ReaxFF Reactive Force-Field Study of Molybdenum Disulfide (MoS₂)
Abstract
New ReaxFF for MoS₂ fit to Jaguar B3LYP/LAV3P** molecular data and VASP PBE-PAW periodic data; applications to vacancies, migration barriers, 2H↔1T pathways, ripplocations, and elastic response with PBC uniaxial/biaxial strain.
Summary¶
The letter develops a Mo/S/H ReaxFF trained against DFT: Jaguar optimizations (B3LYP with the LAV3P basis) for molecules in the training set and VASP (PBE, PAW, 400 eV cutoff, Γ-centered k-grids, forces <0.02 eV/Å), including bilayer spacing scans with constrained Mo planes and NEB barriers where needed (see Methods and SI). ReaxFF energies combine bonded, torsion, lone-pair, over/under-coordination, vdW, and Coulomb terms (eq. 4 in paper). The fitted potential reproduces bond/angle dissociation and condensed-phase benchmarks tabulated against DFT/experiment (Table 1–2). MD applications use periodic boundary conditions in-plane; uniaxial/biaxial strain increments extract 2D elastic constants and ultimate strengths (~24–26 GPa for SL MoS₂ depending on direction). 2H↔1T pathways and vacancy formation/migration energetics, ripplocation defects, and interplay with S vacancies are explored at ReaxFF level. The parameterization targets mechanochemistry and defect physics in TMDs where fixed-bond potentials are insufficient.
Methods¶
Force-field training / fitting. Mo/S/H ReaxFF is fit to Jaguar B3LYP/LAV3P** molecular training data and VASP GGA-PBE PAW periodic data (400 eV plane-wave cutoff, Γ-centered k-grids, force threshold <0.02 eV Å⁻¹), including bilayer spacing scans with constrained Mo planes and nudged elastic band (NEB) barriers where cited. Optimization follows the standard ReaxFF least-squares / genetic workflow with SI tables (S1, S2, S4) and trainset.in listings in 2017ostadhossein-venue-microsoft-word-3 documenting weights and geometries.
MD application (atomistic dynamics). Engine / code: LAMMPS is used for ReaxFF molecular dynamics and force-field NEB (the letter explicitly contrasts CI-NEB at DFT in VASP with NEB at ReaxFF level in LAMMPS for the 2H↔1T pathway). System size & composition: Monolayer and few-layer MoS₂ supercells for defect, phase, ripplocation, and mechanical tests—exact atom counts vary by figure and are N/A — not duplicated tabularly on this wiki page (see article structures). Boundaries / periodicity: In-plane periodic MoS₂ cells under uniaxial/biaxial strain protocols described in the letter (PBC implied for extended sheets). Ensemble / thermostat / timestep / duration: N/A — the JPCL main text available to this pass emphasizes energy models, barriers, and elastic extraction workflows without restating a single consolidated NVT/NVE table, explicit timestep (fs) settings, or nanosecond-scale production clocks—confirm per-figure simulation cards in the PDF/SI when auditing trajectories. Barostat: N/A — not highlighted for the mechanical strain excerpts summarized here. Temperature: thermal equilibration and MD segments are referenced figure-by-figure, but a unified 300 K (or other) thermostat table is not consolidated on this wiki page—see VOR/SI. Pressure / electric field / enhanced sampling: N/A — not central to the summarized strain / defect / NEB protocol beyond standard NEB restraint.
Static QM / DFT. Jaguar and VASP calculations above supply QM training and validation energies/structures tabulated against ReaxFF in Tables 1–2.
Review / non-simulation framing. N/A — primary JPCL parameterization + application letter.
Findings¶
Outcomes & mechanisms. The fitted Mo/S/H ReaxFF reproduces selected DFT and experimental structural/elastic benchmarks for 2H MoS₂ while enabling reactive sampling of vacancy formation and migration, 2H↔1T pathways, ripplocations coupled to sulfur vacancies, and nonlinear mechanics under strain.
Comparisons. ReaxFF barriers and energy ordering are compared explicitly to DFT (e.g., 2H–1T energetics and barrier scaling discussed in the text) and to nanoindentation/AFM mechanical trends cited by the authors.
Sensitivity & design levers. Strain direction (uniaxial vs biaxial) and defect content modulate elastic response and failure strengths reported near ~24–26 GPa ultimate tensile strengths for SL MoS₂ depending on orientation (see letter).
Limitations & outlook (as authored). DFT band-gap limitations carry over to defect-level interpretations; absolute kinetic rates from ReaxFF require the down-scaled barrier caveats noted in the discussion.
Corpus / KB honesty. SI parts 2017ostadhossein-venue-microsoft-word, 2017ostadhossein-venue-microsoft-word-2, 2017ostadhossein-venue-microsoft-word-3 hold machine-readable training artifacts; refresh MD numerics from the VOR PDF if any figure–text mismatches appear after corpus updates.
Limitations¶
Standard DFT band-gap limitations noted; structural energetics focus avoids excited-state physics. Charge transport and excitonic effects in MoS₂ devices are therefore outside the scope of the mechanical and defect benchmarks emphasized here.
Relevance to group¶
Core group ReaxFF publication for TMD chemistry and mechanics. Subsequent corpus pages on 2D layered materials frequently cite this letter as the MoS₂ parameterization anchor.
Citations and evidence anchors¶
- DOI:
10.1021/acs.jpclett.6b02902