A new approach to the determination of column overload characteristics in reversed-phase liquid chromatography

Column overloading is very common during the separations of basic analytes in analytical scale reversed-phase liquid chromatography (RPLC). Due to the complex interactions of ionic analytes with stationary and mobile phases, only a very small amount of ionized sample compared to the amount of nonpolar solute can be injected before the peak shape is distorted by non-linear chromatographic processes. Often the amount that can be injected before overload is observed is so small that the signal is quite noisy, thereby making the measured plate count imprecise and possibly inaccurate. The purpose of the present study was to develop a practical method for the precise measurement of the plate count and a column overload parameter using a simple but mathematically rigorous model of Langmuirian non-linear chromatography. An "overload profile", i.e. a plot of apparent plate count versus amount injected, is characterized by two parameters: the limiting plate count (N₀) and the column sample loading capacity (ω_0.5). The limiting plate count is the plate count that should be observed when the amount of sample injected is so small that a linear isotherm pertains. The column sample loading capacity, which is taken as the sample load that leads to a plate count equal to half of the limiting plate count, is a measure of the maximum amount of sample that can be injected into that column. The approach was tested by applying it to the study of cationic analytes in RPLC. We show that N₀ under constant conditions (column length, flow rate, mobile phase composition, etc.) is almost independent of column type (manufacturer); however, different column types (at the same length, diameter and flow rate) exhibit clear differences in their sample loading capacity (ω_0.5). We believe that for most well packed type B columns, the column sample loading capacity and not the limiting plate count is the more important property that accounts for most of the apparent differences in peak width when different types of columns are examined.