The Ab Initio Method, and Simplifications of it

When one wants to determine the structure of a molecule by calculation, as opposed to by experiment, there are a variety of ways in which one may proceed. The most direct way in principle would be by solution of the Schrodinger equation for a given nuclear configuration, followed by an adjustment of nuclear configuration so as to minimize the energy of the molecule. This can be done 

A semi-empirical method which takes a somewhat different track is the MINDO method of Dewar (Bingham and Dewar, 1972). This method differs from previously mentioned semi-empirical methods in that it seeks to reproduce not the ab initio results which are in some respects inaccurate anyway, but rather the experimental results. The formalism is similar to that used in a CNDO calculation. However, by adjusting other quantities in the calculation, an effort is made to compensate for the lack of explicit inclusion of electron correlation. There are both strengths and weaknesses to this method. When suitably parameterized, it may give results which arc more useful and dependable than those obtained by ab initio calculation. At present this can be done very well in some respects, not so well in other respects. The computation time is similar to that required by the CNDO method. 

The extended Hiickel method (Hoffmann, 1963) might be mentioned here. This is a simple, very fast method, which can be used to gain qualitative insights into problems, but, because of the extensive approximations made, it cannot compete with the other methods mentioned in terms of general reliability and accuracy. It has the attractive feature that it enables reasonably straightforward interpretations to be made of various phenomena, but it is not generally applicable to the kind of problem being discussed here. 

---