Ab initio modeling of superconducting alloys

Abstract

Designing new, technologically relevant superconductors has long been at the forefront of solid-state physics and chemistry research. However, developing efficient approaches for modeling the thermodynamics of superconducting alloys while accurately evaluating their physical properties has proven to be a very challenging task. To fill this gap, we propose an ab initio thermodynamic statistical method, the Extended Generalized Quasichemical Approximation (EGQCA), to describe off-stoichiometric superconductors. Within EGQCA, one can predict any computationally accessible property of the alloy, such as the critical temperature in superconductors and the electron-phonon coupling parameter, as a function of composition and crystal growth conditions by computing the cluster occurrence probabilities that minimize the overall mixing Gibbs free energy. Importantly, EGQCA incorporates directly chemical ordering, lattice distortions, and vibrational contributions. As a proof of concept, we applied EGQCA to the well-known Al-doped MgB$_2$ and to niobium alloyed with titanium and vanadium, showing a remarkable agreement with the experimental data. Additionally, we model the near-room temperature sodalite-like Y$_{1-x}$Ca$_x$H$_6$ superconducting solid solution, demonstrating that EGQCA particularly possesses a promising potential for designing in silico high-$T_{\text{c}}$ superhydride alloys. Our approach notably enables the high-throughput screening of complex superconducting solid solutions, intrinsically providing valuable insights into the interplay between synthesis, thermodynamics, and physical properties.