Association and discrimination of diesel fuels using chemometric procedures

Abstract

Five neat diesel samples were analyzed by gas chromatography-mass spectrometry and total ion chromatograms as well as extracted ion profiles of the alkane and aromatic compound classes were generated. A retention time alignment algorithm was employed to align chromatograms prior to peak area normalization. Pearson product moment correlation coefficients and principal components analysis were then employed to investigate association and discrimination among the diesel samples. The same procedures were also used to investigate the association of a diesel residue to its neat counterpart. Current limitations in the retention time alignment algorithm and the subsequent effect on the association and discrimination of the diesel samples are discussed. An understanding of these issues is crucial to ensure the accuracy of data interpretation based on such chemometric procedures.

Automated optimization and construction of chemometric models based on highly variable raw chromatographic data

“…To reduce the chromatogram to a relatively small number of meaningful variables, several approaches are available. The use of total ion chromatograms (TICs) [18,19] or extracted ion chromatograms (EICs) [18,20] is relatively straightforward, but these approaches may eliminate useful information (using TICs results in the loss of detail in the mass spectral domain, and EICs impose the analyst's preconceived notions of variable relevance on the model). Other methods such as analysis of variance (ANOVA) [9,21,22] or discriminating variable (DIVA) tests [23] are more complex.…”

Microbial degradation of gasoline in soil: Effect of season of sampling

“…The software used to perform PCA was XLSTAT (AddinSoft), an add-in for Microsoft Excel. Principal Components Analysis (PCA) can be used to elucidate trends in the data while taking into account any correlations between the variables [8,[26][27][28][29]. Prior to conducting PCA, the data was normalized and autoscaled to correct for variation in peak areas due to compound concentration as well as variability in peak areas due to differences in the variance of the ions that were used to construct EIPs.…”

Factors Affecting the Forensic Examination of Automotive Lubricating Oils*

“…To fully understand the potential for discriminating lubricating oils and associating a specific lubricating oil with a suspect vehicle, future HTGC/MS research should be pursued. Further studies that may be warranted include (i) the use of chemometrics in conjunction with retention time data from the TIC (18), (ii) the use of SIM mode to increase the sensitivity of PAH detection, (iii) lower electron impact voltages in the MS to create larger analyte fragments, (iv) analyzing more “mock case” scenarios, and (v) developing blind tests for evaluation. Research in the aforementioned areas will provide greater insight into the forensic analysis of lubricating oils and the conclusions that can be reached from their examination.…”