Skip to content
cccskills
FreePublished on GitHub

adsorbate-placement

Use when the user asks to place an adsorbate molecule on a surface, find adsorption sites, or set up a surface+adsorbate model for DFT.

The English summary is being prepared. Showing the original description.

Hello-QM/catgo-LRG208 stars(GitHub 208 + cccskills 0)

Run score Not tested

See how to install

Files included (1)

  • SKILL.md11.9 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Adsorbate Placement

Overview

The adsorbate_place task type places adsorbate molecules on surface slabs. It uses ferrox (Rust) find_adsorption_sites to locate surface sites, then the CatGo placement engine (utils/adsorbate_placement.py) for Rodrigues rotation, overlap detection, and multi-dentate support.

Task Type: adsorbate_place

  • Type: adsorbate_place (local task, no HPC needed)
  • Engine: ferrox site finder + CatGo placement engine
  • Outputs: structure (slab+adsorbate as JSON)

Parameters

ParameterTypeDefaultDescription
structureJSONrequiredSlab structure input
speciesstr"OH"Adsorbate species name
sitestr"ontop"Site type: "ontop", "bridge", "hollow", or "all"
heightfloat2.0Height above surface in Angstroms
site_indexint0Which site of the given type to use (0 = first)

Supported Adsorbate Species

SpeciesAtomsBinding AtomNotes
OHO, HOHydroxyl, O-H = 0.96 A
OOOAtomic oxygen
OOHO, O, HO1Hydroperoxo, key OER intermediate
HHHAtomic hydrogen
H2OO, H, HOWater molecule
COC, OCCarbon monoxide, C-O = 1.13 A
CO2C, O, OCCarbon dioxide, linear
N2N, NNDinitrogen, N-N = 1.10 A
NHN, HNImide
NH2N, H, HNAmino
NH3N, H, H, HNAmmonia
CHOC, H, OCFormyl
COOHC, O, O, HCCarboxyl

Site Types

Siteferrox TypeCoordinationDescription
ontopatop1-foldDirectly above one surface atom
bridgebridge2-foldBetween two surface atoms
hollowhollow33-foldAbove threefold hollow site
allatop (default)1-foldAuto-selects ontop

Discussion Checkpoints

🔴 Must discuss with user:

  • Adsorbate species — determines the chemistry being studied; wrong species = wrong intermediate in the reaction pathway
  • Adsorption site (ontop/bridge/hollow) — different sites have different binding energies; for screening, test all three and report the most stable

🟡 Recommend confirming:

  • Height above surface (default: 2.0 A) — too close triggers repulsion during geo_opt, too far causes adsorbate to fly away; use 1.5-1.8 A for atomic O, 2.0-2.5 A for molecular species
  • Site index (default: 0) — which specific site of the given type; call catgo_analyze(action="adsorption_sites") first to see available sites
  • Multi-dentate orientation — for OOH, COOH, and other multi-atom adsorbates, the binding orientation matters; verify with catgo_view after placement

🟢 Safe defaults:

  • Collision detection enabled (ferrox automatic)
  • Automatic site finding via ferrox find_adsorption_sites
  • Binding atom orientation follows species convention (O down for OH, C down for CO)

MCP Workflow: Place Adsorbate on Slab

Step 1: Generate slab (or use existing)

{"tool": "catgo_fetch", "arguments": {
  "action": "crystal", "formula": "Pt", "provider": "mp"
}}
{"tool": "catgo_structure", "arguments": {
  "action": "slab",
  "miller_index": [1, 1, 1],
  "min_slab_size": 12.0,
  "min_vacuum_size": 15.0
}}
{"tool": "catgo_structure", "arguments": {
  "action": "supercell",
  "scaling": [2, 2, 1]
}}

Step 2: Find adsorption sites

{"tool": "catgo_analyze", "arguments": {
  "action": "adsorption_sites"
}}

This returns all available sites (ontop, bridge, hollow) with coordinates.

Step 3: Build workflow with adsorbate placement

Using the workflow engine with adsorbate_place node:

{"tool": "catgo_workflow", "arguments": {
  "action": "batch",
  "workflow_id": "wf_123",
  "operations": [
    {"op": "add_node", "node_type": "slab_gen", "label": "slab1",
     "params": {"miller": [1, 1, 1], "layers": 4, "vacuum": 15.0}},
    {"op": "add_node", "node_type": "adsorbate_place", "label": "ads1",
     "params": {"species": "OH", "site": "ontop", "height": 2.0, "site_index": 0}},
    {"op": "add_node", "node_type": "geo_opt", "label": "go1",
     "params": {"software": "vasp", "ENCUT": 520, "freeze_mode": "layers", "freeze_layers": 2}},
    {"op": "connect", "from_id": "<structure_input_id>", "to_id": "slab1"},
    {"op": "connect", "from_id": "slab1", "to_id": "ads1",
     "from_handle": "structure", "to_handle": "structure"},
    {"op": "connect", "from_id": "ads1", "to_id": "go1",
     "from_handle": "structure", "to_handle": "structure"}
  ]
}}

Step 4: PENDING_REVIEW -- verify adsorbate position

The user should inspect the structure before submitting geo_opt. Check:

  • Adsorbate is at the correct site (ontop/bridge/hollow)
  • Height above surface is reasonable (1.5-2.5 A for most species)
  • No atom overlaps or unrealistic bond lengths
  • Binding atom orientation is correct (e.g., C down for CO, O down for OH)
{"tool": "catgo_view", "arguments": {"action": "get_state"}}

Step 5: Submit for DFT optimization

{"tool": "catgo_workflow", "arguments": {
  "action": "run",
  "workflow_id": "wf_123",
  "run_config": {"cluster": "expanse", "partition": "shared", "walltime": "04:00:00"}
}}

Python API

from catgo.workflow import Workflow

wf = Workflow("OH on Pt(111)")

# Bulk input
inp = wf.add_task("structure_input", structure=pt_bulk_json)

# Cut slab
slab = wf.add_task("slab_gen",
    structure=inp.output.structure,
    miller=(1, 1, 1),
    layers=4,
    vacuum=15.0)

# Place adsorbate
ads = wf.add_task("adsorbate_place",
    structure=slab.output.structure,
    species="OH",
    site="ontop",
    height=2.0,
    site_index=0)

# PENDING_REVIEW: user should verify adsorbate position before geo_opt

# Geometry optimization
opt = wf.add_task("geo_opt",
    structure=ads.output.structure,
    software="vasp",
    ENCUT=520,
    freeze_mode="layers",
    freeze_layers=2)

wf.submit()

Complete OER Workflow Example

The oxygen evolution reaction (OER) has four intermediates: *OH, *O, *OOH, and clean slab. Each needs geo_opt + freq + gibbs_energy.

from catgo.workflow import Workflow

wf = Workflow("OER on IrO2(110)")

# Shared bulk input
inp = wf.add_task("structure_input", structure=iro2_bulk_json)
slab = wf.add_task("slab_gen",
    structure=inp.output.structure,
    miller=(1, 1, 0), layers=4, vacuum=15.0)

# --- Clean slab branch ---
slab_opt = wf.add_task("geo_opt", structure=slab.output.structure,
    software="vasp", ENCUT=520, system_name="clean_slab",
    freeze_mode="layers", freeze_layers=2)
slab_freq = wf.add_task("freq", structure=slab_opt.output.structure,
    software="vasp", freeze_mode="layers", freeze_layers=2)
slab_gibbs = wf.add_task("gibbs_energy",
    energy=slab_opt.output.energy,
    frequencies=slab_freq.output.frequencies,
    phase="adsorbed")

# --- *OH branch ---
oh_ads = wf.add_task("adsorbate_place", structure=slab.output.structure,
    species="OH", site="ontop", height=2.0)
oh_opt = wf.add_task("geo_opt", structure=oh_ads.output.structure,
    software="vasp", ENCUT=520, system_name="OH_ads",
    freeze_mode="layers", freeze_layers=2)
oh_freq = wf.add_task("freq", structure=oh_opt.output.structure,
    software="vasp", freeze_mode="layers", freeze_layers=2)
oh_gibbs = wf.add_task("gibbs_energy",
    energy=oh_opt.output.energy,
    frequencies=oh_freq.output.frequencies,
    phase="adsorbed")

# --- *O branch ---
o_ads = wf.add_task("adsorbate_place", structure=slab.output.structure,
    species="O", site="ontop", height=1.8)
o_opt = wf.add_task("geo_opt", structure=o_ads.output.structure,
    software="vasp", ENCUT=520, system_name="O_ads",
    freeze_mode="layers", freeze_layers=2)
o_freq = wf.add_task("freq", structure=o_opt.output.structure,
    software="vasp", freeze_mode="layers", freeze_layers=2)
o_gibbs = wf.add_task("gibbs_energy",
    energy=o_opt.output.energy,
    frequencies=o_freq.output.frequencies,
    phase="adsorbed")

# --- *OOH branch ---
ooh_ads = wf.add_task("adsorbate_place", structure=slab.output.structure,
    species="OOH", site="ontop", height=2.0)
ooh_opt = wf.add_task("geo_opt", structure=ooh_ads.output.structure,
    software="vasp", ENCUT=520, system_name="OOH_ads",
    freeze_mode="layers", freeze_layers=2)
ooh_freq = wf.add_task("freq", structure=ooh_opt.output.structure,
    software="vasp", freeze_mode="layers", freeze_layers=2)
ooh_gibbs = wf.add_task("gibbs_energy",
    energy=ooh_opt.output.energy,
    frequencies=ooh_freq.output.frequencies,
    phase="adsorbed")

# --- Gas-phase references (H2O, H2) ---
h2o_inp = wf.add_task("structure_input", structure=h2o_gas_json)
h2o_opt = wf.add_task("geo_opt", structure=h2o_inp.output.structure,
    software="vasp", ENCUT=520, ISMEAR=0, KPOINTS=[1,1,1],
    system_name="H2O_gas")
h2o_freq = wf.add_task("freq", structure=h2o_opt.output.structure,
    software="vasp")
h2o_gibbs = wf.add_task("gibbs_energy",
    energy=h2o_opt.output.energy,
    frequencies=h2o_freq.output.frequencies,
    phase="gas")

h2_inp = wf.add_task("structure_input", structure=h2_gas_json)
h2_opt = wf.add_task("geo_opt", structure=h2_inp.output.structure,
    software="vasp", ENCUT=520, ISMEAR=0, KPOINTS=[1,1,1],
    system_name="H2_gas")
h2_freq = wf.add_task("freq", structure=h2_opt.output.structure,
    software="vasp")
h2_gibbs = wf.add_task("gibbs_energy",
    energy=h2_opt.output.energy,
    frequencies=h2_freq.output.frequencies,
    phase="gas")

# Free energy diagram
fed = wf.add_task("free_energy_diagram",
    gibbs_values={
        "clean": slab_gibbs.output.gibbs,
        "OH": oh_gibbs.output.gibbs,
        "O": o_gibbs.output.gibbs,
        "OOH": ooh_gibbs.output.gibbs,
        "H2O": h2o_gibbs.output.gibbs,
        "H2": h2_gibbs.output.gibbs,
    },
    step_order=["clean", "OH", "O", "OOH", "O2"])

wf.submit()

DAG Structure (Single Adsorbate)

bulk_crystal --> slab_gen --> adsorbate_place --> [PENDING_REVIEW] --> geo_opt

DAG Structure (OER)

                                /--> *OH  --> geo_opt --> freq --> gibbs --\
bulk --> slab_gen --> slab -----+--> *O   --> geo_opt --> freq --> gibbs ---+--> free_energy_diagram
                          \    \--> *OOH --> geo_opt --> freq --> gibbs --/
                           \--> clean_slab --> geo_opt --> freq --> gibbs -/

Comparing Multiple Sites

To compare adsorption at different sites (ontop, bridge, hollow), create separate branches from the same slab:

for site_type in ["ontop", "bridge", "hollow"]:
    ads = wf.add_task("adsorbate_place",
        structure=slab.output.structure,
        species="OH",
        site=site_type,
        height=2.0,
        site_index=0)
    opt = wf.add_task("geo_opt",
        structure=ads.output.structure,
        software="vasp", ENCUT=520,
        system_name=f"OH_{site_type}")

Common Pitfalls

  1. Always build the slab and supercell BEFORE placing adsorbates. Adding adsorbates to a 1x1 slab gives unphysically high coverage.
  2. The initial height matters: too close triggers repulsion, too far may cause the adsorbate to fly away during geo_opt. Use 1.5-2.5 A for most species.
  3. OOH tends to dissociate into O + OH during relaxation on some surfaces. Use tight EDIFFG and monitor the trajectory.
  4. For bridge and hollow sites, the adsorbate is placed between atoms automatically by the ferrox site finder. Do not manually calculate midpoints.
  5. After placement, always verify with catgo_view that no atoms overlap and the geometry looks reasonable before submitting to DFT.
  6. When using site_index, call catgo_analyze with action: "adsorption_sites" first to see available sites and their indices.
  7. The site="all" option defaults to ontop (atop) sites. For specific site types, always pass the explicit type name.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

abinit

Free

Generate and manage ABINIT DFT calculations. Use when the user requests ABINIT, or needs DFPT phonons, GW calculations, or BSE optical spectra.

The English summary is being prepared. Showing the original description.

Hello-QM/catgo-LRG208Updated Oct 23, 2026

Use when the user asks for adsorption energy, binding energy, or wants to compare how strongly a molecule binds to a surface.

The English summary is being prepared. Showing the original description.

Hello-QM/catgo-LRG208Updated Oct 23, 2026

Use when the user asks to analyze computational results: Gibbs free energy, OER/HER/CO2RR overpotentials, adsorption energy, convergence tests, DOS/d-band analysis, or Bader charge analysis.

The English summary is being prepared. Showing the original description.

Hello-QM/catgo-LRG208Updated Oct 23, 2026

Use AmberTools antechamber for molecular force field parameterization. Generates GAFF/GAFF2 parameters and AM1-BCC or RESP charges for use in AMBER or LAMMPS classical MD simulations.

The English summary is being prepared. Showing the original description.

Hello-QM/catgo-LRG208Updated Oct 23, 2026

One-shot recipes for adding, deleting, moving, and replacing individual atoms in the active CatGo viewer structure. Use whenever the user asks to "add an atom", "delete atoms", "remove this atom", "move atom N", "replace H with Li", or any single-atom edit — including "add a hydrogen on top of atom 12" / "delete atoms 5–9" / "swap the oxygen at site 3 for fluorine". Skips the discovery loop (no need to list structure tools or check adsorption sites first); goes straight to the matching `catgo_structure` action. Triggers in Chinese on 加原子, 删原子, 减原子, 移动原子, 替换原子, 增加氢, 删除氧, 把 ... 换成 ....

The English summary is being prepared. Showing the original description.

Hello-QM/catgo-LRG208Updated Oct 23, 2026

Diagnose and fix the CatGo backend. Use whenever the user asks to "check the backend", "diagnose the backend", "fix the backend", "why isn't X working", reports connection problems, a blank/failing UI, "router not loaded", "deps missing", or "can't reach the server". Reads GET /api/diagnostics, explains each issue from its fix_hint, auto-applies SAFE fixes (reconnect HPC, pip install a missing dep via the terminal), confirms risky actions first, and — if diagnostics itself is unreachable — tells the user how to start the backend instead of pretending to fix one that is down.

The English summary is being prepared. Showing the original description.

Hello-QM/catgo-LRG208Updated Oct 23, 2026

See all skills by Hello-QM

Report a problem with this skill