From chemistry-request@server.ccl.net  Thu Dec  9 18:37:44 1999
Received: from darkwing.uoregon.edu (genghis@darkwing.uoregon.edu [128.223.142.13])
	by server.ccl.net (8.8.7/8.8.7) with ESMTP id SAA11074
	for <chemistry@ccl.net>; Thu, 9 Dec 1999 18:37:44 -0500
Received: from localhost (genghis@localhost)
	by darkwing.uoregon.edu (8.9.3/8.9.3) with SMTP id OAA08750
	for <chemistry@ccl.net>; Thu, 9 Dec 1999 14:28:27 -0800 (PST)
Date: Thu, 9 Dec 1999 14:28:27 -0800 (PST)
From: "Dale A. Braden" <genghis@darkwing.uoregon.edu>
To: cclpost <chemistry@ccl.net>
Subject: NBO vs AIM
Message-ID: <Pine.GSO.3.96.991209134415.11039A-100000@darkwing.uoregon.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

Dear CCL,

I would like to solicit opinions about the relative merits and demerits of
using the natural bond orbital (NBO) method of population analysis versus
the atoms-in-molecules (AIM) method.  We reported both quantities in a
paper, leaving it to the reader's preference, and found ourselves in the
delightful position of being told by one reviewer that NBO is not reliable
and to use AIM only, and the other of whom said not to trust AIM, but
stick with NBO.  Catch-22.

As I understand it, the main argument against NBO is that against any
orbital-based method: orbital-based methods leave artifacts because the
occupancies of diffuse orbitals are weighted towards the atom on which
they are centered.  Another argument is that an NBO analysis is based on
the best localized set of orbitals for the molecule, and if the molecule
is in fact very delocalized, then the NBO results may not be very
sensible.

At least one argument against AIM is that it often gives larger values of
atomic charges (+ or -) for atoms connected by bonds that are expected to
be covalent. 

I also understand that there is no right answer to the contrived problem
of population analysis, but I would like to hear opinions from those who
have made or seen lots of comparisons between these methods.  For what
it's worth, and to help provoke responses, I think AIM makes more sense
for quantifying the total charge or spin around a nucleus, and if the
total wave function describes the kind of charge distribution seen in an
X-ray diffraction experiment, then an AIM analysis would be the way to
break it up into atomic contributions, and the unobservable orbitals need
not be considered.  If AIM charges are not what is expected, then it could
be a sign that AIM is not reliable, or it could equally well be a sign
that traditional notions of covalency/ionicity are in need of revision. 
If AIM describes ethane (in broken symmetry) as an ionic species (CH3-,
CH3+), then I will change my mind. 

What do you think?  I will certainly summarize the arguments for and
against each method.  If you want anonymity (to avoid flame wars), respond
to me personally, because I will just summarize the arguments pro and con,
without names.

Best wishes to all,

Dale

Dale Braden
Department of Chemistry
University of Oregon
Eugene, OR 97403-1253
genghis@darkwing.uoregon.edu