Empirical valence bond reliability

QM/MM (AM1/CHARMM) or empirical valence bond methods, respectively164 168), for forcing the conformation of chorismate in solution into the more restricted conformation found in the enzyme. This equates to a catalytic benefit of only around 40-55% of the total AA G between enzyme and solvent. The good agreement between these findings, which applied completely different theoretical methods, is striking and suggests that this is a reliable result. 

During our search for reliable methods for studies of enzymatic reactions it became apparent that, in studies of chemical reactions, it is more physical to calibrate surfaces that reflect bond properties (that is, valence bond-based, VB, surfaces) than to calibrate surfaces that reflect atomic properties (for example, MO-based surfaces). Furthermore, it appears to be very advantageous to force the potential surfaces to reproduce the experimental results of the broken fragments at infinite separation in solution. This can be effectively accomplished with the VB picture. The resulting empirical valence bond (EVB) method has been discussed extensively elsewhere [3, 24], but its main features will be outlined below, because it provides the most direct microscopic connection to PT processes. 

Numerical values of thermochemical bond energy terms, derived from heats of atomization of molecules, are given in this chapter. An attempt is made to assess the reliability of the experimental information on which they are based. Only molecules for which an unambiguous valence bond structure can be written are discussed so that the question of empirical resonance energies is not treated. 

---