ReaxFF based molecular dynamics simulations of ignition front propagation in hydrocarbon/oxygen mixtures under high temperature and pressure conditions
Summary¶
Continuum combustion models rely on empirical correlations for ignition delay, flame speed, and pollutant formation, yet atomistic simulations with explicit bond rearrangements can stress-test those closures when experimental data at extreme pressures are sparse. Ashraf et al. apply ReaxFF molecular dynamics to estimate ignition-front propagation speeds in supercritical hydrocarbon/oxygen mixtures at 55 MPa with an unburned temperature of 1800 K—thermodynamic conditions chosen in the abstract to accelerate oxidative chemistry so that fronts develop within tractable MD windows. Representative alkyne, alkane, and aromatic fuels are compared under matched states, and the extracted MD front speeds are juxtaposed with continuum reactive-flow solutions computed with consistent thermodynamic assumptions.
Methods¶
Reactive MD (LAMMPS, §3.1). ReaxFF CHO-2008 (papers/Ashraf_PCCP_2016_flame_propagation_online.pdf) drives ignition-front studies in a 12 × 12 × 2000 Å 3D PBC cell containing 240 n-C₄H₁₀ and 1560 O₂ molecules (ϕ = 1 baseline) elongated along z—order 10⁴ atoms once hydrogen counts are included. The unburned mixture starts at T_u = 1800 K and 55 MPa to accelerate chemistry. Workflows begin with conjugate-gradient minimization, 25 ps of NVT equilibration at 1800 K using Δt = 0.25 fs with C–O/H–O reactive terms disabled, Berendsen thermostat (100 fs damping), then NVE production at 1800 K with Δt = 0.1 fs once reactions activate. O–O bonds at the z ends are stretched to seed O radicals while keeping them ≥ 2.5 Å from neighbors. The quoted protocol is constant-volume (NVT then NVE; no NPT); electric fields and replica / enhanced sampling are not used.
Force-field training: N/A — applies CHO-2008 without new fitting in this article.
Continuum benchmark (§2.2). NGA integrates Navier–Stokes plus species/energy transport with CaltechMech augmented by JetSurf 2.0 n-butane chemistry, invoking an ideal-gas EOS after a compressibility-factor argument at the stated reduced P, T.
Findings¶
ReaxFF trajectories reproduce the qualitative reactivity ordering among alkyne / alkane / aromatic fuels studied, matching broad ignition propensity trends seen experimentally in the combustion literature referenced by the authors. Ignition-front speed depends strongly on equivalence ratio, analogous to laminar flame speed sensitivities but evaluated at 55 MPa/1800 K. Compared to Navier–Stokes solutions with matched thermochemistry, the continuum solver tracks the MD trends and lands near the tabulated comparison margins, supporting ReaxFF as a qualitative mechanistic probe while acknowledging uncertainty in quantitative rates. Sensitivity to ϕ mirrors classical flame-speed behavior but at supercritical conditions chosen for computational tractability. Limitations: the protocol is a computational convenience, not a direct map to atmospheric flames; future work should treat agreement as diagnostic rather than as universal calibration for engineering burners.
Limitations¶
The supercritical protocol is a computational convenience and does not map directly to atmospheric flames; extrapolation requires separate validation. ReaxFF omits electronically excited states and non-Born–Oppenheimer channels that may matter for minor species and emissions.
Continuum comparisons inherit uncertainties in transport models and numerical stiffness at 55 MPa that are independent of the force field; readers should treat agreement as diagnostic rather than as a universal flame-speed calibration for engineering burners.
Ignition-front diagnostics in the article tie spatially resolved heat-release or species-gradient measures to an effective front speed; reproduce those definitions carefully and consistently before comparing to external flame-speed databases.
Relevance to group¶
Penn State van Duin-group ReaxFF combustion application paper extending flame/ignition front concepts.
Citations and evidence anchors¶
- DOI:
10.1039/C6CP08164A.