Skip to content

Atomistic adsorption of oxygen and hydrogen on platinum catalysts by hybrid grand canonical Monte Carlo/reactive molecular dynamics

Summary

The study combines hybrid grand canonical Monte Carlo with reactive MD (GCMC/RMD) and a Pt/O/H ReaxFF to predict coverage-dependent adsorption of oxygen and hydrogen on Pt facets, reconstructed Pt(110), and Pt nanoparticles across pressure–temperature conditions relevant to operando catalysis. The study reports isotherm-style behavior, subsurface/bulk penetration at aggressive chemical potentials, and spot validation of O binding on Pt(321) outside the primary training set. Industrial coauthorship (ExxonMobil) signals application targeting reactor-relevant adsorbate structure preparation for follow-on reaction MD.

Methods

MD application (hybrid GCMC/RMD + ReaxFF): Grand canonical Monte Carlo moves insert or delete O and H under imposed gas-phase chemical potentials while ReaxFF forces relax Pt(111), unreconstructed and reconstructed Pt(110), and several Pt nanoparticle morphologies. Pressure–composition style sweeps span roughly \(10^{-20}\)–\(10\) atm in the abstract’s statement. A Pt(321) spot test probes O binding on a kinked facet outside the primary training set. Boundaries / PBC: 3D periodic supercells for extended Pt facets and nanoparticle models as in standard LAMMPS-style setups (JPCC computational section—N/A — cell vectors not on the short extract). Temperature setpoints for GCMC/RMD segments: N/A — not on the short corpus extract; the abstract emphasizes gas-phase potential sweeps rather than tabulating K-resolved schedules here. Total trajectory time per chemical-potential point / production-run duration (ps or ns): N/A — not on the short corpus extract; use papers/Gai_PtHO_JPC_2016_proof.pdf and SI for equilibration vs production staging. Engine, timestep, thermostat/barostat: N/A — not on the short corpus extract; use papers/Gai_PtHO_JPC_2016_proof.pdf and SI.

Force-field training: The article reports a Pt/O/H ReaxFF and ties GCMC/RMD behavior to training-set coverage (QM program, functional/basis, weighting, and optimizer details are in the JPCC computational section—N/A — not transcribed on this page).

Static QM / DFT: Validation against DFT and literature experiment as cited in the paper; functional/basis/k-mesh: N/A — see article rather than this summary.

Findings

  • Adsorption isotherms map how O and H populate surface, subsurface, and bulk regions as a function of imposed gas-phase potential.
  • Pt(321) tests indicate transferable qualitative performance for step/kink sites not explicitly in the training data (as claimed in the abstract narrative).
  • Results are cross-compared to DFT and experiment where available in the article body.
  • The abstract-level message is that hybrid GCMC/RMD can prepare realistic adsorbate ensembles on complex Pt morphologies before follow-on reaction MD that would be too costly if starting from bare surfaces.

O and H fill surface, then subsurface, then bulk-like regions as the imposed potential becomes more oxidizing, so uptake competes with subsurface transport. Where DFT and experiment are cited, the authors argue the Pt/O/H parametrization is adequate for these coverage models. Sensitivity to the 10⁻²⁰–10 atm pressure window (abstract) moves the dominant adsorption branch. Gas-phase grand-potential control is not a full electrochemical double layer; quantitative isotherms and final pagination belong in the journal PDF, not this note.

Limitations

  • GCMC/RMD still inherits ReaxFF uncertainties for oxidized, hydroxylated, and reconstructed Pt under electrochemical potentials not identical to the gas-phase grand potential used here.
  • Proof PDF path may differ cosmetically from the final issue layout.
  • Electrochemical interfaces with explicit solvent and applied bias may require extensions beyond the gas-phase grand-potential protocol emphasized in the abstract-level summary.
  • Coverage isotherms should be read with the same pressure scales and reference states used in the article to avoid misinterpreting chemical potential units.

Relevance to group

Showcases time-accelerated sampling + ReaxFF for Pt catalysis, a recurring theme in collaborations between Penn State and industry partners.

Citations and evidence anchors

  • Abstract and metadata in papers/Gai_PtHO_JPC_2016_proof.pdf; DOI: 10.1021/acs.jpcc.6b01064.