Development of a ReaxFF reactive force field for ammonium nitrate and application to shock compression and thermal decomposition
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¶
ReaxFF for ammonium nitrate (AN) is built by extending a nitramine/TATB parameterization with QM-targeted fits to dissociation barriers, heats of formation, and crystal properties of AN phases. The field predicts the isothermal P–V response of phase IV AN within about 10% of DFT and experiment over the compression range considered, while the unreacted principal Hugoniot comes out stiffer than experiment by roughly ~17%. Thermal decomposition simulations on heating to 2500 K reproduce pathway trends consistent with experiment. Adri C. T. van Duin coauthors with Sandia (Shan, Thompson).
Methods¶
Local sources: the PDF at papers/Shan_Thompson_AN_proof_JPCA_2014.pdf is present in this workspace; the abstract and introduction are captured in normalized/extracts/2014shan-venue-research_p1-2.txt and duplicated in normalized/extracts/2014shan-venue-jp408397n_p1-2.txt (same DOI text).
Force field development: the authors extend the nitramine/TATB ReaxFF and reoptimize parameters against electronic-structure targets for NH\(_4\)NO\(_3\)—including barriers for hydrogen transfer between NH\(_4^+\) and NO\(_3^-\), heats of formation, and crystal data for AN phases—using NWChem for key quantum calculations and VASP PBE DFT with PAW potentials (550 eV cutoff, 6×6×6 k-mesh) for bulk phase IV P–V references (Section 2 of the article). ReaxFF parameter fitting uses the standalone ReaxFF code; production MD is run in LAMMPS. Existing C/H/O/N ReaxFF sets are first shown inadequate for AN (e.g., spontaneous NH\(_4^+\)+NO\(_3^-\)→NH\(_3\)+HNO\(_3\) in NPT tests), motivating the new fit.
Applications. Isothermal compression and unreacted Hugoniot states for phase IV use isotropic cell compression in MD (the article flags MSST/uniaxial shock as future work where relevant). Thermal decomposition uses 4×4×4 phase IV supercells heated from 0 K to 500–2500 K over 40 ps at fixed volume, then held at the target temperature for 1000 ps, with species tracked from bond orders (sampled every 1 fs and time-averaged over 100 fs windows). Additional PBC vectors, thermostat coupling for ramp versus hold, and any NPT segments used for EOS/Hugoniot benchmarks are only fully tabulated in papers/Shan_Thompson_AN_proof_JPCA_2014.pdf (N/A for complete damping/staging transcription on this page).
Findings¶
The new ReaxFF reproduces lattice parameters for AN phases I, IV, and V within about ±3% vs. DFT. The isothermal P–V curve for phase IV agrees with DFT and experiment within about 10% over the range reported. Unreacted US–uP and P–V Hugoniot points track experimental scatter below U\(_P\)\(\approx 1\,\mathrm{km/s}\) and P\(\approx 15\,\mathrm{GPa}\) but the model is too stiff at higher compression (about 17% vs. experiment on the principal Hugoniot in the abstract’s summary; the discussion ties part of the error to isotropic compression vs. experiment and density offsets). Thermal decomposition at 2500 K shows NH\(_4\)NO\(_3\) dissociation to NH\(_3\) and HNO\(_3\) followed by HNO\(_3\) chemistry, with major products after 1 ns including H\(_2\)O, N\(_2\), NO\(_2\), and OH, qualitatively aligned with reported high-temperature routes—while onset temperatures are higher than experiment, attributed to barriers, small cells, fast heating, and defect-free crystals.
Limitations¶
- Proof-class PDF in corpus—confirm pagination and final edited text against the J. Phys. Chem. A issue for the DOI above; the version-of-record reading copy is
[[2014shan-venue-jp408397n]](papers/Shan_Thompson_AN_JPCA_2014.pdf). - Reactive FF uncertainty for detonation-grade multiphysics remains; compare to higher-fidelity kinetics when available.
Relevance to group¶
Core ReaxFF development for energetic nitrate systems tied to safety and shock physics—direct van Duin collaboration with national-lab partners.
Citations and evidence anchors¶
- DOI: https://doi.org/10.1021/jp408397n