Development of AiiDA-AkaiKKR: Automated Workflow and Improved Integration-Energy Scheme for First-Principles Vertex-Corrected Conductivity Calculations

AkaiKKR is a first-principles electronic-structure calculation software capable of evaluating electrical conductivity including vertex corrections. In this study, AiiDA-AkaiKKR was developed as an AiiDA-based workflow to automate AkaiKKR calculations with full provenance tracking and reproducibility. Preparing input files from crystal structures had previously been cumbersome, but in this implementation, inputs can be automatically generated from standard CIF files, greatly simplifying the setup process. Intermediate data are stored on the computation server, reducing network load during additional or restarted calculations. A key numerical issue in the KKR formalism is the determination of the lower integration energy measured from the Fermi energy required in the Green-function formalism, which fundamentally differs from wavefunction-based methods such as VASP or Quantum ESPRESSO. The automatic determination scheme for this energy has been improved, resolving issues identified in the previous implementation. In the next stage, systematic calculation data will be collected and AI techniques will be employed to adaptively optimize this parameter. This work demonstrates how workflow automation, data-driven input generation, and numerical improvements can contribute to the AiiDA software ecosystem and to the advancement of reproducible and efficient first-principles simulations.