Skip to content

Direct observation of realistic-temperature fuel combustion mechanisms in atomistic simulations

Summary

Atomistic simulation can in principle enumerate mechanisms, intermediates, and products of complex reacting flows without embedding a fixed reaction network, but conventional molecular dynamics is limited to nanoseconds, which has forced prior reactive simulations of fuel chemistry to use very high temperatures that poorly represent many practical low-temperature regimes. Bal and Neyts apply collective-variable-driven hyperdynamics (CVHD)—a self-learning scheme that grows a bias on bond-distortion collective variables—together with ReaxFF (Chenoweth et al. parameter set with QEq charges in LAMMPS). Their headline result is that CVHD + ReaxFF reaches effective physical times far beyond brute-force MD: n-dodecane pyrolysis is observed down to 1000 K (up to ~57 ms effective time in their Table 1 example), while fuel-lean oxidation of n-dodecane is pushed down to 700 K with an effective time up to ~39 s and a very large boost factor. Product compositions and dominant pathways are reported to be strongly temperature-dependent and consistent with experiments and kinetic models in the comparisons they present.

Methods

MD application (atomistic dynamics)

Bal and Neyts implement CVHD in LAMMPS with the colvars module, using ReaxFF with the Chenoweth et al. hydrocarbon parameter set and QEq charge equilibration as distributed in LAMMPS. The equations of motion use Δt = 0.1 fs. Initial thermalization uses a Langevin-type thermostat, followed by NVT sampling with a Nosé–Hoover chain (0.1 ps relaxation time quoted in the article) and, where isotropic NPT is required, Martyna–Tobias–Klein barostat dynamics integrated with the Tuckerman et al. scheme (1 ps relaxation time quoted).

System size & composition (approx. atom counts from stoichiometry): Pyrolysis: 24 n-dodecane molecules (912 atoms) in a 50 × 50 × 50 ų three-dimensional periodic box (~0.05 g cm⁻³). CVHD biases C–C and C–H bond strains with r_min/r_max windows (1.55–2.20 Å for C–C; 1.05–1.65 Å for C–H) chosen so high-barrier bond-breaking transition states remain unbiased. Gaussian hills (w = 0.25 kcal mol⁻¹, d = 0.025) are deposited every 0.2 ps; the transition detector uses a waiting time t_w = 1 ps. CVHD pyrolysis runs span 1000–1800 K; unbiased comparison MD is reported at 2500 K.

Lean combustion cells: 5 n-dodecane + 100 O₂ (390 atoms as counted from the explicit molecular formula in the article) in a 40 × 40 × 40 ų periodic box (~0.1 g cm⁻³), with oxygen interactions mapped to the same CV strain parameters as carbon (as stated in the article). CVHD oxidation runs span 700–1800 K at the reported high densities; average pressures rise from ~200 bar at 700 K to ~500 bar at 1800 K in their constant-volume oxidation setup. Additional constant-pressure CVHD sets at 1000 K (10–500 bar) use a longer Gaussian deposition stride (0.5 ps) in NPT to limit bias buildup when collision frequency drops.

Electric field: N/A — not used. Replica / parallel tempering / umbrella sampling: N/A — acceleration is CVHD, not replica exchange or umbrella sampling.

Force-field training

N/A — not a ReaxFF refit paper; the work applies an established ReaxFF description for hydrocarbon chemistry together with CVHD.

Static QM / DFT

N/A — not a DFT production study for the reported accelerated trajectories; higher-level QM enters as literature context on reactive potential accuracy.

Findings

  • Accessible timescales: Table 1 in the article summarizes the lowest temperatures reached for pyrolysis (1000 K, ~57 ms longest effective time, large boost) versus combustion (700 K, ~39 s longest effective time, ~1.3×10⁹ boost vs prior long ParRep alkane work they cite). The abstract’s “700 K” headline refers primarily to the oxidation demonstration; pyrolysis at 1000 K still constitutes a large departure from prior >2000 K brute-force reactive MD practice.
  • Pyrolysis chemistry trends: At high T, β-scission-like channels dominate (C₂-rich products, similar to prior >2000 K MD). At lower T, low-barrier isomerizations occur more often, producing more stable secondary radicals and, after eventual β-scission, heavier 1-alkene products (C₃+ fractions much larger at 1000 K than at 2500 K in their Fig. 2 narrative). Propagation by H/CH₃/C₂H₅ abstraction becomes more central at lower temperatures, while unimolecular C–C fission gains weight at high temperature.
  • Oxidation mechanism regimes: The authors contrast low-T oxidation initiated by O₂ hydrogen abstraction and downstream peroxy / hydroperoxyalkyl (•QOOH) radical chemistry with a high-T regime where chemistry begins like pyrolysis (C–C fission / β-scission, C₂H₄ formation) followed by later oxidation—plus intermediate-T mixed behavior they discuss with pathway figures.
  • Validation stance: They report qualitative consistency with experimental product trends and kinetic-model expectations for the staged comparisons they show, while emphasizing CVHD design choices (stride vs pressure, collective variables) that must be validated when porting the method to other gas-phase reactive systems.

Limitations

Second corpus PDF (2016bal-chemical-sci-direct-observation) duplicates the article with a different file hash; cite one consistent PDF for pagination. Acceleration artifacts and bias potential design choices must be validated case by case; CVHD parameters are not universal.

Relevance to group

Demonstrates accelerated reactive MD with ReaxFF for combustion-relevant chemistry—adjacent to group interests in reactive hydrocarbon systems.

Citations and evidence anchors