Elucidation of Molybdenum Trioxide Sulfurization: Mechanistic Insights into Two-Dimensional Molybdenum Disulfide Growth
Summary¶
Two-dimensional MoS₂ growth from MoO₃ and elemental sulfur feeds involves a zoo of gas-phase sulfur allotropes (S₂–S₈) and condensed oxysulfide intermediates whose elementary steps are difficult to isolate experimentally. This J. Phys. Chem. A article combines ReaxFF molecular dynamics in LAMMPS with Gaussian DFT (B3PW91 with LanL2MB-style basis handling as stated) to dissect how smaller sulfur units emerge from larger allotropes on molybdenum oxide templates and how those units participate in sulfurization toward MoS₂-like bonding motifs. The MD portion emphasizes state-to-state connectivity extracted from reactive trajectories, while DFT supplies barriers and intermediates for S₂ and S₃ channels that anchor the MD interpretation.
Methods¶
1 — MD application (ReaxFF, LAMMPS). MoO₃ slabs and S\(_n\) allotropes (and oxysulfide rings) are built in Avogadro, pre-relaxed with MMFF94, then placed in ~10 Å PBC cubic supercells with explicit atom lists. After conjugate-gradient relaxation, each system receives 2 ns of NVT equilibration with the temperature ramp in the paper, then production ReaxFF molecular dynamics with 0.25 fs timestep, Nosé–Hoover thermostat (25 fs damping) on T = 500, 700, 850, 1000, 1150, 1300 K. The ReaxFF force field follows the single-layer MoS₂ mechanical training set cited in J. Phys. Chem. A. Barostat and N/A independent isotropic pressure control in these NVT cells; N/A — external electric field; N/A — umbrella in the reported NVT ramp; N/A to treat as metadynamics. Fragment connectivity graphs summarize reaction pathways from the trajectories.
2 — Force-field training. N/A — the study uses a published ReaxFF for Mo–S–O chemistry; it is not a new parameter fit report.
3 — Static QM (DFT). Gaussian 09 DFT uses B3PW91 with LanL2MB-style ECP/basis handling for molybdenum and light atoms (per Methods), k-mesh / grid settings and tight force convergence for S₂ and S₃ reaction pathways; barriers and frequencies support the ReaxFF-based mechanism story.
Findings¶
Across allotrope starting points, reactive pathways repeatedly factor into S₂- and S₃-like fragments or their combinations, supporting a reduced reaction basis for complex sulfur vapor chemistry near the oxide surface. MoO₃ is argued to catalyze fragmentation of larger allotropes, effectively funneling sulfur toward sizes that can insert into Mo–S networks. DFT barrier analysis for the highlighted channels yields temperature-dependent rates peaking near 1000–1100 K, overlapping typical chemical vapor transport windows used in MoS₂ growth. Together, the MD connectivity picture and DFT barriers provide a mechanistic narrative linking gas-phase speciation to experimentally accessible temperature ranges. Finite simulation cells and short reactive trajectories may omit long-range sulfur transport or gas-phase three-body chemistry; treat absolute reaction rates as order-of-magnitude guides unless validated against experiment or higher-level theory for the specific allotrope mixture. The article’s connectivity diagrams and temperature sweep should be read as mechanistic scaffolds for experimentalists tuning sulfur partial pressures and oxide pretreatment in MoS₂ growth reactors.
Comparisons, sensitivity, corpus honesty. DFT barrier benchmarks complement the ReaxFF connectivity data; the 1000–1100 K window is a key sensitivity to how CVT is tuned. All kinetics are order-of-magnitude unless the user reconciles to measured S\(_2\)/S\(_8\) mixtures.
Limitations¶
Finite cells and short-timescale MD may omit long-range transport or reactor-scale effects; DFT level and ReaxFF transferability bound quantitative barrier accuracy.