Quasiperiodic pinning arrays, as recently demonstrated theoretically and experimentally using a five-fold Penrose tiling, can lead to a significant enhancement of the critical current Ic as compared to "traditional" regular pinning arrays. However, while regular arrays showed only a sharp peak in Ic(Phi) at the matching flux Phi1 and quasiperiodic arrays provided a much broader maximum at Phi<Phi1, both types of pinning arrays turned out to be inefficient for fluxes larger than Phi1. We demonstrate theoretically and experimentally the enhancement of Ic(Phi) for Phi>Phi1 by using non-Penrose quasiperiodic pinning arrays. This result is based on a qualitatively different mechanism of flux pinning by quasiperiodic pinning arrays and could be potentially useful for applications in superconducting micro-electronic devices operating in a broad range of magnetic fields.
Reinforcement learning-guided optimization of critical current in high-temperature superconductors
Quantum Measurement Group, MIT, Cambridge, MA 02139, USA（MIT量子测量组）；Center for Computational Science and Engineering, MIT, Cambridge, MA 02139, USA（MIT计算科学与工程中心）；Department of Materials Science and Engineering, MIT, Cambridge, MA 02139, USA（材料科学与工程系）；Department of Nuclear Science and Engineering, MIT, Cambridge, MA 02139, USA（核科学与工程系）；Department of Physics, MIT, Cambridge, MA 02139, USA（物理系）
