COMPUTER-ASSISTED PREDICTION OF NORMAL BOILING POINTS OF FURANS, TETRAHYDROFURANS, AND THIOPHENES

Computer-assisted methods are employed for the development of predictive equations which relate molecular-based structural features to the normal boiling points for a large number of compounds containing furan, tetrahydrofuran, and thiophene ring systems. Predictive models are described for a set containing all three ring classes, as well as for a set containing only furans and tetrahydrofurans, and another set comprised only of thiophene-related compounds. The fit error for the combined data set is 4.9% of the mean boiling point for the data set, while the fit error for the furan/THF subset was 5.8% and for the thiophene subset was 3.8% of the mean boiling points of the respective data sets. Similar results are obtained for the prediction of new external data sets for each model. The models developed here are examined to gain insight into the relationship between structural features and normal boiling points for the three ring classes described.