Classification and replacement test of HPLC systems using principal component analysis

Theoretical plate numbers and symmetry factor values have been measured for three solutes (benzoic acid, N,N-dimethyl-aniline and Vancomycin) in various chromatographic systems (stationary phases and different mobile phase compositions) with the aim of grouping the systems and studying the possibilities for replacement of columns and eluent compositions. The mobile phases consisted of acetonitrile or methanol modifier and KH2PO4-H3PO4 or triethyl amine-H3PO4 aqueous buffer. The organic content of mobile phases was adjusted so that the log ks (logarithms of retention factor) of tested compounds vary in the range of 0. 1-1.3 in case of all chromatographic systems and test compounds. Plate numbers and symmetry factors at all retention factor values (between log k=0.1-1.3 with 0.1 units) were subjected to principal component analysis (PCA). The input data were arranged in 13 x 28 data-matrices in case of all compounds separately for plate numbers and symmetry factors. Two principal components were retained in the model, which carry not less than 98% of the total variance. Principal component analysis showed which chromatographic systems (stationary and mobile phases) gave the same (similar) chromatographic parameters using three different kinds of compounds. Replacement of chromatographic systems is possible on the basis of measured chromatographic parameters using principal component analysis. The monofunctional test compounds provide only approximate information on the usefulness, classification and replacement of HPLC systems, if the aim is the separation of multifunctional solutes. Although the theoretical plate number (column efficiency) is negatively correlated with the symmetry factor, both are necessary for proper classification and characterization of stationary phases. (c) 2004 Elsevier B.V. All rights reserved.