PREDICTION OF TEMPERATURE-PROGRAMMED RETENTION INDEXES FOR POLYNUCLEAR AROMATIC-HYDROCARBONS IN GAS-CHROMATOGRAPHY

The objective of this study was to relate temperature-programmed GC retention indexes of nonsubstituted PAHs with calculated molecular properties. Prediction of retention was made by use of a partial least squares regression equation using five descriptors, i.e., first-order valence molecular connectivity ionization potential, length, height, and quadrupole moment, Retention index predictability was studied with respect to individual molecular weight ranges. Pericondensed PAHs were more accurately predicted (RMS = 1.6 index units (iu) or coefficient of variation of 0.31%) than catacondensed PAHs (RMS IS = 5.8 iu or coefficient of variation of 1.1%) with the descriptor set employed.