BLIND PREDICTION OF CU(II) SORPTION ONTO GOETHITE: CURRENT CAPABILITIES OF DIFFUSE DOUBLE LAYER MODEL

Abstract

The paper presents examples illustrating the current blind predictive capabilities of the diffuse double layer model (DDLM) as the model requiring the smallest set of parameters and thus being most suitable for substituting even more empiric sorption approaches such as distribution coefficients KD. The general strategy for the selection of numerical data are discussed. Based on the information about the minerals compiled in the sorption database RES3T (Rossendorf Expert System for Surface and Sorption Thermodynamics), first a set of relevant surface species is generated. Then relevant surface complexation parameters are taken from RES3T: the binding site density for the minerals, the surface protolysis constants, and the stability constants for all relevant surface complexes. To be able to compare and average thermodynamic constants originating from different sources, a normalization concept is applied. Our demonstration is based on a blind prediction exercise, i.e., the goal was not to provide optimal fits. The system considered is Cu(II) sorption onto goethite. The predictions are compared with raw data from three independent experimental investigations. The calculations were performed with the FITEQL 3.2 code. In most cases the model predictions represented the experimental sorption values for the sorbed amount of Cu(II), expressed as conventional distribution coefficients, within one order of magnitude or better. We conclude that the application of DDLM can indeed be used for estimating distribution coefficients for contaminants in well defined mineral systems. A stepwise strategy of species selection, data collection, normalization, and averaging is outlined. The SCM database so far assembled within the RES3T project is able to provide the parameter sets. Copyright © 2005 Elsevier Ltd.