Sampling the Materials Space for Conventional Superconducting Compounds.

Area of Science:

• Materials Science
• Condensed Matter Physics
• Computational Chemistry

Background:

• Conventional superconductors are crucial for many technologies.
• Discovering new superconductors with higher transition temperatures is a key scientific challenge.
• High-throughput computational methods can accelerate materials discovery.

Purpose of the Study:

• To develop and apply a machine learning model for predicting superconducting properties.
• To screen a large database of metallic compounds for potential high-temperature superconductors.
• To identify novel superconducting materials with practical applications.

Main Methods:

• Generated a dataset of ~7000 electron-phonon calculations.
• Trained a machine learning model on structural, compositional, and electronic properties.
• Screened ~200,000 metallic compounds for predicted transition temperatures (Tc).
• Validated promising candidates using density-functional perturbation theory.

Main Results:

• Identified 541 compounds with Tc > 10 K.
• Discovered several promising nitrides, hydrides, and intermetallic compounds.
• Predicted LiMoN2 in its trigonal phase to have Tc > 38 K.
• LiMoN2 in this phase has been previously synthesized.

Conclusions:

• Machine learning significantly accelerates the search for novel superconducting materials.
• The study identified numerous promising candidates for future research and applications.
• LiMoN2 represents a particularly promising material for high-temperature superconductivity.