Mapping the Separation Landscape in Two-Dimensional Liquid Chromatography: Blueprints for Efficient Analysis and Purification of Pharmaceuticals Enabled by Computer-Assisted Modeling.

Recent developments in two-dimensional liquid chromatography (2D-LC) now make separation and analysis of very complex mixtures achievable. Despite being such a powerful chromatographic tool, current 2D-LC technology requires a series of arduous method development activities poorly suited for a fast-paced industrial environment. Recent introductions of new technologies including active solvent modulation and a support for multicolumn 2D-LC are helping to overcome this stigma. However, many chromatography practitioners believe that the lack of a systematic way to effectively optimize 2D-LC separations is a missing link in securing the viability of 2D-LC as a mainstay for industrial applications. In this work, a computer-assisted modeling approach that dramatically simplifies both offline and online 2D-LC method developments is introduced. Our methodology is based on mapping the separation landscape of pharmaceutically relevant mixtures across both first (1D) and second (2D) dimensions using LC Simulator (ACD/Labs) software. Retention models for 1D and 2D conditions were built using a minimal number of multifactorial modeling experiments (2 × 2 or 3 × 3 parameters: gradient slope, column temperature, and different column and mobile phase combinations). The approach was first applied to online 2D-LC analysis involving achiral and chiral separations of complex mixtures of enantiomeric species. In these experiments, the retention models proved to be quite accurate for both the 1D and 2D separations, with retention time differences between experiments and simulations of less than 3.5%. This software-based concept was also demonstrated for offline 2D-LC purification of drug substances.

Retention modelling in hydrophilic interaction chromatography.

The retention behaviour of acidic, basic and quaternary ammonium salts and polar neutral analytes has been evaluated on acidic, basic and neutral hydrophilic interaction chromatography (HILIC) stationary phases as a function of HILIC operating parameters such as MeCN content, buffer concentration, pH and temperature. Numerous empirical HILIC retention models (existing and newly developed ones) have been assessed for their ability to describe retention as a function of the HILIC operating parameters investigated. Retention models have been incorporated into a commercially available retention modelling programme (i.e. ACD/LC simulator) and their accuracy of retention prediction assessed. The applicability of HILIC modelling using these equations has been demonstrated in the two-dimensional isocratic (i.e. buffer concentration versus MeCN content modelling) and one-dimensional gradient separations for a range of analytes of differing physico-chemical properties on the three stationary phases. The accuracy of retention and peak width prediction was observed to be comparable to that reported in reversed-phase chromatography (RPC) retention modelling. Intriguingly, our results have confirmed that the use of gradient modelling to predict HILIC isocratic conditions and vice versa is not reliable. A relative ranking of the importance of the retention and selectivity of HILIC operating parameters has been determined using statistical approaches. For retention, the order of importance was observed to be organic content > stationary phase > temperature ≈ mobile phase pH (i.e. pH 3-6 which mainly effects the ionization of the analyte) ≈ buffer concentration. For selectivity, the nature of the stationary phase > mobile phase pH > buffer concentration > temperature > organic content.