Molecular dynamics simulation and topological analysis of the network structure of actinide-bearing materials

Abstract

Actinide waste production and storage is a complex problem, and a whole-cycle approach to actinide management is necessary to minimize the total volume of waste. In this dissertation, I examine three actinide-bearing materials relevant to both the front end and back end of the nuclear fuel cycle: light water reactor (LWR) spent fuel stored in a crystalline ceramic medium (zircon), LWR spent fuel stored in a glassy medium (alkali borosilicate glass), and three molten salt systems (LiF-BeF2, LiF-ThF4 , and LiF-UF4). I model these materials using molecular dynamics (MD) simulations, and then perform a range of material-dependent analyses - including structural evaluation, species segregation, solubility limits, and assessment of transport properties - to examine their suitability as actinide-bearing materials. The initial portion of this work focuses on actinide waste storage media, examining the microstructural changes induced in zircon and alkali borosilicate glass doped with uranium. Alpha-decay of the uranium changes the structure of the host material, inducing amorphousness, recrystallization, and microcracking, among other structural changes. My work on actinide waste storage shows the utility of topological methods for quantifying the intermediate-range structure of amorphous systems. In many cases, the intermediate-range structure correlates with larger-scale properties, such as density and viscosity. I then identify three molten salt systems of interest - LiF-BeF2 , LiF-ThF4, and LiUF4 - as a focus for analysis. LiF-BeF2 is a coolant salt, and LiF-ThF4 and LiF-UF4 are fuel salts used on the front end of the nuclear fuel cycle in molten salt reactors (MSRs). MSRs can, in some configurations, achieve extremely high actinide bum-ups. Some molten salt reactors can also be fueled by the actinides in spent fuel produced by LWRs. While MSRs have many advantages, research into new designs often proceeds slowly because of gaps in available experimental data for the molten fuel and coolant salts. I use MD simulations to evaluate the transport properties and structure of these salts, and show that these simulations can be used reliably to augment the existing body of experimental data describing the salts' material properties. Furthermore, I examine how the structure of the salt correlates with its material properties, in particular its viscosity. I use network topology-based algorithms to describe the amorphous structure quantitatively. Network-based topological methods have never before been applied to molten salts, and many new insights can be gained from the analysis.