A reactive force field study on the interaction of lubricant with diamond-like carbon structures
Summary¶
Hard-disk interfaces rely on nanometer-thin perfluoropolyether (PFPE) lubricants to limit wear and friction between flying heads and rotating media; heat-assisted recording and tighter head–media spacing increase sensitivity to lubricant desorption and chemical degradation. This Journal of Physical Chemistry C article applies ReaxFF molecular dynamics to the Demnum-class PFPE D4OH interacting with diamond-like carbon (DLC) films intended to mimic protective overcoats, comparing neat DLC with hydrogenated DLC and nitrogen-containing hydrogenated DLC (DLC:H:N) variants. Industrial motivation is underscored by coauthors from Western Digital, while Adri C. T. van Duin leads the reactive force-field modeling at Penn State. The narrative ties lubricant spreading kinetics and thermal/chemical degradation networks to the local bonding character of each DLC class, arguing that surface functionalization changes not only mechanical contact but also the radical and fragmentation chemistry accessible during reactive trajectories.
Methods¶
1 — MD application (ReaxFF). Diamond-like carbon (DLC) films are built by melting an initial carbon-diamond arrangement in an Ar-filled cell at 7500 K, cooling to 3000 K, then running constant-pressure (NPT) MD to relax volume and internal stress. Hydrogenated DLC (DLC:H) is grown by pyrolyzing ethylene as a carbon source with Ar present. DLC:H:N is obtained by heating DLC:H with N\(_2\), tuning Ar and N\(_2\) inventories so that the model approaches reported H/N/C composition and sp\(^2\)/sp\(^3\) targets. The abstract reports final sp\(^3\):sp\(^2\) ratios of 27.3% (DLC) and 31.7% (DLC:H:N), with H and N contents of 17.9% and 13.7%, respectively, in the nitrogen-functionalized film. Lubricant simulations place a droplet of nine D4OH PFPE strands on DLC and DLC:H:N substrates and follow spreading and degradation chemistry relative to bulk lubricant reference runs. Timestep, thermostat and further temperature/pressure schedules after film preparation, production lengths, electrostatic settings, and any shear protocol for spreading are N/A — not captured in the indexed excerpt; use pdf_path. Electric fields and enhanced sampling are N/A — not indicated in that excerpt.
2 — Force-field training. N/A — the study applies ReaxFF for C/H/O/F plus DLC chemistry; it is not a new parametrization paper.
3 — Static QM / DFT. N/A — headline results are classical reactive MD.
Cells use PBC with Ar during melt/quench; post-build NPT relaxation adjusts volume and stress before lubricant-on-substrate stages. Molecular dynamics after film build continues with ReaxFF as in the article; NVT/NPT staging after relaxation, timestep (fs), equilibration/production run lengths (ps/ns) for droplet spreading and degradation, thermostat and barostat/pressure targets, and target temperature (K) schedules are N/A — not captured in the indexed excerpt—confirm in pdf_path. Electric field driving and umbrella/metadynamics/replica-exchange sampling are N/A — not indicated there.
Findings¶
Outcomes. D4OH spreads faster on the nonfunctionalized DLC surface than on DLC:H:N under the reported protocols. Both substrates change PFPE degradation chemistry compared to bulk lubricant-only simulations, so substrate composition enters degradation networks alongside temperature and any mechanical driving treated in the article.
Mechanistic attribution. The authors connect faster spreading on DLC versus DLC:H:N to differences in surface–lubricant interaction and functionalization (see discussion in J. Phys. Chem. C).
Industrial DLC microstructures and deposition histories are richer than the melt-quench training cells; treat reported sp\(^2\)/sp\(^3\) and composition as model targets aligned with the paper’s narrative, not universal device descriptors.
Limitations¶
Sputtered or plasma-deposited industrial DLC films exhibit broader structural disorder and impurity chemistry than melt-quench training cells; extrapolating every degradation channel to device lifetimes requires validation against experiment. ReaxFF captures ground-state reactive events but not electronic excitations or laser heating paths relevant to some recording physics scenarios.
Relevance to group¶
Combines ReaxFF with tribology/storage materials problems coauthored by van Duin.
Citations and evidence anchors¶
- DOI:
https://doi.org/10.1021/acs.jpcc.6b09729(papers/Lotfi_2016_DLC_paper.pdf).