Self-modeling curve resolution by factor analysis of a continuous series of pyrolysis mass spectra

A self-modeling technique for curve resolution of overlapping processes is presented. This method uses factor analysis combined with the variance diagram technique to resolve the total ion current (TIC) curves from time-resolved pyrolysis mass spectrometry data into chemical component curves and their spectra. Examples are presented of time-resolved pyrolysis mass spectrometry data from a biopolymer mixture (deoxyribonucleic acid, bovine serum albumin, and glycogen), douglas fir wood, and a rubber copolymer. The analysis of each of these samples resulted in the resolution of the TIC's into chemical component curves. The spectra corresponding to the chemical components were clearly similar to reference spectra.