Spectral transformations for deconvolution methods applied on gas chromatography–mass spectrometry data

Abstract In hyphenated chromatography, overlapping chromatographic peaks can be resolved into pure spectra and pure chromatographic profiles by several multivariate deconvolution techniques. In general, these methods require bilinearity, which implies that the spectrum of each analyte is constant. The slow scan speeds normally used in gas chromatography–mass spectrometry (GC–MS) will destroy bilinearity and introduce systematic noise in the data because the concentration in the detector changes during the scan. This effect, described as the scan effect , may hinder successful resolution by multivariate deconvolution. In selected ion monitoring (SIM) GC–MS, the scan effect may be removed by simple transformations of the mass spectra. The effects of different transformations are demonstrated both on pure chromatographic peaks and on difficult resolution problems where there are small differences between the spectra of the analytes.