Multi-scale modeling of gas-phase reactions in metal-organic chemical vapor deposition growth of WSe₂ (uncorrected proof)
Summary¶
This page documents an uncorrected proof PDF (papers/Xuan_J_Cryst_growth_WSe2_multiscale_proof.pdf) for the Journal of Crystal Growth article (DOI 10.1016/j.jcrysgro.2019.125247) on multi-scale modeling of gas-phase chemistry during metal–organic chemical vapor deposition (MOCVD) growth of tungsten diselenide from tungsten hexacarbonyl and hydrogen selenide precursors, with Adri C. T. van Duin among the co-authors. The scientific program—as fully curated on the version-of-record page 2019xuan-journal-of-c-multi-scale-modeling—combines density functional theory to identify reaction classes and train ReaxFF for W/H/C/O/Se, ReaxFF molecular dynamics to extract pathways and activation energies feeding Arrhenius parameters, and reacting computational fluid dynamics of a cold-wall reactor to connect chemistry with mass and heat transport. The proof file exists in the corpus alongside the final-layout PDF to preserve publisher-state provenance during ingest; readers should assume figures, pagination, and minor wording may differ from the journal PDF until cross-checked.
Methods¶
Methodologically, the workflow is identical to 2019xuan-journal-of-c-multi-scale-modeling: QM provides reference data for gas-phase bond rearrangements and thermochemistry, ReaxFF supplies an empirical reactive model consistent with those references at MD-accessible cost, and a compiled kinetic mechanism is embedded in a CFD solver capturing inlet conditions, buoyancy, heating, and species transport in the MOCVD chamber used experimentally. Validation emphasizes comparing gas-phase concentrations of tungsten-containing chalcogenide species near the substrate to spatial trends in measured film thickness, treating those species as effective growth feedstock proxies. Because this slug tracks proof bytes, the authoritative Methods section numbering, supplementary pointers, and figure labels should be read from the VOR wiki page and its PDF rather than from the proof PDF alone.
ReaxFF + DFT + CFD (same science as the VOR). Engine: LAMMPS-style reactive molecular dynamics for W/H/C/O/Se gas chemistry in periodic cells; DFT reference data for training; CFD for cold-wall reactor transport. System: gas-phase precursor mixtures with W(CO)₆ and H₂Se-derived species; N/A for full stoichiometry tables on this proof page. Ensemble: NVT-like ReaxFF MD with Nose–Hoover-class thermostat when NVT is used (per VOR), plus NVE-like segments if the article uses them; timestep and ns-scale duration in the VOR Methods. Temperature fields follow reactor heating in CFD; atomistic K for ReaxFF are in the main text. Barostat for non-periodic open reactor flow: N/A in the MD leg; CFD handles pressure gradients. Shear/strain, electric field, umbrella sampling: N/A for the MOCVD gas chemistry focus in the abstract-level summary. Detail source: 2019xuan-journal-of-c-multi-scale-modeling and the final J. Cryst. Growth PDF, not the uncorrected proof bytes alone.
Findings¶
The published study reports that the coupled gas-phase model reproduces experimental trends linking precursor transport and local chalcogenide availability to deposited WSe₂ thickness across the wafer, supporting the paper’s argument that full-chamber chemistry–transport coupling is necessary when experimental gas-phase maps are sparse. The authors also position the framework as transferable to other CVD chemistries by swapping mechanism blocks while retaining the ReaxFF → kinetics → CFD pipeline. For this proof-ingest page, the wiki does not restate every numerical comparison; instead, it directs readers to 2019xuan-journal-of-c-multi-scale-modeling for sentence-level findings and quantitative benchmarks aligned with the final article text. Corpus honesty: this file is for provenance of the uncorrected proof PDF; cite VOR figures and tables in downstream work.
Limitations¶
Uncorrected proofs may include watermarks, layout differences, and non-final figure resolution; citations requiring stable page or figure numbers should use the version-of-record path. Surface growth chemistry itself is not resolved atomistically in the CFD layer—gas-phase fluxes supply boundary conditions to continuum growth interpretations rather than a fully coupled ReaxFF-on-surface multiscale loop.
Relevance to group¶
Multiscale ReaxFF + CFD workflow for 2D TMD MOCVD (van Duin co-author); this page is ingest provenance for the proof file.
Citations and evidence anchors¶
DOI: 10.1016/j.jcrysgro.2019.125247