Computationally efficient method for calculating electron-phonon coupling for high-throughput superconductivity screening

High temperature superconductivity is one of the most sought-after properties in high-throughput computational screening due to its large potential impact. However, calculation of the superconducting transition temperature (T_c) from first principles remains computationally expensive, limiting the system sizes that can be screened. This makes it unfeasible to evaluate systems like heterostructures, interfaces or compounds with large primitive cells. This is of interest due to 2D systems like FeSe on SrTiO₃ which shows a large enhancement of T_c due to the substrate.

We present a computationally inexpensive method (https://github.com/od-qmul/lambda_calc_pub) for estimating electron-phonon coupling (EPC) using density functional theory (DFT) calculations of frozen phonons in materials with large unit cells from the perturbation of electronic bands near the Fermi energy. The Allen-Dynes equation can then be used to estimate T_c. We compare our estimations of the EPC in a variety of materials, including hydrides, to values calculated using full Eliashberg methods, showing good agreement with previous results. This demonstrates the reliability of our method which can now be used to screen any materials or complex interfacial systems calculable with DFT, meaning systems up to ~1,000 atoms are now feasible.