Skip to content

Ignition in an Atomistic Model of Hydrogen Oxidation

Abstract

PuReMD-GPU ReaxFF NVE trajectories seed H₂/O₂ ignition statistics across equivalence ratios; shortest ignition delay at ϕ = 0.5 with H₂O₂ precursor and anticorrelation of ignition time with energy dissipation rate above 200 MPa.

Summary

Reactive NVE molecular dynamics with ReaxFF in PuReMD-GPU studies H₂/O₂ ignition statistics. Mixtures use equivalence ratio ϕ from 0.08 to 3.31; stoichiometric boxes use 66 H₂ and 33 O₂ in an 8 nm cubic cell with 99 molecules total at other ϕ. Simulations use Δt = 0.1 fs and 3 ns duration (sufficient to convert >80% reactants at reported conditions). Initial state Tᵢ = 1500 K, ρᵢ = 250 kg/m³ with 100 independent runs per ϕ; an additional 1000 runs at 1000 K for stoichiometric gas. Chemistry is seeded with one OH radical to reduce ignition delay cost. Pressure and temperature evolve during exothermic reaction (NVE). Ignition times are extracted from kinetic energy traces; H₂O₂ mole fraction peaks precede ignition (~100 ps warning window); at high pressure (>200 MPa) ignition time anticorrelates with energy dissipation rate, implicating thermal feedback.

Methods

Force-field training / fitting. The study employs a published ReaxFF parametrization for H/O chemistry (literature references 30–33, 39 in the article); no new QM-based refit is reported in this J. Phys. Chem. A contribution.

MD application (atomistic dynamics). Engine / code: PuReMD-GPU drives ReaxFF reactive trajectories. System size & composition: H₂/O₂ boxes at multiple equivalence ratios ϕ spanning 0.08–3.31; the manuscript quotes a stoichiometric reference cell with 66 H₂ and 33 O₂ molecules in an ~8 nm cubic periodic supercell, with analogous stoichiometry adjustments at off-stoichiometric ϕ. Boundaries / periodicity: Three-dimensional periodic gas-phase supercells (as standard for these homogeneous ignition ensembles). Ensemble: NVE microcanonical reactive dynamics (temperature and pressure evolve with chemistry and energy release). Timestep: Δt = 0.1 fs. Duration: 3 ns trajectories (reported sufficient to consume >80% of reactants under the stated conditions). Replica / statistics: 100 independent replicas per ϕ at initial Tᵢ = 1500 K and ρᵢ = 250 kg/m³, plus 1000 replicas for the stoichiometric mixture at 1000 K; each trajectory seeds chemistry with one OH radical to shorten ignition delay while preserving subsequent growth chemistry. Thermostat / barostat: N/A — NVE protocol without stochastic thermostat or hydrostatic barostat control. Target temperature: initial conditions 1500 K (additional 1000 K stoichiometric batch as above); not a constant-T production thermostat. Pressure: evolves under NVE reactive heating (authors analyze high-P states exceeding ~200 MPa in the statistics). Electric field: N/A — not applied. Enhanced sampling: N/A — brute-force MD ensembles rather than umbrella sampling or metadynamics. Diagnostics: ignition times from energy traces; H₂O₂ mole-fraction transients; correlation of ignition delay with energy dissipation rate at elevated pressures.

Static QM / DFT. N/A — DFT is not the time-integration engine for the ignition statistics summarized here.

Review / non-simulation framing. N/A — primary reactive MD research article (not a narrative review).

Findings

Outcomes & mechanisms. Ignition delay is shortest near ϕ = 0.5 (equal H₂ and O₂ counts), consistent with the authors’ comparison to a cited reduced chemical kinetic model. H₂O₂ accumulates ahead of ignition, providing a ~100 ps “early warning” window in the reported traces. At P > 200 MPa, ignition time anticorrelates with the energy dissipation rate, supporting a thermal feedback interpretation in the high-pressure portion of the NVE ensembles.

Comparisons. Atomistic ignition statistics are positioned against continuum kinetic expectations and the referenced reduced model (see article figures/tables for quantitative agreement).

Sensitivity & design levers. Trends are mapped versus equivalence ratio, initial temperature (1500 K vs 1000 K stoichiometric batch), and the evolving pressure state reached under NVE exothermic compression.

Limitations & outlook (as authored). OH seeding omits some natural initiation channels; the ultrahigh-pressure regime explored is more representative of detonation-related physics than ambient flames—interpret accordingly.

Corpus / KB honesty. Claims follow the indexed PDF at pdf_path; refresh numerics after any corpus PDF swap.

Limitations

Seeded OH radical omits some initiation channels; high-pressure regime explored is extreme vs ambient combustion but informative for detonation-related physics. Repository automation maps this stable paper_id to normalized/papers/2017mohammad-alaghemandi-j-phys-chem-jp7b00249.json and the repo-relative pdf_path. Where extraction_quality is partial, the tracked PDF and DOI remain the quantitative authority over short local extracts.

Relevance to group

Demonstrates GPU ReaxFF ignition statistics workflow relevant to reactive MD benchmarking.

Citations and evidence anchors

  • DOI: 10.1021/acs.jpca.7b00249