Machine learning guided study of BCS superconductors

Superconductors, materials with zero electrical resistance and perfect diamagnetism below a critical temperature, hold promise for advancing technology. However, traditional materials discovery methods, that could identify promising superconductors, are time-consuming and resource-intensive. This study presents a machine learning (ML) framework to expedite the identification of novel BCS superconductors. A comprehensive database, amalgamating the Crystallographic Open Database and the Handbook of Superconductivity, was curated.

ML models were trained on these data to accurately predict superconducting properties. Our results demonstrate remarkable prediction accuracy, effectively distinguishing superconducting and non-superconducting materials. Additionally, the ML models uncovered hidden patterns within the high-dimensional materials data, revealing intricate relationships between crystal structure and electronic properties. This research offers a promising avenue for accelerating the discovery of new superconductor materials, paving the way for transformative applications in fields such as energy, transportation, and quantum computing.