Use of a semi-empirical topological method to predict the chromatographic retention of branched alkenes

The new semi-empirical topological index (IET) previously developed by Heinzen, Soares, and Yunes has been extended to branched alkenes. The main advantage of this method was excellent predictive power and high statistical quality using one topological descriptor only, in comparison with those achieved with a variety of descriptors by multivariate regression techniques and recognition patterns. Another advantage is that this molecular descriptor is related to structural features of the molecule, which enables thecis andtrans structures of branched alkenes to be distinguished and highlights the propyl effect—a special steric effect resulting from branching at carbon atoms β to the double bond. The stability and validity of the model generated on squalane (80°C) was tested by the cross-validation (r2CV=0.997) technique.