Advisor(s)
Cheng-Chien Chen
Committee Member(s)
Da Yan
Renato Camata
Wenli Bi
Yogesh Vohra
School
College of Arts and Sciences
Document Type
Dissertation
Department (new version)
Physics
Date of Award
9-11-2025
Abstract
Superconductors traditionally have been discovered through trial-and-error efforts, as a general understanding of how to engineer their superconducting transition temperatures ($\Tc$) is lacking. In this dissertation, we develop a combined first principles and data-driven framework to uncover design principles for enhancing $\Tc$ in phonon-mediated superconductors. We begin by training machine-learning models to predict key materials properties and rigorously test a recently proposed empirical bound on the $\Tc$ of phonon-mediated superconductors. The results of this study allow us to hypothesize several design principles. In particular, we perform \textit{ab initio} calculations to study recently discovered superhard boron-carbon-nitrogen metals, revealing structural motifs and bonding environments which promote strong electron-phonon coupling and emergent $\Tc$ comparable to that of magnesium diboride, the highest known $\Tc$ phonon-mediated superconductor at ambient pressure. We then investigate the phenomenon of robust superconductivity in niobium-based alloys and high-entropy alloys, showing how compositional disorder and structural distortion can stabilize superconducting states under extreme conditions. Finally, I introduce ElphonPy, an open-source software package that automates the workflow for electron-phonon coupling calculations, enabling reproducible, transparent, and scalable computational studies of superconducting materials. Our efforts bridge computational materials science, machine learning, and workflow automation to provide a predictive roadmap for designing next-generation phonon-mediated superconductors.
ProQuest ID
Recommended Citation
Smith, Adam Dudley, "Integrating First-Principles Simulations And Machine Learning To Uncover Design Principles In Phonon-Mediated Superconductors" (2025). All ETDs from UAB. 7364.
https://digitalcommons.library.uab.edu/etd-collection/7364
