Skip to content

Effect of surface chemistry on water interaction with Cu(111)

Summary

Classical molecular dynamics with the third-generation charge-optimized many-body COMB3 potential studies water structure and dynamics on Cu(111), comparing COMB3 adsorption energetics and geometries to DFT. Nanoscale water droplets are simulated on bare, oxidized reconstructed, and hydroxylated Cu(111) at 20 K, 130 K, and 300 K to capture temperature-dependent wetting and spreading, including spreading exponents and final base radii. The motivation is nanoscale wetting on metal electrodes and corrosion interfaces where native oxides and hydroxyls alter hydrophilicity relative to atomically clean Cu.

Methods

MD application (atomistic dynamics)

  • Engine / code: LAMMPS with COMB3 for coupled Cu / O / H chemistry and flexible water (implementation notes in §2.2 of the article).
  • System size / composition (droplet spreading benchmark): 576 H₂O molecules prepared as an ice I_h-derived droplet with initial diameter ~2.82 nm (orthorhombic melt-preparation cell yields a slightly non-spherical initial cluster); Cu(111) substrate 142 Å × 143 Å with 10,752 Cu atoms in three metal layers (Z = surface normal).
  • Droplet preparation: NPT melt of ice-I_h at 300 K for 200 ps to obtain an amorphous liquid droplet configuration before placing on Cu (§3.1).
  • Boundaries / periodicity: 3D periodic slab with ≥10 Å vacuum in earlier adsorption benchmarks; droplet-on-slab images in the article are standard periodic-slab setups with vacuum separation along Z (see figures and §2.4 slab description: 31×31 Å nine-layer slab with 10 Å vacuum for adsorption minimizations—distinct cell from the large droplet slab).
  • Ensemble: NVT (constant-volume, constant-temperature) for the spreading runs described in §3.1 (“constant-volume, constant-temperature ensemble at 300 K” for the room-temperature case; analogous setups at 130 K and 20 K).
  • Thermostat: Langevin thermostat on the mobile water + top two Cu layers with damping 100.0 ps; bottom Cu layer fixed. The manuscript documents an effective-temperature calibration trick: input 370 K for water to obtain ~300 K output, while Cu uses input 300 K to stabilize charges/vibrations (§3.1).
  • Timestep: N/A — an integration Δt is not stated in the Langmuir 2016, 32, 8061–8070 article body text extracted from pdf_path (full-PDF text search did not locate an explicit fs timestep); confirm in Supporting Information or the publisher PDF if strict reproduction requires this parameter.
  • Duration: Droplet spreading statistics are accumulated after defining interfacial / precursor-film / surface / bulk regions as in §3.1 (use the article’s stated equilibration/spreading windows for exponent fits).
  • Barostat: N/A — spreading production is NVT (barostat appears only in the 200 ps droplet-preparation melt noted above).
  • Pressure: N/A — not a controlled variable in the quoted NVT spreading segment.
  • Electric field: N/A — not used.
  • Replica / enhanced sampling: N/A — brute-force classical MD on the reported droplet trajectories.

Force-field training (COMB3 parameter context)

The article documents COMB3 O/H fitting with POSMat against molecular and ice I_h benchmarks and couples that water description to existing Cu–hydrocarbon and Cu₂O COMB3 parameter files (§2.2–2.3). This page does not duplicate the parameter tables—use the paper/SI for numeric parameter sets.

Static QM / DFT (benchmarking against COMB3)

  • Program / functional: VASP 5.3.5, PBE-GGA, PAW cores; DFT-D3 dispersion with Becke–Johnson damping (§2.1).
  • Cutoffs / smearing: 450 eV plane-wave cutoff; Methfessel–Paxton smearing (σ = 0.2 eV for Cu systems, 0.003 eV for isolated water in the quoted setup text).
  • k-sampling: 5×5×1 Monkhorst–Pack mesh for Cu(111) slabs in §2.1.
  • Structures / targets: Quasi-Newton relaxations to 0.02 eV Å⁻¹ force tolerance; adsorption energies for monomer/dimer/hexamer water configurations on Cu(111) compared to COMB3 (§2.4, Table 3).

Findings

  • DFT vs COMB3 adsorption: COMB3 reproduces the trend of stronger adsorption with increasing cluster coverage for the dimer/hexamer cases tabulated, while monomer site ordering (atop vs hollow) can disagree with DFT because of transferable Cu–O/Cu–H parameters not refit on these specific adsorption points (authored caveat in §2.4).
  • Spreading exponents (300 K): Bare Cu(111) shows R₀ ≈ t^0.16 with final base radius ~3.5 nm in the abstract-level summary (simulation setup uses the ~2.82 nm initial droplet diameter in §3.1).
  • Low temperature: At 20 K and 130 K, droplets remain compact with minimal spreading, consistent with STM-reported clustering / limited mobility discussed in the article.
  • Surface chemistry: Oxidized reconstructed and OH-covered Cu(111) reduce both the spreading exponent (≈t^0.14) and the final base radius (≈3.0 nm oxidized, ≈2.5 nm hydroxylated, as reported in the abstract), linking oxide/hydroxide termination to less efficient nanoscale wetting on Cu.

Limitations

COMB3 is distinct from ReaxFF; do not conflate with reactive bond-order ReaxFF parameterizations in downstream retrieval.

Relevance to group

PSU co-authors (Janik); complements metal–aqueous interface and corrosion-adjacent simulation themes.

Citations and evidence anchors