2D autocovariance function for comprehensive analysis of two-way GC-MS data matrix: application to environmental samples.

This paper describes a signal processing method for comprehensive analysis of the large data set generated by hyphenated GC-MS technique. It is based on the study of the 2D autocovariance function (2D-EACVF) computed on the raw GC-MS data matrix, extending the procedure previously developed for 1D to 2D signals. It appears specifically promising for GC-MS investigation, in particular to single out ordered patterns in complex data: such patterns can be simply identified by visual inspection from deterministic peaks in the 2D-EACVF plot. A case of order along the retention time axis (x=t(R)) is represented by a horizontal sequence of peaks, located at the same interdistance Δt(R)=b(x), e.g., b(x) is the CH(2) retention time increment between subsequent terms of an homologous series. The order along the fragment mass axis (y=m/z) contains information on analyte fragmentation patterns. Deterministic peaks appear in the 2D-EACVF plot at Δm/z values corresponding to the most abundant ion fragments - dominating fragments in MS spectrum - or to ions generated by repetitive loss of the same ion fragment, i.e., Δm/z=14 amu produced by the [CH(2)]() group loss in n-alkanes. Method applicability was tested by processing GC-MS data of organic extracts of atmospheric aerosol samples: attention is focused on identifying and characterizing homologous series of organics, i.e., n-alkanes and n-alkanoic acids, since they are considered molecular tracers able to track the origin and fate of different organics in the environment. Copyright Â