Skip to content

A reactive force field study on the interaction of lubricant with diamond-like carbon structures

Summary

ReaxFF MD constructs amorphous DLC and hydrogen/nitrogen-functionalized DLC films by high-temperature melt/quench recipes, targets experimental sp²:sp³, H, and N compositions, then simulates perfluoropolyether (D4OH) droplet spreading and chemical degradation on DLC vs DLC:H:N surfaces relevant to hard-disk lubrication. The study connects atomistic wettability and scission chemistry to industry-relevant PFPE fluids used near slider interfaces where tribochemical stability is a yield limiter.

Methods

Film construction uses a C–H–N ReaxFF extension on published carbon parameters for DLC generation [61] with EEM charges [60] (Section 2.1). Starting from a 4×4×1 diamond supercell (128 C) in 7.27×7.27×20 ų with Ar-filled headspace, an NVT melt at 7500 K (115k iterations, Δt = 0.1 fs, Berendsen thermostat, 250 fs damping) is followed by cooling 7500 → 3000 K at 0.004 K per iteration, NPT relaxation to relieve stress, and in-plane expansion to reduce symmetry artifacts; DLC:H comes from ethylene pyrolysis in Ar and DLC:H:N from heating with N₂ (Section 2.2 onward). Composition targets near ~30% sp³:sp², ~20% H, and ~15% N yield sp³:sp² = 27.3% (DLC) versus 31.7% (DLC:H:N), with 17.9% H and 13.7% N for DLC:H:N in the abstract.

LAMMPS ReaxFF then places a nine-strand D4OH (perfluoropolyether) droplet on DLC (12,800 atoms) and DLC:H:N (9,850 atoms) slabs. The droplet is equilibrated in an 80 × 80 × 140 ų box at 1200 K for 1,000,000 steps (100 ps at 0.1 fs), deposited on each surface, and equilibrated under NVT at 300, 1000, and 1500 K for 500,000 steps (50 ps) before spreading and degradation analysis (Section 2). Three-dimensional periodic boundaries apply. Berendsen damping (250 fs) is documented for the melt/quench DLC build; thermostat details for every D4OH segment should be confirmed in Section 2 if not quoted here. NPT appears during DLC stress relaxation; the droplet-on-surface equilibration lines above are NVT. External electric fields and umbrella or replica metadynamics are not used. Section 3.1 documents further C–H–N ReaxFF training against ab initio QM for C/N bonds and angles—supporting parametrization for DLC:H:N chemistry beyond the abstract’s application focus.

Findings

Less functionalized DLC shows faster D4OH spreading than DLC:H:N in the simulations summarized in the abstract. Both surfaces alter PFPE degradation chemistry relative to gas-phase expectations, showing interfacial participation in scission pathways. The work contrasts DLC versus DLC:H:N wettability and decomposition using the composition metrics above. Ar versus N₂ processing and ethylene pyrolysis stoichiometry are the main film-construction levers used to approach industrial H/N/C and sp²/sp³ targets. Extended HDD context and outlook appear in the full article.

Corpus note. Local pdf_path is an ACS proof PDF—confirm pagination against the J. Phys. Chem. C issue pages.

Limitations

Corpus PDF is an ACS proof; cite the J. Phys. Chem. C issue for final pagination. - Timescales for degradation may still be short relative to HDD service lifetimes—qualitative chemistry insight. - Industrial lubricant blends include additives beyond the PFPE strand model shown here; transferability to full formulations requires additional validation not claimed by the abstract alone. - Shear rates and contact pressures in MD may still exceed drive testing conditions; map atomistic events to wear metrics using the scaling discussion in the full article where provided.

Relevance to group

van Duin co-authorship; ReaxFF application to DLC + fluoropolyether tribochemistry tied to data-storage industry context (Western Digital co-authorship).

Citations and evidence anchors