| name | adsorbate-and-intermediate-generation |
| description | Use this skill for generating adsorbates and reaction intermediates, standardizing molecular inputs, and preparing structures for adsorption placement. |
adsorbate-and-intermediate-generation
Overview
Use this skill to turn SMILES-level adsorbate requests into stable 3D molecule files for later slab placement.
Quick Start
- Normalize the requested species name and canonical SMILES before generation.
- Choose
fmt intentionally: poscar for slab workflows, xyz for inspection, both when both are useful.
- Set
output_path as a path prefix, not a final filename.
- Carry forward the returned
xyz_file_rel or poscar_file_rel instead of reconstructing paths by hand.
Allowed tools
create_molecule_from_smiles
Workflow
1. Normalize chemistry first
- Resolve ambiguous protonation, charge, or radical assumptions before calling the tool.
- If the SMILES is chemically wrong, the generated structure will still be wrong.
2. Generate one canonical molecule file per species
create_molecule_from_smiles builds a 3D conformer with deterministic embedding.
fmt=poscar writes a boxed .vasp; fmt=xyz writes .xyz; fmt=both writes both.
box_padding controls the cubic POSCAR box size for isolated-molecule references.
3. Use returned paths as the handoff contract
- The tool returns
formula, natoms, xyz_file_rel, poscar_file_rel, and box_size.
- Downstream adsorption workflows should use the returned molecule path directly.
Method-critical defaults
- Keep protonation, charge, and radical assumptions explicit because they change the generated molecule, not just the file name.
- Use the same
fmt convention across a comparison set so the downstream placement workflow sees a consistent input type.
Output Contract
Return:
- canonical molecule file path(s)
- chosen
fmt
- any charge or species assumptions that remain unresolved
References
- For slab placement, hand off the returned molecule file to
adsorption-site-screening first, then promote the screening set to adsorption-screening.