Skip to content

Role of Bilayer Graphene Microstructure on the Nucleation of WSe2 Overlayers

Non-primary PDF

The corpus pdf_path is a galley PDF (Bachu_ACS_Nano_2023_WSe2_graphene_galley.pdf). For final pagination, copy-edited text, and issue placement, prefer a version-of-record ACS Nano PDF for DOI 10.1021/acsnano.2c12621 when ingested. Galley PDFs may retain line numbers and differ slightly from the published article layout.

Summary

Wafer-scale growth of transition-metal dichalcogenides such as WSe\(_2\) on graphene templates is sensitive to how strain and defects distribute in the underlying carbon layers. This work combines chemical-vapor-deposited bilayer graphene (BLG) substrates, scanning/transmission electron microscopy (S/TEM) characterization of stacking and defects, and ReaxFF molecular dynamics to connect Bernal-stacked versus twisted bilayer regions—and interlayer dislocations—to nucleation density of WSe\(_2\) islands. The central experimental observation is that Bernal BLG can host interlayer dislocations and associated localized buckling that are largely absent in the twisted bilayer graphene (tBLG) cases examined, and that Bernal regions show higher WSe\(_2\) nucleation density. ReaxFF simulations interpret these trends by comparing strain relaxation pathways: Bernal stacking permits localized out-of-plane distortion that concentrates adsorption sites for WSe\(_x\) species, whereas twist tends to spread strain more uniformly, yielding fewer concentrated trapping sites for precursor condensation in the model framing of the paper.

The galley PDF registered here (Bachu_ACS_Nano_2023_WSe2_graphene_galley.pdf) may differ in pagination and line breaks from the issue PDF; cite methods and figures against a VOR file when available, while using this slug for internal provenance of which PDF was ingested.

Methods

Experiments (CVD, microscopy)

  • Growth: CVD-grown bilayer graphene under conditions in the ACS Nano article; dark-field TEM / related imaging distinguishes Bernal vs twisted stacking and maps interlayer dislocations and buckling.

ReaxFF molecular dynamics (B)

  • Cells: BLG models with Bernal vs twist; strain relaxation protocols per Methods.
  • Analysis: Locate out-of-plane buckles and regions of enhanced interaction with WSe\(_x\)-like adsorbates (force field and partial-charge treatment in the publication).
  • Reproducibility: Buckling is sensitive to initial interlayer spacing, relaxation length, and boundary conditions—mirror the paper’s settings when rerunning; compare stackings under the same strain protocol.

Experiment–simulation coupling

Qualitative mapping between simulated preferential adsorption/trapping sites and experimental WSe\(_2\) island nucleation density—not claimed as quantitative barrier matching from ReaxFF alone.

MD application (bilayer graphene relaxation)

Engine / code: LAMMPS-compatible ReaxFF (per ACS Nano). BLG supercells (Bernal vs twist), 3D PBC; relaxation to capture out-of-plane buckling; optional WSe\(_x\)-type adsorbate motifs for site comparison. Timestep, thermostat and Langevin or Nose/Berendsen-style coupling (as named), Temperature, NVT/NVE (or as stated), ps/ns stages in the VOR/SI; N/A — not a CVD dynamics run of full WSe\(_2\) deposition in the RMD as summarized. N/A — no NPT barostat unless the VOR uses isotropic pressure; N/A — no in-plane static electric field in the reactive relax protocol summarized; N/A — no replica/metadynamics; Coulomb and QEq in the VOR.

Findings

Microstructure vs nucleation

Bernal BLG shows interlayer dislocations (and associated buckling) under imaging conditions where tBLG largely does not; Bernal regions correlate with higher WSe\(_2\) nucleation density.

Simulation interpretation

ReaxFF relaxation associates Bernal stacking with localized buckles and twist with more uniform strain partitioning; buckles are proposed as favorable traps for WSe\(_x\) species, qualitatively consistent with higher nucleation on Bernal areas.

Scope note. Trends are demonstrated for the reported CVD recipe; precursor chemistry, carriers, and substrates can change which rate-limiting step controls nucleation—reconcile WSe\(_2\) kinetics to the VOR when citing T, partial pressure, or substrate strain beyond the RMD scope summarized here.

Limitations

ReaxFF remains approximate for metal chalcogenide chemistry and precursor coverages relevant to CVD; galley PDFs complicate precise figure and page citation until a VOR file replaces the corpus path.

Relevance to group

Penn State van Duin-group ReaxFF coupled to 2DCC / Redwing-led epitaxy of WSe\(_2\) on graphene.

Citations and evidence anchors

Reproducibility and corpus locators

This note documents where to find primary evidence in-repo; it does not add new scientific claims beyond the cited publication.

Normalized layer. When present, normalized/papers/{slug}.json mirrors manifest hashes, bibliography fields, and extraction pointers; if pdf_path or PDF bytes change, follow AGENTS.md and docs/PHASE3_RUNBOOK.md to re-profile rather than editing PDFs in place.

Authority chain. For numerical settings (cutoffs, timesteps, ensembles, kinetics), use the peer-reviewed PDF (and publisher Supporting Information) as the authoritative source; this wiki summarizes for navigation and retrieval.