Simulation of neptunium migration as a function of clay mineralogy and redox conditions

Abstract

Migration of neptunium, a minor component of high-level nuclear waste but with a long half-life (237Np: 2.14·106 years) and high radiotoxicity, is assessed for safety of potential nuclear waste disposal sites. Argillaceous rocks, like the Opalinus Clay (OPA), are preferred host rocks because of their low hydraulic conductivity and high sorption capacity. Numerical simulations are required to quantify radionuclide migration in the context of safety assessments.One-dimensional diffusion simulations are conducted with PHREEQC to quantify the general neptunium migration processes. Sorption of Np(V) on the clay minerals is taken into account using surface complexation models. Carbonate minerals, mainly calcite, govern the pH, while the redox potential (pe) is controlled by pyrite. Therefore, both are considered in the mineral assemblage for atmospheric pCO2 conditions. Numerical simulations as a function of mineralogy and the associated changes in redox state and pore water composition are compared for two diffusion laboratory experiments conducted by Fröhlich et al. (2013) and Wu et al. (2009). Although both experiments were performed with the same setup, the determined transport parameters differ by one order of magnitude. Np(V) was applied via a synthetic pore water with pH 7.6 and pe 6 under atmospheric conditions, but to different OPA core samples. The Fröhlich experiment has already been modelled, and therefore this numerical setup is now applied to the Wu experiment to assess quantitatively whether the differences are due to variations in clay mineralogy or geochemistry.Impact of clay mineralogy is quantified for maximum and minimum weight percentages of kaolinite and illite. Our results revealed that mineralogical variations only have a minor impact on the migration, and therefore cannot explain the difference between the two experiments. Np speciation is highly sensitive to pe and is partitioned between Np(IV) and Np(V). In the experiments, a pe of 6 is applied resulting in a redox disequilibrium within the sample compared to in-situ conditions with a pe of -3.8. Simulated migration is underestimated at high pe due to increasing sorption and overestimated at low pe, where Np(V) does not sorb. Results are consistent with the Fröhlich experiment when the pe is in an intermediate range between in-situ and experiment. Applying different redox conditions compared to Fröhlich, results also coincide with the Wu experiment. Therefore, variations in neptunium sorption are mainly attributed to different redox conditions.We conclude that neptunium migration is governed less by clay mineralogy and more by redox. Therefore, redox conditions need to be accurately controlled in laboratory experiments as they determine neptunium speciation, and hence sorption. Distribution coefficients (Kd) can vary significantly, for instance, by one order of magnitude between two experiments with the same set up. This might hinder the applicability of experimentally determined Kd to assess neptunium migration for in-situ conditions. ReferencesFröhlich, D. R., et al. (2013). Radiochimca Acta, 101(9), 553-560. DOI:10.1524/ract.2013.2059Wu, T., et al. (2009). Environmental Science & Technology, 43(17), 6567-6571. DOI: 10.1021/es9008568