Development and validation of a ReaxFF reactive force field for Fe/Al/Ni alloys: molecular dynamics study of elastic constants, diffusion, and segregation
Evidence and attribution¶
Authority of statements
Prose below summarizes the publication identified by doi and pdf_path. The PDF is a galley; verify pagination against the version of record.
Summary¶
A ReaxFF parametrization for Fe/Al/Ni binary alloys uses QM training on bulk phases and (100)/(110)/(111) surface energies and adatom binding. Validation includes elastic constants (300–1100 K) for FeAl, FeNi₃, Ni₃Al (abstract: trends vs experiment), Al/Ni diffusivity in composition-gradient layers at 1000 K (abstract contrasts trace vs pure-metal ends), and surface segregation in L1₂ clusters at 2500 K (Al strongest in Fe₃Al, weakest in Ni₃Al; depletion layers adjacent to segregation).
Methods¶
Force-field training¶
Parent field / elements: ReaxFF for Fe/Al/Ni binaries, extending prior metallic alloy parameterization practice described in the article.
QM reference & training set: Fe–Al, Fe–Ni, and Ni–Al unlike-pair terms are fit to QM equation-of-state curves for B2, L1₂, DO₃, and L1₀ phases across 25–75 at.% composition sweeps, augmented by (100)/(110)/(111) surface energies and adatom binding data (Section III.A; papers/Yun_JPCA_2012_Alloys_galley.pdf / extract).
Optimization & validation split: Parameters are optimized against that QM training set; MD simulations below are validation trajectories of the fitted field.
MD application (atomistic dynamics)¶
Engine / code: ADF/ReaxFF for the reported workflows (Methods in pdf_path).
Elastic constants (finite-T): NPT molecular dynamics with Δt = 0.25 fs, Berendsen thermostat (100 fs coupling) and barostat (500 fs coupling). Supercells: (5×5×5) (500 atoms) and 45°-rotated 512-atom cells for L1₂ Ni₃Al and FeNi₃; (6×6×6) (432 atoms) plus a rotated 500-atom cell for B2 FeAl. >1 ps production averaging after equilibration (article Methods).
Diffusion: fcc Ni matrix in (4×4×11) unit cells with a linear Al composition gradient (n−1)/10 per layer, open x/z surfaces with ~30 Å vacuum, 1000 K, Berendsen thermostat 100 fs, Δt = 0.25 fs, 1 ns trajectory with sampling every 0.25 ps.
Surface segregation: (10×10×10) L1₂ supercells (4000 atoms) with ~11 nm vacuum padding; 300 K equilibration 5 ps; 300 → 2500 K ramp at 0.05 K/step; 2500 K hold 60 ps; cool to 300 K at −0.0125 K/step.
Boundaries / periodicity: 3D PBC in bulk elastic runs; slab-like open surfaces normal to z in the diffusion geometry; vacuum-padded clusters for segregation (Methods).
Ensemble: NPT for elastic constants; diffusion and segregation blocks use constant-temperature MD as specified (NPT where stated for elastic block; verify labels for diffusion in pdf_path).
Pressure: NPT elastic runs imply hydrostatic coupling; N/A — target pressure value not transcribed here—confirm in PDF.
Electric field / enhanced sampling: N/A — not used.
Findings¶
Outcomes: Cᵢⱼ(T) for FeAl, FeNi₃, and Ni₃Al soften with increasing T from 300 K to 1100 K, tracking experimental trends for C₁₁, C₁₂, and C₄₄ (abstract). Al diffusivity at 1000 K rises by roughly two orders of magnitude from trace layers to the Al-rich end of an Al/Ni gradient, whereas Ni diffusivity changes only modestly between trace and pure-Ni-like regions at T far below Ni melting (abstract). L1₂ Fe₃Al, Fe₃Ni, and Ni₃Al clusters at 2500 K show Al segregation strongest in Fe₃Al and weakest in Ni₃Al, with depletion zones adjacent to segregating species (Al in Fe₃Al/Ni₃Al, Ni in Fe₃Ni) as in the abstract’s summary.
Comparisons: Elastic and segregation patterns are compared to experimental literature cited in the article.
Sensitivity: Results depend on temperature, stoichiometry, and composition-gradient endpoint vs trace placement.
Limitations / outlook: Galley PDF (papers/Yun_JPCA_2012_Alloys_galley.pdf); confirm pagination against the version of record [[2012yun-venue-jp-2012-08507x]]. 2500 K segregation is an aggressive thermal stress test—interpret alongside experimental windows cited in the paper.
Corpus honesty: Numerics above are taken from the galley Methods text in-repo; if pdf_path is updated to the ASAP PDF only, re-verify identical protocol wording.
Limitations¶
Galley PDF; high-T segregation simulations are extreme compared to many laboratory conditions; ReaxFF alloy accuracy is limited by the QM training scope. For pagination and any publisher updates, compare with [[2012yun-venue-jp-2012-08507x]] (same DOI).
Relevance to group¶
Adri C. T. van Duin senior author line on metallic ReaxFF validation (NETL-PSU collaboration).
Citations and evidence anchors¶
- DOI 10.1021/jp308507x — J. Phys. Chem. A (volume/pages per final issue).
- PDF:
papers/Yun_JPCA_2012_Alloys_galley.pdf; extract:normalized/extracts/2012yun-venue-research_p1-2.txt.