Force Field Development and Validation for Liquid Metal Plasma-Facing Materials

Abstract

Several liquid metals are being considered as plasma-facing materials in tokamak fusion reactors. Liquid metal plasma-facing materials have several advantages over conventional solid materials, such as the ability to self-replenish during reactor operation. Liquid lithium, tin, and lithium-tin alloys are some of the most promising candidates. Before these materials can be utilized, a sound understanding of basic properties relevant to plasma-facing applications needs to be established. While many experimental studies have been performed with this goal in mind, there is still a noticeable lack of important data. Classical molecular simulation techniques present an alternative method for studying relevant properties and can complement existing experimental studies. However, before simulation studies can take place, careful development and validation of liquid metal force fields needs to be done. In this thesis, embedded-atom method force fields for lithium and tin were compared to experimental and first-principles data in order to assess their ability to predict properties of liquid metal plasma-facing materials. The wetting properties of liquid lithium on solid molybdenum were also studied using molecular simulations.

Six lithium force fields found in the literature were tested by examining their ability to predict a variety of liquid and coexistence properties. It was found that the force field developed by Cui et al. [Modell. Simul. Mater. Sci. Eng. 20, 015014 (2012)] is the most robust due to its accurate prediction of the melting temperature and liquid-phase data. Further simulation studies of liquid lithium utilized this force field. The initial conclusion is supported by the fact that this force field accurately predicts the self-diffusivity and viscosity of liquid lithium.

A new liquid tin force field was also developed in this work. It was found that the Ravelo and Baskes force field [Phys. Rev. Lett. 79, 2482 (1997)] is suitable for the solid phases of tin, but not the liquid phase. Therefore, a simulated annealing procedure was used to construct the new force field by primarily fitting to experimental liquid data. The new force field accurately reproduces a majority of the experimental data used in the fitting procedure and also accurately predicts liquid data not used in the optimization.

Finally, the wetting properties of liquid lithium on solid molybdenum were examined. A lithium-molybdenum force field was developed by fitting to first-principles data. It was found that liquid lithium perfectly wets the (110) surface of molybdenum, in contrast with experimental data. This suggests the presence of oxygen and surface structure can significantly effect the ability of lithium to wet molybdenum. It was also found that the lithium atoms close to the molybdenum surface exhibit solid-like behavior as evidenced by their low mobility and ordered structure, even at temperatures well above the bulk melting temperature of lithium. This shows that lithium strongly adheres to the molybdenum surface. These results suggest that if the solid walls of the reactor are primarily composed of molybdenum, liquid lithium would be a strong candidate for liquid plasma-facing material due to its ability to wet and adhere to the solid surface.

This thesis highlights the importance of careful force field development and validation when performing molecular simulation studies, especially for systems with limited experimental data. The work done here establishes tools, mostly in the form of force fields, for the further study of plasma-facing materials. It also illustrates the constant interplay between classical simulations with experiments and first-principles calculations. These tools can be used to understand new materials and applications.