- 3:40 PM
304b

Unified Physical Property Estimation Relationships-Upper

Samuel H. Yalkowsky, Collegel of Pharmacy, University of Arizona, 1703 E Mabel, Tucson, AZ 85721

Unified Physical Property Estimation Relationships, UPPER Samuel H. Yalkowsky, College of Pharmacy, University of Arizona, Tucson, AZ 85721

As the synthesis of new compounds becomes more efficient, the importance of estimating environmentally, pharmaceutically, and industrially relevant physical properties of molecules (especially melting point) becomes more valuable. There are a number of techniques available for the estimation of partition coefficient, solubility, melting point, vapor pressure, and other relevant properties. Unfortunately, many of these methods are not compatible with each other, i.e., they are based upon different models or assumptions for the molecule. It is frequently difficult for a user to work with a number of different estimation techniques, and the data obtained from the various methods are sometimes contradictory. This point is illustrated in the compilations of physical property estimation methods by Lyman, Reehl, and Rosenblatt (1982), Joback and Reid (1987) Baum (1998), Reinhard and Drefahl (1999), and Boethling and MacKay (2000). There are four key components to the effective development of a unified physical property estimation scheme: a reliable database of physical property values, a reasonable molecular fragmentation pattern which is based upon a single molecular structure descriptor such as the SMILES string, and the ability to characterize those molecular properties that are not strictly additive and constitutive, and to distinguish among isomers. a thermodynamically sound series of integrated relationships with a minimum number of assumptions, The objective of this proposal is to develop such a scheme.

Figure 1 shows the Unified Physical Property Estimation Relationships (UPPER) scheme. From a SMILES string or a MOLFILE structure 2 additive descriptors (group contributions to the heat of boiling and the heat of melting) and 2 nonadditive descriptors (molecular symmetry and molecular flexibility) are generated. These are used to predict the following:

Heat of boiling, Heat of melting, Energy of boiling, Energy of melting, Entropy of boiling, Entropy of melting, Heat capacity change on boiling, Heat capacity change on melting, Boiling point, Melting point, Vapor pressure, Heat of vaporization, Entropy of vaporization.