Innovative multidimensional gas chromatography mass spectrometry analysis of fatty acids in complex sample matrices

The determination and accurate identification of fatty acids (FAs) in biological samples remain challenging due to the complexity of these compounds and confounding sample matrix effects. This thesis highlights a range of method developments and applications using advanced multidimensional gas chromatography (MDGC) and comprehensive two dimensional gas chromatography (GC×GC) techniques, coupled with mass spectrometry (MS) for FA profiling and structural identification. An investigation of appropriate combinations of stationary phases was studied prior to the subsequent multidimensional analytical method development. Recently introduced ionic liquid (IL) capillary GC stationary phases covering a wide polarity range have been evaluated for fatty acid methyl esters (FAMEs) analysis in terms of their elution patterns and retention indices (e,g. ECL) on the IL columns. The performance of a newly proposed integrated GC×GC / heart-cut (H/C) MDGC-FID method for isomeric PUFA separation was then evaluated. As a practical application, abundant long chain UFA isomers with carbon numbers ranging from 18 to 22 in marine oil, as well as a dairy food product, can almost be fully separated by using GC×GC. Additionally, more than 7 other FA compounds were found in the same region by switching the system to H/C MDGC-FID under a modified GC condition. Both these applications employed fast analysis conditions. Next, GC×GC was coupled to the quadrupole-accurate mass time of flight mass spectrometry (QTOFMS). This was done to achieve accurate detection and structural confirmation of individual known or unknown branched FA obtained through predicted empirical formula and accurate mass information. The approach was applied to phospholipid fatty acids (PLFAs) from complex forest soil samples in order to investigate the microbial community. Several high abundance branched hydroxyl- (OH-), iso-/anteiso- and cyclopropyl- (cy-) FAME were clearly determined. Tentative identities of trace level OH-FAME and unusual epoxidised FAME were found according to their predicted empirical formulae and elemental compositions. Epoxidised FA were previously found in the metabolic pathway of a Gram-positive soil bacterium Bacillus megaterium. The findings indicate that this approach has great potential towards fingerprinting and FA biomarker discovery in environmental research. Lastly, comprehensive fast GC×MS with a positive chemical ionisation mode (PCI) was proposed for FA analysis in medical studies. Method performance was evaluated by application to different types of FAs including PUFAMEs and bacterial FAMEs (e.g. branched FAMEs) as well as to real biological samples (e.g. wild type and transgenic mouse brains associated with Alzheimer’s disease). Methanol was selected as the most appropriate PCI chemical reagent for total FA profiling. A comparison of both rapid comprehensive GC×PCI-QMS and GC×GC approaches, for fast screening capabilities and the rapid quantitative determination of long chain PUFA isomers in mice brain associated with Alzheimer’s disease, was conducted in this study. An evaluation of both methods was based on the comparison of their 2D presentations, calibration linearity and minimum levels of detection. Fast GC×MS data demonstrated the good quantification potential of this approach towards biological FA analysis, especially in medical studies.