Discovery of high critical magnetic field superconductors
ORAL
Abstract
Superconducting materials are of significant technological relevance for a broad range of applications, and intense research efforts aim at enhancing Tc. Intriguingly, while numerous studies have explored different computational and machine-learning routes to predict Tc, the fundamental role of the critical field Hc has so far been overlooked.
Here we present a database of critical fields Hc, Hc1, and Hc2 for 7,000 compounds covering distinct materials classes, calculated consistently from first principles and focusing on electron-phonon-paired superconductors. A theoretical framework is developed that combines α2F(ω) spectral functions [1,2] and highly accurate Fermi surfaces from density functional theory with clean-limit Eliashberg theory to obtain the coherence lengths, London penetration depths, and Ginzburg-Landau parameters. The results provide a unique statistical perspective on these key observables and uncover an unexpectedly large number of Type-I superconductors. Finally, we present a first crystal graph convolutional neural network that predicts the critical fields exclusively from the crystal structure. These results pave the way towards foundational AI models that realize the concept of inverse materials design for high-Tc and high-Hc superconductors.
[1] Gibson et al., arXiv:2401.16611 [cond-mat.supr-con]
[2] Cerqueira et al., Adv. Mater. 36, 2307085 (2024)
Here we present a database of critical fields Hc, Hc1, and Hc2 for 7,000 compounds covering distinct materials classes, calculated consistently from first principles and focusing on electron-phonon-paired superconductors. A theoretical framework is developed that combines α2F(ω) spectral functions [1,2] and highly accurate Fermi surfaces from density functional theory with clean-limit Eliashberg theory to obtain the coherence lengths, London penetration depths, and Ginzburg-Landau parameters. The results provide a unique statistical perspective on these key observables and uncover an unexpectedly large number of Type-I superconductors. Finally, we present a first crystal graph convolutional neural network that predicts the critical fields exclusively from the crystal structure. These results pave the way towards foundational AI models that realize the concept of inverse materials design for high-Tc and high-Hc superconductors.
[1] Gibson et al., arXiv:2401.16611 [cond-mat.supr-con]
[2] Cerqueira et al., Adv. Mater. 36, 2307085 (2024)
–
Presenters
-
Benjamin Geisler
University of Florida
Authors
-
Benjamin Geisler
University of Florida
-
Philip M Dee
Oak Ridge National Laboratory, University of Florida
-
James J Hamlin
University of Florida, Department of Physics, University of Florida
-
Gregory R Stewart
University of Florida, Department of Physics, University of Florida
-
Richard G Hennig
University of Florida, Department of Materials Science and Engineering, University of Florida
-
Peter J Hirschfeld
University of Florida, Department of Physics, University of Florida