Developing a Complete AI-Accelerated Workflow for Superconductor Discovery

Abstract

The quest to identify new superconducting materials with enhanced properties is hindered by the prohibitive cost of computing electron-phonon spectral functions, severely limiting the materials space that can be explored. Here, we introduce a Bootstrapped Ensemble of Equivariant Graph Neural Networks (BEE-NET), a machine-learning model trained to predict the Eliashberg spectral function and superconducting critical temperature with a mean-absolute-error of 0.87 K relative to DFT-based Allen-Dynes calculations. Intriguingly, BEE-NET achieves a true-negative-rate of 99.4\%, enabling highly efficient screening for the rare property of superconductivity. Integrated into a multi-stage, AI-accelerated discovery pipeline that incorporates elemental-substitution strategies and machine-learned interatomic potentials, our workflow reduced over 1.3 million candidate structures to 741 dynamically and thermodynamically stable compounds with DFT-confirmed $T_{\mathrm{c}}>5$ K. We report the successful synthesis and experimental confirmation of superconductivity in two of these previously unreported compounds. This study establishes a data-driven framework that integrates machine learning, quantum calculations, and experiments to systematically accelerate superconductor discovery.

10 pages, 5 figures

References in corpus (20)

• Quantum ESPRESSO toward the exascale
• Electron-phonon interactions from first principles
• Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals
• A Universal Graph Deep Learning Interatomic Potential for the Periodic Table
• Machine learning modeling of superconducting critical temperature
• Ab-initio theory of superconductivity - I: Density functional formalism and approximate functionals
• Exploration of new superconductors and functional materials and fabrication of superconducting tapes and wires of iron pnictides
• Ab-initio theory of superconductivity - II: Applications to elemental metals
• Ironing out the details of unconventional superconductivity
• Deep Learning Model for Finding New Superconductors
• High-throughput discovery of high-temperature conventional superconductors
• Room Temperature Superconductivity: the Roles of Theory and Materials Design
• Predicting novel superconducting hydrides using machine learning approaches
• Predicting new superconductors and their critical temperatures using unsupervised machine learning
• Functional Form of the Superconducting Critical Temperature from Machine Learning
• Materials informatics based on evolutionary algorithms: Application to search for superconducting hydrogen compounds
• A recipe for an effective selection of promising candidates for high-temperature superconductors among binary hydrides
• Automatic extraction of materials and properties from superconductors scientific literature
• Accelerating superconductor discovery through tempered deep learning of the electron-phonon spectral function
• Charting the landscape of Bardeen-Cooper-Schrieffer superconductors in experimentally known compounds