Classical and reactive molecular dynamics: Principles and applications in combustion and energy systems
Summary¶
Molecular dynamics has become a standard engineering-facing tool for predicting material properties and reactive flows when sufficient computational resources are available. Mao, Feng, Jiang, Ren, Luo, and van Duin review how classical and reactive MD are used in combustion and energy systems, spanning gas-, liquid-, and solid-fuel oxidation and pyrolysis, catalytic and heterogeneous combustion, electrochemistry, nanoparticle and flame synthesis, heat transfer, phase change, and nanoscale fluid phenomena. The article first presents MD methodology for nonreactive and reactive potentials, emphasizing ReaxFF development trained on quantum chemistry and/or experiments. It then discusses numerical methods, boundary conditions, post-processing, and computational cost, followed by selected application vignettes that illustrate how reactive MD exposes pathways and energetics in complex environments. The review explicitly frames MD’s expansion from fundamental science into engineering design loops enabled by petascale computing, while warning that predictive use still hinges on validated potentials and adequate sampling.
Methods¶
Review structure (D)¶
Narrative synthesis of classical and reactive MD for combustion/energy: integrators, ensembles, timestep selection, thermostats/barostats, nonequilibrium driving, reaction-network statistics.
Reactive force fields (A)¶
Bond-order ReaxFF-style formalisms and training workflows at textbook level with pointers to primary parametrization studies.
Boundary conditions and applications (B)¶
Confined flows, interfaces, reactive walls relevant to energy devices.
Traceability rule¶
Any specific numerical claim requires the cited primary paper—not this review alone.
4 — Reviews, perspectives, non-simulation (primary genre)¶
Literature scope, comparison logic, and vignette pointers replace a single simulation table. For reactive and classical MD (LAMMPS-class engines, NVT/NPT/NVE choices, 0.25 fs order-of-magnitude for some ReaxFF work, thermostats/barostats, PBC in condensed phases, K-scale temperatures in reactor-like phenomena, ns-length reactive trajectories in turbulent-flame surrogates): the review teaches what categories to document but does not fix one atom count, one kPa/GPa target, or one rare-event sampling run—N/A in that single-study sense. FF training and static QM subsections in cited primary work must be read per reference; the review synthesizes ReaxFF development concepts without substituting a de novo training recipe here.
Findings¶
Capabilities highlighted (abstract themes)¶
ReaxFF MD used to map pyrolysis/oxidation pathways, soot/nanoparticle phenomena in flames, and catalytic/interfacial chemistry where bond rearrangement matters; nanoscale flow/heat/phase phenomena complement continuum models.
Validation framing¶
Emphasizes force-field error and sampling limits as ongoing constraints.
Pedagogical use¶
Useful as a syllabus map; project workflows still need potential validation against target chemistry.
The Prog. Energy Combust. Sci. scope statement (see Introduction) also situates MD within petascale engineering workflows: reactive simulations are increasingly used to inform kinetic submodels and material selection, but the review repeatedly cautions that predictive deployment still requires validated potentials and sufficient sampling of rare events in heterogeneous environments.
Limitations¶
Breadth limits depth per application; quantitative conclusions must be traced to original studies. Force-field accuracy remains system dependent.
Readers should treat the application vignettes as illustrative: each cited primary study will specify ensemble, boundary conditions, and validation metrics that this review cannot reproduce in full without duplicating entire bibliographies.
Relevance to group¶
Adri C. T. van Duin is a co-author; central reference for ReaxFF in combustion within the corpus.