Determination of competitive adsorption isotherm of enantiomers on preparative chromatographic columns using inverse method

Precise determination of competitive adsorption isotherm is of great importance for model-based optimization and scale-up of chromatographic separation processes, particularly for the simulated moving bed (SMB) systems due to its intrinsic complexity. In this study, five parameters of a bi-Langmuir model for adsorption of ketoprofen enantiomers on a preparative Chiralpak AD column at 25 °C were determined by the application of inverse method (IM). Transport-dispersive (TD) model parameters, axial dispersion and mass transfer coefficients were independently estimated first by tuning a series of linear pulse experiments. Competitive adsorption isotherm model parameters were then determined independently by fitting the experimental data. A new numerical scheme that couples non-dominated sorting genetic algorithm (NSGA) and Levenberg–Marquardt algorithm (LMA) was proposed for the data fitting. NSGA was capable of rapidly locating the optimal solution in the parametric space while LGA allowed speedy convergence given a good initial guess. Independently measured band profiles and simulated moving bed (SMB) experiments were later used to validate the model parameters. The experimental results were in good agreement with the model predictions.

► The bi-Langmuir parameters for adsorption of ketoprofen on a preparative column were determined.
► Model parameters, DL and k, were independently estimated using a series of pulse experiments.
► The pulse experiments gave reliable Henry's constant, thus reducing number of fitting parameters.
► A numerical scheme that couples genetic algorithm and Levenberg–Marquardt algorithm is proposed.
► The method allowed to locate optimal solution in a wide parametric space with fast convergence.