PREDICTION OF GAS CHROMATOGRAPHIC RETENTION INDICES USING VARIABLE CONNECTIVITY INDEX

Milan Randić, Subash C. Basak, Matevž Pompe, Marjana Novič

Department of Mathematics and Computer Science, Drake University, Des Moines, IA 50311, USA
Natural Resources Research Institute, University of Minnesota at Duluth, Duluth, MN 55811, USA
Faculty of Chemistry and Chemical Technology, University of Ljubljana, Aškerčeva 5, 1000 Ljubljana, Slovenia
National Institute of Chemistry, Hajdrihova 19, Ljubljana,  Slovenia
 
 

ABSTRACT

We have re-examined gas chromatographic retention indices of alkanes (48 compounds), and alcohols (31 compounds), combining all molecules into a single set (n=79) using variable connectivity index 1Xf. By varying the weight for oxygen atom we obtained    the regression characterized by the correlation coefficient r = 0.9933, the standard error s=14.24 retention time units, and Fisher ratio F = 5695. Use of the simple connectivity index 1?, which does not differentiate carbon and oxygen atoms, gives regression with the standard error four times larger.