Skip to content

Impact of three-body interactions in a ReaxFF force field for Ni and Cr transition metals and their alloys on the prediction of thermal and mechanical properties

Scope

Adds three-body (valence-angle) terms to Ni/Cr ReaxFF, fixing C\(_{12}\)≠C\(_{44}\) elastic behavior and stacking-fault energy for fcc Ni; melting via hysteresis MD; Ni\(_3\)Cr alloy stability vs DFT.

Summary

Standard ReaxFF metal descriptions without explicit three-body angle terms yield equal C\(_{12}\) and C\(_{44}\) for fcc nickel and a negative stacking-fault energy that spuriously accelerates fcc→hcp transformation. This work augments the Ni and Cr ReaxFF parameter sets with three-body interactions for the transition metals, reproducing experimental elastic constants for fcc Ni and bcc Cr at finite temperature. Temperature-dependent lattice expansion from 0 to 1700 K follows expected trends (higher expansion for Ni versus Cr). Melting points from hysteresis molecular dynamics with large cells and voids land near 1698 K for Ni (~1.7% of experiment) and 2410 K for Cr (~10%). A combined Ni/Cr alloy parameterization predicts, consistently with DFT, negative heat of formation for fcc-like Ni\(_3\)Cr and positive heats for bcc solid solutions, suggesting phase separation tendencies. Elastic constants at 0 K match DFT within about 5–10% except for a larger deviation in C\(_{33}\) as noted in the abstract.

Methods

1 — MD application (LAMMPS, ReaxFF). Molecular dynamics in LAMMPS with the fitted ReaxFF for fcc Ni and bcc Cr and Ni–Cr alloys uses 3D PBC supercells containing 10⁴+ atoms in void-containing melting setups. NVT trajectories with a thermostat cover thermoelastic properties and hysteresis heating/cooling melting; lattice expansion vs temperature (0–1700 K) and T\(_m\) (≈1698 K for Ni, 2410 K for Cr in the reported hysteresis runs) are extracted from the article. Timestep (fs), equilibration and production lengths in ps–ns, and optional NPT / Parrinello–Rahman barostat use are in Comput. Mater. Sci. N/A — external electric field; N/A — umbrella in the main metal benchmark suite.

2 — Force-field training. The ReaxFF for Ni/Cr is augmented with valence angle (three-body) terms; the parrex-style parameter fit optimization targets DFT reference data (elastic tensors at 0 K, stacking fault energies, equation-of-state-type reaction subsets) and experimental validation for melting/elastic behavior. EEM-style QEq with a ~5 Å bond cutoff is discussed versus EAM-style metal potentials in the introduction.

3 — Static QM (validation). DFT supplies 0 K elastic data and alloy formation energies compared to the fitted ReaxFF; C\(_{33}\) shows a larger benchmark error in some cases.

4 — Experiments. N/A — the paper compares to published experimental melting and elastic data for metals from the literature, not a new laboratory program.

Findings

  • Three-body terms remove the unphysical C\(_{12}\)=C\(_{44}\) degeneracy and improve stacking-fault energetics for Ni relative to the prior ReaxFF treatment described.
  • Predicted thermal expansion and melting temperatures align semiquantitatively with experiment under the hysteresis setups employed.
  • Ni/Cr alloy energetics from ReaxFF follow the DFT trend favoring ordered fcc Ni\(_3\)Cr over disordered bcc mixtures in the tests shown.

Comparisons and sensitivity. ReaxFF is compared to DFT and experimental elastic/melting data; temperature-dependent lattice response and hysteresis-based T\(_m\) are central levers. Corpus: confirm C\(_{33}\) and any NPT usage in the full PDF.

Limitations

Remaining discrepancies (e.g., C\(_{33}\)) and melting hysteresis sensitivity to simulation setup are discussed in the article.

Relevance to group

Core ReaxFF metals methodology from the group extending transferable metal descriptions for alloy and corrosion workflows.

Citations and evidence anchors

Reader notes (navigation)

These sections summarize what the checked-in extraction and abstracts support; they are not a substitute for the full PDF. For theme-level retrieval, see paper-index-by-domain and hubs linked from canonical_tags in the front matter above. Operators updating chunks should reconcile numbers with normalized/extracts/ and the version-of-record PDF when available.