Fundamental Considerations in One and Two-Dimensional Chromatography for Improved Chemometric Analysis

Abstract

One and two-dimensional chromatography offer well-established instrumental platforms that produce separations of chemical compounds in complex samples in a variety of fields and applications and when coupled with spectral detection, results in powerful instruments capable of producing informative data that facilitates quantitative results with high chemical selectivity. However, chemometric data analysis is often required to fully extract useful information from the data. This dissertation presents several research investigations on the interconnectedness between the fundamental theoretical, experimental, and instrumental considerations for one and two-dimensional chromatography, particularly gas chromatography, and the application of advanced chemometric data analysis methods. First, a new method to predict the probability of quantitative success for one-dimensional chromatography based on sample complexity, separation peak capacity, nc, and the algorithm specific limit of chemometric resolution using a simulation based study of multivariate curve resolution alternating least squares for demonstration. For comprehensive two dimensional gas chromatography the relationship of phase ratio, β, between the primary and secondary separation dimensions and the implications of β on 2D peak capacity, nc,2D, were examined. The β ratio, βR = 1β/2β, is defined as a quantitative metric to facilitate this study. Overall, βR substantially affected nc,2D by influencing retention factors on the 2D column and thereby changing the modulation period necessary for proper 2D column separations. Modulation period selection is then further investigated as a part of a new method to determine the true modulation ratio, MR, from the measurable effective modulation ratio, MR*, in comprehensive two-dimensional gas chromatography, GC×GC, which can be generalized to other comprehensive two-dimensional chromatographic methods. The method was developed through the investigation of modulator induced band broadening, as a function of 1Wb and the selected modulation period, PM. Finally, a new quantitative metric, trilinearity deviation ratio (TDR), is introduced to describe the trilinearity of two-dimensional chromatography data with spectral detection for the purpose of predicting the performance of the chemometric method parallel factor analysis (PARAFAC). It is found that use of a modulation period in the 1 s to 2 s range simultaneously optimizes 1nc, nc,2D, and TDR to facilitate low quantitative errors with PARAFAC.