Page 25 - Strategies and Applications in Quantum Chemistry From Molecular Astrophysics to Molecular Engineer
P. 25

10                                                                J. TOMASI

                            The concept  of atom,  for  example,  must be  accompanied by  a definition of  "valence
                            states",  based on  the  definition of hybrid  orbitals,  which play  a prominent role in the
                            definition of VB structures and also constitute an important component in models based on
                            localized orbitals (LO). Localized orbitals represent a bridge between partitions of a
                            molecule based on molecular orbitals and partitions based on a dissection of the charge
                            distribution into local subunits. LO give an alternative view of electron correlation holes,
                            and in a different context offer a good starting point for semiclassical segmental partitions
                            of the charge distribution. The alternative concept of charge partitioning according to the
                            values of some physical quantities  (e.g. the Bader's approach) challenge the traditional
                            partition into atoms or into LOs, offering new views of some concepts and permitting new
                            admixtures of tools.

                            I have selected few examples taken from the theory of molecular bonding, but this aspect
                            of competition and transference of concepts  is present in all the fields  of chemistry  -and
                            they are quite numerous- in which quantum theory has given contributions.
                            Another important point  is the connection  with in-depth calculations. We  have already
                            remarked that this connection is extremely important for the elaboration of interpretations.
                            There are two main lines, in my opinion.

                            The first could be called a constructive (synthetic)  approach. Pragmatic  and theoretical
                            reasons suggest to start computations appropriate for the model one has selected at a low
                            level of the theory; the goal is not to obtain the most complete report on the problem under
                            scrutiny but to reach a satisfactory degree of confidence in the interpretation obtained by
                            analyzing the results, and a set of reports at increasing levels of the mathematical model
                            may be of noticeable help in assessing the quality of the description. The constructive
                            approach may be also employed with the aim of getting accurate reports. In this case an
                            interpretative model is (tentatively) adopted, and the mathematical model is formulated
                            accordingly. The goal now is to get accurate results at a lower computational cost than
                            using brute force methods. Many formulations of this strategy have been elaborated, some
                            addressed to specific  problems, others of more general character. The quotation of an
                            example always make more clear what a general statement means; to this end we recall the
                            recent  generalized  multi-structural wavefunction  method (GMS) [21].  This approach
                            remind me  the  suggestion expressed by Dirac  in the last  lines of the  above  reported
                            quotation. It is worth to remark that the modern reformulations of the VB theory play an
                            important role in this field [22-23]. On the whole, the constructive approach is the most
                            potent and versatile method to connect chemical concepts to in-depth calculations,

                            The second  approach is  addressed to  elaborate methods  able to  derive  from accurate
                            calculations the points of interest for the interpretation  The strategy, in general, consists in
                            the adoption  of  a  simpler  model  (the mathematical  aspects of the model  are  again
                            concerned) and the task consists in reducing the information coming from the full in-depth
                            calculation  (not the the numerical  values of observables and other statutory  quantities
                            alone) to the level of the simpler model. For example accurate calculations may be reduced
                            at the level of a simple VB theory (Robb, Hiberty) or of a simple MO perturbation scheme
                            (Bernardi) making more transparent the interpretation.
                            Both  approaches are  subjected to criticisms  and to errors. The stipulation of the  model
                            constitutes a bias in favour of one among several possible alternatives, and the conclusions
                            will suffer if the choice was not appropriate. All the models, by definition, are subjected to
                            failures of this kind, and it rests with the users to exert their acumen to decide if the model
                            is applicable to the case under examination.
   20   21   22   23   24   25   26   27   28   29   30