Development of a Charge-Implicit ReaxFF for C/H/O Systems
Scope
Charge-implicit ReaxFF for C/H/O removes explicit charge equilibration, runs faster than the parent parametrization, and is validated on densities and on trehalose bombardment by water clusters (SIMS context).
Summary¶
ReaxFF achieves accurate reactive chemistry in condensed organics but charge equilibration (QEq) is costly. This work extends the earlier hydrocarbon charge-implicit ReaxFF to oxygen-containing systems: Coulomb contributions are folded into other ReaxFF terms so explicit electrostatics are omitted, targeting faster simulations without losing key thermochemistry and density behavior. The motivation includes SIMS and organic matrix problems where high-energy collisions require stable integration and reactive fragmentation without paying for full QEq every step.
Methods¶
1 — MD application (atomistic dynamics)¶
- Engine / code: LAMMPS (ReaxFF pair style) for all validation MD; VMD for visualization (J. Phys. Chem. Lett.).
- System & protocols: (i) Density and thermodynamic benchmarks for C/H/O liquids and tests in Table 1 (e.g. 1,5-pentanediol in systems up to 6.2×10⁵ atoms as tabulated) and an o-xylene oxidation stunt at 2,500 K with protocol in SI; (ii) SIMS-style (H\(_2\)O)\(_n\) cluster bombardment of a ~60 nm-scale hemispherical trehalose target; (H\(_2\)O)\(_n\) projectiles with 2, 2.85, and 5 eV per H\(_2\)O in the main-text examples (larger-n and energy sweeps in SI; C\(_{60}\)-on-ice crater figures use ~40 nm samples, 20 keV, 40° incidence in the main text for comparison to prior fields). 3D PBC (periodic boundary conditions) in the liquid/ice and bombardment cells (full cell vectors in SI as cited in JPCL).
- Ensemble & thermostating: Constant-volume NVT-class control for equilibrated liquids as implied by the density studies; high-T oxidation and impulsive bombardment trajectories are NVE-like during impact in standard SIMS-style setups—exact NVT/NVE switches and thermostat damping are in SI where the authors separate preequilibration from projectile impact.
- Timestep / barostat / pressure / field / enhanced sampling: N/A — the main JPCL pages excerpted here do not restate a single fs timestep; N/A — no NPT barostat-driven stress control in the summarized validations; N/A — no static electric field; N/A — no metadynamics / replica exchange (kinetics from conventional MD).
2 — Force-field training (charge-implicit ReaxFF-CHO)¶
- Parent / elements: Charge-implicit ReaxFF for C/H/O built from ReaxFF-2008 (Chenoweth et al.) data; Coulomb is folded into other terms so explicit QEq is omitted in production runs.
- QM reference & training set: DFT-level curves for bonds, reactions, valence, and dihedral hurdles; augmented table-based targets for liquid/vapor densities, water heat of vaporization, and amorphous ice cohesive energy (average density deviation ~5% on the fit set per Table 1; ZBL splicing for ultra-short-range repulsion per Ziegler–Biersack–Littmark in SI).
- Optimization: Successive single-parameter refinement for bonding; iterative nonbonding retuning to match Table 1 to ~10% on energies/densities as stated; ReaxFF-lg used as a density comparator in the article.
- Reference validation (throughput): Table 2 reports μs (timestep)\(^{-1}\) atom\(^{-1}\) cost—ci-ReaxFF-CHO is >2× faster in step time than the parent explicit ReaxFF in the tabulated large-cell cases (JPCL Table 2).
Findings¶
Speed / fidelity: The ci-ReaxFF-CHO parametrization achieves >2× wall-time or per-step gains vs conventional ReaxFF in the documented Table 2 systems while retaining bond/angle/torsion fidelity against DFT training data comparable to ReaxFF-2008 in mean deviation on training reactions (~17% vs ~16% in the main text for the H-abstraction test). Densities and reactions for C/H/O materials remain usable; ZBL stitching supports keV-scale impacts in C\(_{60}\)-on-ice comparators.
SIMS mechanism: In (H\(_2\)O)\(_n\)-on-trehalose runs mirroring experiment, higher molecular ion yields track ejection of trehalose–water complexes at certain per-molecule energies (Figure 4 in JPCL), matching the peaked experimental sensitivity trend for massive water projectiles (≥~order-of-magnitude gains cited in the intro and ref** [4]).
Limitations¶
Transferability to elements or chemistries outside the fitted C/H/O space requires additional extension; long-lived polarization effects may require explicit polarizable treatments in some solvents. SIMS-relevant impact conditions probe high-energy collisions that stress short-range repulsion terms; users should confirm that their projectile mass and velocity ranges remain within the training spirit of the trehalose benchmark cases highlighted in the article. Charge-implicit runs omit long-range Coulomb forces; do not use them for electrolyte systems where explicit solvent polarization dominates reaction barriers without additional validation. Trehalose benchmarks are illustrative organic matrices, not exhaustive coverage of oxygenate chemistries.
Relevance to group¶
Co-authored method paper on accelerated ReaxFF for oxygenated organics and surface bombardment.