Using molecular dynamics simulations with a ReaxFF reactive force field to develop a kinetic mechanism for ammonia borane oxidation
Evidence and attribution¶
Authority of statements
Prose below summarizes the publication identified by doi, title, and pdf_path.
Summary¶
Ammonia borane (NH₃BH₃) combines high hydrogen content with propellant-relevant chemistry; this work uses ReaxFF MD of AB oxidation to build an elementary kinetic mechanism for continuum models without prior experimental rate data for every step. The abstract’s pathway picture is sequential H₂ loss from gas-phase AB, reaction of H₂ with O₂, oxygen attack on boron-bearing fragments, B–N cleavage, and equilibrium products including H₂O, HOBO, and N₂. DFT supplies missing thermochemistry, collision theory gives first rate estimates, and CHEMKIN closed-reactor constant-pressure, constant-energy runs are reported consistent with the atomistic trajectories.
Methods¶
ReaxFF for NH₃BH₃ oxidation uses the prior parameterization (ref. [11] in the article). The introduction contrasts nonreactive FFs (e.g., MM3, general DREIDING-class models) that cannot break bonds with ReaxFF’s bond-order updates each timestep as interatomic distances evolve—enabling large reacting cells that remain intractable for fully quantum dynamics. Reactive MD identifies elementary sequences: two H₂ eliminations from AB, O₂ attack on boron-bearing fragments, B–N cleavage, and H₂O / HOBO / N₂ products. DFT supplies missing thermochemical data; collision theory gives first rate estimates; Jaguar B3LYP/6-311G locates the transition state for the second H₂ loss from H₂NBH₂ (Eq. 9). A reduced mechanism is implemented in CHEMKIN for a closed reactor at fixed pressure/energy. Illustrative NVT heating ramps use 20 AB + 45 O₂ in a 2.5 nm periodic cube at 0.00522 K/fs (Figure 3).
MD application (ReaxFF oxidation trajectories + illustrative ramps)¶
Engine / code: ReaxFF molecular dynamics; N/A — specific MD engine string not recovered from normalized/extracts/2012weismiller-proceedings-using-molecular_p1-2.txt—verify pdf_path.
System size & composition: Illustrative gas-phase cells such as 20 NH\(_3\)BH\(_3\) + 45 O\(_2\) in a 2.5 nm periodic cube (Figure 3 protocol on this page); larger reacting cells are discussed qualitatively in the introduction.
Boundaries / periodicity: Three-dimensional periodic cube for the Figure 3 ramp example.
Ensemble: NVT for the documented heating-ramp illustration; broader MD stages in the mechanism survey follow pdf_path.
Timestep: N/A — Δt not recovered from the indexed excerpt; verify pdf_path.
Duration / stages: Multi-stage temperature ramps at 0.00522 K/fs for the Figure 3 illustration; overall MD segment lengths for the full mechanism survey are in pdf_path.
Thermostat: N/A — thermostat details for all MD segments are not excerpted on pages 1–2; verify pdf_path.
Barostat / pressure control: N/A — NPT not stated for the quoted NVT ramp demo.
Temperature: NVT heating ramp control as in Figure 3; absolute temperature ranges for all production segments are tabulated in the article.
Pressure / stress: CHEMKIN closed-reactor demos use constant pressure/energy controls as stated in the abstract; N/A — this is continuum-level rather than atomistic stress control.
Electric field: N/A — not used.
Replica / enhanced sampling: N/A — not used.
Force-field training (prior ReaxFF for NH\(_3\)BH\(_3\) oxidation)¶
Parent FF / elements: Uses the existing NH\(_3\)BH\(_3\) oxidation ReaxFF parametrization cited as ref. [11] in the article.
QM reference: DFT (Jaguar B3LYP/6-311G) for selected thermochemistry and a TS search for the second H\(_2\) loss from H\(_2\)NBH\(_2\) (Eq. 9).
Training set / reference data: N/A — full training-set listing is not excerpted here; it lives in the prior parametrization reference.
Optimization: N/A — new ReaxFF optimization is not reported in this proceedings article.
Reference data used: DFT thermochemistry, collision-theory rate estimates, ReaxFF trajectories, and CHEMKIN reactor consistency checks as described in the abstract.
Findings¶
Outcomes / mechanisms: ReaxFF trajectories support a sequential picture: two H\(_2\) eliminations from AB, O\(_2\) attack on B-bearing fragments, B–N cleavage, and equilibrium products including H\(_2\)O, HOBO, and N\(_2\).
Comparisons: CHEMKIN closed-reactor constant-pressure, constant-energy runs are reported consistent with the atomistic MD observations; DFT supplies missing thermochemistry relative to experimentally sparse elementary rates.
Sensitivity / design levers: Temperature ramps (K/fs protocol in Figure 3) modulate which intermediates dominate along the oxidation sequence.
Limitations / outlook: ReaxFF combustion chemistry accuracy and the scope of the reduced CHEMKIN mechanism remain limitations acknowledged implicitly by the multi-method workflow.
Corpus / KB honesty: Grounded in pdf_path and normalized/extracts/2012weismiller-proceedings-using-molecular_p1-2.txt (front-matter heavy); Proc. Combust. Inst. 34 pages hold complete tables.
The introduction additionally contrasts AB with pyrophoric boranes as a solid, easily stored hydrogen carrier for propulsion studies, motivating the combined DFT + MD + CHEMKIN workflow when elementary experimental rates are sparse.
Limitations¶
ReaxFF combustion chemistry accuracy; reduced mechanism scope; Proceedings page year 2013 vs submission 2012.
Relevance to group¶
Adri C. T. van Duin coauthored; ReaxFF/combustion kinetics for boron-nitrogen fuels.
Citations and evidence anchors¶
- DOI 10.1016/j.proci.2012.06.030 — Proc. Combust. Inst. 34, 3489–3497 (2013).
- Extract:
normalized/extracts/2012weismiller-proceedings-using-molecular_p1-2.txt.