mat-dft-electron-phonon

Files (4)

• SKILL.md
• README.md · examples/silicon/
• si_flow.json · examples/silicon/
• generate_inputs.py · scripts/

SKILL.md (first lines)

# mat-dft-electron-phonon

## Goal
To determine the impact of electron-phonon coupling on the electronic eigenstates of a solid. At $T=0$ K, quantum fluctuations (zero-point motion) slightly perturb the geometric symmetry of a perfectly static lattice, causing a contraction known as zero-point renormalization (ZPR). As temperature increases, higher phonon modes dictate further eigenenergy shifts.

## Background
Standard DFT predicts bandgaps under the Born-Oppenheimer limit (fixed infinite massive ions). Computing true temperature-dependent optical properties (photoluminescence shifting, exciton broadening) mandates adding back the phonon response. The `ElectronPhononMaker` calculates phonon modes first (via `phonopy`), generates properly thermalized stochastic structural snapshots respecting the true classical/quantum Bose-Einstein occupancies, and computes the static gap for each snapshot.

## Instructions

### 1. Construct the Electron-Phonon Workflow
Generating the inputs uses the `ElectronPhononMaker`. You only need to provide the target primitive structure and the temperature list you want dynamically sampled.

```bash
${CLAUDE_SKILL_DIR}/../../venv/run cpu python ${CLAUDE_SKILL_DIR}/scripts/generate_inputs.py --output elph_flow.json
```

### 2. Job Execution
The default script serializes the Directed Acyclic Graph (DAG) logic. Because calculating robust phonon displacements involves constructing potentially hundreds of large supercell single-point DFT calculations, ensure you map this to an established HPC worker infrastructure (`jobflow` or Fireworks) rather than executing interactively locally.

### 3. Parse Output
The termination node evaluates the mean and variance of the bandgap/band edges from all stochastically distributed geometric snapshots at a given temperature, returning the renormalized gap.

## Examples

Run the DAG generation for pristine primitive Silicon.

```bash
cd ${CLAUDE_SKILL_DIR}/examples/silicon
${CLAUDE_SKILL_DIR}/../../venv/run cpu python ../../scripts/generate_inputs.py --output si_flow.json
```

## Constraints
- **Environments**: Scripts require the `cpu` environment, which contains `phonopy`.
- **Phase Stability**: The material must be strictly stable. If imaginary modes exist in the phonon branch, thermal displacement mapping will fail catastrophically since occupations of negative frequencies diverge.
- **Supercells**: Accuracy is critically bound to taking a large enough supercell (`supercell_matrix`) to capture long-wavelength phonons correctly.

## References
- Zacharias, M., & Giustino, F. "One-shot calculation of temperature-dependent optical spectra and phonon-induced band-gap renormalization", *Phys. Rev. B*, 94, 075125 (2016). [DOI](https://doi.org/10.1103/PhysRevB.94.075125)

---

**Author:** Bowen Deng
**Contact:** [GitHub](https://github.com/learningmatter-mit)

From learningmatter-mit/AtomisticSkills, MIT.

More from learningmatter-mit/AtomisticSkills