PhD Theses - Eleanor Jay

Atomic Scale Modelling of Phosphate Mineral Phases for Nuclear Waste Form Development

Abstract

Phosphate minerals such as β-tricalcium phosphate and fluorapatite are abundant, exhibit a large chemical variability and samples of both phases are often used for geochronometry; so their stability over thousands of years is known. This makes them attractive as nuclear waste forms as they are stable over extended periods of time and are able to accommodate a range of waste species. In particular, due to the difficulty of incorporating halides in conventional waste forms (in particular glass), fluorapatites are considered as potential hosts for radioactive waste streams containing both actinides and halides. β -tricalcium phosphate, a structure related to fluorapatite, is not only used as a precursor for preparing apatites, but could itself be used as a host for radioactive waste. To better understand the structure and stability of these phases, atomic scale computer simulation has been employed (both dynamic and static), using classical pair potentials. One site in the β -tricalcium phosphate structure has previously been assigned a half occupancy from fitting to Rietveld power diffraction experimental data. The half occupancy gives rise to an ordering effect where the Ca2+ ions are arranged over the half occupied Ca(4) sites, in many different configurations. A comparison of different cell sizes (and increasing configuration number) is given and reported in terms of lattice energy and structural parameters. The largest cell size considered here generates the most stable structures, which have a symmetry group related to the experimentally derived average primitive unit cell. Simulated X-ray diffraction patterns indicate a difference in spectra for these low energy configurations. However, experimental X-ray diffraction patterns fail to differentiate between low energy structures. The results of the experiment and small difference in lattice energy for the most stable structures, indicate that domains containing different low energy structures are likely to exist. The configurations described here are also discussed in terms of statistical analysis. Substitutions of a range of isovalent and trivalent cations have been carried out (and compared) at Ca2+ sites in both fluorapatite and β-tricalcium phosphate structures. The defect and solution energies were calculated for both structures and compared, in order to predict the partitioning across the two phases. For the isovalent defects investigated, the defects segregate to the β-tricalcium phosphate lattice. The trivalent defects present a different trend such that defects with a smaller ionic radii than Ca2+ have an energy preference for β-tricalcium phosphate, but for those with atomic radii closer to that of the host cation the preference for either structure is indistinguishable. The ramifications of cation partitioning are discussed. Cation and anion migration in fluorapatite is considered, where the overriding feature of migration in this structure is that ionic transport (of either cations or anions) occurs preferably along the c-axis. Consideration of fluorine transport yields a more sophisticated migration mechanism than that reported previously. A “concerted mechanism” for fluorine ion transport in the lattice is discussed fully and described using fluorine density plots, over a range of temperatures. Radiation damage effects in these minerals and their ability to recover and resist damage is a crucial consideration when designing a nuclear waste host. This thesis compares the Kinchin- Pease model of threshold displacement energies to full radiation cascades. Furthermore, the effect of radiation damage on the lattice is considered by virtue of incident damage, defect types, phosphate group response and recovery of the lattice. These results indicate that the fluorapatite lattice is “relatively tolerant” to radiation damage, especially with regard to the phosphate tetrahedra, but more cascades should be considered to obtain a more statistically reliable prediction. Finally, loss of material from the waste form is most likely to occur at a surface and understanding the processes by which this occurs is important. Therefore, it is necessary to first model the surfaces, which involves static simulations of surfaces to predict surface formation energies. These energies are used as the basis for predictions of particle morphology and the effect of isovalent defect incorporation near the surface of fluorapatite.

Full Text

Journal Articles Resulting from this Thesis