Analysis of a Li-Ion Nanobattery with Graphite Anode Using Molecular Dynamics Simulations

Abstract

Classical MD in LAMMPS models a full nanoscale Li-ion cell: B3PW91-parameterized EC/LiPF₆, UFF+LJ mixing, Tersoff graphite, composite cathode/SEI/insulator regions, external electric fields for charging—reports energies, dipoles, drift, conductivity vs field.

Summary¶

Classical MD with LAMMPS models a Li-ion nanobattery: graphite anode, LiCoO₂ cathode, 1.15 M LiPF₆ in ethylene carbonate electrolyte, Cu- and Au-like collectors, SiOₓ-like insulator, optional SEI (Li₂CO₃-like stoichiometry per article). Packmol builds electrolyte starting coordinates (21 ion pairs, 310 EC molecules in 1.15 M cell excerpt). Force fields: CFF93 base with Gaussian-09 B3PW91/cc-pVTZ-fitted bond/angle/dihedral parameters for EC and LiPF₆; UFF LJ mixing rules with pair-specific adjustments to prevent unphysical penetration; Mulliken/NBO charges for Coulombics; Tersoff (Lindsay et al.) for in-plane graphite plus LJ interlayer; cathode/electrode regions parameterized as described in Tables 1–3. Equilibration: 1 ns at 5 K, ramp to 293 K in 4 ns, 10 ns NVT; charging applies external electric fields along the cell (0.2–1.5 V/Å range in reported runs, with varying simulated duration). Observables include energy, temperature, volume, polarization, MSD, Li flux to anode, conductivity/resistivity vs field, and SEI permeation behavior. The setup is intentionally a nanoscale testbed with large effective fields relative to macroscopic cells.

Methods¶

1 — MD application (atomistic dynamics)¶

Classical MD in LAMMPS models a nanoscale Li-ion cell with graphite anode, LiCoO₂ cathode, Cu-/Au-like current collectors, a SiOₓ-like insulator region, and 1.15 M (also 3.36 M) LiPF₆ in ethylene carbonate electrolyte built with Packmol (example stoichiometry in the article: 21 LiPF₆ ion pairs with 310 EC molecules for the 1.15 M cell caption). Periodic boundary conditions apply to an initial orthorhombic cell 166.6 × 27.0 × 58.2 Å³ (Fig. 1). Integration uses the Verlet leapfrog algorithm. Equilibration follows 1 ns at 5 K, a 4 ns ramp to 293 K, then 10 ns NVT at 293 K before charging trajectories that impose external electric fields along the cell axis (field strengths, polarization diagnostics, and segment lengths are tabulated in the article body beyond the indexed p1–2 excerpt).

Engine / code: LAMMPS classical MD with Verlet integration (J. Phys. Chem. C 2017, 121, 12959–12971).
System size & composition: Graphite (Tersoff + interlayer LJ), LiCoO₂ cathode block, EC + LiPF₆ electrolyte, optional SEI region modeled with a Li₂CO₃-like stoichiometry, and AX₂-like insulator spacer—full parameter tables in Tables 1–3 of the article.
Boundaries / periodicity: 3D PBC on the nanobattery supercell (Fig. 1b–c).
Ensemble: NVT for the 10 ns 293 K equilibration segment described in the figure caption; charging stages continue under field-driven dynamics as detailed in Methodology (see PDF for thermostat specifics beyond the indexed excerpt).
Timestep: N/A — integration Δt (fs) is not stated on the indexed p1–2 pages; confirm in the full Methodology section.
Duration / stages: 1 ns at 5 K, 4 ns heating ramp to 293 K, 10 ns NVT equilibration at 293 K, followed by field-on charging segments of lengths reported with each field strength in the article.
Thermostat: N/A — explicit thermostat identity/damping not reproduced from the indexed excerpt; confirm in the PDF.
Barostat: N/A — equilibration segment quoted is NVT without NPT barostat control in the indexed caption.
Temperature: 5 K start, ramp to 293 K equilibration target as in Fig. 1 caption.
Pressure: N/A — isochoric NVT equilibration; no barostat summary on indexed pages.
Electric field: Static external fields along the cell axis mimic voltage sources during charging; field magnitudes and polarization diagnostics are central observables in the article.
Replica / enhanced sampling: N/A — brute-force MD with applied E-field.

2 — Force-field training / classical parameterization¶

Non-reactive classical potentials: intramolecular EC parameters from Gaussian B3PW91/cc-pVTZ optimizations (Table 1 excerpt); LiPF₆ bonded terms similarly DFT-fitted; LJ cross terms largely UFF-based with pair-specific adjustments to avoid unphysical overlap; Mulliken/NBO-derived partial charges for Coulombics; Tersoff (Lindsay et al.) in-plane graphite plus LJ interlayer; cathode/current-collector regions use the literature parameterization summarized in Tables 1–3. Not a ReaxFF reparameterization study.

3 — Static QM / DFT-only¶

B3PW91/cc-pVTZ DFT supplies bond/angle/dihedral reference data for EC (and related) intramolecular terms; not AIMD production.

Findings¶

Outcomes and mechanisms¶

Charging under applied electric fields drives Li⁺ migration from cathode toward graphite, with time traces for energy, temperature, volume, polarization, MSD, and Li flux reported in the article. Polarization of the solvent rises approximately linearly with accumulated charge until a saturation regime where additional charging stalls because energy input from the external source drops to very low levels—interpreted as the nanobattery reaching an effectively fully polarized electrolyte state.

Comparisons¶

The authors compare simulated conductivity / resistivity, current density, and related transport metrics against available experimental conductivity data for EC/LiPF₆ solutions, reporting qualitative consistency within the limitations of the classical model.

Sensitivity / design levers¶

Electrolyte concentration (1.15 M vs 3.36 M) and external field magnitude modulate Joule-like heating, Li⁺ drift, and conductivity; at high fields the 3.36 M cell shows stronger heating than 1.15 M under otherwise similar conditions (article discussion).

Limitations and corpus honesty¶

Model is classical and non-reactive—no bond-making/breaking for SEI chemistry beyond the simplified Li₂CO₃-like layer used to probe Li⁺ permeation. Field strengths are nanoscale testbed magnitudes and are not direct macroscopic cell voltages. Indexed excerpt covers early methodology pages; quantitative tables/figures should be taken from the DOI-linked PDF when citing numbers beyond this summary.

Limitations¶

Classical non-reactive force fields—no bond-breaking electrolyte decomposition; fields are large compared to laboratory cells (nanoscale testbed). Charge transfer beyond fixed partial charges and electrochemical polarization at true voltages are not represented at full QM fidelity.

Relevance to group¶

Non-ReaxFF battery MD reference in corpus; contrasts with reactive workflows elsewhere.

Citations and evidence anchors¶

DOI: 10.1021/acs.jpcc.7b04190

batteries-interfaces-reaxff