Path-integral simulations of solid and liquid atomic hydrogen

For decades, the characterization of bulk hydrogen and its phases has challenged experimentalists and theorists alike. Solid atomic hydrogen, which requires extremely high pressures and has yet to be produced in the lab, seems just within reach. Accurate calculations of the solid phase are necessary to inform the next generation of experiments. For the theorist, hydrogen’s unique and erratic behavior requires novel and powerful techniques in simulation. We calculated the phonons of solid atomic hydrogen with reptation quantum Monte Carlo (RQMC). This is the first phonon calculation of this kind, made possible with a simple trick that we devised. We also simulated the lattice dynamics of LaH10, an analogue of solid atomic hydrogen which displays high- temperature superconductivity, with path-integral molecular dynamics. The resolution of our simulations allowed us to identify an intrinsic distortion of the superconducting structure which was previously thought to be extrinsic. Further investigation of this structural transformation also provided some insight into the design of hydrogen-rich superconductors. The improved resolution was enabled by the development of a machine-learned model that learned ab-initio calculations. Finally, we simulated the melting of solid atomic hydrogen. By combining RQMC and thermodynamic path-integral Monte Carlo, we believe that our results are the most accurate to date. Once again, these simulations were made possible by a machine-learned model. We found the melting line to be higher than previously suggested, though still below the projected superconducting temperature. We also found a clear decrease in the melting line with increasing pressure, reviving hopes that the liquid may be stable down to very low temperatures.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Physics

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Ly, Kevin Kim

Contributors dc:contributor

• Ceperley, David M
• Schleife, André
• Lorenz, Virginia O
• Gammie, Charles F

• Quantum Monte Carlo
• Atomic Hydrogen
• Machine Learning
• Lattice Dynamics

Statement dc:rights

• Copyright 2024 Kevin Ly

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/124476

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F124476