NUMERICAL ALGORITHMS FOR CAPILLARY ELECTROPHORESIS

Three groups of algorithms applied to mass transport equations in electrokinetic processes are described and compared. They are used for solving a set of partial differential equations expressing the conservation of mass and charge laws, the dissociation equilibria and electroneutrality. These equations should be able to predict the shape and structure of the boundaries and peak diffusion and anomalies as the analytes are driven in the electric field past the detector port. Three different numerical algorithms have been proposed: by Mosher et al. (The Dynamics of Electrophoresis, VCH, Weinheim, 1992), by Dose and Guiochon [Anal. Chem., 63 (1991) 1063-1072] and by Ermakov et al. [Electrophoresis, 13 (1992) 838-848]. The first two algorithms offer numerical solutions which can only be implemented at unrealistically low current densities, two to three orders of magnitude lower than normally adopted in practical electrophoresis. When applied to real experimental conditions, both previous algorithms break down: the separated analyte zones, as obtained by simulation, decay into several solitary waves (solitons), resulting from the action of numerical dispersion. In contrast, the numerical algorithms proposed by Ermakov et al. still allow the prediction of peaks of with the correct shape, with only minor non-physical spikes.