to calculate this
energy difference? (Assuming Born-Oppenheimer approximation,
and dealing with size-consistent methods, e.g. HF level of theory)

My own answer would be:
 If one accepts a view that energies of both RC and TS
 (relative to isolated reactants) are affected by BSSE,
 one should correct both according to the CP scheme, and now:
 E_act = E(TS,TS_basis_in_TS_geometry)-E(RC,RC_basis_in_RC_geometry)-
        E(A,TS_basis_in_TS_geometry) -E(B,TS_basis_in_TS_geometry)+
        E(A,RC_basis_in_RC_geometry) +E(B,RC_basis_in_RC_geometry)

 Obviously, the number of basis functions in TS_basis and RC_basis
 is the same, but the geometrical position of ghost orbitals relative
 to partner molecule are different in TS_geometry and RC_geometry.
 So, the last four terms do not cancel out.

This seems a right way to account for BSSE, but the above method
was quoted as "misinterpretation of the BSSE concept" by Dr. J. A. Sordo
in "Comment on "Ab initio investigation of internal rotation in the
ethylene-sulfur dioxide dimer"" (J. Chem. Phys., 106(14), 6204).

Could somebody kindly explain me how do I misinterpret the BSSE
concept here? Or do we need to be concerned about BSSE when
calculating ab initio activation barriers?

Thanks a lot, Kalju

P.S. I will summarize your responses.

P.S.2. Happy Valentine's Day!

******************************************************************
Kalju Kahn				kalju@bioorganic.ucsb.edu
Chemistry Department			tel: (805)-893-7158
University of California Santa Barbara
Santa Barbara, CA 93106
##################################################################