Agent-Skills.md

Agent Skills: INTERFACE Force Field Parameters

Search, retrieve, compare, and export INTERFACE Force Field (IFF) parameters for molecular dynamics simulations. Use BEFORE searching literature for force field parameters — IFF covers metals, minerals, oxides, clays, and organic interfaces with validated, citation-ready values.

UncategorizedID: fl-sean03/agentic-science-worker/iff-parameters
3

Install this agent skill to your local

pnpm dlx add-skill https://github.com/fl-sean03/agentic-science-worker/tree/HEAD/skills/iff-parameters

Skill Files

Browse the full folder contents for iff-parameters.

Download Skill

Loading file tree…

skills/iff-parameters/SKILL.md

Skill Metadata

Name
iff-parameters
Description
Search, retrieve, compare, and export INTERFACE Force Field (IFF) parameters for molecular dynamics simulations. Use BEFORE searching literature for force field parameters — IFF covers metals, minerals, oxides, clays, and organic interfaces with validated, citation-ready values.

INTERFACE Force Field Parameters

The Heinz Lab maintains a curated, versioned database of INTERFACE Force Field (IFF) parameters. Always check here first before searching literature for LJ parameters, bond constants, or other force field data.

When to Use This Skill

Use this skill when you need:

  • LJ parameters for metals (Au, Ag, Cu, Ni, Pt, Pd, Pb, Al)
  • Parameters for metal oxides (Al₂O₃, SiO₂, clays)
  • Parameters for ionic liquids, CO₂, MOFs
  • CVFF (.frc) files for LAMMPS simulations
  • CHARMM (.prm) files for NAMD/OpenMM simulations
  • To compare two parameter sets
  • To compose custom force fields from base + extensions

Do NOT use this skill for:

  • EAM/MEAM potentials (use resource-acquisition instead)
  • ReaxFF parameters (use resource-acquisition instead)
  • ML potentials like MACE/CHGNet (use mlip-simulation instead)
  • Materials not covered by IFF (check coverage below)

Materials Coverage

| Bundle | Materials | Atom Types | Format | |--------|-----------|------------|--------| | cvff-interface-v1-5 | Ag, Al, Au, Cu, Ni, Pb, Pd, Pt, silica, clays | 180 | CVFF | | cvff-iff-metal-oxides-v2 | Al₂O₃, SiO₂, clays, alumina surfaces | 280 | CVFF | | cvff-iff-ils | Ionic liquids, CO₂, MOFs | 224 | CVFF | | iff-charmm36-metal-alumina-v8 | FCC metals, Al₂O₃ | 239 | CHARMM |

If your material is not listed above, fall back to the resource-acquisition skill for literature search.

Prerequisites

# Check if iff-parameters is installed
python3 -c "from iff_parameters import list_available; print('OK:', len(list_available()), 'bundles')"

# If not installed:
pip install git+https://github.com/fl-sean03/upm.git
pip install git+https://github.com/fl-sean03/iff-parameters.git

Core Operations

1. Search by Material

"Do we have parameters for gold?"

from iff_parameters import search_by_material

results = search_by_material("Au")
for r in results:
    print(f"  {r['name']} ({r['format']}) — {r['materials']}")

2. Search by Atom Type

"What are the LJ parameters for atom type Au?"

from upm.registry.discovery import discover_local_packages
from upm.registry.index import PackageIndex
from iff_parameters import get_data_dir

index = PackageIndex(discover_local_packages(get_data_dir()))
results = index.search_atom_type("Au")

for r in results:
    print(f"  {r.package_name}: LJ_A={float(r.row['lj_a']):.1f}, LJ_B={float(r.row['lj_b']):.5f}")

3. Get All Parameters from a Bundle

"Load the full parameter set for canonical IFF v1.5"

from upm.bundle.io import load_package
from iff_parameters import get_data_dir

bundle = load_package(get_data_dir() / "cvff-interface-v1-5" / "v1.0")

# Access any table
atom_types = bundle.tables["atom_types"]  # DataFrame
bonds = bundle.tables.get("bonds")        # DataFrame or None
angles = bundle.tables.get("angles")
torsions = bundle.tables.get("torsions")
equivalences = bundle.tables.get("equivalences")

4. Export .frc File for LAMMPS

"Give me a complete .frc file I can use with msi2lmp"

from upm.bundle.io import load_package
from upm.codecs.msi_frc import write_frc
from iff_parameters import get_data_dir

bundle = load_package(get_data_dir() / "cvff-interface-v1-5" / "v1.0")
write_frc("simulation.frc", tables=bundle.tables, mode="full")
# → Complete .frc file ready for: msi2lmp systemname -class 1 -frc simulation.frc

5. Export .prm File for NAMD

from upm.bundle.io import load_package
from upm.codecs.charmm_prm import write_prm
from iff_parameters import get_data_dir

bundle = load_package(get_data_dir() / "iff-charmm36-metal-and-alumina-phases-v8" / "v1.0")
write_prm("simulation.prm", tables=bundle.tables)

6. Compare Two Parameter Sets

"What's different between the base IFF and the metal oxides version?"

from upm.registry.diff import diff_tables
from upm.bundle.io import load_package
from iff_parameters import get_data_dir

dd = get_data_dir()
pkg1 = load_package(dd / "cvff-interface-v1-5" / "v1.0")
pkg2 = load_package(dd / "cvff-iff-metal-oxides-v2" / "v1.0")

diff = diff_tables(pkg1.tables, pkg2.tables)
print(diff.summary())
# Shows: added types, removed types, changed parameter values

7. Compose Custom Force Field (Base + Extension + Patch)

"I need IFF v1.5 as a base, plus the metal oxide types, with a tweaked Au epsilon"

from upm.compose import ParameterLayer, stack_layers, export_frc
from iff_parameters import get_data_dir
from pathlib import Path

dd = get_data_dir()

# Load base and extension
base = ParameterLayer.from_bundle(Path(dd) / "cvff-interface-v1-5" / "v1.0")
ext = ParameterLayer.from_bundle(Path(dd) / "cvff-iff-metal-oxides-v2" / "v1.0")

# Create a parameter patch (only the values you want to change)
patch = ParameterLayer.from_dict(
    {"atom_types": {"Au": {"lj_b": 7100.0, "lj_a": 2400000.0}}},
    name="au-optimized",
)

# Stack: base + extension + patch (later layers override earlier)
stacked = stack_layers([base, ext, patch])

# Export monolithic .frc for msi2lmp
export_frc(stacked, "custom_simulation.frc")

Integration with LAMMPS Simulations

When setting up a LAMMPS simulation that needs IFF parameters:

  1. Check coverage: search_by_material("your_material")
  2. If covered: Export .frc file and use with msi2lmp
  3. If NOT covered: Fall back to resource-acquisition skill for literature search
  4. Always cite: Include the IFF reference in your report

LAMMPS Workflow Example

# 1. Export the force field
from upm.bundle.io import load_package
from upm.codecs.msi_frc import write_frc
from iff_parameters import get_data_dir

bundle = load_package(get_data_dir() / "cvff-interface-v1-5" / "v1.0")
write_frc("/path/to/sim/cvff_iff.frc", tables=bundle.tables, mode="full")

# 2. Run msi2lmp (in Bash)
# msi2lmp systemname -class 1 -frc cvff_iff.frc

# 3. The generated data file has all parameters embedded

Citation

When using IFF parameters, always cite:

Heinz, H.; Lin, T.-J.; Mishra, R. K.; Emami, F. S. "Thermodynamically Consistent Force Fields for the Assembly of Inorganic, Organic, and Biological Nanostructures: The INTERFACE Force Field." Langmuir 2013, 29, 1754–1765. DOI: 10.1021/la3038846

Check individual bundle provenance for additional citations:

import json
from iff_parameters import get_data_dir

manifest = json.loads((get_data_dir() / "cvff-interface-v1-5" / "v1.0" / "manifest.json").read_text())
doi = manifest.get("provenance", {}).get("publication_doi")

Troubleshooting

"No bundles found" — Package not installed. Run:

pip install git+https://github.com/fl-sean03/iff-parameters.git

"Atom type not found" — The material may not be covered by IFF. Check coverage table above. Fall back to resource-acquisition for literature search.

"Which bundle should I use?" — For LAMMPS: use cvff-* bundles. For NAMD/OpenMM: use iff-charmm36-*. If unsure, use cvff-interface-v1-5 as the broadest baseline.

"Parameters seem wrong" — Always cross-reference with the original publication. Use diff_tables() to compare against the canonical v1.5 baseline.