Graphene healing mechanisms: A theoretical investigation
Summary¶
Defect engineering and repair of graphene are central to nanocarbon electronics and microscopy, yet vacancy healing involves bond rearrangements that are difficult to capture with non-reactive potentials. Botari et al. use reactive molecular dynamics with a ReaxFF-style carbon field to follow large perforated graphene motifs as they evolve under high-temperature annealing with an explicit carbon feedstock and under room-temperature scenarios where electron-beam energy deposition mimics scanning transmission electron microscopy (STEM) conditions reported experimentally. The study emphasizes transient non-hexagonal intermediates—linear carbon chains and five- and eight-membered rings in planar and non-planar arrangements—as obligatory waypoints on routes back toward hexagonal graphene.
Methods¶
MD application (atomistic dynamics)¶
LAMMPS reactive MD uses ReaxFF (Chenoweth et al. (2008) C/H/O, Carbon). The model is single-layer graphene with a circular hole (~3.2 Å radius), in-plane PBC, and perimeter atoms tied by virtual springs (K = 30.0 kcal mol⁻¹ Å⁻²); in the electron-beam scenario those edge atoms also receive temperature fixing to dissipate excess energy (Carbon Methods/figures). Scenario (i) anneals 300–2000 K with carbon adatoms injected at random positions every 500 fs with random kinetic energies (single-atom depositions). Scenario (ii) keeps the sheet at 300 K but applies cylindrical local heating in a moving zone to mimic STEM-like excitation, with the same 500 fs deposition schedule. Integration uses 0.1 fs steps under NVT with a Nose–Hoover chain thermostat (parameters in Carbon). Selected SCC-DFTB (DFTB+, Slater–Kirkwood dispersion) calculations benchmark ReaxFF. Barostat, applied electric fields, and enhanced sampling beyond the stated heating/deposition: N/A — not used. Hydrostatic pressure control: N/A — not used (no NPT barostat). Total trajectory lengths per scenario are tabulated in the PDF.
Force-field training¶
N/A — uses published ReaxFF and DFTB parametrizations; no new QM refit is reported as the primary contribution.
Static QM / DFT¶
N/A — SCC-DFTB appears as a comparative electronic-structure tool, not as the main production method.
Findings¶
With high temperature and carbon supply, reactive trajectories show the hexagonal lattice can close when edges and adatoms surmount rearrangement barriers; oxidation-adjacent rearrangement and decomposition of strained rim sites precede closure. Healing routes pass through non-hexagonal intermediates rather than a single zippering step. At 300 K, complete healing in the model is tied to localized heating reminiscent of STEM conditions; without that localization, healing is kinetically limited on accessible MD timescales—an explicit limitation of pure room-temperature annealing in the setup. Authors highlight transient linear carbon chains and five- and eight-membered rings (including Stone–Wales-like motifs) as recurring waypoints before sixfold coordination dominates again. Coverage and defect distributions differ from experiment; quantitative barriers and full trajectory budgets should be taken from the PDF rather than this summary alone.
Limitations¶
The beam model is a phenomenological local heating + deposition scheme, not a first-principles electron-scattering treatment; single-atom carbon deposition simplifies realistic plasmas/microscopes where small hydrocarbon fragments may dominate.
Relevance to group¶
Connects ReaxFF reactive MD to defect healing in nanocarbon, a recurring theme alongside other graphene reactive simulations in the corpus.