Np and Pu Sorption to Manganese Oxide Minerals

Abstract

Manganese oxide minerals are a significant component of the fracture lining mineralogy at Yucca Mountain (Carlos et al., 1993) and within the tuff-confining unit at Yucca Flat (Prothro, 1998), Pahute Mesa (Drellack et al., 1997), and other locations at the Nevada Test Site (NTS). Radionuclide sorption to manganese oxide minerals was not included in recent Lawrence Livermore National Laboratory (LLNL) hydrologic source term (HST) models which attempt to predict the migration behavior of radionuclides away from underground nuclear tests. However, experiments performed for the Yucca Mountain Program suggest that these minerals may control much of the retardation of certain radionuclides, particularly Np and Pu (Triay et al., 1991; Duff et al., 1999). As a result, recent HST model results may significantly overpredict radionuclide transport away from underground nuclear tests. The sorption model used in HST calculations performed at LLNL includes sorption to iron oxide, calcite, zeolite, smectite, and mica minerals (Zavarin and Bruton 2004a; 2004b). For the majority of radiologic source term (RST) radionuclides, we believe that this accounts for the dominant sorption processes controlling transport. However, for the case of Np, sorption is rather weak to all but the iron and manganese oxides (Figure 1). Thus, we can expect to significantly reduce predicted Np transport by accounting for Np sorption to manganese oxides. Similarly, Pu has been shown to be predominantly associated with manganese oxides in Yucca Mountain fractured tuffs (Duff et al., 1999). Recent results on colloid-facilitated Pu transport (Kersting and Reimus, 2003) also suggest that manganese oxide coatings on fracture surfaces may compete with colloids for Pu, thus reducing the effects of colloid-facilitated Pu transport (Figure 1b). The available data suggest that it is important to incorporate Np and Pu sorption to manganese oxides in reactive transport models. However, few data are available for inclusion in our model. A survey of published data found only single-point (Triay et al., 1991; Kersting and Reimus, 2003; Keeney-Kennicutt and Morse, 1984; 1985) and qualitative (Duff et al., 1999; Dyer et al., 2000a; 2000b) Np and Pu sorption information. This report describes recent experiments that quantified the sorption and desorption of Np(V) and Pu(IV) onto three manganese oxide minerals as a function of pH and time. The three manganese oxides (pyrolusite, birnessite, and hollandite) have all been observed on fracture surfaces at Yucca Mountain and are likely to predominate at the NTS. Pyrolusite, birnessite, and hollandite comprise both a range of manganese oxide structure (framework, layered, and tunnel, respectively) and composition and a range of observed manganese oxide mineralogies. The pH range of 3-10 used in these experiments covers the range of pH observed in NTS groundwater (Rose et al., 1997).