Optimization of the Reax force field for the lithium–oxygen system using a high fidelity charge model
Non-primary PDF
AIP proof PDF (papers/OHearn_JCP2020_Li2O_ACKS2_proof.pdf). For final layout and pagination, prefer [[2020o-hearn-j-chem-phys-optimization-reax]] when that slug references the accepted / VOR PDF. See docs/corpus/NON_PRIMARY_ARTICLE_PAPER_SLUGS.md (proof duplicate pattern).
Summary¶
Lithium oxide chemistry underpins solid-state electrolytes, battery interfaces, and ceramic fracture phenomena where charge redistribution during bond breaking challenges classical QEq/EEM schemes within ReaxFF. This J. Chem. Phys. article replaces EEM/QEq-style charge equilibration with ACKS2 (atom-condensed Kohn–Sham DFT to second order) embedded in ReaxFF for the Li–O system. Parameters are optimized with genetic algorithms against DFT, MRCI, and MRCI+Q bond-breaking training data, then validated in LAMMPS builds with ACKS2 enabled. The new fit improves charge evolution along Li–O dissociation versus a 2016 QEq parameter set, yields qualitatively different Li\(_2\)O melt behavior, but still does not reproduce full brittle fracture in showcased notched slab tests.
The proof PDF path flags pagination risk; operators should mirror numerical benchmarks against [[2020o-hearn-j-chem-phys-optimization-reax]] using the accepted/VOR PDF when reconciling figure labels and Supporting Information references.
Methods¶
Aligned with [[2020o-hearn-j-chem-phys-optimization-reax]]: OGOLEM / genetic optimization; LAMMPS with ACKS2; NPT equilibration of Li\(_2\)O bulk at 300 K and 0 pressure; timestep stability checks (0.25 fs in NVE); NVT plane-strain fracture of notched slabs at 300 K. Full tabulated settings and convergence criteria reside on the VOR page.
Reproducibility detail. ACKS2 integrations need a compatible LAMMPS build; mirror [[2020o-hearn-j-chem-phys-optimization-reax]] and the article inputs.
1 — MD application (atomistic dynamics). Engine / code: LAMMPS with ACKS2-enabled ReaxFF (staged protocols summarized on the VOR page). System size & composition: Li₂O bulk melt cubes and notched fracture slabs; counts and box sizes on [[2020o-hearn-j-chem-phys-optimization-reax]]. Boundaries / periodicity: PBC for bulk; plane-strain fracture geometry as in the article figures (VOR). Ensemble / timestep / duration: NVE checks at 0.25 fs timestep; NPT preequilibration of bulk Li₂O at 300 K near 0 pressure; NVT plane-strain fracture at 300 K. Thermostat: ensemble-control details (including any thermostat and damping) are in the VOR pdf_path; this proof stub does not duplicate them. Barostat / pressure: NPT only for bulk preequilibration; NVT fracture stage has no barostat. Electric field: N/A — not used. Replica / enhanced sampling: N/A — not used in the melt/fracture showcases.
2 — Force-field training. ReaxFF for Li–O reoptimized with ACKS2 against DFT and MRCI/MRCI+Q reference data (genetic / OGOLEM-style workflow in the paper); see [[2020o-hearn-j-chem-phys-optimization-reax]] for citable parameter and validation tables.
Findings¶
ACKS2-coupled ReaxFF reproduces training-set charge evolution and Li–O bond-breaking targets more faithfully than the 2016 QEq parameter set. NPT melt simulations of Li\(_2\)O show different amorphization and RDF signatures than QEq on the same benchmarks. Fracture tests on notched Li\(_2\)O slabs remain partly ductile relative to expected brittleness—i.e. improved electrostatics along dissociation does not by itself fix long-range crack physics in the classical model. Headline numbers, tables, and thermostat/barostat labels for reproduction: [[2020o-hearn-j-chem-phys-optimization-reax]] (VOR). Corpus honesty: this wiki slug tracks the AIP proof pdf_path; pagination and figure placement may differ from the VOR—reconcile QEq-2016 vs ACKS2 comparisons on the non-proof PDF when precision matters.
Limitations¶
Proof PDF: page numbers and figure placement may differ from the final issue PDF—reconcile locators on the VOR sibling.
Relevance to group¶
Duplicate ingest for AIP proof workflow (van Duin co-author); charge fidelity for Li–O ceramic and battery-adjacent chemistry.
Reader notes (navigation)¶
[[2020o-hearn-j-chem-phys-optimization-reax]]Proof duplicate. Reconcile timestep, ACKS2 build flags, and fracture slab geometries with[[2020o-hearn-j-chem-phys-optimization-reax]]using the non-proof PDF when available.