Predictive steady-state chromatography. 1. Algorithms for leading and trailing edges of resolved and unresolved peaks in liquid chromatography.

Abstract

This paper describes mathematical models and curve-fitting procedures that permit steady-state saturation signals to be computed accurately from data along leading and trailing edges of liquid chromatograms. This new approach to quantitative chromatography is called predictive steady-state chromatography (PSSC). It is shown that the computed saturation signals are virtually the same when determined from data along leading and trailing edges and they vary linearly with analyte concentration. Most importantly, the computed saturation signals for a given analyte concentration are virtually independent of experimental variables such as sample volume and flow rate. For example, for sample volumes between 25 and 45 microL, the average computed saturation signal for a 0.025 mM solution of theophylline was 0.11 V with a standard deviation of 0.00097 V (RSD = 0.8%); similar results were found for other concentrations and for changes in flow rate. Dependencies of the PSSC method on sample volume and flow rate were compared with dependencies for peak-height and peak-area methods by using relative error coefficients. Dependencies on sample volume were 0.04%/microL for the PSSC method and 3 and 4%/microL for peak-height and peak-area methods, respectively. Dependencies on flow rate were 2%/mL/min for the PSSC method and 17 and 120%/mL/min for the peak-height and peak-area methods, respectively. Thus, the predictive steady-state method is 10-100-fold more rugged than peak-height and peak-area methods.