Selecting optimal columns for clarithromycin impurity analysis according to the quantitative relationship of hydrophobic subtraction model

Abstract

Hydrophobic subtraction model (HSM) is widely applied to select columns of equivalent or different selectivity compared with a reference column, but its application in identifying optimal columns for specific separations of real samples is rare. In this work, a column selection method was proposed by firstly directly correlating separation selectivity of different pairs of solutes to column parameters based on the quantitative relationship of HSM and then selecting the optimal columns according to the predicted selectivity in consideration of the total separation of all critical pairs of solutes. Three critical pairs of solutes in clarithromycin impurity analysis were evaluated as examples. Starting with the analysis of clarithromycin impurities on 15 columns with different selectivities, ten optimal columns were finally identified for clarithromycin impurity analysis from the HSM column characterization database containing nearly 600 columns and two of them were validated with satisfactory separations for all critical peak pairs. The proposed methodology was also compared to the traditional column selection procedure based on calculations of scalar measures of the Euclidean distance between chromatographic columns. Results showed that our method provides an effective way to find the desired columns that may be overlooked by the traditional column selection due to selection of an inappropriate reference column or overestimation of column similarity, such as Fs introduced in HSM.