Mechanistic models for the interaction of Curium(III) and Europium(III) with crystalline rocks

In the 21st century the safe disposal of highly radioactive nuclear waste remains a major challenge to humanity. For the first 300 years nuclear fission products will be the main source of large amounts of the radioactivity, however, over a period of 100,000s of years most of the radiotoxicity of nuclear waste will originate from transuranium elements such as Pu, Am, and Cm. Because of the reducing conditions within a potential deep underground repository Am and Cm will exclusively – and Pu at least partially – be present in their trivalent state. For the safety case of a repository, it will be of importance to understand the mobility and retention behavior of those trivalent actinides within the host rock environment. Here, the focus is set on crystalline rock, a host rock under consideration in many countries such as Finland, Sweden, Czech Republic, and also Germany. Research in the last two decades has focused on determining the chemical speciation behavior of radionuclides on the molecular scale using primarily pure, individual mineral phases in their pulverized form. Real crystalline rock systems, on the km scale of a repository, are however much more complex, physically large and bulky, and heterogeneous. Parameters, such as heterogeneity of topography or composition, and competing surface processes have to be considered adequately in transport modelling as well as the safety assessment and thus large-scale experiments without the loss of molecular level information are needed. In this work, the focus was set on the assessment of parameters affecting chemical speciation and retention of Cm(III) and its chemical homologue Eu(III) at 5·10-7 M to 10-4 M concentrations on surfaces of crystalline rocks, relevant to potential future repositories. Correlative spectromicroscopy as a combination of spatially resolved techniques, was applied to study mineralogy, surface topography, quantitative metal ion uptake, and surface metal speciation on the same thin sections in the mm to cm scale with μm resolution. Cm(III) sorption experiments on cleaved K-feldspar crystals at pH 5.5 and 6.9 using autoradiography in conjunction with micro-focus time-resolved laser-induced fluorescence spectroscopy (μTRLFS), and vertical scanning interferometry, showed that sorption uptake and speciation is dependent on pH and surface roughness. In regions expressing a high surface roughness quantitatively more Cm(III) sorbed to the surface than in smoother regions. Further, the formed Cm(III) surface complexes are more strongly bound in regions with a high roughness, since they possess a higher number of strong sorption sites. The speciation between smooth and rough surfaces does not differ at pH 5.5. In contrast, at pH 6.9 in rougher areas, stronger species form favorably in addition to weaker surface sorption species. For comparison, Cm(III) sorption at pH 7.3 was conducted on a granodiorite crystalline rock thin section from the Grimsel Test Site, Switzerland with an inclusion of a large K-feldspar mineral grain. The spectromicroscopy approach was extended by Raman-microscopy to identify the surface mineralogy. From calibrated autoradiography of the entire thin section, it could be assessed that most sorption occurs on mica minerals rather than on feldspar. However, the K-feldspar grain was investigated in detail for a comparison of Cm(III) speciation on crystals and natural grains. On the grain, the sorption was heterogeneously distributed mainly occurring in rough surface regions. Since the overall sorption uptake was lower, the dominant species formed are bound strongly to the surface occupying strong sorption sites. Complexes were observed that were either the result of ternary complexation on the surface or incorporation. While sorption on K-feldspar crystals was higher than on the mineral grain, no ternary complexation could be observed, therefore the formation of this species is likely a result of the heterogeneous mineralogy. The formation of those complexes with carbonate is likely, whereas silicate does not seem to contribute. The exact complex structure determination will be the focus of future work. Increasing the complexity of the system, the next step was to study Cm(III) sorption at pH 8.0 on entire crystalline rock thin sections and analyzing speciation within and between different mineral gains. The focus was set on granite from Eibenstock, Germany and gneiss from Bukov, Czech Republic. On granite, sorption was highest on mica as well as feldspar and smallest on quartz as measured on the sub-mm to cm scale by quantitative autoradiography. On quartz regions, where surface roughness was higher, sorption quantities close to those on feldspar and mica could be observed. A detailed μTRLFS analysis shows highest sorption quantities on topaz, followed by feldspar, and only minor uptake was detected on quartz. Assessment of Cm(III) uptake on mica was not possible due to strong quenching of the Cm(III) luminescence through Fe as part of the mineral structure. Sorption on topaz, feldspar, and quartz occurred preferably in regions with a high surface roughness, such as surface pits, cracks, or mineral grain boundaries between the grains. Surface complexes in high roughness regions are bound more strongly than in smoother regions. A process that could either be Cm(III) ternary complexation or incorporation was exclusively observed on feldspar and quartz, likely because more sorption sites allowed for surface incorporation after which ternary complexation can occur. The experiments showed that mineralogy is the most important parameter when it comes to surface sorption, however it is closely followed by the surface roughness. On gneiss the overall mineralogy was different, which in turn affected the surface Cm(III) uptake. Combining the results from Raman microscopy and autoradiography, sorption was found almost exclusively on amphibole and mica, while little to no sorption was observed on feldspar and quartz. Due to the high Fe content of amphibole and mica, quenching hindered Cm(III) luminescence detection using μTRLFS. However, on feldspar and quartz μTRLFS allowed for uptake and speciation analysis. While no uptake was seen in smooth regions, uptake was increased in rougher regions. This highlights that in a competitive sorption environment, on low sorbing mineral phases, sorption is controlled by surface roughness. Cm(III) uptake and speciation analysis on mica using μTRLFS proved to be difficult because of luminescence quenching induced by structurally incorporated Fe in the minerals. However, to tackle this challenge, Cm(III) sorption experiments were performed on a granitic pegmatite from Olkiluoto, Finland at adjusted experimental settings, i.e., using a five times higher metal concentration than in previous experiments. Autoradiography as well as μTRLFS showed that Cm(III) sorption mainly occurred on mica, while uptake on feldspar and quartz was minor. Mica itself can be found as part of small cracks and pits, or as large grains. Inherently mica showed a high surface roughness, however Cm(III) uptake and speciation differed between the differently sized grains. On smaller grains, uptake was lower than on the larger grains, resulting in primarily stronger inner-sphere sorption species that formed. The high uptake on larger grains lead additionally to the formation of weaker inner-sphere sorption species. To compare the obtained Cm(III) results Eu(III) sorption was performed on a granite thin section. It was observed that Eu(III) uptake and speciation not only occurs heterogeneously between different minerals, but also within single grains particularly close to mineral grain boundaries. This implies that surface roughness, next to mineralogy, influences the sorption process. However, the detection of Eu(III) was hindered by surface precipitation, naturally incorporated Eu(III), and thus vague luminescence peak analysis. Spatially resolved correlative spectroscopy workflows that were further optimized in this thesis proved to be universally applicable in the range of mineral crystals to different crystalline rock thin sections and luminescent metals (Cm, Eu) and can be taken in the future as references for further studies with other luminescence metals such as Am(III) or U(VI). The derived findings show that in future assessments of the mobility of trivalent radionuclides in reactive transport modelling, parameters need to be selected carefully. Additional processes and parameters not considered before, like the surface roughness, will influence the retention of radionuclides within the geosphere. Those processes and parameters need to be quantified and implemented in the models to represent the deep repository system more reliable