Quantitative relationships between structure empirical correlations

Quantitative structure/activity relationships (QSARs) for hydrolysis are based on the application of linear free energy relationships (LFERs) (Well, 1968). An LFER is an empirical correlation between the standard free energy of reaction (AG0), or activation energy (Ea) for a series of compounds undergoing the same type of reaction by the same mechanism, and the reaction rate constant. The rate constants vary in a way that molecular descriptors can correlate. 

The attempt to correlate biological activity with chemical structure in quantitative terms assumes that a functional dependence exists between the observed biological response and certain physicochemical properties of molecules. Without a mechanistic model one obtains it by an empirical correlation. A rational way to define such empirical relationships is within the extrathermo-dynamic approach (1). Although extrathermodynamic relationships... 

The relationship of the selectivity of an electrophile to its reactivity is a separate issue, because the above quantitative correlations of reactivity can be used empirically, without accepting that they allow comment about the reactivity of electrophiles. There is no direct evidence for the view that differences in the selectivities of electrophiles are related in a simple way to their different reactivities. Indeed, it is difficult to grasp the meaning of comparisons attempted between electrophiles of very different structures, which bring about reaction under disparate conditions by different mechanisms. 

This article describes the current capabilities for predicting materials properties using atomistic computational approaches. The focus is on inorganic materials including metals, semiconductors, and insulators in the form of bulk solids, surfaces, and interfaces. Properties of isolated molecules, liquids. and organic polymers are treated as separate entries. Besides a computational approach based on physical laws, materials properties can also be predicted by empirical rules and statistical correlations between chemical composition, bonding topology, and macroscopic properties. These very useful and quick approaches, which include so-called quantitative structure-property relationship (QSPR) methods, are covered in other entries of this encyclopedia (see Quantitative Structure-Property Relationships (QSPR)). 

---