Responses of core–shell Al/Al\(_2\)O\(_3\) nanoparticles to heating: ReaxFF molecular dynamics simulations
Summary¶
Reactive MD with ReaxFF in LAMMPS follows core–shell Al / amorphous Al\(_2\)O\(_3\) nanoparticles heated to 2000 K to probe oxidation, melting, and mass ejection relevant to Al combustion. Aluminum cores ~4 nm are coated with amorphous alumina shells of thickness δ = 0.6–1.0 nm (~2334 Al atoms in fcc core; ~14.8k atoms total for δ = 1 nm, ~6.6 nm diameter) with an initial ~0.3 nm void between core and shell. Simulations show oxygen in-diffusion through the shell, Al migration, alumina melting (~1153 K for 1 nm shell, higher for thicker shells), and ejection of small (AlO)\(_n\) clusters (n ≈ 3–5) interpreted as nano-detonation-like events. Mean-square displacement analysis highlights O motion as the early rate-limiting ingress step. The article also discusses how electric fields may couple to diffusion-limited steps relevant to ignition of passivated Al powders—connecting atomistic transport to practical combustion contexts beyond thermal heating alone. Readers should treat cluster ejection counts and melting temperatures as model-dependent outcomes that must be cross-checked against the full results section for exact numerical thresholds and statistical sampling windows.
Methods¶
Force-field training (ReaxFF Al/O). The study uses published ReaxFF parameters for Al/O chemistry from Shin et al. and Rahaman et al. with QEq-style dynamic charges, as cited in J. Phys. Chem. C; N/A — new parameter fit in this paper — the work applies an existing parametrization to core–shell nanoparticle heating.
MD application (core–shell nanoparticle heating). Simulations are ReaxFF molecular dynamics in LAMMPS on finite, nonperiodic core–shell Al / amorphous Al\(_2\)O\(_3\) nanoparticles (order 10⁴ atoms for a ~1 nm shell: ~2334 fcc Al core atoms and ~14.8k atoms total in the abstract’s δ = 1 nm example). Ensemble: NVT with a Nosé–Hoover thermostat (20 fs damping). Timestep: 0.2 fs. Thermal protocol: relaxation near 300 K, then a 100 K/ps ramp to 2000 K with a 100 ps hold at the peak temperature (staging and any equilibration edits must be confirmed in pdf_path). Barostat: N/A — NPT — isolated NP models are heated at constant volume in NVT. Pressure: N/A — target stress (bar/GPa) — not imposed beyond thermostat control on the finite cluster. Electric field: the article discusses electric-field–coupled diffusion as ignition context for passivated Al powders; N/A — explicit EFIELD-driven MD protocol in the indexed summary unless the full PDF documents biased dynamics separate from the thermal ramp. Shear / shock: N/A. Replica / enhanced sampling: N/A — umbrella sampling, metadynamics, or replica exchange for these trajectories.
Analysis. Mean-square displacement and related diagnostics compare O versus Al transport; melting and cluster ejection thresholds are tracked versus oxide shell thickness δ.
Experiments. N/A — wet-lab experiments — the publication is computational; experimental comparisons appear as literature context for Al combustion and passivation.
Findings¶
Outcomes and mechanisms. Inward O diffusion is the early rate-limiting ingress step relative to Al mixing; mean-square displacement analysis supports O motion ahead of extensive core reaction. (AlO)\(_n\) clusters (n ≈ 3–5) eject at later stages, interpreted in the article as nano-detonation-like events. Alumina melting is reported near 1153 K for a 1 nm shell, with higher apparent melting temperatures for thicker δ. A void opens when the shell melts, discussed alongside experimental evidence cited in the article.
Comparisons and sensitivity. Interface diffusivities during the ramp depend weakly on shell thickness δ, whereas post-heating diffusivities decrease as δ increases. Trends are framed relative to prior diffusion-limited versus melt-dispersion ignition pictures in the literature reviewed in the paper.
Limitations and corpus honesty. ReaxFF training scope for Al/O and rapid 100 K/ps heating may omit slower quasi-equilibrium pathways. Quantitative reuse should follow cluster definitions and sampling windows in the peer-reviewed PDF (pdf_path); the local extract normalized/extracts/2018huadong-zeng-j-phys-chem-responses-core_p1-2.txt aligns with the abstract but is not a substitute for Results tables.
Limitations¶
ReaxFF training scope for Al/O; rapid heating may skip quasi-equilibrium pathways; nonperiodic NP models omit bulk oxidation environments. Cluster ejection diagnostics depend on visual thresholds for (AlO)\(_n\) fragments—confirm cluster definitions in the article before quantitative reuse.
Relevance to group¶
Al combustion / core–shell oxidation ReaxFF application adjacent to energetic Al literature and nano-aluminum ignition studies. The nonperiodic nanoparticle setup mirrors other finite-cluster oxidation simulations in the corpus where surface-to-volume ratios dominate burn phenomenology.
Citations and evidence anchors¶
- DOI:
10.1021/acs.jpcc.8b01088—papers/ReaxFF_others/Zeng_AlO_JPCC_2018.pdf; extractnormalized/extracts/2018huadong-zeng-j-phys-chem-responses-core_p1-2.txt.