SORPTION OF EU(III)/CM(III) ON CA-MONTMORILLONITE AND NA-ILLITE. PART 2: SURFACE COMPLEXATION MODELLING

Abstract

Sorption edges and isotherms for Eu(III) uptake on Ca-montmorillonite and Na-illite in 0.066 mol/L Ca(C1O 4) 2 and 0.1 mol/L NaC1O 4 background electrolytes, respectively, were modelled using a quasi-mechanistic sorption model (the two site protolysis non electrostatic surface complexation and cation exchange (2SPNE SC/CE) model). For both clay minerals the Eu sorption edges could be quantitatively modelled in the pH range ~3 to ~10 using cation exchange reactions for Eu 3+/Na + and Eu 3+/Ca 2+ and three surface complexation reactions on the strong sorption sites forming ≡S S OEu 2+, ≡S SOEuOH + and ≡S SOEu(OH) 2 ° inner sphere complexes which appear successively with increasing pH. Time resolved laser fluorescence spectroscopy (TRLFS) measurements of Cm(III) loaded Ca-montmorillonite and Na-illite were available from Part 1 of this work. De-convolution of the normalised fluorescence spectra measured at different pH values indicated three distinct Cm surface complexes, Cm complexes 1, 2 and 3 for both clay minerals, in agreement with model predictions, but with different distribution functions for the individual species. Under the assumption that Eu and Cm exhibit essentially the same hydrolysis and sorption behaviour, the Eu surface complexation constants were used to predict surface species distribution functions for Cm under the same experimental conditions used in the TRLFS measurements. Comparison of modelled and experimentally deduced species distributions indicated that for both clay minerals peak heights and widths of the three peaks did not correspond particularly well. It is shown that the calculated species distribution functions are sensitive to the values of the hydrolysis constants used in the calculations, whereas modelling the sorption edge measurements by applying the 2SPNE SC/CE approach is much less sensitive. By modifying the values of the hydrolysis constants within their uncertainty range and re-modelling the sorption edges, considerably better correspondence between the modelled and TRLFS species distribution functions was found. In particular, peak positions, heights and widths for the model predicted peaks for the ≡S SOCm 2+ and ≡S SOCmOH + species distribution, and those for Cm complexes 1 and 2 derived from TRLFS, were found to be very close for both clay minerals. However, discrepancies were still apparent between the profile for the calculated ≡S S OEu(OH) 2 2 surface species and the Cm complex 3 species, especially in the case of Na-illite. Copyright © 2005 Elsevier Ltd.