Development of theory, instrumentation, and chemometric data analysis tools for comprehensive two- and three-dimensional gas chromatography

Abstract

Comprehensive two-dimensional gas chromatography coupled with time-of-flight mass spectrometry (GCÃ GC-TOFMS) is now recognized as a powerful analytical platform for the separation and analysis of complex mixtures containing volatile and semi-volatile components. However, two major impediments have limited widespread adoption of the technique. First, commercial instrumentation is cost-prohibitive and is not optimized for high peak capacity. Second, the technique generates large and complex data, requiring many man-hours of analysis by an expert user(s) to generate chemically meaningful information. This dissertation presents several developments in instrumentation, data analysis methods, and fundamental theory in order to address these challenges. First a new modulation method, termed dynamic pressure gradient modulation (DPGM), is described, and shown to produce high peak capacity separations in a short time frame using mostly off-the-shelf components. Next, the three-dimensional gas chromatograph - TOFMS is improved to provide total analyte transfer between all dimensions by employing DPGM as a modulator in series with cryogenic modulation. As a result of tandem modulation with total-transfer, exceptional signal enhancement from the first dimension to the third dimension was observed. To verify and study this observation, equations which predict signal enhancement for GC are derived and experimentally validated. In second part of this dissertation, the data analysis tool tile-based F-ratio analysis is investigated to provide suggestions on user-inputs and broaden the scope of the method. Using elemental bromine, olefins are derivatized in gasoline, followed by analysis of the original and brominated gasoline. Tile-based F-ratio analysis is usedto quickly locate and identify the selectively derivatized analytes in the complex gasoline matrix. Finally, computer simulations are used to generate GCÃ GC-MS separations with realistic run-to-run retention time shifting using low-frequency "shift functions". The simulated data are used to evaluate the performance of tile-based F-ratio as a function of the amount of shifting along with chromatographic saturation and peak area RSD. The results of this study provide the user community with parameter recommendations (e.g., tile size) depending on the amount of shifting, saturation, and RSD observed.