maxine-1520/openair_materials_skills/search-material

MP Materials Search Skill

This skill enables intelligent search and filtering of 2D semiconductor materials from the Materials Project database. The agent automatically determines optimal search parameters based on user requirements.

Core Search Function

Use the bundled filter_materials.py script to search Materials Project:

python skills/mp-material-search/scripts/filter_materials.py [options]

Search Parameters

How to Search

Step 1: Understand User Requirements

Ask clarifying questions if needed:

• What elements should the material contain?
• What band gap range is acceptable?
• Is stability (low energy above hull) important?
• Any specific properties needed? (piezoelectric, magnetic, etc.)

Step 2: Execute Search

Run the filter script with appropriate parameters:

cd <project-root>
export MP_API_KEY=<api_key>
python skills/mp-material-search/scripts/filter_materials.py \
  --elements <elements> \
  --band-gap-min <min> \
  --band-gap-max <max> \
  --max-results <num>

Step 3: Present Results

The search returns a JSON array with comprehensive material data:

• Basic info: material_id, formula, band_gap, energy_above_hull
• Structure: lattice parameters, spacegroup, crystal_system
• Properties: dielectric, piezoelectric, magnetic, elastic anisotropy
• Recommendation score (0-1, higher is better)

Present top results with:

1. Formula and material ID
2. Band gap and stability
3. Special properties (piezoelectric, magnetic, etc.)
4. Recommendation score

Example Searches

Search for C-based 2D semiconductors

python skills/mp-material-search/scripts/filter_materials.py \
  --elements C \
  --band-gap-min 0.5 \
  --band-gap-max 2.0

Search for piezoelectric 2D materials

python skills/mp-material-search/scripts/filter_materials.py \
  --band-gap-min 0.5 \
  --band-gap-max 3.0 \
  --max-results 20

Search with specific target band gap

python skills/mp-material-search/scripts/filter_materials.py \
  --elements Mo S \
  --target-band-gap 1.5 \
  --max-results 15

Output Format

Results are saved to results.json in the project root. The output includes:

• material_id: MP material ID
• formula: Chemical formula
• band_gap: Band gap in eV
• energy_above_hull: Thermodynamic stability
• recommendation_score: Overall score (0-1)
• property_scores: Breakdown of scores by category
• Structure and property details

API Key Setup

Users need a Materials Project API key:

1. Sign up at https://next-gen.materialsproject.org/
2. Set MP_API_KEY environment variable or use --api-key argument

Bundled Scripts

• scripts/filter_materials.py: Main search script (adapted from Step1_new_materials)
• scripts/setup_env.sh: Environment setup helper