Electronic Structure Calculations

Overview

This skill group covers first-principles and machine-learning-potential-based electronic structure calculations. Two main approaches are available:

1. MACE (via ASE) -- Fast, ML-potential-based. Good for geometry optimization and quick screening. Cannot produce electronic properties (band structure, DOS) because it is a force field, not an electronic structure method.
2. Quantum ESPRESSO (QE) -- Full DFT. Required for any electronic property (band structure, DOS, charge density, etc.) and for publication-quality total energies.

Sub-Skills

Method Decision Guide

Need electronic properties (bands, DOS, charge density)?
  YES --> Use Quantum ESPRESSO DFT
  NO  --> Is high accuracy required (publication, reaction energies)?
            YES --> Use Quantum ESPRESSO DFT (SCF or vc-relax)
            NO  --> Use ASE + MACE for fast relaxation / screening

Common Prerequisites

• Pseudopotentials: QE calculations require pseudopotential files. All sub-skills show how to download SSSP pseudopotentials automatically.
• Structure files: Start from a CIF, POSCAR, or build with pymatgen/ASE. All sub-skills show how to convert formats.
• Python environment: pymatgen, ASE, MACE-torch, numpy, scipy, matplotlib are pre-installed. Install extras with pip install seekpath sumo phonopy as needed.