Pith. sign in

REVIEW 5 minor 33 references

Hexagonal Stacking Maximizes Proton Configurational Entropy among Ice-I Polytypes

T0 review · 0 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read By encoding every uniform-registry ice-I stacking as a word in a nonnegative transfer matrix and its transpose, this paper proves that hexagonal stacking maximizes the proton configurational entropy among all periodic polytypes, with…

desk verdict A rigorous, honest proof that hexagonal stacking maximizes proton configurational entropy among periodic uniform-registry ice-I polytypes; the load-bearing transpose identity is persuasive but not machine-checked. read the letter →

arxiv 2608.07613 v1 pith:MOZM2LDO submitted 2026-08-06 cond-mat.stat-mech math-phmath.MP

Zhengyue Chen , Sheng Ran This is my paper · ORCID
classification cond-mat.stat-mechmath-phmath.MP MSC 82B2005C3015A60
keywords iceIpolytypesprotonconfigurationalentropyruletransfermatrixhexagonalcubicEulerianorientationsrigorousbounds
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper proves that, within the equal-weight two-in, two-out ice rule, hexagonal stacking maximizes the number of proton configurations among all periodic ice-I polytypes built from the same puckered layer. The proof encodes every uniform-registry stacking sequence as a word in a nonnegative layer transfer matrix $K$ and its transpose, then applies a standard matrix-norm inequality to show that the alternating hexagonal word has the largest trace at every finite cross-section. This gives a thermodynamic statement: the configurational entropy constant $w$ of any periodic polytype is no larger than that of hexagonal ice. The paper also supplies certified bounds, $1.503360395535 \leq w \leq 1.540195787172$, with cubic ice bounded by $1.52769873835$, all valid in the infinite-crystal limit without finite-size extrapolation.

What carries the argument

The central object is the layer transfer matrix $K$, whose entries count two-in, two-out assignments between consecutive layers in one registry; the opposite registry is represented by its exact transpose $K^T$. Every uniform-registry stacking becomes a word in these two matrices, and the number of configurations for a periodic stack with matching closure is the trace of that word. The core identity is that the two registry operators are exact transposes, which forces every factor in a word to have identical singular values, so a standard norm inequality bounds the trace of any even-length word by the trace of the alternating word $(K K^T)^m$, namely the hexagonal stack. The certified bounds come from two further constructions: a positive even-subgraph expansion restricted to disjoint blocks for the lower endpoint, and a prism tiling with replica operators bounded by exact positive-vector certificates for the spectral radius for the upper endpoints.

What would settle it

For the 2×3 transverse section, construct the six-by-six block $K$ displayed in Eq. (2), enumerate all 16 words of length 4, and compare each trace with $\operatorname{tr}[(K K^T)^2]$; if any word exceeds the alternating trace, the central inequality is false.

Watch

Extended reading notes

Core claim

For a fixed transverse section $Q$ and any even-length word $\sigma \in \{+1,-1\}^{2m}$, the number of ice-rule configurations is $Z_Q(\sigma) = \operatorname{tr}(A_{\sigma_1} \cdots A_{\sigma_{2m}})$, where $A_{+1}=K$ and $A_{-1}=K^T$. The paper proves that $\operatorname{tr}(A_{\sigma_1} \cdots A_{\sigma_{2m}}) \leq \operatorname{tr}[(K K^T)^m] = Z_Q(\mathrm{Ih})$ for every such word, using the fact that a matrix and its transpose share the same singular values and applying a standard norm inequality for products of matrices. Because every uniform-registry polytype corresponds to such a word, passing to the thermodynamic limit yields $w(\sigma) \leq w(\mathrm{Ih})$ for every periodic polytype. The lower endpoint is obtained by restricting a positive expansion over even subgraphs to disjoint exactly enumerated blocks, and the upper endpoints by tiling the transverse section with prisms whose replica counts are bounded by exact positive-vector certificates for the spectral radius, giving $1.503360395535 \leq w \leq 1.540195787172$ and $w(\mathrm{Ic}) \leq 1.52769873835$.

Load-bearing premise

The proof assumes that switching from one layer registry to the other exactly transposes the layer transfer matrix; if the transpose relation between the two registries failed, the hexagonal word would not be the guaranteed maximum.

Editorial extensions

If this is right

  • For every periodic uniform-registry ice-I polytype, the configurational entropy constant satisfies $w(\sigma) \leq w(\mathrm{Ih})$; no cubic or mixed periodic stacking can beat hexagonal ice within the equal-weight ice-rule model.
  • The certified interval $1.503360395535 \leq w \leq 1.540195787172$ is rigorous in the thermodynamic limit, shrinking the previous general interval by a factor of about 4.1 for the whole stacking class and about 6.2 for cubic ice.
  • Cubic ice has the tighter certified ceiling $w(\mathrm{Ic}) \leq 1.52769873835$, but the order of the true constants $w(\mathrm{Ic})$ and $w(\mathrm{Ih})$ is not decided; the bounds are consistent with equality and with strict inequality.
  • The same transfer-word comparison applies to any layered constraint model whose two registry operators are nonnegative exact transposes on a common state space, so the hexagonal maximum is a general combinatorial phenomenon rather than an ice-specific accident.
  • The bounds are certified without diagonalizing the relevant transfer operators; they rely on exactly enumerated blocks and on positive rational vectors, so they can be independently verified by exact integer arithmetic.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The certified spread of the bounds (about 0.2 J mol⁻¹ K⁻¹ in entropy) suggests that within the ice-rule model, proton configurational entropy is nearly degenerate across polytypes; observed stacking preferences are therefore more likely determined by energetic, vibrational, or kinetic selection than by the entropy term alone.
  • If a future proof establishes $w(\mathrm{Ic}) = w(\mathrm{Ih})$, cubic stacking would also sit at the entropy maximum, meaning stacking disorder carries essentially no protonic entropy penalty, and the observed prevalence of hexagonal ice would need another explanation.
  • One could probe the tightness of the inequality by exhaustive enumeration on slightly larger transverse sections: if some long word approaches $\operatorname{tr}[(K K^T)^m]$ much more closely than cubic does, that word would mark a near-degenerate polytype worth studying with numerical methods.
  • The prism-tiling upper-bound construction is restricted to sections tileable by at least two prisms in each transverse direction; adapting it to arbitrary cross-sections or laterally varying fault networks would require a new argument, since the degree-two incidence property is essential to the inequality.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

0 major / 5 minor

Summary. The manuscript proves a rigorous extremal statement for proton configurational entropy in ice I. It identifies every cyclic uniform-registry stacking word with a trace of a product of a nonnegative layer transfer matrix K and its transpose, and applies the Schatten–Hölder inequality to show that for every finite transverse section Q and every even-length word σ, Z_Q(σ) ≤ Z_Q(Ih) = tr[(KK^T)^m] (Eq. 3). From this it concludes that among periodic uniform-registry ice-I polytypes the hexagonal (ABAB) stacking maximizes the ice-rule configurational constant. The paper also supplies certified numerical endpoints: a common lower bound 1.503360395535 from Nagle's positive even-subgraph expansion restricted to exactly enumerated disjoint blocks, and upper bounds 1.540195787172 (common) and 1.52769873835 (cubic) from Finner's degree-two hypergraph Hölder inequality and rational Collatz–Wielandt certificates. Aperiodic sequences are treated along specified prism-tileable exhaustions, with the limitations stated explicitly.

Significance. If the result holds, it is a substantial advance over the previous rigorous endpoint comparisons: it covers the full periodic uniform-registry class rather than only the cubic and hexagonal ideal limits. The proof rests on a clean transfer-word identification and a short matrix inequality; the exact rational certificates, the explicit disclosure of the tileability restriction, and the archived verification code and certificates are notable strengths. I do not see the stress-test concern about a hidden boundary permutation in the transpose identity as a defect: S2.1 gives an argument for K^{(-1)}=(K^{(+)})^T and the displayed 2×3 block makes the transpose concrete. The honest treatment of what remains open (equality versus strict inequality of w(Ic) and w(Ih), uniqueness, exhaustion independence for aperiodic sequences) is a further strength.

minor comments (5)
  1. [S2.1] The exact transpose identity K^{(-1)}=(K^{(+)})^T is the load-bearing geometric fact behind Eq. (3). The current proof is compressed: reversal transposes the incidence 'while the mirror is a relabelling of Q that fixes the set of admissible pairs. Hence ... K^T.' Please expand this into an explicit coordinate-level bijection in the co-moving basis of S1.2, showing for arbitrary boundary states (η,η') that the opposite-registry count equals (K^{(+)})_{η'η}. The 2×3 finite check is helpful evidence but finite; an algebraic derivation would remove any residual concern about a surviving boundary permutation.
  2. [§III, Periodic polytypes] When the period word has nonzero net registry shift modulo 3, a cyclic stack of length an odd or even multiple of p is a screw closure rather than an ordinary torus. For the thermodynamic comparison to the infinite periodic stack, please state that an even multiple of three periods (length 6pr, e.g. the word σ^{6r}) is used, so the registry shift vanishes and the finite graph is an ordinary torus; this factor is harmless in the rate. The sentence 'doubling the period first if p is odd' does not by itself remove a nonzero shift and should be adjusted.
  3. [S2.3] The claim d_Q=2^{s_Q/2} deserves one explanatory sentence: since s_Q counts oxygen sites per layer and there are two sites per cell, the one-bit-per-cell boundary encoding gives 2^{#cells}=2^{s_Q/2}. As written, a reader may infer d_Q=2^{s_Q} from 'sites per layer'.
  4. [§III] Typo: 'apply (3) to 2r players' should read 'periods' or 'layers'.
  5. [Title/Abstract] The abstract already says 'periodic uniform-registry polytypes,' but the title 'among Ice-I Polytypes' is broader. Since the later limitation statements are explicit, a small qualifier in the title or first sentence would avoid over-reading.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central Hölder bound is a direct proof, not a renamed fit or self-citation chain.

full rationale

The derivation is self-contained. The central comparison (Eq. 3) is a direct application of Schatten–Hölder: for any word σ∈{±1}^{2m}, Z_Q(σ)=|tr(A_{σ1}...A_{σ2m})| ≤ ∥A_{σ1}∥_{2m}...∥A_{σ2m}∥_{2m}=∥K∥_{2m}^{2m}=tr[(KK^T)^m]=Z_Q(Ih). Each equality and inequality is justified in the text; the alternating hexagonal word is not assumed maximal but is shown to attain the same Hölder bound. The only load-bearing structural premise, the transpose identity K^{(-1)}=(K^{(+)})^T, is proved in Sec. S2.1 independently of the maximization conclusion and is checked on the explicit 6×6 block (Eq. 2); attributing it to Li et al. is an external citation, not a self-citation. The lower bound is a true restriction of Nagle's positive expansion: discarding positive terms cannot overcount, and the Perron-rate identification is standard. The upper bounds are exact Collatz–Wielandt certificates with explicitly exhibited positive rational vectors; they do not fit the target constants in advance but are verified by rational arithmetic. No equation reduces by definition to its own input and no prediction is a renamed fit.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the standard ice-rule model, the exact transpose structure of the two registry operators, and standard inequalities. The only hand-chosen quantities, such as block size, prism dimensions, and comparison vectors, are proof certificates: they are not fitted to any measurement or numerical estimate of w, and the inequalities they certify are exact algebraic statements. No new physical entities are postulated.

assumptions (5)
  • domain assumption Bernal-Fowler ice rule: every oxygen has two near and two distant protons, equivalently every vertex of the 4-regular oxygen graph has Eulerian in-degree 2.
    This defines the object whose entropy is counted; real ice adds energetic and phonon effects that the paper explicitly excludes.
  • domain assumption The two registry layer operators are exact transposes, K(-1)=K^T, and every cyclic uniform-registry stack satisfies Z_Q(sigma)=tr(A_{sigma1}...A_{sigma2m}) in co-moving coordinates.
    Used in Eq. (1) and Sec. S2.1; if this geometric identity failed, the Holder upper bound would not have alternating hexagonal stacking as its value.
  • standard math Nagle's positive even-subgraph expansion Z(G)=(3/2)^N sum_F 3^{-n2(F)} with only positive terms.
    Cited from [8,21]; positivity licenses the block-restriction truncation in Sec. III and Sec. S3.
  • domain assumption Benjamini-Schramm convergence plus the Bencs-Borbenyi-Csikvari limit theorem for Eulerian orientations gives a single thermodynamic configuration constant for the periodic families.
    Cited [18] and invoked in Sec. II; the paper states its lower and upper bounds are independent of this theorem, so a failure here would not destroy the bracketing inequalities.
  • standard math Schatten-Holder, Finner's degree-two hypergraph Holder inequality, Perron-Frobenius, Fekete's lemma, and Collatz-Wielandt are valid as used.
    Standard theorems invoked for the Schatten-norm comparison, prism product inequality, and exact certificate bounds.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hexagonal Stacking Maximizes Proton Configurational Entropy among Ice-I Polytypes." pith.science (2026). https://pith.science/paper/MOZM2LDO

@misc{pith2026260807613,
  author       = {Pith},
  title        = {Pith review of: Hexagonal Stacking Maximizes Proton Configurational Entropy among Ice-I Polytypes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MOZM2LDO}},
  note         = {Machine review of arXiv:2608.07613}
}
abstract

Ice I admits cubic, hexagonal, and mixed layer stackings, but rigorous entropy comparisons have focused on the two ideal endmembers. We represent every cyclic uniform-registry stacking by a word in a nonnegative transfer operator K and its transpose. For every such even-length word, applying the Schatten-H\"older inequality proves that alternating hexagonal stacking maximizes the ice-rule count at every common finite cross-section; the configuration constant is therefore maximal among all periodic uniform-registry polytypes. We obtain the lower endpoint by restricting Nagle's positive even-subgraph expansion to exactly enumerated disjoint blocks. Finner's degree-two hypergraph H\"older inequality and rational Collatz-Wielandt certificates for two-replica prism transfer operators give the upper endpoints. These constructions yield $1.503360 \le w \le 1.540196$, with $w(\mathrm{Ic}) \le 1.527699$.

Figures

Figures reproduced from arXiv: 2608.07613 by the authors.

Figure 1
Figure 1. FIG. 1. Layer stacking, transfer words, and the local ice rule. Cubic and hexagonal ice are built from the same puckered layer [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Certified intervals for the configurational-entropy constant. The previous rigorous interval extends from Pauling’s lower [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. How the lower bound is constructed. (a) The quantity counted is the number of two-in, two-out orientations. (b) [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FIG. 4. How the upper bounds are constructed. (a) The transverse torus is tiled by open prisms running through every layer [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 28 canonical work pages

  1. [1]

    W. F. Giauque and J. W. Stout, The entropy of water and the third law of thermodynamics. The heat capacity of ice from 15 to 273◦K, J. Am. Chem. Soc.58, 1144 (1936)

    work page 1936

  2. [2]

    Pauling, The structure and entropy of ice and of other crystals with some randomness of atomic arrangement, J

    L. Pauling, The structure and entropy of ice and of other crystals with some randomness of atomic arrangement, J. Am. Chem. Soc.57, 2680 (1935)

    1935

  3. [3]

    T. L. Malkin, B. J. Murray, A. V. Brukhno, J. Anwar, and C. G. Salzmann, Structure of ice crystallized from supercooled water, Proc. Natl. Acad. Sci. USA109, 1041 (2012)

    work page 2012

  4. [4]

    T.L.Malkin, B.J.Murray, C.G.Salzmann, V.Molinero, S. J. Pickering, and T. F. Whale, Stacking disorder in ice I, Phys. Chem. Chem. Phys.17, 60 (2015)

    work page 2015

  5. [5]

    del Rosso, M

    L. del Rosso, M. Celli, F. Grazzi, M. Catti, T. C. Hansen, A. D. Fortes, and L. Ulivi, Cubic ice Ic without stack- ing defects obtained from ice XVII, Nat. Mater.19, 663 (2020)

    work page 2020

  6. [6]

    Komatsu, S

    K. Komatsu, S. Machida, F. Noritake, T. Hattori, A. Sano-Furukawa, R. Yamane, K. Yamashita, and H. Kagi, Ice Ic without stacking disorder by evacuating hydrogen from hydrogen hydrate, Nat. Commun.11, 464 (2020)

    work page 2020

  7. [7]

    Huang, L

    X. Huang, L. Wang, K. Liu, L. Liao, H. Sun, J. Wang, X. Tian, Z. Xu, W. Wang, L. Liu,et al., Tracking cubic ice at molecular resolution, Nature617, 86 (2023)

    2023

  8. [8]

    J. F. Nagle, Lattice statistics of hydrogen bonded crys- tals. I. The residual entropy of ice, J. Math. Phys.7, 1484 (1966)

    work page 1966

Show all 33 references
  1. [9]

    Kolafa, Residual entropy of ices and clathrates from Monte Carlo simulation, J

    J. Kolafa, Residual entropy of ices and clathrates from Monte Carlo simulation, J. Chem. Phys.140, 204507 (2014)

    2014

  2. [10]

    Nothing here contradicts those estimates

    reports cubic and hexagonal values that are numer- ically indistinguishable. Nothing here contradicts those estimates. What changes is that the enclosing interval is now certified in the thermodynamic limit and is lattice- specific, so a future proof ofw(Ic) =w(Ih)or of strict...

  3. [11]

    Enrico Fermi

    L. Onsager and M. Dupuis, The electrical properties of ice, Rend. Scuola Int. Fis. “Enrico Fermi”10, 294 (1960)

    1960

  4. [12]

    Xu, T.-Y

    X.-Z. Xu, T.-Y. Lin, and G.-M. Zhang, Equivalence of residual entropy of hexagonal and cubic ices from tensor network methods, Phys. Rev. B113, 214416 (2026)

    2026

  5. [13]

    Li, Y.-J

    D.-Z. Li, Y.-J. Cen, X. Wang, and X.-B. Yang, Residual entropy of ice: a study based on transfer matrices, Phys. Rev. B110, 195414 (2024)

    2024

  6. [14]

    Both endpoints of the pre- vious rigorous interval are therefore lattice-independent, which is precisely what the present certificates give up in order to gain lattice specificity

    applies to any4-regular graph of that girth and so cannot separate polytypes. Both endpoints of the pre- vious rigorous interval are therefore lattice-independent, which is precisely what the present certificates give up in order to gain lattice specificity. Among numerical wo...

  7. [15]

    E. H. Lieb and F. Y. Wu, Two-dimensional ferroelectric models, inPhase Transitions and Critical Phenomena, Vol. 1, edited by C. Domb and M. S. Green (Academic Press, London, 1972) pp. 331–490

    1972

  8. [16]

    Las Vergnas, An upper bound for the number of Eu- lerian orientations of a regular graph, Combinatorica10, 61 (1990)

    M. Las Vergnas, An upper bound for the number of Eu- lerian orientations of a regular graph, Combinatorica10, 61 (1990)

    1990

  9. [17]

    L. Lupi, A. Hudait, B. Peters, M. Grünwald, R. G. Mullen, A. H. Nguyen, and V. Molinero, Role of stacking disorder in ice nucleation, Nature551, 218 (2017)

    2017

  10. [18]

    Hudait, S

    A. Hudait, S. Qiu, L. Lupi, and V. Molinero, Free energy contributions and structural characterization of stack- 7 ing disordered ices, Phys. Chem. Chem. Phys.18, 9544 (2016)

    2016

  11. [19]

    E. A. Engel, B. Monserrat, and R. J. Needs, Anharmonic nuclearmotionandtherelativestabilityofhexagonaland cubic ice, Phys. Rev. X5, 021033 (2015)

    2015

  12. [20]

    Bencs, M

    F. Bencs, M. Borbényi, and P. Csikvári, Number of Eu- lerian orientations for Benjamini-Schramm convergent graph sequences, arXiv:2409.18012 (2024)

    2024 arXiv

  13. [21]

    Bhatia,Matrix Analysis, Graduate Texts in Mathe- matics, Vol

    R. Bhatia,Matrix Analysis, Graduate Texts in Mathe- matics, Vol. 169 (Springer, New York, 1997)

    1997

  14. [22]

    See the Supplemental Material, appended below after the references, for the ice-I geometry and transfer-word rep- resentation, the stacking-word comparison and its ape- riodic scope, the Nagle expansion and block restriction, the transfer operators and Perron rate, the exact l...

  15. [23]

    Borbényi and P

    M. Borbényi and P. Csikvári, Counting degree- constrained subgraphs and orientations, Discrete Math. 343, 111842 (2020)

    2020

  16. [24]

    Finner, A generalization of Hölder’s inequality and some probability inequalities, Ann

    H. Finner, A generalization of Hölder’s inequality and some probability inequalities, Ann. Probab.20, 1893 (1992)

    1992

  17. [25]

    N. J. Calkin and H. S. Wilf, The number of indepen- dent sets in a grid graph, SIAM J. Discrete Math.11, 54 (1998)

    1998

  18. [26]

    Y.-B. Chan, Upper bounds on the growth rates of hard squares and related models via corner transfer matrices, inDiscrete Mathematics and Theoretical Computer Sci- ence Proceedings, FPSAC 2015(2015) pp. 793–804

    2015

  19. [27]

    T. Hart, T. Hansen, and W. F. Kuhs, A Markov theo- retic description of stacking-disordered aperiodic crystals including ice and opaline silica, Acta Crystallogr. A74, 357 (2018)

    2018

  20. [28]

    Wang and F

    B.-Y. Wang and F. Zhang, Words and normality of ma- trices, Linear Multilinear Algebra40, 111 (1995)

    1995

  21. [29]

    Chen and S

    Z. Chen and S. Ran, Verification code and certificates for rigorous bounds on the proton configurational en- tropy of ice I polytypes, Zenodo (2026), 10.5281/zen- odo.21753932. 8 Supplemental Material Hexagonal Stacking Maximizes Proton Configurational Entropy among Ice-I Poly...

    2026 doi

  22. [30]

    every vertex lies in exactly one prism

  23. [31]

    every prism vertex carries three intra-layer and one interlayer bond

  24. [32]

    Then the two-replica operatorRP of the prism satisfies w≤ρ(R P )1/(2ps), s= 2L uLv, withpthe number of layers per longitudinal period

    every bond is internal to one prism or joins exactly two distinct prisms, and every severed bond is intra-layer. Then the two-replica operatorRP of the prism satisfies w≤ρ(R P )1/(2ps), s= 2L uLv, withpthe number of layers per longitudinal period. Condition (iii) is the hypoth...

  25. [33]

    Commutation with each generator makes the orbit-constant subspace, meaning vectors constant on eachG8-orbit, invariant underR

    The full open3×3cubic replica operator commutes withG8; consequently the orbit-constant subspace is invariant, and a positive vectorxon the quotient satisfying ¯Rx≤µxlifts to a positive vector˜xon the full space satisfyingR˜x≤µ˜x, with no relaxation. Commutation with each gene...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.