Skip to content

Interfacial Bonding Controls Friction in Diamond-Rock Contacts

Summary

Macroscale tribometry compares diamond tips on limestone (calcite-rich) versus granite (quartz-rich) in humid air and water, while nonequilibrium molecular dynamics with newly trained ReaxFF parameters for calcite-related chemistry links friction coefficients to rates and types of interfacial bond formation during sliding.

Methods

MD application (nonequilibrium LAMMPS)

  • Engine / code: NEMD in LAMMPS; velocity Verlet; 0.25 fs timestep.
  • System size & composition: Diamond tip and rock (calcite- vs quartz-dominated) substrate models; interfacial H₂O content varied (0 / 50 / 150 molecules) as reported in the figures. Maximum Hertz-scale contact pressures for the scanned loads span about 1.9–3.9 GPa (diamond–calcite) and 2.0–4.1 GPa (diamond–quartz).
  • Boundaries / periodicity: Periodic in x and y; bottom substrate layer(s) frozen; a reflective boundary to retain desorbed species. N/A—full lateral cell vectors and atom counts: see pdf_path for the reported supercells.
  • Ensemble / control: fix langevin-style Langevin thermostat at 300 K (about 25 fs damping) on selected tip/substrate layers, consistent with a driven NEMD (non-NVE) sliding setup. N/A—NPT barostat: inhomogeneous, load-controlled contact rather than bulk isotropic pressure control in the same sense as NPT bulk equilibration.
  • Equilibration / production: Minimization; 0.1 ns equilibration with 0.1 nN normal load ramp, then the target normal load (about 2.5–40 nN) and vₓ = 10 m·s⁻¹ sliding in the main dataset (1–10 m·s⁻¹ check shows similar friction trends in the text). 0.75 ns sliding trajectories; time-averaged steady friction from the last ~200 ps. Bond order tracked at 1.0 ps intervals (cutoff 0.3 in the author protocol).
  • Sliding vs experiment: Shear in MD at m·s⁻¹ speeds, far above the tribometry order (~0.1 m·s⁻¹); the article discusses scaling/trends accordingly.
  • Long-range / electrostatics in MD: N/A—full cutoff/ReaxFF QEq schedule in pdf_path/SI if needed for reproduction.

Experiments (tribology)

  • Point-contact tribometry on granite vs limestone; diamond indenter/flat-on-rock geometry; humid and water-immersed conditions with friction vs load reporting.

Force-field training (ReaxFF and VASP for calcite)

  • Parent / scope: ReaxFF extension for calcite-related O/Ca/O interactions combined with pre-existing quartz/water subsets in the ReaxFF formalism, as cross-referenced in the article.
  • QM / DFT (VASP): PBE pseudopotentials (PAW), planewave cutoff about 520 eV, spin-polarized, k-mesh for calcite and equations of state training points.
  • Training set and optimization: Bulk and reaction-relevant data for calcite; targeted ReaxFF reoptimization of key off-diagonal and valence terms, then condensed-phase (equation of state, heats) checks as described in the paper. CMA-ES-style/standard ReaxFF optimization is referenced in the J. Phys. Chem. C text.

Findings

Experiments vs composition: In humid/water-rich conditions, limestone-bearing contacts show higher friction than granite-bearing ones despite the lower hardness ranking—K-family roughness/phase identity in the main text is part of the experimental comparison picture (see JPC C and figures).

NEMD trends: In dry NEMD, the diamond–calcite/quartz trend can differ from the humid experiment ordering; with interfacial water in the simulation, diamond–calcite can recover higher friction than diamond–quartz, which the paper ties to more persistent interfacial bonds in the calcite system (C–O–Ca-like and H-bond networks).

Kinetics and scaling: Bond-formation rates are reported as exponential in the effective contact stress/pressure with small activation volumes, and time-averaged friction can scale about linearly with the interfacial bond count under steady sliding. At >~2 GPa contact pressure (both MD and experiments with water), the friction–load curve shows superlinear growth that the paper associates with a bonding-mediated friction regime. Caveat: the corpus pdf_path is a galley-style PDF; verify figure/page alignment against a version-of-record copy for citation-grade numbers (see ## Limitations).

Limitations

The corpus PDF is a galley proof (header shows placeholder page/volume in places). Simulation sliding speeds greatly exceed laboratory speeds; the rigid projectile-like tip model omits diamond wear chemistry.

Relevance to group

A. C. T. van Duin and N. Dasgupta are co-authors; ReaxFF parameterization for calcite interfaces supports the tribology study.

Citations and evidence anchors