Source capture
Authors Omri Lesser, Yanjun Liu, Natalie Maus, Aaditya Panigrahi, Krishnanand Mallayya, Albert Gong, Anmol Kabra, Scott B. Lee, Sudipta Chatterjee, Amira Merino, Kilian Q. Weinberger, Leslie M. Schoop, Jacob R. Gardner, Eun-Ah Kim
arXiv / Publisher https://arxiv.org/abs/2510.07373
PDF https://arxiv.org/pdf/2510.07373
Relevance score 4.529
Primary category cond-mat.supr-con
Published 2026-09-01
Research paradigm Both
Sample form Thin Film

Summary

Predicting superconducting transition temperature (Tc) remains challenging. This study proposes an interpretable, structure- and chemistry-aware Gaussian process model, GP-Tc, that encodes local bonding environments via graphlet histograms and constructs an effective kernel using Earth mover’s distance, enabling Tc prediction with uncertainty. Analysis shows that Tc across different superconducting families can be predicted using only the distribution of electron affinity differences between adjacent atoms, interatomic distances, and a small number of elemental features, revealing that electron affinity difference is a key chemical parameter linking local bonding to macroscopic superconductivity in a mechanism-agnostic manner. The model reproduces the experimental Tc range of the infinite-layer nickelate Nd0.8Sr0.2NiO2 and predicts superconductivity in stoichiometric PtPb3Bi, which is experimentally confirmed with Tc≈3 K. In addition, GP-Tc is openly accessible through a web interface and identifies high-priority candidate materials such as SrNiO2 and K(PRh)2.

Materials

Methods

  • Gaussian process regression
  • graphlet histogram featurization
  • Earth Mover's Distance kernel
  • superconductivity classification
  • high-throughput ICSD screening
  • magnetic susceptibility measurement
  • electronic transport measurement
  • Werthamer-Helfand-Hohenberg analysis

Keywords

  • electron affinity difference
  • superconducting transition temperature
  • structure aware machine learning
  • uncertainty quantification
  • interpretable machine learning
  • local bonding environment
  • high throughput superconductivity screening
  • mechanism agnostic prediction

Highlights

  • PtPb3Bi is a previously unknown stoichiometric superconductor discovered through prospective experimental validation of GP-Tc, with a unique structure type hosting face-sharing dimerized [Pt2Pb8Bi4] bicapped trigonal prismatic units.
  • Electron-affinity difference, though overlooked in favor of Pauling electronegativity, emerges as the most informative descriptor and maps onto charge-transfer energy in the Zaanen-Sawatzky-Allen framework for transition-metal oxides.
  • GP-Tc is made openly accessible through a web interface for crystal-structure-based prediction from standard CIF files.
  • SrNiO2 is predicted to have Tc of about 51.5 K, substantially higher than any known nickelate superconductor, providing a concrete experimental target.

Conclusions

  • The distribution of electron-affinity differences between neighboring atoms, together with interatomic distances and simple elemental features, suffices to predict superconducting transition temperature across disparate superconducting families.
  • GP-Tc reproduces the experimentally reported Tc range of the infinite-layer nickelate Nd0.8Sr0.2NiO2 and predicts and experimentally confirms superconductivity in stoichiometric PtPb3Bi with Tc approximately 3 K.
  • The framework identifies additional high-priority superconducting candidates, including SrNiO2 with predicted Tc of 51.5 K and K(PRh)2.
  • Electron-affinity difference provides a mechanism-agnostic physical basis that captures Tc across conventional and unconventional families, including doped charge transfer insulators.
  • The predictive feature space collapses from 67 descriptors to just four second-order graphlet features, dominated by electron-affinity difference and interatomic distance.

Main claims

Not available in this batch.

Workflow

Not available in this batch.