Chemical and physical origins of friction on surfaces with atomic steps

Summary¶

Macroscale tribology often separates adhesion, roughness, and shearing, but at atomic steps on graphite the relevant physics mixes topography with contact chemistry. Chen, Khajeh, Martini, and Kim report atomic force microscopy (AFM) friction loops for a silica probe sliding on graphite that contains a single-layer graphene step (~0.34 nm height) under dry nitrogen. Terraces exhibit superlubric-like friction, whereas traversing the step edge—especially in step-up geometry—raises dissipation by about two orders of magnitude. Complementary ReaxFF simulations of the silica/graphite contact reproduce the qualitative separation between basal and step sliding and are used to partition lateral forces into mechanical versus bond-making/breaking contributions near the discontinuity. The Science Advances framing connects nanoscale superlubricity on atomically smooth basal planes to engineering contexts where step edges are unavoidable—CVD graphene, exfoliated flakes, or wear particles on graphitic substrates.

Methods¶

AFM experiments (dry N₂). Friction measurements use contact-mode AFM on tape-exfoliated ZYA-grade HOPG with naturally occurring single-layer graphene step edges (~0.34 nm height). Tests use a Si probe (vendor/nominal parameters are listed in the article), in a dry N₂ environment from a nitrogen generator with a reported dew point around −35 °C (corresponding to roughly 200–300 ppm water at atmospheric pressure in the article text). Normal loads span about 7.3–36.7 nN, with contact pressures estimated around ~1.6–2.8 GPa (Hertz) and ~1.9–2.9 GPa (DMT) as stated. Sliding speeds span about 0.25–2 µm/s at room temperature.

Reactive MD (ReaxFF + LAMMPS). Engine / code: simulations are run in LAMMPS with a ReaxFF description cited in the article. System / geometry: the model uses an amorphous silica tip constructed by melt–quench of cristobalite (4000 K melt, slow quench; 0.02 K/fs heating/cooling rate as quoted), shaped as a semicircular disc tip with reported radius/thickness/height (~2.5 / 1.5 / 1.5 nm), with hydroxyl/hydrogen passivation of undercoordinated surface sites and the top 0.5 nm of the tip treated as a rigid body. The graphite substrate includes an armchair monolayer step terminated with alternating –OH/H on the edge, with two mobile graphene layers on each terrace side to reduce cost while retaining step chemistry. Boundary conditions: during sliding, the article reports boundary constraints on the graphite substrate (side edges fixed in all directions as stated in Materials and Methods), consistent with a finite sliding cell setup described around fig. S2.

Ensemble / thermostat / temperature. NVT at ~300 K using a Langevin thermostat on all atoms that are not fixed or rigid.

Protocol stages. Each simulation includes: (i) minimization + equilibration with tip far from substrate; (ii) approach at 10 m/s until the lowest tip atom is 0.2 nm from the top substrate layer; (iii) apply normal load on the rigid tip region and equilibrate 120 ps; (iv) slide at 10 m/s along X using a harmonic spring (6 N/m stiffness). Loads explored include 5, 7.5, 10, 12.5, and 15 nN, mapping to roughly ~2.1–4.8 GPa contact pressure using the article’s real atomic contact area definition. Substrate boundaries: side boundaries of graphite are fixed during sliding. Post-processing: Fourier filtering is applied to noisy lateral-force traces; hydrogen-bond counting follows the Guàrdia et al. logic cited in the article.

Timestep. N/A — an explicit integration timestep was not located in the main-text Methods span extracted for this curation pass (confirm in Science Advances supplementary text if required for reproduction).

Barostat / bulk pressure coupling. N/A — the quoted protocol is NVT + loaded/slid nanoscale contact mechanics rather than bulk NPT servocontrol.

Electric fields / enhanced sampling. N/A — not part of the summarized MD workflow.

Findings¶

On basal regions the coefficient of friction remains near 0.003 for the cited load/adhesion decomposition—consistent with superlow friction on atomically smooth graphite. Step-up sliding yields μ ≈ 0.1, approaching boundary-lubricated organic friction despite dry N₂, highlighting how a single atomic defect can dominate dissipation. Step-down curves exhibit hysteresis and more complex stick–slip structure. The authors argue classical Prandtl–Tomlinson models with only geometric barriers (for example Ehrlich–Schwoebel-type step-crossing penalties) are insufficient without the chemical interaction channel that ReaxFF captures at the step edge. Practically, the paper motivates reactive contact models wherever silica tips encounter undercoordinated carbon at steps, not only for graphite but as a template for oxide–carbon interfaces more broadly. - Dry N₂ conditions isolate tip–surface chemistry from water-mediated pathways; humid AFM extensions would test passivation of dangling bonds.

MD–experiment correspondence. The article reports qualitative agreement between AFM and ReaxFF trends across basal vs step kinematics, and uses simulation-derived shear strain/hydrogen-bond metrics to separate physical vs chemical contributions near the step edge (see the Science Advances figures cited in the PDF).

Limitations¶

Single tip chemistry and dry N2 environment; generalization to humid or reactive lubricants requires further study. Future work that adds water layers at the silica/graphite contact would test whether superlubricity persists when hydrogen-bonded networks can form at the step.

Relevance to group¶

Kim/Martini collaborations; same graphite friction theme as related ACS papers in the corpus.

See [[2019arash-khajeh-acs-effect-ambient]] for adjacent graphite tribology entries in this knowledge base when building comparative retrieval clusters.

Citations and evidence anchors¶

https://doi.org/10.1126/sciadv.aaw0513