In case a factor describes one chemical component in a complex mixture, the factor score can be used as a quantitative measure of the compound. This is demonstrated in the application of Py-MS to the detection of ppm levels of impurities of the toxic, technical polymer diethylaminoethyl(DEAE)-dextran in batches of poliomyelitis virus vaccine as shown in Figure 34. By introducing standard virus preparations, spiked with 20, 40, or 80 ppm of DEAE-dextran in the sample series, the first factor described this polymer. Concentrations in the samples to be checked can be determined down to 20 ppm (ref. 63).
In the pyrolysis mass spectra of complex mixtures, factor spectra often represent correlated changes of a number of chemical components. Generally, this is due to mathematical reasons, e.g. the normalisation of the spectra, and the non-orthogonality of factors. In such cases, specialised factor analysis techniques such as vector rotation and renormalisation (refs. 138, 139, 140) allow the examination of mass spectra from mixtures for the presence of specific components. Other methods suitable for this purpose are target transformation factor analysis as described by Malinowski and McCue (ref. 141) and the probability-based reversed matching method
Was this article helpful?