Investigation of methane oxidation by palladium-based catalyst via ReaxFF Molecular Dynamics simulation
Evidence and attribution¶
Authority of statements
Prose below (Summary, Methods, Findings, etc.) summarizes the publication identified by doi, title, and pdf_path in the front matter. For exact numerical values and figures, use the peer-reviewed article at DOI 10.1016/j.proci.2016.08.037.
Summary¶
Methane oxidation over palladium nanoparticles is studied with ReaxFF molecular dynamics (Pd/C/H/O force field in LAMMPS), comparing bare metal with oxygen-coated surfaces prepared at controlled coverage. Simulations follow O\(_2\) and CH\(_4\) adsorption and dissociation, CH\(_4\)/O\(_2\) mixture reactivity under temperature-ramp NVT conditions, and fixed-temperature runs from 400–1000 K used to extract apparent activation energies for CH\(_4\) dissociation at ~0.3 and ~0.7 ML oxygen coverage. The authors relate atomic-level trajectories to oxygen blocking of methane adsorption, coverage-dependent methane activation, and reported consistency of derived barriers with DFT and experiment.
Methods¶
A — Force-field training / fitting: Pd/C/H/O ReaxFF in LAMMPS from established combustion/catalysis parametrizations (article cites lineage); no new QM refit summarized in this proceedings paper.
B — Molecular dynamics / sampling: NVT reactive MD in a cubic ~80 Å cell with 3D periodic boundary conditions (standard nanoreactor-style setup in the article). CG energy minimization, Δt = 0.25 fs, Nosé–Hoover thermostat (100 fs damping), 10⁵ equilibration steps; ~3 nm Pd\(_{935}\) nanoparticle oxidized at 1500 K with O\(_2\) to ~0.3 and ~0.7 ML oxygen precoverages; CH\(_4\)/O\(_2\) mixture at ~4.85 atm (e.g. 40 O\(_2\) + 20 CH\(_4\) molecules in the cited cell). Temperature ramp 300 → ~3000 K at ~1.08 K ps⁻¹ (10⁷ MD steps) plus fixed-temperature windows (400–1000 K) for CH\(_4\) dissociation kinetics. VMD visualization.
C — DFT / static QM: Literature DFT/experiment cited for barrier comparison—not ab initio MD here.
D — Review / non-simulation framing: Proc. Combust. Inst. application—not a review.
Pressure / barostat: N/A — NPT not used for the quoted NVT nanoreactor protocol; no stated hydrostatic pressure control beyond implicit constant-volume gas-phase cell.
Findings¶
Adsorption and site blocking: O\(_2\) adsorbs more readily than CH\(_4\) on both bare and oxygen-coated Pd; preadsorbed O\(_2\) on oxidized surfaces blocks sites available for CH\(_4\) adsorption.
Dissociation and coverage: At low temperature, adsorptive dissociation of CH\(_4\) is easier on oxygen-coated nanoparticles than on bare Pd. O\(_2\) dissociation requires higher temperature than O\(_2\) adsorption, unlike the relatively rapid CH\(_4\) dissociation after adsorption under the conditions explored. CH\(_4\) dissociation rates increase with temperature and depend on surface oxygen coverage.
Activation energies: From fixed-temperature simulations (400–1000 K), apparent activation energies for CH\(_4\) adsorptive dissociation are reported as 3.27 and 2.28 kcal mol\(^{-1}\) on 0.3 and 0.7 ML oxygen-coated particles, respectively; the authors state these are consistent with DFT and experiments.
Limitations¶
- ReaxFF remains an empirical reactive model: barriers and pathways are approximate relative to QM and may depend on parameterization and training-set scope.
- Nanoparticle size (~3 nm), single morphology, and idealized ML coverage may not capture all experimental catalyst dispersion and support effects.
- Reactive MD uses high gas-phase pressures and temperature ramps (or limited duration at fixed \(T\)) to observe chemistry within accessible timescales—not a direct replica of long isothermal reactor conditions.
Relevance to group¶
Adri C. T. van Duin (co-author, Penn State Mechanical and Nuclear Engineering) contributes ReaxFF methodology context; the work applies the Pd/C/H/O reactive framework to heterogeneous methane oxidation relevant to combustion and catalysis modeling.