Comprehensive two-dimensional gas chromatography time-of-flight mass spectrometry with chemometric analysis

Multidimensional separations, in particular Comprehensive Two-Dimensional Gas Chromatography (GC x GC), have proven to be powerful techniques that have found a niche in complex mixture analysis. Instruments of this type produce second-order data that is applicable to chemometric analysis. This combination of multidimensional separations and chemometrics results in reduced analysis times by relaxing the chromatographic resolution requirements for quantitation. By coupling GC x GC to a Time-of-Flight Mass Spectrometer (TOFMS) a truly third-order technique is produced. One separation on a GC x GC-TOFMS provides retention times on two chromatographic columns and a complete mass spectrum for each component of a mixture. Using the added selectivity of the mass spectrometer combined with chemometric techniques, overlapping components in a complex mixture can theoretically be deconvoluted using only one data set. Second-order data combined with chemometric techniques require a sample and a standard data set to deconvolute overlapping analytes.; In this dissertation, aspects of GC x GC and GC x GC-TOFMS combined with chemometrics are presented. A novel configuration of a valve-based GC x GC is proposed and tested to extend the workable temperature range of the valve-based instrument to analyze semi-volatiles. The data structure was confirmed to be applicable to chemometric analysis. This new configuration is combined with a TOFMS to produce the first valve-based GC x GC-TOFMS that is subsequently evaluated for performance. A methodology involving PARAFAC is developed to deconvolute partially overlapped peaks in this data. The chemometric analysis is shown to work for both valve-based and cryogenically modulated GC x GC-TOFMS instruments. The logistics and performance characteristics of peak deconvolution with PARAFAC are analyzed with a set of butyl benzene isomers. The reproducibility and accuracy of deconvolution is subsequently tested as a function of multivariate selectivity, which is related to chromatographic resolution and mass spectral similarity. The deconvolution results for PARAFAC are compared with those for two common GC-MS deconvolution programs. In addition, a tool to locate compounds or classes of compounds of interest based on mass spectral similarity to aid in the analysis of complex mixtures is presented.