Prediction of Boiling Points of Ketones Using a Quantitative Structure-Property Relationship Treatment
Anna Komasa · Polish Journal of Chemistry · 2003
A data set of 202 ketones was used to study a correlation of over 700 descriptors with boiling point using a Quantitative Structure-Property Relationship (QSPR) technique. The QSPR study was performed with the CODESSA (Comprehensive Descriptors for Structural and Statistical Analysis) program. A large part of the descriptors used in the study, was based on quantum-chemical calculations. A four parameter model with the squared correlation coefficient R 2 = 0.9907 and the Fisher ratio F = 5231 was obtained. The final QSPR model includes one topological and three quantum-chemical descriptors: Kier&Hall index (order 1) Total molecular two-center resonance energy divided by the number of atoms, Maximum net atomic charge for a C atom (Gaussian NBO), and Average valency of a C atom. The four-descriptor relationship can be recommended for prediction of the boiling point of ketones.