Pith. sign in

REVIEW 3 major objections 5 minor 18 references

This paper argues that the Si-phase percolation threshold in annealed nonstoichiometric silicon-oxide films is not a fixed stoichiometry value but shifts from x ≈ 0.85 in ~1.4 nm films to x ≈ 1.35 in films thicker than about 4.2 nm.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-03 16:30 UTC pith:PEOUCZIY

load-bearing objection 3D extension of the authors' 2D MC model: new thickness-dependent percolation threshold, but the 4.2 nm crossover rests on single runs without error bars. the 3 major comments →

arxiv 2512.13071 v1 pith:PEOUCZIY submitted 2025-12-15 cond-mat.mtrl-sci physics.comp-ph

3D lattice Monte Carlo modeling of morphology formation of Si/SiOx nanocomposites during phase separation of nonstoichiometric Si oxide films

Ivan Oliinyk , Andrey Sarikov This is my paper
classification cond-mat.mtrl-sci physics.comp-ph
keywords nonstoichiometric silicon oxidephase separationlattice kinetic Monte Carlopercolation thresholdSi nanocrystalsSi/SiO2 nanocompositefilm thickness effectlargest cluster fraction
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents a three-dimensional lattice Monte Carlo model of phase separation in nonstoichiometric silicon oxide (SiOx) films, in which only oxygen atoms move on a fixed diamond-like silicon lattice. By tracking the largest cluster of fully reduced silicon atoms, it maps when the precipitated Si phase forms isolated nanoparticles versus a connected network. The central result is that this boundary, the percolation threshold, depends strongly on film thickness: quasi-two-dimensional films about 1.4 nm thick show a threshold near x ≈ 0.85, while films thicker than roughly 4.2 nm behave like bulk material with a threshold near x ≈ 1.35. This matters because device performance depends on whether Si is isolated or percolated, and the result gives a composition–thickness design map for SiOx-based nanocomposites.

Core claim

The paper establishes that in the phase separation of SiOx films, the percolation threshold xp—the initial oxygen-to-silicon ratio at which the largest cluster fraction jumps from near 0 to near 1—critically depends on film thickness. For the thinnest films (Lz ≈ 1.4 nm), xp ≈ 0.85, matching quasi-two-dimensional behavior. As thickness increases, off-plane connection pathways form, raising xp until it saturates at approximately 1.35 for thicknesses around 4.2 nm and above. The same simulations show that isolated Si nanoparticles form for x ≥ 1.4 and percolated sponge-like networks for x ≤ 0.8 regardless of thickness. The authors conclude that a film thickness of ~4.2 nm marks the transition

What carries the argument

The central object is a three-dimensional diamond-like lattice of Si atoms with oxygen atoms occupying bond midpoints; oxygen migration is the only kinetic move. The driving force is the change in total penalty energy of Si–OaSi4–a complexes (penalties Δ1 = 0.5 eV, Δ2 = 0.51 eV, Δ3 = 0.22 eV) evaluated via the Metropolis criterion. The percolation diagnostic is the Largest Cluster Fraction (LCF): the fraction of all Si0 atoms (Si with four Si neighbors) that belong to the largest connected cluster, found with a Union-Find algorithm. The threshold xp is identified by the sharp LCF transition.

Load-bearing premise

The load-bearing premise is that phase separation in amorphous SiOx is governed solely by oxygen hopping on a fixed silicon lattice with the chosen bond-energy penalties; if silicon atoms move or the penalty energies misrepresent the free-energy landscape, the predicted percolation thresholds and their thickness shift could change.

What would settle it

Measure the percolation transition in annealed SiOx films with controlled thicknesses (e.g., ~1.5 nm, ~3 nm, ~4.2 nm, ~10 nm) via electrical conductivity or Si-nanocrystal connectivity; the model predicts xp ≈ 0.85 at 1.4 nm rising to ≈1.35 by 4.2 nm. Observing no thickness dependence—or a different crossover thickness—would falsify the central claim.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • For ultrathin films (≤ ~1.4 nm), SiOx with x as low as 0.85 still phase-separates into a connected but flat Si network, while x ≥ 1.4 always gives isolated particles; device designers can use thickness as a control knob for connectivity.
  • Bulk-like films thicker than ~4.2 nm form percolated Si networks at x ≈ 1.35, so keeping x above ~1.4 is necessary to avoid percolation in thick films.
  • The critical thickness ~4.2 nm gives a quantitative boundary for quasi-2D versus 3D behavior in SiOx superlattices and ultrathin layers.
  • Because the threshold saturates by Nz = 6, further increasing thickness beyond ~4.2 nm does not change the Si network morphology at a given x.
  • The model's LCF-based percolation criterion provides a direct simulation observable that can be compared with experimentally measured conductivity or charge-storage transitions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The inferred critical thickness is likely tied to the effective lattice constant (~0.7 nm) and the diamond-lattice connectivity; if the oxide density or bond geometry differs, the crossover could move, giving a testable prediction across materials.
  • The model fixes the Si sublattice and lets only oxygen hop; if silicon diffusion or viscous relaxation becomes active during annealing, the percolation thresholds could shift, so extending the model to include Si mobility would test the robustness of the thickness dependence.
  • The LCF counts only Si0 atoms; a different connectivity metric that includes partially oxidized Si could shift the apparent percolation threshold, so experiments on optical or electrical percolation might not align exactly with the computed xp.
  • The quasi-2D limit xp ≈ 0.85 is close to low-dimensional percolation boundaries; the paper's results suggest a continuous crossover, which could be compared against continuum percolation theories to isolate lattice effects.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper presents a three-dimensional lattice kinetic Monte Carlo model of phase separation in nonstoichiometric SiOx films. Oxygen atoms hop on a diamond-like Si lattice with bond-energy penalty parameters from previous work. The authors compute the largest-cluster fraction (LCF) of Si atoms in the Si0 oxidation state for film thicknesses Lz = 1.4-14 nm and initial stoichiometries x = 0.5-1.7, and extract a percolation threshold xp. They report xp ≈ 0.85 for the thinnest film, an increase with thickness, and saturation at xp ≈ 1.35 for Lz ≥ 4.2 nm, interpreted as the transition to bulk-like behavior. Conclusions: isolated Si particles form at x ≥ 1.4, interconnected networks at x ≤ 0.8, and the percolation threshold depends on film thickness below a critical thickness of about 4.2 nm.

Significance. If the thresholds are robust, the paper provides a useful composition-thickness morphology map for Si/SiOx nanocomposites and extends the authors' earlier 2D model. Strengths include a systematic thickness sweep (Nz = 2-20), the oxygen-migration mechanism grounded in Ref. [13], and visualizations of quasi-2D and 3D morphologies. However, the central quantitative claims rest on LCF curves from single runs without a spanning-cluster criterion, finite-size scaling, or statistical uncertainties. The significance therefore depends on whether the threshold estimates survive a more rigorous percolation analysis.

major comments (3)
  1. [§3, Fig. 1, Eq. (3)] LCF is defined in Eq. (3) as the mass fraction in the largest cluster, and xp is read from Fig. 1 as the x where LCF jumps from ~0 to ~1. This is not a spanning-cluster criterion: a non-spanning cluster can have high LCF in a 125x125xNz box, and the jump location depends on lateral box size. No error bars or independent runs are reported; the claimed saturation at Nz = 6 rests on two adjacent curves (Nz = 5 and Nz = 6), so statistical noise could erase the apparent jump. Please compute a proper spanning probability as a function of x for multiple lateral sizes and independent seeds, and report the threshold with finite-size extrapolation and uncertainties.
  2. [§3, MC step count] The manuscript states that each simulation was performed at the number of Monte Carlo steps up to 7×10^9 but reports no convergence criterion or time evolution. It is not established that the final LCF values correspond to a representative phase-separated state rather than a kinetically arrested or still-evolving state. Different runs from independent random initial configurations may differ. Please provide LCF vs MC step curves or a convergence metric, and repeat key simulations with several seeds to estimate statistical error.
  3. [§2, penalty energies and fixed-lattice model] The model assumes that phase separation is driven solely by oxygen hops on a fixed diamond lattice, with Si atoms immobile, and uses penalty energies Δ1 = 0.5 eV, Δ2 = 0.51 eV, Δ3 = 0.22 eV from Ref. [16] for Si-SiO2 interface bonds. If silicon mobility or structural relaxation contributes, the reported xp values and thickness shift could change. Since no experimental validation or sensitivity analysis with respect to Δa is given, the quantitative predictive power is unclear. I view this as a risk rather than a fatal flaw, but it should be addressed with a parameter scan or comparison to measured percolation thresholds.
minor comments (5)
  1. [§3, Fig. 1 caption] 'LCF (2) - 1.4 nm' formatting is informal; unify notation (e.g. 'N_z = 2 (L_z = 1.4 nm)').
  2. [§3, text near Fig. 2] Typo: 'suppoted' should be 'supported'. Also 'Open VIsualization TOol' should be 'Open Visualization Tool'.
  3. [§2, Eq. (2) and text] Check units: 'Δ2 = 0.51 e' should be '0.51 eV'. Also the Metropolis expression has a subscript formatting problem (ΔE/k_BT appears as TkEr in the extracted text).
  4. [§3, Fig. 3 caption] Caption lacks a period after 'formed'; also figures would benefit from scale bars indicating the lateral size (125 lattice constants) for the reader.
  5. [References] Ref. [18] has a typo in the DOI: 'Modell. Simul. Mater. Sci. Eng.' and the DOI link is malformed. Please verify all DOIs.

Circularity Check

0 steps flagged

No circularity: percolation thresholds are emergent simulation outputs, not fitted or self-referential inputs.

full rationale

The paper's derivation chain is: (1) construct a 3D lattice kinetic Monte Carlo model with oxygen hopping driven by bond-energy penalties; (2) run simulations over stoichiometry x and thickness Nz; (3) define LCF and read the percolation threshold xp from a sharp LCF transition. The penalty energies Δa are taken from an external source (Bongiorno and Pasquarello, Ref. [16]), and the largest-cluster fraction is an output metric, not a fitted target. The oxygen-migration mechanism is justified by the authors' earlier Ref. [13] and the 2D baseline by Ref. [14], but those are prior independent results used as modeling premises, not as fits to the new LCF curves. The central thickness-dependent result — xp increasing from ≈0.85 at Nz=2 to ≈1.35 at Nz=6 — is generated by the simulation and is not equivalent to any input parameter by construction. The comparison of the thinnest-film threshold to the 2D value from Ref. [14] is corroborative rather than load-bearing. The absence of finite-size scaling and error bars is a robustness/correctness concern, not a circularity concern. Thus no significant circularity is present.

Axiom & Free-Parameter Ledger

1 free parameters · 5 axioms · 0 invented entities

No new physical entities are introduced. The central predictions depend on external bond-energy parameters, the oxygen-migration-only mechanism, and an in-house percolation criterion; these dominate the uncertainty rather than fitted constants.

free parameters (1)
  • Si–O complex penalty energies Δa = Δ1 = 0.5 eV, Δ2 = 0.51 eV, Δ3 = 0.22 eV
    Imported from Bongiorno–Pasquarello bond-energy calculations [16]; not fitted to the percolation results here but they set the thermodynamic driving force and therefore the predicted thresholds.
axioms (5)
  • domain assumption Phase separation is driven exclusively by oxygen migration; Si atoms remain fixed on a diamond lattice.
    Invoked in §2 from Ref. [13]; if Si mobility contributes, morphologies change.
  • domain assumption Local bond-energy penalties Δa from Ref. [16] capture the free-energy landscape of SiOx configurations.
    Used in §2 to compute ΔE for Metropolis acceptance.
  • domain assumption Random initial placement of NO = x·NSi oxygen atoms on bond midpoints represents the as-deposited SiOx film.
    §2; spatial correlations or deposition ordering in real films could alter initial phase-separation pathways.
  • ad hoc to paper The sharp transition in LCF defines the percolation threshold without finite-size scaling or a spanning criterion.
    §2 Eq. (3) and §3; threshold values may depend on lattice size 125×125×Nz and on the arbitrary LCF cutoffs.
  • ad hoc to paper MC step count up to 7×10^9 reaches a representative final state; no convergence criterion is reported.
    §3; without convergence or equilibrium checks, LCF values may be trajectory-dependent.

pith-pipeline@v1.3.0-alltime-deepseek · 5369 in / 14923 out tokens · 122748 ms · 2026-08-03T16:30:07.209387+00:00 · methodology

0 comments
read the original abstract

In this paper, a three-dimensional lattice model based on the Monte Carlo approach is presented. This model is developed to investigate the kinetics of morphology change during phase separation in nonstoichiometric Si oxide (SiOx, x < 2) films. The model takes into account the SiOx local atomic structure and probabilistic migration of oxygen atoms driven by the tendency of free energy minimization. The influence of the initial SiOx stoichiometry index x and film thickness on the morphology of the precipitated Si phase in the Si oxide matrix is analyzed. The morphology of the Si phase is shown to critically depend on the initial SiOx stoichiometry. Namely, isolated Si nanoparticles form at low excess Si content (x >= 1.4), while interconnected Si networks always appear at x <= 0.8. A dimensional effect on the morphology of the Si phase is revealed. Namely, reducing the film thickness imposes geometric constraints on the Si network formation. The percolation threshold is found to shift from xp ~= 1.35 for the bulk-like SiOx layers to xp ~= 0.85 for the quasi-two-dimensional films. The transition to the bulk material behavior is observed at a SiOx thickness of approximately 4.2 nm.

Figures

Figures reproduced from arXiv: 2512.13071 by Andrey Sarikov, Ivan Oliinyk.

Figure 1
Figure 1. Figure 1: shows the dependences of the LCF value for the phase-separated Si/Si oxide nanocomposite on the initial SiOx stoichiometry x for each of the studied film thickness Nz. Analysis of these dependences allows us to draw the following conclusions. The LCF value is close to zero (~ 0.01…0.03) when the initial SiOx stoichiometry index x ≥ 1.4. This indicates that at a low relative content of excess Si, nontoichio… view at source ↗
Figure 2
Figure 2. Figure 2: Simulated morphology of the Si phase in a quasi [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: presents visualization of the simulated structures with the initial stoichiometry value x = 1.6 for Nz = 2 and Nz = 8 showing formation of isolated quasi-two-dimensional and purely three￾dimensional Si nanoparticles. This supports the conclusion derived from the analysis of [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

18 extracted references · 14 canonical work pages

  1. [1]

    S. K. Ray, S. Maikap, W. Banerjee , S. Das, Nanocrystals for silicon -based light -emitting and memory devices, J. Phys. D: Appl. Phys. 46, 153001 (2013). https://doi.org/10.1088/0022- 3727/46/15/153001

    doi:10.1088/0022- 2013

  2. [2]

    Conibeer, M

    G. Conibeer, M. Green, R. Corkish, Y. Cho, E. -C. Cho, C.-W. Jiang, T. Fangsuwannarak, E. Pink, Y. Huang, T. Puzzer, T. Trupke, B. Richards, A. Shalav, K.-L. Lin, Silicon nanostructures for third generation photovoltaics, Thin Solid Films 511-512, 654 -662 (2006). https://doi.org/10.1016/j.tsf.2005.12.119

    work page doi:10.1016/j.tsf.2005.12.119 2006

  3. [3]

    Boninelli, F

    S. Boninelli, F. Iacona, G. Franzò, C. Bongiorno, C. Spinella, F. Priolo, Formation, evolution and photoluminescence properties of Si nanoclusters, J. Phys .: Condens. Matter 19, 225003 (2007). https://doi.org/10.1088/0953-8984/19/22/225003

    work page doi:10.1088/0953-8984/19/22/225003 2007

  4. [4]

    Müller, K

    T. Müller, K. -H. Heinig , W. Möller, Size and location control of Si nanocrystals at ion beam synthesis in thin SiO2 films, Appl. Phys. Lett. 81, 3049 (2002). https://doi.org/10.1063/1.1512952

    work page doi:10.1063/1.1512952 2002

  5. [5]

    Laube, S

    J. Laube, S. Gutsch, D. Wang, C. Kübe l, M. Zacharias, D. Hiller, Two -dimensional percolation threshold in confined Si nanoparticle networks, Appl. Phys. Lett. 108, 043106 (2016). https://doi.org/10.1063/1.4940971

    work page doi:10.1063/1.4940971 2016

  6. [6]

    Zacharias, J

    M. Zacharias, J. Heitmann, R. Scholz, U. Kahler, M. Schmidt , J. Bläsing, Size -controlled highly luminescent silicon nanocrystals: A SiO/SiO 2 superlattice approach, Appl. Phys. Lett. 80, 661 (2002). https://doi.org/10.1063/1.1433906

    work page doi:10.1063/1.1433906 2002

  7. [7]

    S. T. H. Silalahi, H. Y. Yang, K. Pita, M. B. Yu, Rapid thermal annealing of sputtered silicon -rich oxide/SiO2 superlattice structure, Electrochem. Solid -State Lett. 12, K29 (2009). https://doi.org/10.1149/1.3074295

    work page doi:10.1149/1.3074295 2009

  8. [8]

    Gutsch, D

    S. Gutsch, D. Hiller, J. Laube, M. Zacharias, C. Kübel, Observing the morphology of single-layered embedded silicon nanocrystals by using temperature -stable TEM membranes , Beilstein J. Nanotechnol. 6, 964 (2015). https://doi.org/10.3762/bjnano.6.99

    work page doi:10.3762/bjnano.6.99 2015

  9. [9]

    Sarikov, Thermodynamic theory of phase separation in nonstoichiometric Si oxide films induced by high-temperature anneals, Nanomanufacturing 3, 293 (2023)

    A. Sarikov, Thermodynamic theory of phase separation in nonstoichiometric Si oxide films induced by high-temperature anneals, Nanomanufacturing 3, 293 (2023). https://doi.org/10.3390/nanomanufacturing3030019

    work page doi:10.3390/nanomanufacturing3030019 2023

  10. [10]

    Babiichuk, A

    M. Babiichuk, A. Sarikov, M. Voitovych, V. Voitovych, Phase composition versus stoichiometry and fabrication temperature of Si oxynitride films by thermodynamic modeling, 2025 IEEE 15th Int. Conf. Nanomaterials: Applications & Properties (NAP), Bratislava, Slovakia, 2025, MTFC08 -1. https://doi.org/10.1109/NAP68437.2025.11216220

    arXiv 2025

  11. [11]

    B. J. Hinds, F. Wang, D. M. Wolfe, C. L. Hinkle , G. Lucovsky, Investigation of postoxidation thermal treatments of Si/SiO 2 interface in relationship to the kinetics of amorphous Si suboxide decomposition, J. Vac. Sci. Technol. B 16, 2171 (1998). https://doi.org/10.1116/1.590302. 8

    work page doi:10.1116/1.590302 1998

  12. [12]

    Friedrich, B

    D. Friedrich, B. Schmidt, K. H. Heinig, B. Liedke, A. Mücklich, R. Hübner, D. Wolf, S. Kölling, T. Mikolajick, Sponge-like Si-SiO2 nanocomposite – Morphology studies of spinodally decomposed silicon-rich oxide, Appl. Phys. Lett. 103, 131911 (2013). https://doi.org/10.1063/1.4820453

    work page doi:10.1063/1.4820453 2013

  13. [13]

    Sarikov, V

    A. Sarikov, V. Litovchen ko, I. Lisovskyy, I. Maidanchuk, S. Zlobin, Role of oxygen migration in the kinetics of the phase separation of nonstoichiometric silicon oxide films during high-temperature annealing, Appl. Phys. Lett. 91, 133109 (2007). https://doi.org/10.1063/1.2790814

    work page doi:10.1063/1.2790814 2007

  14. [14]

    Sarikov, M

    A. Sarikov, M. Semenenko, S. Shahan, Percolation threshold in annealed ultrathin SiOx films by 2D Monte Carlo simulations, CrystEngComm 26, 2836 (2024). https://doi.org/10.1039/D4CE00212A

    work page doi:10.1039/d4ce00212a 2024

  15. [15]

    W. H. Baur, Variation of mean Si–O bond lengths in silicon-oxygen octahedral, J. Solid State Chem. 22, 445 (1977). https://doi.org/10.1016/0022-4596(77)90022-6

    work page doi:10.1016/0022-4596(77)90022-6 1977

  16. [16]

    Bongiorno , A

    A. Bongiorno , A. Pasquarello, Validity of the bond -energy picture for the energetics аt Si -SiO2 interfaces, Phys. Rev. B 62, R16326 (2000). https://doi.org/10.1103/PhysRevB.62.R16326

    work page doi:10.1103/physrevb.62.r16326 2000

  17. [17]

    Metropolis, A

    N. Metropolis, A. Rosenbluth, M. Rosenbluth, A. Teller, E. Teller, Equation of state calculations by fast computing machines, J. Chem. Phys. 21, 1087 (1953). http://dx.doi.org/10.1063/1.1699114

    doi:10.1063/1.1699114 1953

  18. [18]

    Stukowski, Visualization and analysis of atomistic simu lation data with OVITO -the Open Visualization Tool, Modell

    A. Stukowski, Visualization and analysis of atomistic simu lation data with OVITO -the Open Visualization Tool, Modell. Simul. Mat er. Sci. Eng. 18, 015012 (2010). https://doi/org/10.1088/0965-0393/18/1/015012

    doi:10.1088/0965-0393/18/1/015012 2010