On The Use Of A Multi-Site Ion-Exchange Model To Predictively Simulate The Adsorption Behaviour Of Strontium And Caesium Onto French Agricultural Soils

In case of nuclear accident, Sr-90 and Cs-134,Cs-137 are major radionuclides to account. In previous works (Appl. Geochem. 87, 167; ibid 93, 167), a database of ion-exchange parameters allowing the description of the Sr2+ and Cs+ adsorption on purified illite and smectite was developed for a multi-site ion-exchange (MSIE) model. In this study, the adsorption behaviours of Sr2+ and Cs+ were obtained with <150 mu m fractions of French soil samples: a cambisol fluvic, a calcosol, and a cambisol typic. The <2 mu m fractions of the soil samples were analysed by X-ray diffraction to estimate their clay minerals proportions that were then approximated to an illite/smectite mixture, in consistency with the CEC of the <150 mu m fractions. The database was implemented with K-illite and -smectite parameters to account for the amendment of K+ in agricultural soils. The isotherms of Sr2+ and Cs+ on the three soils - at 0.033 mol kg(w)(-1) CaCl2 (I = 0.1 mol kg(w)(-1)) and at the pH value of the water equilibrated with the soils - were then compared with simulations obtained using ion-exchange parameters from the database for the MSIE model. This simulation approach, based on the additive adsorption properties between several reactive phases, allowed to describe satisfactorily the adsorption of Sr2+ and Cs+ in most cases. In order to highlight the limiting parameters of the modelling predictive ability, different treatments were made on soil samples. The removal of the natural organic matter did not change significantly the adsorption behaviour of either Sr2+ or Cs+. The removal of the exchangeable aluminium from the cambisol typic allowed a better simulation of the adsorption isotherm in the case of Sr2+. Finally, in the case of the calcosol, the satisfactory modelling of the decrease in adsorption of Sr and Cs using a synthetic CaCl2 pore water with increasing concentrations of KNO3 allowed to verify the robustness of the MSIE model and exchange parameters from the database.