Skip to content

Oxygen interactions with silica surfaces: Coupled cluster and density functional investigation and the development of a new ReaxFF potential

Evidence and attribution

Authority of statements

Prose sections below (Summary, Methods, Findings, etc.) are curated summaries of the publication identified by doi, title, and pdf_path in the front matter above. They are not new primary claims by this wiki.

For definitive numerical values, reaction schemes, and interpretations, use the peer-reviewed article (and optional records under normalized/papers/ when present)—not this page alone.

Summary

The paper combines high-level QM (CCSD(T)-F12b) and Minnesota density functionals to study O and O₂ approaching cluster models of nondefective and defective silica surfaces (undercoordinated Si, nonbridging oxygen, ring defects, etc.). QM consensus motivates binding preferences on defects. The work then extends ReaxFF–SiO (bulk silica parametrization cited from van Duin et al., JPCA 2003) to a gas–surface interaction (GSI) form, ReaxFF–SiO–GSI, reporting strong agreement between ReaxFF and DFT interaction energies for the training motifs. Motivation spans microelectronics etch/CVD, silica nanoparticle oxidation, plasma O recombination on silica walls, and aerospace thermal protection where heterogeneous O recombination contributes significantly to heating.

Methods

3 — Static QM / cluster benchmarks: The authors compute interaction energies for atomic and molecular oxygen approaching cluster models of realistic silica surfaces, including nondefective reconstructions (sites above fully coordinated Si and bridging O) and defective motifs (under-coordinated Si, nonbridging O, ring structures). High-level CCSD(T)-F12b and Minnesota density functionals are used; the Minnesota functionals are employed for binding energies across clusters with singlet/triplet vs doublet/quartet spin assignments for nondefective vs defective models, respectively (J. Phys. Chem. C 2013, 117, 258–269; abstract; papers/MURI_team_SiOx_JPCC_2013.pdf; normalized/extracts/2013muri-venue-jp3086649_p1-2.txt).

2 — Force-field training (ReaxFF–SiO–GSI): Parent FF: extends the empirical ReaxFF\(_\mathrm{SiO}\) parametrization for bulk silica polymorphs (van Duin et al., J. Phys. Chem. A 2003, 107, 3803–3811, cited in the abstract) to gas–surface interactions (ReaxFF\(_\mathrm{SiO}\) GSI). Training / validation data: DFT interaction energies for the defective and nondefective cluster set; the abstract reports very good agreement between ReaxFF\(_\mathrm{SiO}\) GSI and the Minnesota DFT energies. Optimization: standard ReaxFF fitting against the QM database (details in the article; not re-derived here).

1 — MD application (downstream use): The abstract states ReaxFF\(_\mathrm{SiO}\) GSI is intended for reactive large-scale molecular dynamics of oxygen–silica gas–surface processes such as oxide growth and heterogeneous oxygen recombination. N/A — this paper’s indexed opening emphasizes QM training clusters and FF extension rather than reporting a specific production MD protocol: no NVT/NPT/NVE production settings, timestep, slab supercell atom counts, PBC details, thermostat/barostat, target temperature/pressure, run duration, electric field, or enhanced sampling for those large-scale trajectories are extracted here—see later sections of papers/MURI_team_SiOx_JPCC_2013.pdf if the article reports demonstration MD.

Findings

Outcomes and mechanisms: DFT and CCSD(T)-F12b agree for the benchmarked O/O\(_2\) approaches in the abstract’s accounting; defect motifs are energetically favorable binding sites for oxygen relative to the nondefective motifs sampled. The ReaxFF\(_\mathrm{SiO}\) GSI extension reproduces the DFT interaction energies well for the training cluster set, enabling the claimed reactive MD applications on realistic surfaces.

Comparisons: QM methods are cross-checked (CCSD(T)-F12b vs Minnesota functionals); ReaxFF is compared to the DFT interaction energies for the same clusters.

Sensitivity / design levers: Defect type (under-coordinated Si, NBO, ring) vs nondefective sites modulates binding preferences in the QM data summarized in the abstract.

Limitations and outlook: Cluster models may not exhaust amorphous or polycrystalline silica diversity; large-scale MD applications still require validation on extended surfaces and under operational temperature/flux conditions beyond the cluster QM set.

Corpus honesty: This note summarizes papers/MURI_team_SiOx_JPCC_2013.pdf and normalized/extracts/2013muri-venue-jp3086649_p1-2.txt; reaction barriers, full functional list, and complete training matrix should be taken from the PDF body/SI.

Limitations

  • Cluster models and selected defects may not exhaust real amorphous silica diversity; MD applications need validation on extended surfaces.

Relevance to group

Direct ReaxFF SiO–GSI lineage paper for oxygen + silica reactive simulations tied to MURI-style collaborative development.

Citations and evidence anchors

  • Abstract and Sec. 1 (J. Phys. Chem. C 2013, 117, 258–269; PDF pp. 1–2 per extract).

Reader notes (navigation)