Skip to content

Investigation of ethanol oxidation over aluminum nanoparticle using ReaxFF molecular dynamics simulation

Summary

ReaxFF molecular dynamics is used to follow ethanol oxidation in contact with aluminum nanoparticles prepared with different oxidation states (bare metal versus thin or thick oxide shells prepared by controlled O₂ exposure in NVT LAMMPS runs). The abstract reports that nanoparticles can lower the initial ethanol reaction temperature to 324 K on pure Al, contrasts OH-abstraction vs H-abstraction pathways by oxide coverage, and quotes Eₐ ≈ 4.58 kcal/mol for adsorptive dissociation on the thin-oxide particle, benchmarked against DFT.

Methods

1 — MD application (ReaxFF, Fuel 234 (2018) 94–100, §2–3). Simulations use ReaxFF reactive molecular dynamics implemented through the REAXC package in LAMMPS. System size & composition: an Al nanoparticle ~2.8 nm in diameter (856 atoms) is prepared first, then placed in a cubic cell 50 × 50 × 50 Å with PBC in all directions. Oxide variants: 300 or 600 O₂ molecules are distributed around the nanoparticle for NVT oxidation at 298 K (velocity Verlet, timestep 0.2 fs, 2 × 10⁶ steps, 0.4 ns total) until oxygen uptake and potential energy plateau; resulting oxide shells are ~0.76 nm (APO300) and ~1.03 nm (APO600) thick, with pure Al denoted AP. Ethanol oxidation cells: 20 ethanol and 60 oxygen molecules (stoichiometric C₂H₅OH/O₂ mixture) surround each nanoparticle; initial velocities are Maxwell–Boltzmann at 298 K. Ensemble: NVT is used for the ethanol–oxygen runs described in §3. Temperature program: systems are heated from 298 K toward reaction temperatures (the C₂H₅OH/O₂ gas mixture without nanoparticles is ramped to 3000 K at 2 K/ps in §3.1; nanoparticle-bearing systems use 298 → 500 K at the same 2 K/ps rate before higher-T stages in §3.2). Thermostat / barostat / pressure: NVT runs are constant-volume thermal sampling; N/A — hydrostatic pressure is not an independent control variable in those NVT segments (no NPT barostat in the ethanol–O₂ production protocol described in §3). Duration: oxidation prep ~0.3 ns to equilibrate oxygen uptake; ethanol–surface studies extend to 2000 K in the narrative examples in §3.2. Electric field: N/A — not applied. Enhanced sampling: N/A — no umbrella, metadynamics, or replica exchange; the authors additionally report selected NVE microcanonical checks for nanoparticle systems starting at 298 K (§3.3). Electrostatics / QEq: handled within the standard ReaxFF formulation in §2.

2 — Force-field training. N/A — application paper using the Al/C/H/O ReaxFF of Hong et al. as cited in §2 (not a new parameterization).

3 — Static QM / DFT-only. DFT references are used for barrier comparisons (e.g., O–H / Cα–O cleavages on Al(111) and Al₂O₃(0001) surfaces) alongside bond-restraint ReaxFF barrier extraction in §3.2.

Analysis. Track adsorption vs dissociation, temperature-dependent rates, and gas-phase fragments; activation energy for adsorptive dissociation on APO300 is reported as 4.58 kcal/mol, compared with DFT in the abstract.

Findings

  • O₂ vs ethanol adsorption: Oxygen adsorbs more readily than ethanol on the Al-based surfaces under the simulated conditions (abstract).
  • Oxidation state steers chemistry: OH-abstraction dominates on pure Al, whereas H-abstraction is more prevalent on oxidized particles (abstract).
  • Oxides block dissociation: A surrounding oxide layer hinders adsorptive dissociation of ethanol relative to bare pathways; H from hydroxyl couples to Al vs O depending on oxide thickness (abstract).
  • Products vs uptake: With a thick oxide, volatile products such as H₂O and CO appear, whereas for thin or no oxide, much of the C/H/O from ethanol diffuses into the nanoparticle rather than desorbing as CO/H₂O (abstract).
  • Kinetics: Dissociation rates increase with temperature; reported Eₐ ≈ 4.58 kcal/mol for the thin-oxide adsorptive dissociation channel matches DFT within the authors’ comparison (abstract).

Limitations

ReaxFF remains empirical; barriers and product channels should be spot-checked with QM when extrapolating outside the trained Al/C/H/O chemistry window. Industrial ignition involves ms–s timescales and continuum transport not captured in these nanosecond cells. Coauthorship note: Adri C. T. van Duin is a coauthor on the Fuel article; treat quantitative reuse like any external ReaxFF application and cite the version-of-record PDF for any reproduction work.

Relevance to group

Combustion/nanofuel atomistic study in the corpus; useful for linking Al nanoparticle oxidation state to ethanol decomposition channels, separate from the Penn State ReaxFF parameterization lineage unless a coauthorship update is recorded in the version-of-record PDF.

Citations and evidence anchors

  • DOI: 10.1016/j.fuel.2018.06.119
  • Full Methods in the article lists integrator settings, temperatures, and nanoparticle construction beyond the abstract-level kinetics quoted on this page.