Atomistic Insights into Nucleation and Formation of Hexagonal Boron Nitride on Nickel from First-Principles-Based Reactive Molecular Dynamics Simulations

Evidence and attribution¶

Authority of statements

Prose below summarizes the publication identified by doi, title, and pdf_path in the front matter. For exact barriers, supercell dimensions, and run schedules, use the peer-reviewed article (and Supporting Information where cited).

Summary¶

Multiscale modeling combines periodic DFT and ReaxFF-based reactive MD to study how a continuous atomically thin hBN lattice forms from elemental B and N on Ni substrates. DFT targets adsorption and reaction energetics for B, N, and small B\(_x\)N\(_y\) (x, y = 1, 2) on Ni(111) and Ni(211), and diffusion paths for B and N on terraces and in the Ni sublayer. The work reports that B can diffuse competitively on the surface and in the sublayer, whereas N diffuses on the surface only. Those QM data inform Ni–B and Ni–N terms within ReaxFF. Reactive MD then resolves elementary nucleation and growth of an hBN monolayer from deposited B and N; nucleation proceeds from linear BN chains toward branched and hexagonal motifs, with additional DFT used in the paper to check intermediate structures and the consistency of the DFT + ReaxFF workflow. The framing ties this picture to crystal quality, temperature, and substrate effects in CVD-relevant growth.

Methods¶

1 — MD application (LAMMPS rMD on Ni slabs). Reactive molecular dynamics of hBN growth from elemental B and N uses LAMMPS. The Ni(111) substrate is a rectangular five-layer slab with 12×12 surface atoms (720 Ni atoms). Periodic boundary conditions apply in all directions; ~90 Å vacuum separates periodic images along the surface normal. The bottom (fifth) Ni layer is fixed to mimic bulk, while the top four layers relax vertically to represent surface + sublayer. Equal numbers of B and N atoms (200 each) are deposited sequentially in pairs from the gas phase at random (x,y) positions every 0.25 ps, with a minimum initial B–N separation ≥1.90 Å to suppress premature bond formation; deposition targets the relaxed top side only. Production runs explore 900, 1100, 1300, and 1500 K using a Nosé–Hoover thermostat (cited thermostat reference in the article). Each reported trajectory is ≥6 ns with Δt = 0.25 fs and velocity Verlet integration. N/A — barostat / NPT during growth: the published setup describes thermostatted rMD on a slab + vacuum geometry without an explicit NPT barostat for the growth segment summarized here. N/A — external electric field in the rMD protocol description.

2 — Force-field training / fitting. Periodic DFT on Ni(111) and Ni(211) supplies adsorption, diffusion, and small B\(_x\)N\(_y\) reaction data used to develop Ni–B and Ni–N terms within ReaxFF; the full parameter table is pointed to Supporting Information in the article.

3 — Static QM / DFT. Follow-up DFT checks selected intermediates along the rMD growth pathway for energetic consistency (multiscale validation described in the Abstract and Results).

4 — Review / non-simulation framing. N/A: primary ACS Nano study. Galley duplicate PDF: [[2017liu-venue-proof-2-pdf]].

Findings¶

Outcomes and mechanisms. DFT shows B can diffuse on the surface and in the Ni sublayer, whereas N diffuses on the surface only—this asymmetry motivates different transport roles during growth. rMD shows nucleation beginning from linear BN chains, progressing to branched motifs and ultimately hexagonal hBN patches; subsequent DFT on intermediate structures supports the energetic plausibility of the ReaxFF trajectory class and the authors’ self-consistency argument.

Comparisons. The work is positioned relative to prior DFT studies of hBN/precursor chemistry on transition metals and the need for larger-scale dynamics than QM alone allows.

Sensitivity and design levers. Temperature (900–1500 K in the rMD campaign), deposition cadence, and substrate facet ((111) slab focus in the excerpted setup) are the main knobs discussed for crystal quality and growth morphology.

Limitations and outlook (as authored). The ASAP PDF may differ slightly from the final paginated issue; use the version-of-record for pagination when citing figures.

Corpus / PDF honesty. Protocol details above are taken from the ingested ASAP PDF text; the short local extract alone is insufficient for reproduction—prefer pdf_path.

Limitations¶

ASAP PDFs can differ slightly from the final paginated ACS Nano issue; for pagination and figure numbering in citations, prefer the journal version of record once available in your library. Hydrogen and realistic precursor chemistry in CVD are not the main focus of the abstract-level summary—see the full text for scope.

Relevance to group¶

Adri C. T. van Duin and Diana M. van Duin (RxFF Consulting) co-authored the Ni/B/N ReaxFF development and hBN-on-Ni rMD study with Kansas State collaborators; this is a core corpus bridge between 2D hBN growth and ReaxFF lineage work (see also [[2019song-liu-nanoscale-20-predicting-preferred]]).

Citations and evidence anchors¶

DOI: 10.1021/acsnano.6b06736 — PDF: papers/Liu_ACS_Nano_BN_Nickel_2017_ASAP.pdf.
Duplicate galley PDF: [[2017liu-venue-proof-2-pdf]].

Non-primary sibling: [[2017liu-venue-proof-2-pdf]] registers the galley PDF for the same DOI.