Benchmark and application of density functional theory for superconductors
Density functional theory for superconductors (SCDFT) is a method to compute superconducting transition temperature (Tc) without empirical parameters. We formulate a method to compute the spin-fluctuation mediated interaction with the ultrasoft pseudopotentials and implement this method into our first-principles code Superconducting-Toolkit [1]. This implementation enables us to calculate Tc using carefully constructed pseudopotentials such as the Standard Solid-State Pseudopotentials. We also implement the recently proposed Eliashberg-combined electron-phonon functional and plasmonic mass-renormalization. We perform the benchmark calculations to investigate all the above contributions in typical superconducting materials. We find that the latest SCDFT predicts Tc with an error of 30% for various materials (V, Nb, H3S, CaC6, V3Si, etc.), an underestimation of 40-50% for MgB2 and YNi2B2C, and an overestimation of 60%-300% for low- Tc (below 1.2 K) materials (Cd, Zn, and Al). We also report the estimated Tc for newly predicted quaternary hydrides [2].
[1] M. Kawamura, Y. Hizume, and T. Ozaki, Phys. Rev. B 101, 134511 (2020).[2] R. Koshoji, M. Fukuda, M. Kawamura and T. Ozaki, arXiv:2206.04971.