Skip to content

Hydrogen Storage in Palladium Hollow Nanoparticles

Just Accepted PDF

The checked-in file is an ACS Just Accepted manuscript (web 26 Sep 2016). ACS boilerplate warns that copy-editing, formatting, and figure resolution may still change before the version of record.

Summary

Classical molecular dynamics in LAMMPS explores hydrogen uptake in palladium hollow nanoparticles (hNPs) built as fcc Pd ⟨100⟩ shells with spherical voids. EAM (Zhou parametrization) treats Pd–Pd mechanics and initial H loading, while gas-phase H\(_2\) chemistry uses ReaxFF (Senftle parametrization) because EAM alone allows unphysical H clustering. Atomic H is inserted into the cavity every 100 fs to mimic rapid charging. Simulations report hydride formation on the inner wall, subsequent H\(_2\) generation inside the void, rising intracavity pressure (up to ~7 GPa before mechanical failure in the narrative of the abstract), and maximum H/Pd ratios of ~1.21 (cavity-only loading) versus ~1.70 when H is also supplied to the outer surface—about 25% larger than earlier reports cited in the abstract—followed by particle rupture and H\(_2\) ejection when loaded beyond those regimes.

Methods

1 — MD application (classical + reactive segments). Engine: LAMMPS. Boundary conditions: three-dimensional periodic boundaries on the simulation boxes described in Methods (bulk-like embedding of each hNP). Pd skeleton: EAM potential with Zhou parameters; hNPs are carved from bulk fcc Pd ⟨100⟩ with external radii \(6a_0 \le R_\mathrm{ext} \le 60a_0\) (\(a_0 = 3.89\) Å) and shell thicknesses \(a_0 \le w \le 6a_0\) as stated in Methods. Initial relaxation: alternation of FIRE and conjugate-gradient minimization. Thermal ramp: 0 → 300 K in steps of 20 K every 0.2 ns with a Nosé–Hoover thermostat and 1 fs timestep for the stability survey. Hydrogen charging: single H atoms added to the cavity every 100 fs (kinetic protocol motivated in the article). H\(_2\) gas treatment: ReaxFF (Senftle parameters) replaces EAM for H–H interactions when molecular hydrogen forms, as EAM is noted to cluster H unphysically in the gas phase. Ensemble labels for all production segments: follow the manuscript’s staged protocols (NVT-like thermal control is explicit for the 300 K ramp; subsequent uptake stages are described sequentially in Methods—read the PDF for any additional ensemble switches not summarized here). Barostat for charging runs: N/A — not emphasized in the excerpted protocol text reviewed for this page. Pressure control: N/A — no external hydrostatic barostat; intracavity H\(_2\) pressures reported in Results emerge from inserted H loading. Electric field: N/A — not used. Replica / enhanced sampling: N/A — not used.

2 — Force-field training. N/A — the study uses published EAM and ReaxFF parameterizations.

3 — Static QM. N/A — not a DFT paper.

4 — Experimental comparison. Stability diagrams are compared against cited TEM/synthesis literature for hNP dimensions (discussed around Fig. 1).

Findings

Outcomes. MD predicts three stability classes for hollow geometries (stable, half-stable, collapsed) versus \(R_\mathrm{ext}\) and \(w\) when ramped to 300 K, bracketed against experimental size/thickness claims in the article’s stability diagram. H insertion first builds a Pd hydride layer on the inner surface, lowers the uptake rate, and eventually yields H\(_2\) gas inside the cavity with pressures peaking near 7 GPa prior to mechanical breakdown (abstract narrative). Maximum H/Pd reaches ~1.21 for inner-only loading and ~1.70 when both surfaces are charged (25% higher than prior computational reports quoted in the abstract), beyond which the hNP fractures and ejects gas.

Comparisons. Simulation stability regions are mapped against multiple experimental datasets referenced in the Results discussion (e.g., 7 nm and 40 nm examples).

Sensitivity / levers. Response depends strongly on wall thickness, outer radius, and whether H accesses one or both surfaces.

Limitations (authored). Inner-only insertion is a computational expedient versus experimental external charging; classical models omit quantum H effects and electronic stopping relevant to some beam experiments (N/A for this article’s focus).

Corpus honesty. This page was drafted from the Just Accepted PDF at pdf_path; prefer the typeset VOR for pagination and final figure labels when available.

Limitations

  • Just Accepted text may differ slightly from the issue PDF.
  • Charging protocol is idealized (discrete H insertions); map numerical pressure/H/Pd peaks to figures in the article before reuse in benchmarks.

Relevance to group

Adjacent corpus material for Pd–H nanostructures and mechanical failure under hydride formation—useful when contrasting EAM metal models with ReaxFF covalent gas chemistry in multiphysics workflows.

Citations and evidence anchors

  • reaxff-family
  • Palladium hydrides and hydrogen storage (see primary PDF for quantitative tables)