Predictive kinetic optimisation of hydrophilic interaction chromatography × reversed phase liquid chromatography separations: Experimental verification and application to phenolic analysis.

Method development and optimisation for comprehensive two-dimensional liquid chromatography (LC × LC) is complex, since this involves the intricate relationships between a large number of experimental parameters with the aim of achieving three conflicting goals: maximising separation performance (peak capacity), minimising analysis time and minimising dilution. This is especially true for the on-line combination of hydrophilic interaction chromatography (HILIC) and reversed phase LC (RP-LC) due to the relative elution strengths of the mobile phases used in these modes, which has severe implications for the modulation process and dilution. In this study we report a predictive kinetic optimisation tool for on-line HILIC × RP-LC which is based on theoretical relationships between the optimisation goals, the target analyte properties and chromatographic parameters (column dimensions, flow rates, mobile phases, injection volumes, etc.), allowing all chromatographic parameters to vary simultaneously within defined ranges. Experimental restrictions, such as pressure limits, flow rates, etc., are enforced to ensure all results are practically achievable on a given instrumental configuration. A Pareto-optimality approach is then used to obtain optimal sets of experimental conditions, from which the one(s) best satisfying the requirements in terms of time, dilution and/or peak capacity can be chosen. To overcome the challenges associated with mobile phase incompatibility in the coupling of HILIC and RP-LC, splitting of the first dimension HILIC flow, dilution with an aqueous make-up, or a combination of both, were investigated to establish the best approach to minimise total dilution and maximise performance. The validity of the methodology is demonstrated by deriving optimal conditions for the HILIC × RP-LC separation of procyanidins on selected columns and subsequent experimental verification of the performance for the separation of a cocoa extract.