Skip to content

Molecular adsorbate effects on graphite–silica superlubricity: A ReaxFF investigation

Summary

Graphite exhibits ultra-low friction on the basal plane, but ambient adsorbates are known to degrade superlubricity. This study uses ReaxFF-based reactive molecular dynamics of a graphite basal surface sliding against silica to compare phenol, pentanol, and water as interfacial adsorbates. Three different ReaxFF parameter sets—each trained with emphasis on friction-related observables—are compared to show that the optimization objective strongly affects predicted tribological behavior. Beyond sliding simulations, adsorption and binding-energy analyses are combined with a quantitative measure of interfacial roughness to test whether friction tracks interfacial structure more closely than binding energy or adsorbate commensuration with graphene.

Local corpus text is a short front-matter extract; load, velocity, thermostat, and cell construction details should be confirmed from the full PDF or supporting information.

Methods

  • Interactions: ReaxFF (bond-order reactive force field) with three distinct parameterizations optimized using friction-relevant training targets.
  • System: Graphite–silica sliding interface with phenol, pentanol, or water as adsorbates.
  • Simulations: LAMMPS reactive Molecular dynamics of sliding friction at set temperature in K (value in SI); normal load and sliding directionality (relative to the graphite lattice) sweeps in PBC in-plane periodic slab supercells for graphite–silica contact (full lattice vectors in SI).
  • Analysis: Adsorption and binding-energy calculations; quantitative interfacial roughness metric for each adsorbate, related to computed friction coefficients. N/A on this page: LAMMPS (or other) version, NVE/NVT label, time step (fs), ps/ns sliding duration, PBC cell vectors and atom counts, thermostat/barostat details, and GPa/bar normal pressure protocol (see Friction article/SI). N/A — replica exchange in the rare-event sense; N/A — external E-field beyond the load/contact pressure already noted in the normal load sweeps.

Findings

  • The choice of ReaxFF training objective materially changes predicted coefficients of friction and overall tribological response for the same nominal interface chemistry.
  • Normal load and sliding directionality modulate friction in ways consistent with the parameter-set comparison.
  • Friction trends align more directly with the inferred interfacial molecular structure / roughness than with binding energy or simple commensuration arguments alone; the roughness-based interpretation accounts for the calculated friction coefficients across adsorbates. Comparisons across the three ReaxFF parameter lines are internal to the manuscript; use the PDF for absolute μ values.

Limitations

Galley PDF in corpus; extract does not preserve full simulation protocol tables. ReaxFF accuracy remains tied to the chosen training data and may not capture all long-range polarization or quantum effects relevant to some oxide–carbon contacts.

Relevance to group

Adri C. T. van Duin is a co-author; work demonstrates ReaxFF parameter sensitivity in tribology and environment-dependent superlubricity on graphitic carbon.

Citations and evidence anchors

Reader notes (navigation)

  • The corpus uses a galley pdf_path; prefer the final Friction layout PDF for exact figure and parameter tables when available.
  • paper-index-by-domain