majiayu000 name: mechanical-properties description: Mechanical Properties (5 sub-skills: angular-mechanics, elastic-constants, energy-strain-method, equation-of-state, stress-strain-method)
Mechanical Properties
Overview
This skill group covers calculations of mechanical properties of crystalline materials. Two main approaches are available:
- MACE (via ASE) -- Fast ML-potential-based calculations. Good for screening, rapid estimation of elastic constants and equations of state. Seconds to minutes per structure.
- Quantum ESPRESSO (QE) -- Full DFT. Required for publication-quality results and when MACE accuracy is insufficient for the system of interest.
Both approaches follow the same physical workflow (inspired by atomate2's elastic and EOS flows): relax the structure, apply systematic deformations, compute response properties, and fit constitutive models.
Sub-Skills
| Sub-Skill | Directory | Description |
|---|---|---|
| Elastic Constants | elastic-constants/ |
Full elastic tensor via stress-strain method, Voigt-Reuss-Hill moduli, stability criteria |
| Equation of State | equation-of-state/ |
E-V curves, Birch-Murnaghan / Vinet / Murnaghan EOS fits, bulk modulus and its pressure derivative |
Method Decision Guide
Need publication-quality elastic constants?
YES --> Use Quantum ESPRESSO DFT (stress-strain with pw.x)
NO --> Is the material well-represented by MACE training data?
YES --> Use ASE + MACE (fast, ~seconds)
NO --> Use Quantum ESPRESSO DFT
Common Prerequisites
- Structure: Start from a CIF, POSCAR, or Materials Project query. Symmetrize before deformation workflows to reduce the number of independent deformations.
- Pseudopotentials: QE calculations need pseudopotential files (SSSP library recommended).
- Python packages: pymatgen, ASE, mace-torch, numpy, scipy, matplotlib are pre-installed.