What problem does it solve?
This Skill calculates lattice thermal conductivity of materials by running anharmonic lattice dynamics using Machine Learning Interatomic Potentials (MLIPs), producing temperature-dependent conductivity and the required third-order force-constant data.
Core Features & Use Cases
- Insulator/semiconductor guardrail: Checks whether the material is electronically insulating/semiconducting (via a band gap step using the electronic-structure Skill) and stops being meaningful for metals.
- Phonon validation prerequisite: Requires phonon properties (via the mat-phonon Skill) and advises rejecting structures with imaginary phonon frequencies before proceeding.
- MLIP-driven anharmonic workflow: Uses MLIP wrappers (MACEWrapper, MatGLWrapper, or FAIRCHEMWrapper) to compute thermal conductivity from phonon-phonon interactions with phono3py/Phonon3Calc.
- Actionable outputs: Writes a summary JSON (including approximate kappa at 300 K and 100 K) and a
phonon3.yaml file containing third-order force constants and supercell metadata.
Quick Start
Run the thermal conductivity calculation by executing the provided script with a structure file and an MLIP model name, for example: use the command python .agents/skills/mat-lattice-thermal-conductivity/scripts/calculate_thermal_conductivity.py --structure Si.cif --model_type mace --model_name MACE-OMAT-0-small --output_dir research/Si_thermal_conductivity.