Development and evaluation of an alternative method for the resolution of overlapped chromatographic responses

Dihua Jin, Purdue University

Abstract

This research focuses on the development and evaluation of an alternative method for the deconvolution of overlapped chromatographic peaks that intended to improve the accuracy of results and the flexibility of curve-fitting models. The approach studied here involves the integration of overlapped peaks and resolution of the unidirectional area data using curve-fitting methods analogous to those used to resolve overlapped kinetic responses resulting from simultaneous chemical reactions. Two empirical four-parameter mathematical models, a segmented sequential sigmoid (SSS) model and an unsegmented sequential sigmoid (USS) model, were developed for the sigmoid shaped area vs. time data. The performance characteristics of the proposed models were first evaluated by fitting them to single-component responses for barbiturates and five types of peak shapes simulated by published chromatographic models with asymmetry ranging from 0.37 to 7.33. Results show that the two sequential sigmoid models are more rugged to adapt to a variety of peak shapes and degrees of asymmetry (both tailing and fronting). They give smaller errors of peak area and better fitting quality than the EMG model for peaks with moderate to high degrees of asymmetry. Subsequently, this approach was evaluated on artificially overlapped and real unresolved multicomponent chromatograms for barbiturates and nitrotoluenes. Results demonstrated the feasibility of using peak areas rather than peak data to resolve components that are not fully resolved in chromatographic process. The degree of peak overlap has less impact on the sequential sigmoid models than on the EMG models. For large degrees of overlap (resolution, R s, < 0.7), the SSS and USS models are less susceptible to the errors than the EMG models.

Degree

Ph.D.

Advisors

Pardue, Purdue University.

Subject Area

Analytical chemistry

COinS