Skip to content

Chemical effects on subcritical fracture in silica from molecular dynamics simulations

Summary

ReaxFF (USER-REAXC, Yeon & van Duin 2015 Si/O/H) MD in LAMMPS evaluates mode-I loading of silica with three scenarios: vacuum dynamic fracture, water-filled crack dynamic fracture, and water with static subcritical load plus dissolution tracking. The central question is whether humidity primarily sharpens crack tips through surface chemistry or also softens the bulk process zone through water-mediated Si–O scission. Water reduces K\(_{IC}\)-like toughness metrics by ~25%, consistent with experiment; water plasticizes the process zone broadly, not only surface sites, lowering Si–O rupture thresholds without requiring large dissolution flux. The JGR framing connects geophysical subcritical fracture to atomistic stress corrosion language used in materials science.

Methods

  • Engine / code: LAMMPS molecular dynamics with USER-REAXC ReaxFF integration.
  • Potential: ReaxFF Si/O/H parameterization Yeon & van Duin (2015) via USER-REAXC; prior surface energy/modulus benchmarks discussed in-paper.
  • Thermostat: Nosé–Hoover (or equivalent NVT thermostat parameters) as specified in J. Geophys. Res. Methods for the 300 K production segments.
  • Crack geometry: LEFM mode-I displacement field applied to far-field atoms; frozen boundary band; interior mobile; nonperiodic in-plane after initial relaxation.
  • Protocol A (mechanical): Increment K\(_I\) by 0.071 MPa·√m steps; NVT 300 K, 0.1 fs, 5 ps per step; 100 steps to K\(_I\)≈0.65 MPa·√m fracture.
  • Protocol B (chemical): Fixed load; add water in crack at 0.5 g/cm³; reflective walls; 500 ps NVT at 300 K, 0.1 fs.
  • Protocol C (coupled): Iterative loading + GCMC water resupply to maintain coverage as volume opens (details in paper).
  • Diagnostics: Track Si–O bond populations, coordination defects, and dissolution-like events separately so mechanical failure can be distinguished from corrosion-like mass loss.

Additional controls. PBC / geometry: mode-I LEFM displacement on far-field atoms with a frozen boundary band and mobile interior; nonperiodic opening in the crack plane after relaxation (see article). Barostat / bulk pressure: N/A — NPT hydrostatic control not used for the quoted NVT fracture stages. Electric field: N/A — bias not applied. Enhanced sampling: N/A — umbrella / metadynamics / replica exchange not reported.

Findings

Outcomes and mechanisms. Water lowers effective toughness metrics by ~25% versus inert environments while plasticizing the process zone through Si–O scission pathways that are not limited to the immediate crack tip; dissolution can be minor yet subcritical failure accelerates because solvation lowers bond rupture thresholds.

Comparisons. Trends are discussed against experimental stress-corrosion literature for silica and against purely surface-roughening pictures.

Sensitivity. Stress level, water presence (0.5 g/cm³ in the static-chemistry cell), and loading protocol (K\(_I\) stepping vs fixed load + GCMC resupply) change how decomposition vs mechanical damage partition.

Limitations and PDF grounding. Absolute ReaxFF fracture numbers and GCMC+MD coupling approximations limit transferability; confirm coordination definitions in the JGR PDF (pdf_path).

Limitations

ReaxFF fracture absolute values; GCMC+MD coupling approximations; 2D slit crack idealization. Real geological silica networks include disorder, microporosity, and ionic impurities that can couple to subcritical crack growth in ways a single ideal crack cannot represent, so the reported toughness shifts are best interpreted as mechanistic trends benchmarked against the experimental window cited in the paper.

Relevance to group

Pairs with 2018jcp-silica-hydrophil-venue-paper on silica reactivity; uses van Duin Si/O/H parameter line.

Citations and evidence anchors

  • DOI: 10.1029/2018JB016120.

Reader notes (navigation)