Skip to content

Atomic-scale mechanisms of plasma-assisted elimination of nascent base-grown carbon nanotubes

Evidence and attribution

Authority of statements

Prose sections below (Summary, Methods, Findings, etc.) are curated summaries of the publication identified by doi, title, and pdf_path in the front matter above. They are not new primary claims by this wiki.

For definitive stage labels, structural assignments, and comparisons to DFT, use the Carbon article and Supplementary material—not this page alone.

Summary

Hydrogen-rich plasmas can etch nascent carbon nanotubes (CNTs) while growth is still ongoing, yet nanoscale sequencing of cap versus sidewall chemistry is poorly resolved experimentally. This paper combines ReaxFF-based reactive MD with temperature-accelerated force-bias Monte Carlo (tfMC) to follow atomic hydrogen attack on base-grown (5,5) and (10,0) motifs supported on Ni nanoclusters with a virtual Al substrate, intended to mimic catalyzed base growth. The simulations resolve a multi-stage hydrogenation and elimination pathway through intermediate carbon nanosheet, nanowall, and polyyne-like motifs, and contrast etch onset for caps versus tube segments with literature DFT and TEM where cited.

Methods

1 — MD application (reactive MD + tfMC). The authors combine reactive molecular dynamics with temperature-accelerated force-bias Monte Carlo (tfMC) to follow atomic hydrogen attack on base-grown carbon nanotube motifs. Temperature is held at 1600 K using the canonical Bussi thermostat. Interatomic interactions use ReaxFF with the Ni/C/H parameter set of Mueller et al. (as cited in the article). System size & composition: finite Ni nanoclusters (55 Ni atoms or 147 Ni atoms) plus CNT cap/tube carbon (e.g. 40 carbon atoms in the (5,5) cap example in Results) and variable H loadings maintained in the gas-phase reservoir. Boundary conditions: finite cluster models on a virtual Al substrate (non-extended bulk lattice); N/A — explicit 3D periodic supercell vectors for the full catalyst+gas assembly are not spelled out on the indexed pages—confirm in SI. Initial geometries include (5,5) and (10,0) caps and tubes on Ni\(_{55}\) or Ni\(_{147}\) clusters physisorbed on a virtual Al substrate, with tangential versus perpendicular tube/cluster arrangements intended to mimic catalyzed base growth. Gas phase uses atomic H only (not a full plasma mixture); H concentration in the MD stage is kept constant. After each H adsorption event, structures are relaxed with tfMC while no new H impinges; gas-phase etching products are removed every 10⁶ MD steps to limit pyrolysis side chemistry. Duration / stages: alternating MD impingement and tfMC relaxation segments with periodic product purges; N/A — a single quoted production length in ns is not tabulated in the excerpted computational section—use the article/SI for aggregated times. N/A — timestep (fs) and engine label beyond “reactive MD”: the indexed Carbon computational section does not state an integration timestep or name a specific MD package; confirm in the full PDF/SI if needed. N/A — barostat / hydrostatic pressure: the described protocol is canonical (constant-temperature) MD with no stated NPT or stress control. N/A — electric field in the simulation cell: the model isolates atomic H chemistry rather than imposing an external field in the MD setup (plasma fields are discussed only at the literature level in the Introduction).

2 — Force-field training. N/A — new fit in this paper: the study employs published ReaxFF Ni/C/H parameters (Mueller et al.) validated in prior work cited by the authors; there is no reported QM refit or CMA-ES/ParReX optimization campaign in this article.

3 — Static QM / DFT. N/A — on-the-fly DFT: comparisons to literature DFT and TEM are cited for context, but the production trajectories are ReaxFF + tfMC only.

4 — Review / non-simulation framing. N/A: primary application article, not a methods review.

Findings

Outcomes and mechanisms. For cap elimination, hydrogenation proceeds in staged fashion: a (5,5) cap on a carbon-containing Ni cluster first hydrogenates (stage I), then evolves toward a carbon nanosheet (stage II). Stage II marks the onset of net carbon loss by etching in the idealized cap trajectory, whereas defective caps can begin losing carbon earlier so that stages I and II merge. The nanosheet may convert to nanowalls or vertical graphene patches before simplifying to rings and short polyyne chains, then full removal from the cluster (stages III–IV in the authors’ division). For tubes, the authors report that etch onset differs from caps: hydrogenation/dehydrogenation compete first, with H coverage on long tubes reaching roughly 30% before etching starts in their analysis; the first C–C bond cleavage often appears on the sidewall rather than the cap region, with details depending on chirality and strain as developed in the Results section and figures.

Comparisons. The article frames cap versus tube etching sequences against available theoretical and experimental evidence on hydrogen-driven restructuring of carbon nanostructures (see Introduction and Discussion in the Carbon article).

Sensitivity and design levers. Temperature (1600 K in the simulations), cap versus tube morphology, defectiveness of the cap, chirality, and strain in the tube scenarios all shift where hydrogenation, carbon loss, and intermediate motifs (nanosheet, nanowall, polyyne-like fragments) appear in the reported trajectories.

Limitations and outlook (as authored). The model isolates atomic H rather than a real plasma composition; rates are illustrative of mechanistic pathways, not quantitative plasma replicas. The virtual substrate and nanocluster construction simplify supports, mass transport, and catalyst restructuring relative to experiment.

Corpus / PDF honesty. This page is grounded in the peer-reviewed Carbon text and the short local extract; quantitative kinetics and any SI-only protocol refinements should be checked in the full PDF/SI if operators extend beyond the indexed pages.

Limitations

  • Only atomic hydrogen is injected, not a full plasma mixture (ions, electrons, radical diversity), so rates and selectivities are mechanistic rather than quantitative plasma replicas.
  • 1600 K is far above many CVD operating windows; it accelerates chemistry and may access rearrangements less relevant at growth temperatures.
  • Cluster + virtual substrate models simplify real supports, gas transport, and catalyst restructuring during long experiments.

Relevance to group

Antwerp PLASMANT work on ReaxFF + tfMC for plasma–carbon etching; complements the knowledge base’s reactive carbon and CNT literature without van Duin group authorship.

Citations and evidence anchors

  • DOI: 10.1016/j.carbon.2017.03.068
  • Text pointers: normalized/extracts/2017khalilov-carbon-118-2-atomic-scale-mechanisms_p1-2.txt; article figures for stage-resolved cap and tube scenarios.

Reader notes (navigation)