Recent Advances for Improving the Accuracy, Transferability, and Efficiency of Reactive Force Fields (accepted proof PDF)
Evidence and attribution¶
Authority of statements
Prose below summarizes the publication identified by doi and the journal-layout PDF on [[2021itai-leven-j-chem-theor-recent-advances]].
Summary¶
This corpus slug records an accepted proof (papers/Leven_JCTC_2021_proof.pdf) for the same J. Chem. Theory Comput. Perspective that is fully curated on [[2021itai-leven-j-chem-theor-recent-advances]] using papers/Leven_JCTC_2021.pdf. The article surveys how practitioners improve accuracy, transferability, and efficiency for ReaxFF-class reactive MD, covering charge models, eReaxFF-related ideas, energy conservation in coupled schemes, hybrid partitions, and throughput (e.g. LAMMPS-oriented acceleration). The scientific text should match the VOR file on the non-proof slug; page and figure locators for external use should come from the typeset article or [[2021itai-leven-j-chem-theor-recent-advances]], not this proof alone.
Methods¶
1 — MD application (atomistic dynamics)¶
N/A — the Perspective is a synthetic review; for any concrete LAMMPS-style timestep, ensemble, and system definition, follow primary sources cited in the text and the narrative on [[2021itai-leven-j-chem-theor-recent-advances]].
2 — Force-field training¶
N/A as a new ReaxFF reoptimization in this file; the manuscript synthesizes training strategies from cited parametrization studies—see the VOR page for the full survey prose.
3 — Static QM / DFT¶
N/A as a standalone DFT campaign; QM references enter through cited work.
4 — Reviews, perspectives, or non-simulation studies¶
Proof vs VOR: content parallels [[2021itai-leven-j-chem-theor-recent-advances]]; this PDF may differ in pagination and line breaks. Corpus / KB honesty: do not rely on this slug alone for citation-ready page numbers.
Findings¶
Synthesis in the JCTC Perspective (same DOI on both proof and VOR files) is summarized on [[2021itai-leven-j-chem-theor-recent-advances]]: the authors connect EEM/QEq limitations, eReaxFF and hybrid-embedding threads, and practical ReaxFF + LAMMPS workflows for charge-bottlenecked simulations, always via citations to application and parametrization primary literature. This proof-PDF page is suitable for in-repo bibliographic tracking but not a separate empirical result set. Comparisons between modeling choices are narrated in the full VOR text; this stub avoids retyping those subsections so the knowledge base has one canonical narrative and avoids duplicated errors when the galleys differ only by layout. Limitations (corpus / KB): if pagination in the proof and VOR diverge, any “p. …” locator in teaching materials must be reconciled against the journal PDF on the non-proof slug or the DOI HTML; uncertainty on line-level diffs is an inherent hazard of duplicate PDF stages in the corpus tree.
Limitations¶
Proof PDFs are not a substitute for the version-of-record when citing page / figure; method citations should use the VOR entry. Survey content ages; follow post-2021 MLIP and Reaxff literature for current defaults.
Relevance to group¶
Adri C. T. van Duin coauthored the Head-Gordon-led overview—a spine document alongside application papers in the wiki.
Citations and evidence anchors¶
Primary entry: 2021itai-leven-j-chem-theor-recent-advances. https://doi.org/10.1021/acs.jctc.1c00118