Mass spectrometry detection versus flame ionisation detection in pyrolysis-gas chromatography: spurious pyrolysis peaks and possible errors in estimating relative product yields

In the first part of this paper it is demonstrated that spurious pyrolysis peaks can arise when a mass spectrometer (m.s.) is used as the detector in pyrolysis-gas chromatography (g.c.). The peaks can be of significant size, and have retention times which are comparable with those of some genuine gaseous pyrolysis products, such as hydrogen chloride or carbon dioxide, towards which Flame Ionisation Detectors (FID) are almost totally insensitive. If the spurious peaks were mistakenly taken as pyrolysis products they would be associated with unusual kinetics, and an activation energy more commensurate with adsorption than thermal degradation (23 kJ mol(-1)). The spurious peaks are shown to consist of a higher concentration of the known impurities (principally air gases) which are present in the cylinder of helium carrier gas. It is also demonstrated that the phenomenon which causes the peaks occurs at the head of the chromatography column, where the pyrolysis unit is situated. The results indicate that desorption of the impurity gases from the filament when it is fired, although occurring, is not the factor principally responsible for the spurious peaks. The major factor is the preferential expansion and back-diffusion of the helium from the pyrolysis chamber when the filament is fired. This leaves a preponderance of impurity gases in the chamber, which are pushed into the chromatography column as a slug when the filament is switched off. The mass spectrometer then records this change of ratio of carrier gas and impurity gases as a peak. This understanding of the phenomenon has led to a modification of the equipment which has reduced the spurious peaks to a size commensurate with the baseline noise level. The second part of the paper is concerned with assessing relative amounts of components present from their gas chromatographic peak areas. Even when there is no specific sensitivity discrimination between components in either FID or m.s. detection, it is not expected that measurements of component ratios, based on peak area ratios, will be the same from FID and m.s. data. This is because (to a first approximation) the FID measures weight ratios, whereas the m.s. measures number ratios. This situation is illustrated by measured oligomer/monomer ratios in polyisobutylene pyrolysis products. The raw peak area ratios from FID and m.s. are widely different, but close correspondence is obtained if the m.s. results are converted to weight ratios. (C) 1999 Elsevier Science Ltd. All rights reserved.