Skip to content

Reactive Potentials for Advanced Atomistic Simulations

Evidence and attribution

Authority of statements

The canonical scholarly reference is Annu. Rev. Mater. Res. 2013, 43, 109–129 (DOI in front matter). The PDF filename includes 2016 metadata from download; cite the Annual Reviews volume/issue for bibliography.

Summary

This Annual Review compares variable-charge reactive empirical potentials, focusing on ReaxFF and COMB (charge-optimized many-body) families, and surveys applications in heterogeneous environments where bond-order reactivity and electrostatic flexibility matter: catalysis on metals and oxides, oxidation, electrochemical interfaces, and complex multicomponent solids. The review’s pedagogical goal is to clarify how these potentials encode charge equilibration alongside reactive bond energetics, and how practitioners typically fit them to QM databases.

The article contrasts ReaxFF-style formulations with COMB approaches at the level of energy partitioning, charge models, and typical training practices, then uses literature examples to illustrate where variable-charge reactive models outperform fixed-bond force fields for chemically evolving systems.

Methods

This peer-reviewed Annual Review article is not a single new simulation study; it synthesizes how variable-charge reactive empirical potentials—principally ReaxFF and COMB—partition bond-order-mediated reactivity versus many-body corrections, how electrostatics are treated (self-consistent charge equilibration versus charge-optimized many-body ideas), and how practitioners historically optimize parameters against QM databases cited throughout the chapter. Vignettes steer readers toward metal/oxide interfaces, oxidation, and electrochemical/multicomponent solids where bond-making and charge reorganization matter together. Because each primary reference brings its own LAMMPS/VASP/CP2K recipe, the review does not consolidate a single benchmark MD protocol—those details remain in the cited application papers.

MD / reactive simulations (review-level). Because the article is not one production study, LAMMPS/CP2K/VASP-class engine choices, supercell/slab sizes, PBC details, NVT/NPT labels, timestep (fs), equilibration/production duration (ps/ns), temperature (K) ramps, barostat/pressure targets (bar/GPa), and thermostat damping are all N/A at review level—read each cited primary paper. Electric field: N/A — not a unified theme. Umbrella / metadynamics / replica exchange: N/A — not surveyed as a methods spine.

Force-field training (review-level). Parent potentials: ReaxFF (van Duin lineage) vs COMB (Sinnott/Phillpot lineage). QM reference data: discussed generically with citations to DFT training corpora used historically in the literature. Training sets and optimization workflows (least squares, genetic/CMA-style searches where cited) vary by application; the review maps philosophies rather than publishing a new fit. Experimental benchmarks: cited where prior literature couples ReaxFF/COMB to measurements.

Findings

The review argues that variable-charge reactive potentials are a workable compromise when simulations must reach large system sizes and nanosecond scales yet still allow topology changes. It contrasts ReaxFF and COMB not as interchangeable black boxes but as families with different electrostatic and fitting philosophies; transferability remains training-limited in both cases. Comparisons: the abstract frames a side-by-side of ReaxFF vs COMB for multicomponent problems. Sensitivity / levers: performance depends on how QM training breadth, charge models, and bond-order treatments match the application—details vary by cited study rather than a single knob table here. Limitations / outlook: reviews age; readers must pull validation evidence from primary sources rather than treating this article as a parameter file. Corpus honesty: this wiki entry aligns to Annu. Rev. Mater. Res. 2013, 43, 109–129 (DOI in front matter); the corpus PDF imprint shows a Penn State download stamp from 2016—bibliography should follow the journal volume/issue, not the filename year.

Limitations

Being anchored in 2013 literature, later parameterization advances, software implementations, and best practices must be updated from primary sources. Reviews do not replace validation for a specific parameter file.

For operators maintaining this wiki: use this review as a map, then jump to application-specific ReaxFF papers (oxide–water, organics, metals) for quantitative performance. The COMB comparison is especially useful when a simulation needs charge behavior beyond a minimal QEq treatment— but compatibility between COMB and ReaxFF databases is not automatic.

Relevance to group

van Duin-group coauthored comparison of ReaxFF with COMB—often used as a graduate-level entry to reactive potentials in materials simulations.

When MkDocs readers land here from a wikilink, steer them next to application notes on oxidation, electrolytes, or oxide interfaces, because the review’s value is orientation, not simulation-ready parameters.

Citations and evidence anchors

MAS / retrieval

Use this review as a hub pointer, not as a source of parameter numbers; chunk text still indexes useful terminology (COMB, variable charge, ReaxFF) for hybrid search.