CONTINUOUS MODEL FOR COMPLEX MIXTURE ADSORPTION

Adsorption of mixtures containing several ill-identified solutes can be studied by discrete or continuous models. In the first case, all the solutes are lumped into a few pseudocomponents. More than one pseudocomponent is generally required to obtain a reliable description of the system, and the computational effort increases with the number of pseudocomponents used. In the second, the mixture composition is described by means a continuous distribution function. If the feed distribution is known, the distribution in the liquid phase at equilibrium depends solely on the degree of total solute removal. Under some hypotheses on the feed distribution function, which are largely borne out in real systems, a limiting distribution is achieved for a high degree of solute removal. The difference between the feed distribution and the limiting distribution is significant when the variance of the former is large. The overall adsorption isotherms can be predicted if the feed distribution function is known. An identification method is proposed for a fairly complete statistical characterization of the feed distribution function extracting from the experimental data.