Skip to content

Atomic structure of graphene subnanometer pores

Evidence and attribution

Authority of statements

Summaries follow ACS Nano (DOI in front matter). This work is AC-TEM + DFT + supporting tight-binding dynamics—not a ReaxFF study.

Summary

Sub-nanometer pores in graphene matter for desalination, gas separation, and biopolymer translocation, but their atomic rim structure is hard to image before beam-induced healing or contamination closes the gap. Robertson et al. combine aberration-corrected TEM at 80 kV with in situ heating (500–800 °C) to stabilize open pores long enough for picometer-scale imaging, reporting non-ideal, disordered edge configurations rather than pristine zigzag/armchair filaments alone. DFT (VASP, PBE) relaxes vacancy supercells and supports measured bond lengths; tight-binding molecular dynamics (environment-dependent formulation referenced in the article) explores ring rearrangement kinetics along pore edges. Together, the data emphasize five-membered ring projections decorating evolving pore perimeters and connect imaging-derived bond metrics to relaxed atomic models.

Methods

Experimental microscopy and sample handling. Graphene is grown by CVD on Cu, transferred with PMMA, and cleaned on TEM grids using a 350 °C bake to remove polymer residues (Methods narrative in the article). AC-TEM uses a JEOL 2200MCO instrument at 80 kV with a monochromator (energy spread quoted <300 meV in the extract) and a DENS heating holder. High temperature (500–800 °C) suppresses pore filling by mobile surface carbon during prolonged imaging. Flux / dose estimates and detailed drift/calibration steps are given in pdf_path and SI.

Static QM / DFT. Functional: PBE GGA in VASP for vacancy and pore-edge supercells as described in Methods. Dispersion: vdW corrections are discussed where edge adsorbates matter; the wiki does not assign a universal DFT-D label without the SI line—see pdf_path. Basis / potentials: Plane-wave PAW expansion at 400 eV cutoff (article Methods). k-point / Brillouin sampling: Γ-centered (2×2×1) k-mesh for the 448-atom graphene supercells summarized on this page. Structures / properties: relaxations of vacancy and reconstructed rim motifs; multislice image simulation matched to experimental intensity profiles for bond-length extraction (Results).

Tight-binding MD (supporting kinetics). Environment-dependent tight-binding MD explores five-membered ring migration along pore perimeters (code references in the article/SI). Timestep / duration / thermostat for TBMD: not duplicated here—see pdf_path.

MD application (classical ReaxFF/AIMD production): N/A — not a classical reactive MD paper.

Force-field training: N/A.

Findings

Imaging outcomes: With heating, sub-nm pores remain open long enough for atomic-resolution imaging; diameters in one reported range are ~0.5–0.8 nm alongside atomic models in the article’s Figure 1 narrative (extract).

Edge chemistry and kinetics: Pores grow under the beam by sequential sputtering; undercoordinated perimeter atoms are removed preferentially, often leaving five-membered rings projecting outward. DFT + image simulation support reconstructed C–C separations at the rim (e.g., ~169 pm reconstructed vs non-reconstructed contrasts in the Results discussion—exact numbers on pdf_path).

Comparisons / levers: Temperature and beam exposure control whether sub-nm pores survive versus self-healing/contamination filling; 80 kV operation is chosen near the knock-on threshold for sp² carbon to balance milling and imaging (Introduction/Results).

Limitations

The electron beam drives chemistry and heating; PBE omits some dispersion subtleties for certain edge adsorbate configurations. TBMD approximates π electronic structure; highest-accuracy barriers may require DFT NEB for specific elementary steps. Quantitative DFT convergence settings, TBMD integration parameters, and figure-level distance metrics should be taken from pdf_path and SI rather than from this summary alone.

Relevance to group

Graphene nanopore structural science for separation and sensing—complements reactive carbon simulation elsewhere in the corpus but is primarily microscopy + QM.

Citations and evidence anchors

MAS / retrieval

Canonical tags emphasize 2D + DFT + experiment integration; do not add method:reaxff from this page alone.