Optimal experimental design (OED) – think first, then act!
Ion-exchange, hydrophobic interaction or mixed mode chromatography – model predictions can be improved by creating particularly reliable models of common chromatographic purification processes. ChromX allows to identify optimal experimental set-ups for a fast and robust model calibration.
Chromatography models contain many parameters which are difficult to measure experimentally. These model parameters are for instance activity coefficients or interaction coefficients. Especially in multi-component settings it is hard to predict the optimal experimental set-up to determine these parameters for model calibration. The concept of optimal experimental design takes one step back:
Parameter sensitivity analysis to predict optimal estimation experiments. In an OED approach, first the experimental set-up with highest information content is identified. Hereto, uncertainties of parameters are accessed by taking into account their parameter confidence intervals. Driven by a minimization of these, new experiments are then suggested. Resulting ideal experiments for model building may then look completely different to typical Design-of-Experiments runs. This way, model parameters with maximal statistical significance were found using the least numbers of experiments. The approach respectively needs less time and material compared to manual experimental design.
Improved reliability of isotherm parameters. The presented approach can be successfully used for automated process development. The reliability of isotherm parameters as well as the robustness of downstream processing can be enhanced. Nevertheless, expert knowledge is still necessary to set an appropriate parameter space.