|
(T/MF156) Prediction of vapour pressures from molecular structure. Altschuh, Joachim1, Rehfeldt, Florian1, Lenoir, Dieter2, 1 2 ABSTRACT- The vapour pressure, in addition to partition coefficient data, is a key property in the estimation of fate and distribution of chemicals in the environment. Reliable methods for the determination of vapour pressures are therefore needed. Experimental vapour pressure measurements are often time-consuming and expensive. Consequently, methods to predict vapour pressures are of increasing importance. Basically these methods fall into two categories: Quantitative structure-property relationships (QSPRs), that make use of molecular structure descriptors, and property-property relationships (PPRs), that depend on other experimental data, such as boiling points. In this contribution we analyse a data set of about 250 organic chemicals from various chemical classes. The vapour pressures are taken mostly from the literature but include also some newly determined experimental data. Using multilinear regression analysis, we have developed QSPRs based on graph-theoretical indices for the complete data set as well as for distinct chemical classes. Additionally, a PPR, which results from the integration of the Clausius-Clapeyron equation, is analysed. Here melting points and boiling points are required as input data. Both approaches show reliable predicting capabilities for the vapour pressures. Their advantages and disadvantages are discussed. Key words: vapour pressure, prediction, QSPR, graph theoretical indices |
