Skip to content

A ReaxFF force field for sodium intrusion in graphitic cathodes

Summary

A Na–C (graphitic) ReaxFF parameter set is trained against DFT/PBE data on potential energy surfaces, geometries, and charge-related observables, then applied to hybrid grand-canonical Monte Carlo / molecular dynamics (GC-MC/MD) simulations of sodium intrusion in simplified graphitic cathode models motivated by aluminum electrolysis wear and sodium-ion battery adjacency. The PCCP article frames Na chemical potential control as essential for modeling intercalation versus pore filling in disordered carbon microstructures.

Methods

Force-field training (Na–C ReaxFF): QM reference uses DFT/PBE data for Na adsorption and intercalation on PAH models (circumcoronene, stacked coronene pairs, etc.), including gradient-rich PES regions weighted heavily in the error function. Optimization uses Metropolis Monte Carlo moves on ReaxFF parameters with simulated annealing cycles minimizing a weighted least-squares mismatch to QM targets (Table 1 in PCCP). EEM bond-order/charge parameters are trained against MDC-q reference charges on selected Na/PAH configurations.

MD application (GC-MC/MD + ReaxFF): After fitting, hybrid grand-canonical Monte Carlo / MD imposes Na chemical potential on graphitic cathode granule models with distinct ordered vs pore-like inter-domain carbon (PCCP §4 and figures). Timestep, thermostat/barostat, move acceptance rules, and trajectory lengths: N/A — not transcribed numerically here; see article/SI.

Static QM / DFT: Training DFT/PBE settings (dispersion, basis, k-point conventions) are in the PCCP computational section (N/A — not copied into this summary).

GC-MC/MD supercells: Na insertion cells use 3D PBC LAMMPS-style granule models whose composition sets porosity. Imposed bulk hydrostatic pressure during uptake: N/A — chemical potential is the primary control in the summarized protocol.

Findings

  • The fitted force field reproduces training PES features with reasonable accuracy while acknowledging EEM limitations for steep charge vs separation trends.
  • GC-MC/MD applications illustrate Na uptake behavior in graphitic model systems consistent with the article’s qualitative transport picture (see figures for density profiles and insertion statistics).
  • Discussion ties intercalation vs pore diffusion mechanisms to domain type within the cathode microstructure model.

Limitations

  • EEM charges are approximate; authors caution on long-range charge transfer artifacts (mitigated somewhat for alkali + PAH cases).
  • Structural disorder of real industrial cathodes is only coarsely represented.

For density profiles, insertion statistics, and move probabilities in GC-MC/MD, use papers/ReaxFF_others/Hjertenaes_PCCP_2016_Na_graphite.pdf / SI rather than this summary alone.

Relevance to group

ReaxFF parameterization for Na + graphitic carbon overlaps thematically with battery interface and carbon wear interests; complements Li-focused ReaxFF lines elsewhere in the corpus.

Citations and evidence anchors