Based on Theoretical Parameters

Taking advantage of molecular modeling and quantum mechanical techniques to obtain relevant structural parameters and molecular properties such as conformational behavior, ionization potentials, and dipole moments, several researchers have proposed alternative predictive models of lipophilicity. Using parameters derived from the 3D molecular structure should in principle introduce the (missing) third dimension in log P calculations. 

A structural parameter largely dependent on the 3D molecular structure, the solvent-accessible surface area (SASA) is of potential interest in predicting partition coefficients. Although not always explicitly stated, the duality of lipophilicity described by Eq, [8] serves as a guide to these developments. 

Indeed, owing to the constant influence of polarity on the lipophilic behavior in a homologous series of small rigid solutes (Eqs. [4] and [5]), the observed relations between partition coefficients and SASA for hydrophobic compounds ° i or for congeneric series are trivial because variations of lipophilicity in these series depend exclusively on variations in hydro-phobicity. However, even in these simple series, the use of SASA allows one to model the well-known effect of stereochemistry on lipophilicity. Other approaches have been explored to extend the use of SASA to noncongeneric 

The method proposed by Moriguchi et al. is based on the SASA generated by a solvent probe of 1.4 A radius. Standard surface area algorithms were used to calculate this parameter. But because polar groups with appreciable surface areas make negative contributions to log P, several parameters (called Sh see Table 9) were added to encode hydrophilic effects and were evaluated by least-squares techniques using the data of total and observed log P for 138 model compounds (Eq. [21]). 

Although Eq. [21] is statistically good and can be satisfactorily extended to water solubility (Eq. [22]), the parameters Sh are in faa a mere set of fragmental constants encoding the polarity component of lipophilicity (A as described in Eq. [5]). This fact explains why Eqs. [21] and [22] do not appear suited to take into account 3D factors such as proximity effects.  

---