Agent Skills: Materials Database Access

Materials Database Access

You are querying materials databases for structures and properties.

Available Databases

Materials Project (Primary)

• 150,000+ inorganic materials
• DFT-calculated properties
• Crystal structures
• Band gaps, formation energies
• Phase diagrams
• API: https://api.materialsproject.org/

NIST Chemistry WebBook

• Thermodynamic data
• Spectroscopic data
• Phase change data
• URL: https://webbook.nist.gov/chemistry/

PubChem

• Chemical compounds
• Molecular structures
• Properties
• URL: https://pubchem.ncbi.nlm.nih.gov/

AFLOW

• Crystal structure database
• Calculated properties
• URL: http://aflowlib.org/

Crystallography Open Database (COD)

• Experimental crystal structures
• URL: http://www.crystallography.net/cod/

Materials Project API

Using pymatgen (Recommended)

import os
from mp_api.client import MPRester

# Initialize with API key from environment
api_key = os.environ.get("MP_API_KEY")
with MPRester(api_key) as mpr:
    # Get structure by formula
    docs = mpr.materials.summary.search(formula="TiO2")

    # Get by material ID
    structure = mpr.get_structure_by_material_id("mp-2657")

    # Search with properties
    results = mpr.materials.summary.search(
        band_gap=(1.0, 2.0),
        is_stable=True
    )

Direct API Access

# Search for materials
curl -H "X-API-KEY: YOUR_KEY" \
  "https://api.materialsproject.org/materials/summary/?formula=Fe2O3"

# Get specific material
curl -H "X-API-KEY: YOUR_KEY" \
  "https://api.materialsproject.org/materials/mp-19770/"

Common Queries

Get Crystal Structure

1. Search by formula (e.g., "Fe2O3")
2. Get material ID (e.g., "mp-19770")
3. Download structure (CIF, POSCAR, etc.)

Get Properties

• Band gap
• Formation energy
• Density
• Magnetic properties
• Elastic constants

Phase Diagrams

• Stability of compositions
• Competing phases
• Synthesis guidance

Structure File Formats

CIF (Crystallographic Information File)

Standard format for crystal structures:

data_TiO2
_cell_length_a   4.5937
_cell_length_b   4.5937
_cell_length_c   2.9587
_cell_angle_alpha   90.000
_cell_angle_beta    90.000
_cell_angle_gamma   90.000
_symmetry_space_group_name_H-M   'P 42/m n m'
...

POSCAR (VASP format)

TiO2 rutile
1.0
4.5937  0.0000  0.0000
0.0000  4.5937  0.0000
0.0000  0.0000  2.9587
Ti O
2 4
Direct
0.0000  0.0000  0.0000
0.5000  0.5000  0.5000
...

XYZ

Simple atomic coordinates:

6
TiO2 unit cell
Ti  0.000  0.000  0.000
Ti  2.297  2.297  1.479
O   1.396  1.396  0.000
...

Converting Structures

Using ASE (Python)

from ase.io import read, write

# Read CIF, write POSCAR
atoms = read('structure.cif')
write('POSCAR', atoms, format='vasp')

# Read CIF, write LAMMPS data
write('structure.data', atoms, format='lammps-data')

Using pymatgen

from pymatgen.core import Structure
from pymatgen.io.lammps.data import LammpsData

# Read CIF
struct = Structure.from_file('structure.cif')

# Write LAMMPS data file
lammps_data = LammpsData.from_structure(struct)
lammps_data.write_file('structure.data')

Workflow

Getting a Structure for Simulation

1. Search Database

   • Find material by formula or name
   • Check that it's the correct polymorph/phase

2. Download Structure

   • Get CIF or POSCAR format
   • Verify structure looks correct

3. Convert for Simulation

   • Convert to LAMMPS data or QE input format
   • May need to create supercell

4. Add Force Field (for MD)

   • Assign atom types
   • Apply force field parameters

Getting Material Properties

1. Search by Material ID or Formula
2. Check Data Quality
   • Is it experimentally verified?
   • What level of theory (GGA, GGA+U, etc.)?
3. Extract Relevant Properties
4. Document Source and Methodology

Data Quality Notes

Materials Project

• Properties are DFT-calculated (GGA/GGA+U)
• Band gaps are typically underestimated
• Formation energies are referenced to elemental phases
• Stability is based on convex hull analysis

Experimental vs Computed

• Always note whether data is experimental or computed
• Computed properties may differ from experiment
• Cross-reference when possible

Saving Results

Save database queries to:

workspaces/project-name/
├── structures/
│   ├── mp-19770_Fe2O3.cif
│   ├── mp-19770_Fe2O3.vasp
│   └── mp-19770_Fe2O3.data
├── properties/
│   └── Fe2O3_properties.json
└── README.md  # Document sources

Common Materials

Oxides

• TiO2 (rutile: mp-2657, anatase: mp-390)
• Fe2O3 (hematite: mp-19770)
• ZnO (wurtzite: mp-2133)
• Al2O3 (corundum: mp-1143)

Metals

• Fe (bcc: mp-13)
• Cu (fcc: mp-30)
• Al (fcc: mp-134)
• Pt (fcc: mp-126)

Semiconductors

• Si (diamond: mp-149)
• GaAs (zincblende: mp-2534)
• GaN (wurtzite: mp-804)