# Electronic Thesis/Dissertation

## Insight into Nucleation Mechanisms of Tetrahedral Materials from Advanced Molecular Dynamics Simulations Open Access

This dissertation studies the nucleation mechanisms of ice, clathrate hydrate and silicon clathrate which all belong to tetrahedral materials and carry significant importance to modern society. Because of the stochastic nature and the ultra-fine scale of nucleation, the mechanisms through which these important tetrahedral materials form from liquid remain poorly understood. Our goal is to address the current knowledge gap between experiment and theory on the nucleation mechanisms by conducting molecular dynamics (MD) studies. To overcome the rare event nature of nucleation, an advanced sampling method Forward Flux Sampling (FFS) is integrated with classical MD simulations. This integration allows obtaining not only nucleation rate explicitly but also an ensemble of nucleation trajectories with their correct statistical weights. By analyzing the ensemble of trajectories obtained from FFS, we reveal the important details of nucleation at the molecular level, particular at the early stage of nucleation. By combining Backward Flux Sampling (BFS) with FFS, we can also compute the free energy profile of nucleation explicitly, which allows a comparison against the classical nucleation theory (CNT). We began our investigation by studying heterogeneous ice nucleation, which is the most relevant form of ice formation. In this part of study, we aim to understand the key microscopic factors that control ice formation, including surface hydrophilicity, surface crystallinity, and surface geometry. Our simulations reveal that heterogeneous ice nucleation on graphitic surfaces is controlled by the coupling of surface crystallinity and surface hydrophilicity. In particular, our analysis shows that the crystalline graphitic lattice with an appropriate hydrophilicity may indeed template ice basal plane by forming a strained ice layer, thus significantly enhance its ice nucleation efficiency. The templating effect is further found to transit from within the first contact layer of water to the second as the surface hydrophilicity increases, which yields an oscillating distinction between the crystalline and amorphous graphitic surfaces in their ice nucleation efficiency. With this understanding, we then shifted our focus on the role of surface geometry, where we find that an atomically sharp, concave wedge can further promote ice nucleation. Remarkably, our molecular analysis shows a significant enhancement of ice nucleation can emerge both when the geometry of a wedge matches the ice lattice and when such lattice match does not exist. In particular, a 45\si{\degree} wedge is found to greatly enhance ice nucleation by facilitating the formation of special topological defects that consequently catalyse the growth of regular ice. We then investigated the nucleation of gas hydrate, a binary compound composed of water and natural gas, e.g., methane. To facilitate the application of FFS in studying gas hydrate nucleation, we proposed a new order parameter: Half-Cage Order Parameter (H-COP), based on the topological signature of hydrate structure, and conducted a $p_{B}$ histogram analysis to verify the effectiveness of the H-COP order parameter. Our analyses show that the nucleation of M-hydrate (M can be methane or carbon dioxide) starts in the vicinity of water-M interface, and gradually transit from amorphous to crystalline structures. This is the direct support to the proposed two-step nucleation mechanism of methane hydrate. However, within the ensemble of nucleation trajectories, we also identified some direct crystallization pathways without going through a amorphous phase in the nucleation of both M-hydrate and L-hydrate. Remarkably the free energy profile of L-hydrate nucleation, which is obtained independent of nucleation theory, is found to fit well against the CNT, albeit with noticeable differences when the nucleus is small. Therefore, the CNT-like free energy profile and the existence of multiple nucleation pathways indicate the near degeneracy of these pathways in their free energy profiles and highlight the complexity of hydrate nucleation. On the basis of these results, we further proposed a new perspective of gas hydrate nucleation, which can be considered to be an entropically driven, kinetic process that proceeds via multiple pathways that have similar free energy profiles. Finally, inspired by our insight gained in studying gas hydrate nucleation, we proposed a novel synthesis route to obtain inert gas silicon clathrate, which has an attractive opto-electronic property for energy application, but remains as an experimental challenge for synthesis. We thoroughly examined this proposal through high-throughput computational studies that show the novel phases of silicon could form spontaneously from liquid silicon in the presence of noble gases under high pressure and high temperature. In particular, our results show that a medium size of noble gas, e.g., Ar, can trigger the nucleation and growth of inert-gas silicon clathrate, whereas a small noble gas such as He is able to induce the formation of an unconventional, inclusion-type compound $\text{Si}_{2}\text{He}$. Our findings, along with the gained molecular insights, thus strongly suggest it is viable to experimentally synthesize novel silicon phases with noble gas through high pressure and high temperature.