Skip to content

Evaluating the ability of selected force fields to simulate hydrocarbons as a function of temperature and pressure using molecular dynamics

Summary

OPLS-AA, AIREBO-M, and ReaxFF (CHON-2017_weak) are benchmarked for pure liquid hydrocarbons—trans-decalin, n-hexadecane, isocetane, and tetralin—across 293.15–373.15 K at 0.1 MPa and pressures 0.1–40 MPa at 313.15 K. The paper’s aim is to quantify when a reactive hydrocarbon FF can double as an accurate liquid-state EOS model for fuels work. OPLS-AA best reproduces densities; ReaxFF is slightly better for isentropic bulk modulus; AIREBO-M is least accurate for both. Structural diagnostics (RDFs, ARDFs, radius of gyration for n-hexadecane, ring puckering for tetralin) highlight conformational biases, especially tetralin π-stacking propensity in AIREBO-M/ReaxFF vs OPLS-AA.

Methods

MD application (classical liquid benchmarks in LAMMPS). LAMMPS is used to simulate neat liquid hydrocarbons — trans-decalin, n-hexadecane, isocetane, and tetralin — in NPT with 3D PBC for the bulk liquids. The study sweeps temperature 293.15–373.15 K at 0.1 MPa and, separately, pressure 0.1–40 MPa at 313.15 K (as stated in the abstract/Introduction). Timestep: 0.25 fs; AIREBO-M and ReaxFF use unconstrained (non-SHAKEd) C–H dynamics with parameters documented in the paper/SI, while OPLS-AA follow SI Tables S1–S4. Ensemble: NPT; each NPT segment couples a Nosé–Hoover-style or Berendsen-style thermostat with an isotropic Parrinello–Rahman/barostat (or equivalent) for hydrostatic pressure control as in the Energy & Fuels methods. Equilibration + production staging — follow article/SI; N/A here to paste every run length. System size — on the order of a few thousand atoms (SI); exact counts in pdf_path. Electric field, shear, impact, replica/enhanced sampling: N/A. Electrostatics — as defined by each potential: OPLS long-range Coulomb; AIREBO/ReaxFF per their respective LAMMPS pair styles.

Potentials compared. (1) OPLS-AA/AMBER-family parameters for the listed fluids; (2) AIREBO-M via the ch.airebo-m file packaged with LAMMPS; (3) ReaxFF with CHON-2017_weak parameters from the cited 2020 Energy & Fuels work (ref. 20, SI in that paper; parameters reproduced in this work’s SI).

Thermodynamic observables — mass density, isothermal bulk modulus from fluctuation relations (eqs. 2–4) and the isentropic modulus from eq. (1); RDF/ARDF center-of-mass, radius of gyration for n-hexadecane, and puckering metrics for tetralin.

Force-field training and standalone DFT. N/A — the paper is benchmark/validation of published parameter sets, not a new ReaxFF fit. N/A — no ab initio production DFT in the benchmark table unless the article cites a separate subset (not the focus of this work).

Findings

Comparisons vs experiment. For mass density, OPLS-AA is overall most accurate in RMSE/MAE vs experiment across the conditions tested; ReaxFF shows systematic overdensity in some aromatic and branched cases; AIREBO-M is the least accurate in density. For isentropic bulk modulus, ReaxFF marginally outperforms OPLS-AA on the chosen metrics; AIREBO-M is worst. N/A in this short summary to reproduce every table entry — use the article.

Structure. RDF/ARDF and R\(_g\) diagnostics show conformational biases: e.g. tetralin shows stronger π–stacking-like order in AIREBO-M/ReaxFF than in OPLS-AA; n-hexadecane can appear more coiled in ReaxFF compared with OPLS-AA/AIREBO extended-chain sampling. The authors stress that good thermodynamic averages can hide defects in local liquid structure—check RDFs and R\(_g\) alongside ρ and B.

Practical levers (as explored). Sensitivity to T and P in the NPT grids above; N/A — the paper does not claim reactive pyrolysis in these nonreactive liquid benchmarks.

Limitations

Reactive potentials are not exhaustively validated for pyrolysis/combustion here—the focus is non-reactive liquid thermophysics. Cross-comparisons use RMS/MAE metrics reported in the article against experimental densities and bulk moduli; readers extending the benchmark to additional hydrocarbons should repeat the same NPT equilibration and fluctuation formulas so B\(_T\) and B\(_S\) remain comparable across potentials.

Relevance to group

Head-to-head ReaxFF vs OPLS benchmarking for hydrocarbon fuels properties relevant to classical/reactive MD method choice.

Citations and evidence anchors

  • Energy Fuels 35, 5123–5139 (2021); Section 2 (methods) and Section 3 (results).

Reader notes (navigation)