Modelling chromatographic behaviour as a function of pH and solvent composition in RPLC

Summary In this work we establish the basic layout of IONICS, an expert system for optimizing the separation of ionogenic solutes in Reversed-Phase Liquid Chromatography, using the pH and the organic-modifier concentration of the mobile phase as parameters. We also present REMO, a front-end system that automates the retention modelling stage, based on a 9-parameter model. This system uses a scale transformation to suppress several numerical problems previously observed and features a strategy for automatic calculation of an initial approximation to the model optimum. The successful application of this system to a set of seven drugs is described. The final models are accurate and have smaller numerical problems. We also describe the use of a genetic algorithm instead of classical non-linear least-squares for fitting the model to the experimental data. Results indicate that genetic algorithms are a valuable, complementary tool for retention modelling.