Skip to content

Watching (De)Intercalation of 2D Metals in Epitaxial Graphene: Insight into the Role of Defects

Summary

Niefind and colleagues combine photoemission electron microscopy (PEEM) with ReaxFF molecular dynamics and DFT to compare defect-gated (de)intercalation of two-dimensional Ag and Ga sandwiched between bilayer epitaxial graphene (EG) and SiC. PEEM tracks bright Ag-rich features after in situ annealing (e.g., 436 K), with AFM, SEM-EDX, and XPS supporting surface Ag. Ag shows de-intercalation and re-intercalation through roughly circular “windows,” with an estimated intercalation front speed near 0.5 nm s\(^{-1}\) (±0.2 nm s\(^{-1}\)) from combined PEEM/AFM analysis. Ga de-intercalates irreversibly with faster kinetics and non-circular window shapes; MD shows Ga pile-up between graphene sheets before egress, and DFT indicates stronger Ga–graphene binding, consistent with observations.

Methods

Samples and intercalation context. 2D metals are prepared at the EG/SiC interface using confinement heteroepitaxy (CHet)-style processing: oxygen plasma generates graphene defects that act as entry paths for metal intercalation at elevated temperature/pressure conditions (as summarized in the introduction).

PEEM and complementary microscopy. In situ PEEM follows intensity changes associated with Ag redistribution during annealing cycles up to ~575 K (regimes described around 405–494 K in the discussion of ROI traces). AFM confirms 3D Ag particles atop graphene with representative heights ~15.2 nm ± 2.1 nm and steep sidewalls; SEM-EDX and laterally integrated XPS corroborate Ag surface enrichment after annealing. Multiple regions of interest illustrate zero-order-like kinetics where intercalation speed is weakly dependent on Ag structure size, interpreted as rate limitation by window/defect pathways.

1 — MD application (ReaxFF). Engine / code: ReaxFF molecular dynamics in LAMMPS (as stated in the article; see Computational Methods for callouts and PBC). System: 2D Ag and Ga at EG/SiC-type stacks; full atom counts, timestep, ensemble, and run lengths are in the PDF and are not itemized in this short wiki summary. Thermostat algorithm and time constant: N/A on this summary (see VOR PDF; NVT is the typical ensemble for the reported intercalation T). In situ experimental T (e.g. toward ~436 K in discussed traces, stages up to ~575 K) provide the thermal context; mapping every MD isotherm to experiment is in the paper. Barostat / stress / E-field / replica or enhanced sampling: N/A in the high-level description here — confirm whether the MD uses NPT-style cells or static NVT-like protocols in the SI/PDF; no external electric field in the protocol summary on this page.

2 — Force-field training — N/A as a new general ReaxFF parameterization report; the study applies a ReaxFF parameter set to 2D metal/graphene chemistry (see Computational Methods for the exact FF line).

3 — Static QM (DFT support). DFT benchmarks relative binding of Ga vs Ag on graphene (and related support calculations); program, functional, and basis are given in the article. 4 — PEEM/AFM/EM — in situ PEEM and complementary AFM/SEM-EDX/XPS (above) supply observables that the atomistic work interprets; N/A for a single MD “production run” line item list in this summary—use the version-of-record for tables.

Findings

Outcomes, comparisons, and levers. Ag shows semi-reversible (de)intercalation through defect windows after CHet-style processing, with PEEM/AFM-inferred front speeds on the order of 0.5 nm s\(^{-1}\) (±0.2). Ga de-intercalates irreversibly, with faster kinetics and non-circular windows; ReaxFF MD shows Ga pile-up between graphene sheets and DFT favors stronger Ga–graphene binding vs Ag—linking to kinetic/thermodynamic contrast. Defect-mediated paths and metal-dependent graphene restructuring are argued to co-determine kinetics. ROIs with zero-order-like kinetics (weak dependence of rate on Ag structure size) support defect/window-limited transport.

Corpus honesty — see ## Limitations for model scope (ReaxFF/projection of PEEM).

Limitations

Surface spectroscopy/microscopy integrates over projection and surface sensitivity; MD uses ReaxFF approximations to quantum energetics. In situ temperatures and coverage may not map one-to-one to all device-relevant environments.

Relevance to group

Adri C. T. van Duin co-authors the ReaxFF modeling thread alongside NIST/PSU experimental imaging—an integrated 2D metal / graphene defect case study for reactive MD validation against PEEM.

Citations and evidence anchors

  • PEEM/AFM kinetics and velocity estimate: Results and discussion, Fig. 1–2 region (Small 20, 2306554 (2024)); DOI above.