Skip to content

Subsurface structural change of silica upon nanoscale physical contact: Chemical plasticity beyond topographic elasticity

Summary

Nano-FTIR / s-SNOM and ReaxFF MD (LAMMPS) are combined to study fused quartz under nanoindentation and nanoscratch with a diamond indenter, plus 150 H\(_2\)O molecules to mirror ~40% RH lab air. The Si–O stretch near ~1100 cm\(^{-1}\) red-shifts and broadens in the scratch even when height recovers elastically, indicating sub-Å network changes (chemical plasticity) invisible to topography. ReaxFF shows longer mean Si–O bonds in the densified plastic zone than in pristine glass, so volume loss is tied to bond elongation, not contraction.

Methods

1 — MD application (atomistic dynamics)

  • Engine / code: LAMMPS with the ReaxFF implementation for C/H/O/Si (diamond + silica) as in Section 2.3; further parameter notes in Supporting Information per the article.
  • System & composition: Amorphous silica substrate 5184 Si/O atoms; diamond indenter 3515 atoms, spherical tip radius 1.5 nm; 150 H\(_2\)O between tip and substrate.
  • Boundaries / periodicity: 3D PBC in the supercell used for the scratch model (SI construction).
  • Ensemble: NVT (canonical) for the scratch and post-scratch equilibration segments described.
  • Timestep: 0.25 fs throughout the MD in Section 2.3.
  • Duration / stages: 300 K equilibration 100 ps; indenter engaged; nanoscratch at 10 GPa contact pressure, 10 m/s slip speed, 4 nm slide; 50 ps equilibration on the extracted substrate; optional heating 300 K → 1500 K at 10 K/ps, 1 ns hold, cool 5 K/ps to 300 K, then 50 ps equilibration (annealing protocol in the proof text) before structural analysis of Si–O length and O–Si–O angle distributions.
  • Thermostat: Nosé–Hoover, τ = 100 fs in the quoted NVT runs.
  • Barostat / pressure: N/A for the main NVT scratch segment; N/A — pressure not isotropically barostatted in that stage (contact pressure 10 GPa is a load target in the protocol, not a hydrostatic NPT barostat setting).
  • Temperature: 300 K baselines; 10 K/ps and 5 K/ps ramps when the annealing block is used.
  • Electric field: N/A.
  • Replica / enhanced sampling: N/A.

Experiment (nano-IR): Fused quartz; Hysitron triboindenter; Berkovich and cono-spherical diamond tips; 5 mN indent, 3–7 mN scratch loads, 1 µm/s speed; neaSNOM nano-FTIR with MCT detection and interferometric amplitude/phase as in §2.2; conversion to dielectric / refractive index per tip–sample models in SI.

2 — Force-field training

N/A — the paper applies published ReaxFF for C/H/O/Si (and diamond) with details in SI; it does not report a new parameter fit here.

3 — Static QM

N/A as a new DFT production study; interpretation leans on literature vibrational assignments for Si–O modes (e.g. cited discussion of ~1100 cm\(^{-1}\)).

4 — Review

N/A.

Findings

  • Spectroscopy: The main Si–O stretch feature shifts from ~1100–1120 toward ~1050–1070 cm\(^{-1}\) and broadens from pristine pile-up to vertex / track regions, so subsurface chemistry can change without residual roughness from AFM height.
  • MD vs experiment: In the plastic densified region, ReaxFF Si–O bonds are slightly longer than in pristine glass, matching the red-shift / elongation narrative; elastic-looking contact can still show bond-length broadening in MD.
  • Comparisons: Nano-IR line shapes vs simulated network metrics; literature on silica under contact. Limitations: Proof PDF; DOI 10.1016/j.actamat.2021.116694 for the VOR pagination and SI; tip–sample optical models for nano-FTIR add uncertainty as the authors note.

Corpus / KB honesty: This file is a galley in pdf_path; confirm page and figure numbers from the typeset Acta Materialia file when citing externally.

Limitations

The repository PDF is an uncorrected proof; use the DOI-resolved article for final SI and pagination. Nano-FTIR mapping to absorbance uses model approximations for the tip–sample system.

Relevance to group

van Duin-coauthored joint experiment–ReaxFF study of silica tribo-chemistry.

Citations and evidence anchors