Two metrics for measuring orthogonality for two-dimensional chromatography

Abstract

Orthogonality can be used as a selection parameter for two-dimensional chromatography column selection (e.g. in GC × GC or LC × LC) or for method optimization purposes, both aiming for maximal orthogonality for a particular analytical application. In order to improve the concurrence of two-dimensional chromatography expert’s orthogonality grading, two orthogonality metrics, %FIT and %BIN, were developed, evaluated and compared with the Asterisks orthogonality metric. The %BIN is a bin counting approach where the number of bins is fixed at 25 and deviations from the expected average number of peaks per bin is used as the basis for the orthogonality calculation. The %FIT is based on fitting polynomials of degree two, through the xy and the yx data and calculating the average minimal distance and standard deviation of all data points above and below the fitted polynomials. The orthogonality metrics were evaluated by using 14 different types of computer generated xy datasets and two measured LC × LC datasets. Both %FIT and %BIN, were shown to have a larger discriminative power than the Asterisks equations, and are in good agreement with the orthogonality scores for 2D-chromatograms provided by nine experts.