Multivariate optimization strategy for liquid chromatography. II. Exploring and finding optimal conditions in the search area of the multidimensional solvent space

A multivariate three-step strategy for chromatographic optimization is described in this paper. The first step of this strategy is a predictive method for the difficult task of initial solvent selection, which is based on information from the literature or from scouting runs, i.e. targeting the search area in the multidimensional solvent property space. In the second step of the strategy the targeted search area is explored using a highly reduced factorial design, in this step system variables can also be included such as pH, temperature, flow rate, among others. In the third and final step of the strategy a factorial design is used to generate a partial least squares (PLS) model that allows prediction of optimal chromatographic conditions within the explored domain.