Outside the QTAIM realm, the main definitions of atomic charges are Mulliken
charge, electrostatic charge, and natural charge [3]. Values can differ considerably
from one method to another, but the trend with a particular method can provide
useful information. None of these three methods of assigning charge claims to be
unique.
References
- Bader RFW (1990) Atoms in molecules. A quantum theory. Clarendon Press, Oxford, particu-
larly chapter 5 - (a) Mata CF, Bader RFW (2006) J Phys Chem A 110:6365. (b) Bader RFW, Mata CF (2004)
J Phys Chem A 108:8354 - Levine IN (2000) Quantum chemistry, 5th edn. Prentice Hall, Upper Saddle River, NJ,
pp 505–510
Chapter 4, Harder Questions, Answers
Q7
It has been reported that the extended H€uckel method can be parameterized to give
good geometries. Do you think this might be possible for the simple H€uckel
method? Why or why not?
A report of a promising method of inducing theextendedH€uckel method EHM)
to yield good geometries appeared in 1994 [1]. The method was said to give
geometries as good as or better than the popular AM1 method, and to be two to
four times as fast. Unfortunately, further results, and the wide application of this
approach, do not seem to have followed (a possibly related approach [2] and
another fast semiempirical method [3] have been reported). Such a method would
be very useful, because the EHM is very fast, due to its very simple way of
calculating energies and molecular orbitals, and the fact that it is not iterative – a
single matrix diagonalization gives the results.
Recall that in the EHM geometric information is present in the Fock matrix by
virtue of the overlap integrals in the off-diagonal elements. For the simple H€uckel
method (SHM) the situation is completelydifferent.TheSHMdoesnottakeany
account of molecular geometry, as distinct from mere connectivity, with one
hedge: one can vary the adjacenti,jinteraction terms in an attempt to reflect
changes in overlap integrals. This can be done by allowing the terms to move from
#1toward0asabondislengthened,andbymakingtermsproportionaltothe
cosine of the angle of deviation from perfectp–pparallelalignmenttoaccountfor
nonplanarity [4]. Bond lengths and angles ofpsystems could be varied to give the
lowestpenergy. But the SHM method is tied topsystems, severely limiting the
Answers 617