Skip to content

ReaxFF-lg: Correction of the ReaxFF reactive force field for London dispersion, with applications to the equations of state for energetic materials

Evidence and attribution

Authority of statements

Prose sections below (Summary, Methods, Findings, etc.) are curated summaries of the publication identified by doi, title, and pdf_path in the front matter above. They are not new primary claims by this wiki.

For definitive numerical values, reaction schemes, and interpretations, use the peer-reviewed article (and optional records under normalized/papers/ when present)—not this page alone.

Summary

The paper introduces ReaxFF-lg, adding a low-gradient (lg) long-range dispersion correction on top of standard ReaxFF so that molecular crystal densities are not systematically too low—addressing the known limitation that practical DFT training data for solids underbinds London dispersion, which propagates into ReaxFF’s Morse-like vdW terms when those terms were not trained for long-range attraction. Parameters are fit using low-temperature crystal structures for benchmarks including graphite, polyethylene, CO₂(s), N₂(s), and energetic crystals (RDX, PETN, TATB, NM). The abstract reports the average volume error dropping from 18.5% to 4.2% for those systems, and highlights improved α–γ RDX transition pressure/volume compared to experiment.

Methods

1 — MD application (atomistic dynamics)

The article’s primary deliverable is a dispersion-corrected reactive force field and crystal equation-of-state validation rather than a standalone production-trajectory study on pp. 1–2 of normalized/extracts/2011reaxff-venue-acs-jx_p1-2.txt.

  • Engine / code: Reactive molecular dynamics (ReaxFF-RD) is discussed as a major motivation/use-case for accurate crystal densities (Introduction); N/A — a specific MD integrator/package for the EOS validation runs is not named on the indexed excerpt pages.
  • System size & composition: Benchmark molecular crystals and energetic crystals named in the abstract: graphite, polyethylene, CO₂(s), N₂(s), RDX, PETN, TATB, NM (indexed excerpt).
  • Boundaries / periodicity: Molecular crystals imply three-dimensional periodic models for EOS work; N/A — explicit supercell sizes are not stated on the indexed excerpt pages.
  • Ensemble / timestep / duration / thermostat / barostat: N/A — NVT/NPT/NVE schedules, timestep sizes, trajectory segment lengths, and thermostat/barostat algorithms are not stated on the indexed excerpt pages for the EOS validation protocol (the excerpt focuses on energy definitions and fitting inputs).
  • Temperature: Room temperature comparisons are stated for EOS vs experiment in Sec. 2 opening (extract).
  • Pressure / stress: Pressure–volume / phase transition language appears for RDX (α–γ) in the abstract; N/A — full stress-control protocol details are not on the indexed excerpt pages.
  • Electric field: N/A — not indicated in the indexed excerpt.
  • Replica / enhanced sampling: N/A — not indicated in the indexed excerpt.

2 — Force-field training

  • Parent FF / elements: ReaxFF with the standard valence + vdW + Coulomb partition in Eq. (2.2) of the article (extract).
  • QM reference / training philosophy: The introduction states prior ReaxFF development used consistent QM training, commonly B3LYP-flavor DFT with 6-31G** for molecular training data, and notes systematic London dispersion underbinding for molecular solids at practical DFT levels (Introduction, extract).
  • Training set / targets: Low-temperature crystal structures for graphite (P63mc), polyethylene (Pnam), CO₂ (Pa3), N₂ (Pa3) are used to determine lg parameters for “ordinary organic” motifs; densities and heats of sublimation enter the training discussion; parameters are then extended/refined for RDX, PETN, TATB, NM (Sec. 2, extract).
  • Functional form / optimization variables: ReaxFF-lg defines \(E_{\text{Reax-lg}}=E_{\text{Reax}}+E_{\text{lg}}\) with \(E_{\text{lg}}\) as a pairwise long-range correction built from \(r_{ij}^{-6}\)-like terms using \(R_{e,ij}\) and \(C_{\text{lg},ij}\) (Eqs. (2.1)–(2.3), extract). Geometric combination rules apply unless off-diagonals are explicitly listed; \(d=1.0\) is used; \(R_e\) is taken from UFF vdW radii and only \(C_{\text{lg}}\) is fitted (extract).
  • Reference data used for validation: After fitting, the authors report equations of state compared to experiment at room temperature (Sec. 2 opening + abstract, extract).

Findings

Outcomes and mechanisms: ReaxFF-lg reduces the average equilibrium volume error for the listed benchmark systems from 18.5% to 4.2% (abstract, extract). The lg correction is designed to add long-range London dispersion while leaving short-range valence interactions largely intact (Introduction + Eq. (2.3) discussion, extract).

Comparisons: The abstract highlights crystal structures and equations of state in good agreement with experiment after correction, and gives a concrete RDX α–γ transition benchmark (~4.8 GPa and ~2.18 g/cm³ from ReaxFF-lg vs ~3.9 GPa and ~2.21 g/cm³ experimentally in the abstract wording, extract).

Sensitivity and design levers: The correction is parameterized from low-temperature reference crystals and then applied to energetic chemistries; the α–γ RDX case shows sensitivity of transition pressure/density to the dispersion treatment (abstract, extract).

Limitations and outlook (authored tone): The introduction frames the underlying issue as DFT-for-solids dispersion errors propagating into ReaxFF training when vdW terms were not trained for long-range attraction; ReaxFF-lg is proposed as a pragmatic extension (Introduction, extract).

Corpus / KB honesty: Numeric transition values above follow the abstract on the indexed excerpt pages; full tables, weighting schemes, and any additional validation cases require pdf_path beyond pp. 1–2.

Limitations

  • Adds complexity and parameters; users must ensure ReaxFF-lg is consistently applied when comparing to older ReaxFF datasets.
  • Coverage is tied to the training chemistry set; extension to new elements/environments needs explicit validation.

Relevance to group

Core method lineage paper for dispersion-corrected ReaxFF used in condensed-phase organics and energetic materials simulations.

Citations and evidence anchors

  • Abstract and Sec. 2: ReaxFF-lg definition, motivation, benchmark systems (J. Phys. Chem. A 2011, 115, 11016–11022; PDF pp. 1–2 per extract).