Interfacing GW Calculations on Perovskite Systems through the ASE Driver

Many-body perturbation theories (MBPTs) hold great relevance for solving the properties of chemical and material systems, as they effectively capture electron correlations better than methods like Density Functional Theory (DFT). However, a significant challenge in applying these theories is their dependence on multiple software programs. Generally, one must first run a mean-field calculation using software like Quantum Espresso or PySCF to generate an initial electron density and then run software like Wannier90 for post-processing or constructing low-energy models for subsequent quantum embedding methods. As such, this intricate workflow limits democratized access to MBPTs for the broader scientific community. Hence, our work primarily focused on developing and implementing a Python interface using the Atomic Simulation Environment (ASE) package to automate the entire pipeline for generating MBPT and quantum embedding calculations. We also used our code to automate the GW band structure calculations for three perovskites (SrTiO3, SrZrO3, and KTaO3) via the in-house developed COQUÍ software package. Our findings revealed greater band structure dispersion when using GW for all three materials versus DFT.