Targeted analyte deconvolution and identification by four-way parallel factor analysis using three-dimensional gas chromatography with mass spectrometry data

- A four-way PARAFAC method is presented to analyze GC3-TOFMS data.
- The method overcomes challenges of reducing rank prior to PARAFAC analysis.
- All 42 target analyte spectra were successfully resolved and identified.

Comprehensive three-dimensional gas chromatography with time-of-flight mass spectrometry (GC3-TOFMS) creates an opportunity to explore a new paradigm in chemometric analysis. Using this newly described instrument and the well understood Parallel Factor Analysis (PARAFAC) model we present one option for utilization of the novel GC3-TOFMS data structure. We present a method which builds upon previous work in both GC3 and targeted analysis using PARAFAC to simplify some of the implementation challenges previously discovered. Conceptualizing the GC3-TOFMS instead as a one-dimensional gas chromatograph with GC × GC-TOFMS detection we allow the instrument to create the PARAFAC target window natively. Each first dimension modulation thus creates a full GC × GC-TOFMS chromatogram fully amenable to PARAFAC. A simple mixture of 115 compounds and a diesel sample are interrogated through this methodology. All test analyte targets are successfully identified in both mixtures. In addition, mass spectral matching of the PARAFAC loadings to library spectra yielded results greater than 900 in 40 of 42 test analyte cases. Twenty-nine of these cases produced match values greater than 950.

172