REVIEW 1 major objections 4 minor 2 cited by
Soft cells, Kelvin's foam and the minimal surfaces of Schwarz
T0 review · 1 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash
Pith's one-line read The paper classifies all soft space-filling tilings that preserve the vertices and symmetries of the bcc Voronoi tiling, proving there are exactly two (or four under weaker symmetry) and matching two of them to the Schwarz P and D minimal…
desk verdict A worthwhile classification of second-order soft tilings with a convincing core, but Proposition 1 leans on unverified premises and should be tightened before publication. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the nodal set: the four unit vectors $\mathbf{a}, \mathbf{b}, \mathbf{c}, \mathbf{d}$ giving the edge half-tangents at each node of the (e2) tiling. The EEB algorithm lets the fundamental vector $\mathbf{a}$ run over the unit sphere, applies the symmetry group through linear transformations to generate the full nodal set, and turns the softness condition into equations of the form $\mathbf{u}_i \cdot \mathbf{u}_j = -1$ inside every vertex set. These softening equations reduce to great circles and isolated points on the sphere of Euler angles, and second-order equivalence classes are the isolated solutions; requiring at least one planar face adds further great-circle constraints, and the intersections of softness and planarity circles yield the four cells in Table 2. The argument fixes half-tangents but not the actual curves or surfaces of the cells, so the resulting classes are second-order descriptions rather than full cell geometries.
What would settle it
Compute, from a numerical Voronoi decomposition of the skeletal graphs of the Schwarz P and D surfaces, the four unit tangent directions of the edges at a vertex of each cell and compare them with the nodal sets listed for (g2) and (i2). If the sets are not related by the appropriate space-group symmetries, the claimed equivalence fails; a direct enumeration that finds a third full-symmetry soft solution in the same first-order class would refute Theorem 1.
Extended reading notes
Core claim
The central discovery is a complete second-order census of soft tilings in the first-order class of the (e2) tiling, the Dirichlet-Voronoi tiling of the bcc lattice. The Extended Edge Bending (EEB) algorithm computes the possible unit half-tangent vectors at a node consistent with a prescribed symmetry group, then imposes softness: in every cell at a node, some pair of half-tangent vectors must satisfy $\mathbf{u}_1 \cdot \mathbf{u}_2 = -1$, meaning the two edge directions meet smoothly with no sharp corner. Solving these equations yields exactly two inequivalent solutions under the full bcc symmetry group $Im3m$ --- the standard soft cell (f2) and the non-standard soft cell (g2) --- and exactly four solutions when the symmetry is relaxed to the tetrahedral group $Pn3m$ and at least one planar face is required: (f2), (g2), (h2), and (i2). The paper then argues, by elimination, that the Voronoi cells of the Schwarz P and D minimal surfaces belong to the second-order classes of (g2) and (i2), respectively, because they share the vertices of the regular map $\{6,4|4\}$, have planar faces, and have the required symmetries.
Load-bearing premise
The link between minimal surfaces and soft cells depends on the stated facts that the Voronoi cells of the Schwarz P and D surfaces share the same corner positions as the bcc Voronoi tiling, have the claimed symmetries, and have flat faces; if any of those descriptions is inaccurate, the link can fail even if the classification counts are correct.
Editorial extensions
If this is right
- If Proposition 1 holds, the Voronoi tilings of the Schwarz P and D surfaces are not exotic outliers but members of the same second-order soft-cell families as (g2) and (i2), so results about those families transfer to the minimal-surface tilings.
- The two-count in Theorem 1 closes the census for full-symmetry soft tilings in the (e2) first-order class: no undiscovered full-symmetry soft cell exists in this class.
- The four-count in Theorem 2 says that any soft cell in this class that is first-order equivalent to (e2), has at least tetrahedral symmetry, and has at least one planar face must be one of (f2), (g2), (h2), or (i2).
- The one-parameter family through (e2), (f2), (h2), and the Kelvin cell gives a continuous route from the truncated octahedron to standard soft tilings, placing Kelvin's foam as an intermediate configuration on that route.
- The gyroid's soft cell, with the highest computed softness value $\sigma = 0.576$, shows that the EEB method also produces candidates outside the (e2)-centered classification, such as the soft tiling induced by the gyroid structure.
Reading between the lines
- The identification of the Schwarz P and D cells with (g2) and (i2) is made by elimination from external descriptions; a direct numerical computation of the half-tangent sets from parametrized skeletal graphs would remove that reliance and is a natural verification.
- Because the classification is confined to the first-order class of (e2), the gyroid's soft cell, with its degree-3 skeletal graph and nonconvex first-order polyhedron, lies outside the theorems; an analogous EEB census for the gyroid's own first-order class could test whether its unusually high softness value is extremal.
- In the foam direction, one could evaluate total surface area along the one-parameter family that includes the Kelvin cell; if Kelvin's cell is a local area minimum, the soft-cell framework may provide a new coordinate system for Plateau's problem.
- The second-order classification suggests defining softness as a continuous spectrum rather than a binary property, with the parameter $\sigma$ measuring how close a tiling is to the soft limit; this could make the standard-versus-non-standard distinction a quantitative rather than qualitative one.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper introduces an Extended Edge Bending (EEB) algorithm and applies it to the Dirichlet–Voronoi tiling on the bcc lattice, called the (e2) tiling. The main classification results are Theorem 1, stating that exactly two second-order equivalence classes of soft tilings share the full symmetry group of (e2), and Theorem 2, stating that exactly four such classes have at least tetrahedral symmetry and at least one planar face. The paper then claims that the Voronoi cells of the Schwarz P and Schwarz D minimal surfaces belong to the non-standard classes (g2) and (i2) of Theorem 2 (Proposition 1), and it constructs one-parameter families of first-order (e2)-equivalent tilings that include the Kelvin foam and connect the two Schwarz cells. The proofs in Section 3 are largely explicit: three complete sets of softening equations are listed and solved on the unit sphere, and Table 2 gives the resulting cells. The main weaknesses are a concrete error in the stated normal for one planar-face great circle and a gap in the proof of Proposition 1, which relies on an unproved observation about Voronoi partitions of TPMS labyrinths rather than a direct computation of half-tangents.
Significance. If the classification is correct, it provides a useful structural framework for soft tilings at the second-order level and gives a concrete algebraic handle on the half-tangent data of space-filling cells. The connection to Schwarz P and D minimal surfaces is an appealing geometric result with potential relevance to materials science (mesoatoms) and to the theory of triply periodic minimal surfaces. The explicit enumeration of all soft tilings up to second order in the first-order class of the bcc Voronoi tiling is a valuable reference result, and the one-parameter families, including the Kelvin foam, are elegant. However, the paper's central applied claim rests on Proposition 1, whose current proof is not fully self-contained; the classification theorems themselves appear sound and are supported by explicit computation.
major comments (1)
- [§4.1, Observation 1 and Proposition 1] Proposition 1 is the load-bearing link between the abstract classification and the Schwarz minimal surfaces, but its proof is not complete. Observation 1 asserts that a Voronoi partition of a TPMS labyrinth yields a soft tiling because tube cross-sections are smooth; however, softness in Definition 2 is an antipodal-half-tangent condition at nodes, and smoothness of cross-sections is not shown to imply that condition. The proof of Proposition 1 then excludes the standard soft cells (f2) and (h2) by asserting that their planar faces are non-smooth while the Schwarz cells have smooth planar faces, but this is a higher-order geometric property that is not part of the second-order classification in Theorem 2. Since the proof of Proposition 1 does not compute the half-tangent data of the Schwarz P and D Voronoi cells, the elimination from Theorem 2 is not rigorous as written. A direct computation of the half-tangents of these Voronoi cells, or a rigorous derivation from the cited vertex-set and symmetry data, is needed to confirm the identification with (g2) and (i2).
minor comments (4)
- [Table 2 and §3.6] The great-circle labels in Table 2 are not consistent with the notation introduced in §3.4.2: for example, the (g2) row uses 'gadbc' where the text defines g_abcd as the circle containing that solution. Please align the table with the notation in the proof.
- [Throughout] There are several typos and stylistic inconsistencies: 'KEL VIN' in the title, 'polyhedic' for 'polyhedric', 'ahve' in §4.1, and 'monohedric' for 'monohedral' in places. These should be corrected.
- [Reference [2]] The entry for Schoen's technical report appears garbled ('Technical Note, s2-43:NASA TN D–5541'); please verify the exact report number and formatting.
- [§4.1, Proposition 1 proof] The phrase 'both Schwarz Voronoi cells have n smooth, planar faces' is ambiguous: it is unclear whether all faces are planar or whether some are curved, and in what sense the (f2)/(h2) faces are 'non-smooth'. This should be clarified, and the property used for the elimination should be stated precisely.
Circularity Check
No significant circularity: the classification is derived from symmetry/softening equations, and the Schwarz identification rests on cited external geometry plus an unproved but non-circular Observation 1.
full rationale
The central classification (Theorems 1 and 2) does not reduce to its inputs. Section 3 solves the softening equations ui·uj = -1 from Definition 2 for the half-tangent vector a of the (e2) first-order class, using the Im3m or Pn3m symmetry transformations and the planarity constraints (13)-(14); the isolated solutions in Table 2 are derived, not fitted. No target-dependent parameter enters the derivation, and the uniqueness statements are proven by enumeration of the equation systems. The paper's self-citation to [1] supplies terminology, the (e2) label, the previously constructed standard (f2) cell, and the softness-value convention; these are definitions and an independent prior construction, not assumptions equivalent to the new results. Proposition 1 is an elimination argument: it uses Theorem 2 together with external facts from Schoen [2] and Coxeter [12] (the Schwarz P/D cells share the {6,4|4} vertex set and have octahedral/tetrahedral symmetry) and with Observation 1 for softness. Those premises are external or asserted rather than proved; if any is wrong, Proposition 1 would fail, but that is a correctness gap, not a circular reduction. Observation 1 is explicitly an unproved implication from smooth tube cross-sections to the half-tangent softness condition (Definition 2), so it is flagged as missing support, but it is not equivalent to the conclusion by construction. The one-parameter families and Kelvin/PD cells are computed from constraint equations (15)-(18) and are not predictions derived from their own targets. Section 5.4 further limits Proposition 1 to second-order equivalence, weakening rather than smuggling the claim. Hence no step exhibits Eq. X = Eq. Y by construction or a fitted parameter renamed as a prediction.
Assumptions & free parameters
assumptions (4)
- domain assumption The unit cells of the Schwarz P and D surfaces are first-order equivalent to the (e2) tiling and carry the vertices of the regular map {6,4|4} (Schoen [2], Coxeter [12]).
- domain assumption The Voronoi cells of a TPMS labyrinth have planar faces and smooth edges, making the resulting Voronoi partition a soft tiling (Observation 1).
- standard math The space groups Im3m and Pn3m and their point groups (order 48 and 24) are as given in the International Tables [8].
- domain assumption A soft cell is fully characterized to second order by its half-tangent vectors; higher-order shape choices (circular edges, minimal faces) do not affect the equivalence classes (Definitions 1 and 2).
Cite this review
Pith. "Pith review of Soft cells, Kelvin's foam and the minimal surfaces of Schwarz." pith.science (2026). https://pith.science/paper/AKIW5JFX
@misc{pith2026241204491,
author = {Pith},
title = {Pith review of: Soft cells, Kelvin's foam and the minimal surfaces of Schwarz},
year = {2026},
howpublished = {\url{https://pith.science/paper/AKIW5JFX}},
note = {Machine review of arXiv:2412.04491}
}
read the original abstract
Recently, we introduced a new class of shapes, called soft cells which fill space as soft tilings without gaps and overlaps while minimizing the number of sharp corners. We introduced the edge bending algorithm that deforms a polyhedral tiling into a soft tiling and we proved that an infinite class of polyhedral tilings can be smoothly deformed into standard soft tilings. Here, we demonstrate that certain triply periodic minimal surfaces naturally give rise to non-standard soft tilings. By extending the edge-bending algorithm, we further establish that the soft tilings derived from the Schwarz P and Schwarz D surfaces can be continuously transformed into one another through a one-parameter family of intermediate non-standard soft tilings. Notably, by carrying its combinatorial structure, both resulting tilings belong to the first order equivalence class of the Dirichlet-Voronoi tiling on the body-centered cubic bcc lattice, highlighting a deep geometric connection underlying these minimal surface configurations. By requiring identical end-tangents for edges in a first order class, we also define second order equivalence classes among tilings and prove that there exist exactly two such classes among soft tilings which share the full symmetry group of the DV-bcc tiling. Additionally, we construct a one-parameter family of tilings bridging standard and non-standard soft tilings, explicitly including the classic Kelvin foam structure as an intermediate configuration. This construction highlights that both the soft cells themselves and the geometric methods employed in their generation provide valuable insights into the structural principles underlying natural forms. We also present the soft tiling induced by the gyroid structure.
Figures
Figures from the paper (4 more)
Forward citations
Cited by 2 Pith papers
-
Soft cells, Tubular Tilings and the Hidden Phases in Binary Mixtures
Tubular tilings of binary mixtures obey global Euler balance laws that infer hidden-phase topology from the observable phase plus interface geometry, and for d>2 they form a subclass of soft (corner-free) tilings.
-
Prismatic Soft Cubes
Under lattice-direction and planar-edge conditions, exactly 26 soft, space-filling cubes and 68 fundamental domains exist.
Reference graph
Works this paper leans on
-
[1]
G. Domokos, A. Goriely, ´A. G. Horv´ ath, and K. Reg˝ os. Soft cells and the geometry of seashells. PNAS Nexus , 3(9):pgae311, 2024
work page 2024
-
[2]
Infinite periodic minimal surfaces without self-intersections
Alan Schoen. Infinite periodic minimal surfaces without self-intersections. Technical Note, s2-43:NASA TN D–5541, 1970
work page 1970
-
[3]
Gregory M. Grason and Edwin L. Thomas. How does your gyroid grow?a mesoatomic perspec- tive on supramolecular, soft matter network crystals. Physical Review Materials , 7:045603, 2023
work page 2023
-
[4]
H.A. Schwarz. Gesammelte Mathematische Abhandlungen . Berlin: Springer, 1933. SOFT CELLS, KEL VIN’S FOAM AND THE MINIMAL SURF ACES OF SCHW ARZ 19
work page 1933
-
[5]
Lu Han and Shunai Che. An overview of materials with triply periodic minimal surfaces and related geometry: From biological structures to self-assembled systems. Advanced Materials, 30:1705708, 2018
work page 2018
-
[6]
On the division of space with minimum partitional area
William Thomson. On the division of space with minimum partitional area. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science , 24(151):503–514, 1887
-
[7]
G.E. Budd. Morphospace. Current Biology, 31:R1141–R1224, Oct 2021
work page 2021
-
[8]
M.I. Aroyo (ed). International Tables for Crystallography, Volume A: Space-group symmetry. International Union of Crystallography, 2016
work page 2016
Show all 21 references
-
[9]
Alan L. Mackay. Periodic minimal surfaces. Nature, 314:604–606, 2015
2015
-
[10]
B. Riemann. ¨ uber die fl¨ che vom kleinsten inhalt bei gegebener begrenzung. Abhandlungen der K¨ oniglichen Gesellschaft der Wissenschaften zu G¨ ottingen, 13:3–52, 1867
-
[11]
Gandy and J
P.J.F. Gandy and J. Klinowski. Exact computation of the triply periodic schwarz p minimal surface. Chemical Physics Letters , 322:579–586, 2000
2000
-
[12]
H.S.M. Coxeter. Regular skew polyhedra in three and four dimensions and their analogues. Proceedings of the London Mathematical Society , s2-43:33–62, 1936
1936
-
[13]
J.A.F. Plateau. Statique Exp´ erimentale et Th´ eorique des Liquides soumis aux seules Forces Mol´ eculaires. Paris : Gauthier-Vilars, 1873
-
[14]
Kelvin’s foam structure: a commentary
D Weaire. Kelvin’s foam structure: a commentary. Phil. Mag. Letters , 88(2):91–102, 2007
2007
-
[15]
R. Bitsche. Space-filling polyhedra as mechanical models for solidified dry foams. Master’s thesis, Technical University of Vienna, 2005
2005
-
[16]
Daxner, R.D
T. Daxner, R.D. Bitsche, and H.J. B¨ ohm. Space-filling polyhedra as mechanical models for solidified dry foams. Materials Transactions, 47:2213–2218, 2006
2006
-
[17]
Deuxi´ eme note sur les surfaces a lignes de courbure sph´ eriques.C.R
Ossian Bonnet. Deuxi´ eme note sur les surfaces a lignes de courbure sph´ eriques.C.R. Acad. Sci.Paris, 36:389–391, 585–587, 1853
-
[18]
H. Park, S. Jo, B. Kang, K. Hur, S.S. , Oh, D.Y. Ryu, and S. Lee. Block copolymer gyroids for nanophotonics: significance of lattice transformations. Nanophotonics, 11:2583–2615, 2022
2022
-
[19]
Wenpeng Shan and Edwin L. Thomas. Gradient transformation of the double gyroid to the double diamond in soft matter. ACS Nano, 18:9443–9450, 2024
2024
-
[20]
´A. G. Horv´ ath. On the dirichlet—voronoi cell of unimodular lattices. Geometriae Dedicata, 63:183–191, 1996
1996
-
[21]
G.F. Voronoi. Nouvelles applications des param` etres continus ` a la th´ eorie des formes quadra- tiques ii. Geometriae Dedicata, 136:67–181, 1909. G´abor Domokos, HUN-REN-BME Morphodynamics Research Group and Dept. of Morphology and Geometric Modeling, Budapest University of...
1909
Reviewed August 12, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.