Dimerization of Polycyclic Aromatic Hydrocarbon Molecules and Radicals under Flame Conditions
Summary¶
ReaxFF molecular dynamics and supporting quantum chemistry quantify collision efficiencies and dimerization pathways for polycyclic aromatic hydrocarbon (PAH) molecules and radicals relevant to soot nucleation, including oxygenated radicals. The study is positioned as a way to supply microscopic sticking and recombination inputs where full flame chemistry is too expensive to resolve exhaustively at the QM level.
The paper cross-checks ReaxFF barriers for hydrogen abstraction from coronene against B3LYP and M06-2X data, couples ReaxFF MD with 0-D Chemkin kinetics for H/OH pools, and performs large batches of binary-collision ReaxFF MD in LAMMPS (reax/c) for pyrene and coronene species. Gas-phase QC uses Gaussian 09 at B3LYP/6-311G(d,p) geometries with higher single-point energies (M06-2X/B3LYP). MD equilibrates PAHs in NVT (1500–2500 K, Nosé–Hoover, 100 fs damping) before adiabatic NVE collisions at 0.1 fs with 120,000 sampled binary events.
Methods¶
1 — MD application (binary collisions + staged heating). ReaxFF (C/H/O combustion parameterization) is integrated in LAMMPS (fix reax/c) for large batches of binary-collision trajectories: NVT preequilibration of PAH targets at 1500–2500 K with Nosé–Hoover thermostat (100 fs damping quoted in this wiki’s prior curation traceable to the article), followed by adiabatic NVE collisions at 0.1 fs with 120 000 sampled events and 50–80 Å initial separations. System composition: each event pairs pyrene/coronene-family molecules/radicals with ~10²–10³ atoms in the combined periodic cell depending on the selected radical site count and buffer gas content (see J. Phys. Chem. A tables). PBC: large gas-phase cells with three-dimensional PBC as in standard collision studies. Barostat: N/A — NVE/NVT stages without NPT hydrostatic control. Pressure: N/A — not a high-pressure study. Electric fields / enhanced sampling: N/A — statistics come from brute-force collision sampling rather than umbrella/metadynamics.
2 — Static QM. Gaussian 09: B3LYP/6-311G(d,p) optimizations/TSs with M06-2X/6-311G(d,p) single points for barrier benchmarking.
3 — Kinetics coupling. Chemkin-Pro 0-D reactor for H/OH pools (19 species, 27 reactions) parameterized from QM.
4 — Force-field training. N/A — applies an established ReaxFF combustion parameterization with QM spot checks rather than refitting here.
Findings¶
Outcomes / mechanisms: ReaxFF reproduces H-abstraction energetics for coronene between B3LYP and M06-2X trends and tracks 0-D kinetics qualitatively, while some radical–radical recombination efficiencies differ, affecting dimer concentrations. Many radical–radical encounters form weakly bound complexes rather than immediate covalent dimers; oxygenated PAH radicals show reduced propensity toward soot nucleation channels in the modeled scenarios.
Comparisons: explicit QM (B3LYP/M06-2X) and 0-D kinetics comparisons anchor the ReaxFF collision database.
Sensitivity: collision temperature, PAH size, and radical site count modulate sticking vs transient outcomes.
Limitations: high-T gas-phase collisions are not ambient condensed phases; ReaxFF remains approximate for conjugated radicals.
Corpus honesty: protocol numbers follow the curated summary traceable to papers/ReaxFF_others/QianMao_JPCA_2018_PAH_dimer.pdf; confirm any updated values in the VOR.
Limitations¶
High-temperature collision settings target flame chemistry, not ambient condensed phases; ReaxFF remains approximate versus high-level QM for conjugated radicals. Extending the collision database to heavier PAHs or heteroatom-rich radicals would increase statistical cost and parameterization scrutiny without changing the paper’s core conclusion that binary-collision statistics can discriminate sticky versus transient encounters.
Relevance to group¶
Demonstrates ReaxFF + LAMMPS workflows for PAH growth relevant to combustion soot modeling.