VERIFICATION OF THE EXTENDED STATISTICAL-MODEL OF OVERLAP - APPLICATION TO GAS CHROMATOGRAMS CONTAINING VARIABLE PEAK DENSITIES

An extended statistical model of overlap that is fairly insensitive to the density of single-component peaks was applied to gas chromatograms of lime oil. These chromatograms were developed in three different ways. In the first way, chromatograms were developed to contain a constant Poisson density of single-component peaks. In the second way, chromatograms were developed by simple temperature ramps, with no concern about the densities of single-component peaks. In the third way, chromatograms were developed by extreme variations of heating rate to generate single-component peaks having wide density variations. The extended statistical model of overlap was used to predict the number of components in all three types of chromatograms. These numbers were essentially equal to those predicted by an earlier model, which because of theoretical limitations could be applied only to chromatograms developed the first way. This equivalence of component numbers verifies that the extended model is more powerful than its predecessor and, under appropriate circumstances, can be used to interpret chromatograms with little or no concern about the density of single-component peaks.