ReaxFF family
One-paragraph summary¶
ReaxFF is the bond-order-based reactive force field used throughout much of this corpus for systems where bonds break and form, from oxide ceramics and alloys to organic electrolytes, 2D carbon, and MOF/ZIF chemistries at elevated temperature or under mechanical load. The wiki treats “ReaxFF” as a family of parameterizations fit to quantum data and validated for specific element ranges; coverage is not universal and must be checked paper-by-paper.
Functional form and training data¶
ReaxFF expresses energies via bond orders derived from interatomic distances, enabling continuous formation and rupture of bonds without preset connectivity. Parameter sets are trained to DFT (and sometimes experiment) for subsets of chemistry; the group’s publications typically document training reactions, equations of state, and barrier data used in the fit (see cited papers).
Applicable materials and elements¶
Depends on the specific parameterization: examples in this repository include NASICON-type LATP electrolytes, Li–organic electrolyte species, Ni–O metal oxidation, Fe–Al–Ni alloys, BaTiO3 ferroelectrics, hydrocarbon / fuel pyrolysis, and graphene / nanocarbon under extreme mechanics.
Known limitations and failure modes¶
Transferability is limited to chemistries represented in training; quantitative rates and conductivities may require validation against experiment or higher-level theory. Some studies explicitly discuss known FF weaknesses (e.g., moderate-strain elasticity of carbon allotropes) while still using ReaxFF for the relevant coupled chemistry.
Validation benchmarks¶
Illustrative validation hooks referenced in source_refs: ionic conductivity trends for LATP; DFT alignment for electrolyte initiation energetics; qualitative \(E_p^\*\) behavior for graphene impact.
Related parameterizations and software¶
Group workflows commonly pair ReaxFF with LAMMPS-style MD and, where noted, hybrid Monte Carlo / MD schemes for electronic-state or stoichiometry sampling.
How the group uses ReaxFF (corpus-level pattern)¶
Across the wiki, ReaxFF is the default when bond topology changes during a simulation: solid electrolytes, liquid electrolytes at reducing electrodes, metal oxidation, hydrocarbon pyrolysis, graphene impact, and MOF frameworks at high temperature. Papers rarely treat the FF as a black box—most document which elements are covered and which observables were used for sanity checks (ionic conductivity trends, qualitative barrier ordering, mechanical response).
Relationship to other methods in the KB¶
- Classical non-reactive models appear where connectivity is fixed; the KB still cites ReaxFF where reactions gate the science question.
- MLIPs and JAX-differentiable variants are indexed under theme-ml-atomistic-potentials and reaxff-vs-mlip-accuracy.
- Parameterization discipline is summarized in reaxff-parameterization-workflow.
Cross-links (navigation)¶
- Battery-facing work: batteries-interfaces-reaxff.
- Oxide / silica / Ni oxidation: theme-oxides-silica-ceramics.
- Catalysis: theme-catalysis-surfaces.
- Pyrolysis / fuels: theme-pyrolysis-combustion-organics.
- 2D carbon mechanics: graphene-nanocarbon.
Key references¶
- See
source_refsfor entry points; browsepapers/index.mdfor the full parameterization bibliography.