Optimisation of gradient elution in normal-phase high-performance liquid chromatography

Abstract An optimisation approach for linear gradient elution in normal-phase chromatography was suggested. The approach is based on predictive calculations of the retention and of the resolution of the individual pairs of compounds in the sample mixture. Isocratic or gradient-elution retention data, acquired in a few initial runs under different conditions, are employed to determine the parameters of retention equations describing the dependence of the retention factor, k, on the composition of the mobile phase. Unlike earlier procedures, a more complex, three-parameter retention equation can be used as the basis of predictive calculations, if necessary. The approach allows one to use either maximised minimum resolution in the sample mixture or the minimum time necessary to achieve the required resolution as the optimisation criterion. It accounts for the contribution of the initial isocratic elution step induced by the gradient dwell volume, so that it is not necessary to delay the injection with respect to the start of the gradient. Simultaneous optimisation of the gradient slope and of the gradient range is performed, with the gradient volume (time), plate number and other variables as adjustable parameters. The approach is illustrated by examples of gradient-elution separation of phenylurea herbicides on a silica gel and on a bonded nitrile column, with binary gradients of 2-propanol in n-heptane and of dioxane in n-heptane. Dried solvents and the temperature (controlled to ±0.1°C) were used to improve the reproducibility of the retention data.