REVIEW 3 major objections 5 minor 9 references
Apollonian Packing in Polydisperse Emulsions
T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read This paper reports that emulsions made with very little surfactant spontaneously pack their droplets into a scale-invariant Apollonian arrangement at 95% internal-phase volume fraction.
desk verdict A genuinely new experimental result—surfactant-poor HIPEs evolve to near-Apollonian power-law size distributions—that is hampered by a load-bearing dilution step and an undescribed mechanism simulation; still worth refereeing. 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 central object is the Random Apollonian Packing, specifically the Osculatory Random Apollonian Packing algorithm used to generate a disordered space-filling sphere packing. In an ORAP, a point is chosen at random in a void, and the largest sphere that fits there without overlapping its neighbours is inserted; iterating this procedure produces a scale-invariant packing whose size distribution is a power law with fractal dimension about 2.47. In the paper's argument this algorithm serves as the structural model: the authors compare its computed structure factor with the measured SAXS structure factor of the emulsions, and they also use a coalescence-fragmentation simulation in which two coalescing droplets immediately split into several spherical daughters that maximally fill the available space, yielding the same Apollonian exponent.
What would settle it
Measure the droplet-size distribution of an aged surfactant-poor HIPE without dilution, for example by confocal microscopy of a fluorescently labelled oil phase or by in-situ ultra-small-angle scattering, and check whether it still follows a power law with exponent about 2.47 to 2.50. If the undiluted distribution does not match the DLS result, the Apollonian claim fails; likewise, if the specific surface measured by SAXS does not increase with ageing time, the coalescence-fragmentation mechanism is not responsible.
Extended reading notes
Core claim
The central claim is that a class of high-internal-phase-ratio emulsions exists in which oil droplets at 95% volume fraction remain spherical and pack according to Apollonian rules. The discovery is empirical: when the surfactant concentration is reduced to about 0.6 wt% and oil is added dropwise under shearing, the emulsion does not form the usual polyhedral foam-like structure but a flowing dispersion whose droplet sizes spontaneously reorganize, after roughly a month at rest, into a power-law distribution $n(a) \propto 1/a^{d_f+1}$ with $d_f \approx 2.47$ to $2.50$, the Apollonian exponent. The authors support this identification by showing that the experimental structure factor, with its characteristically low main peak $S_{\max} \approx 1.1$ to $1.2$ and no translational order, agrees with that of a numerically simulated oscillatory random Apollonian packing at similar volume fraction. They further argue that coalescence and fragmentation acting together, under the constraints that total volume is conserved and droplets stay spherical, provide the physical mechanism by which such a packing is reached.
Load-bearing premise
The load-bearing assumption is that diluting the aged HIPE in excess continuous phase before dynamic light scattering does not change the droplet-size distribution, so the measured power-law exponent reflects the droplets inside the concentrated emulsion rather than an artefact of dilution-induced coalescence or breakup.
Editorial extensions
If this is right
- High-internal-phase emulsions can be made with a surfactant concentration as low as 0.6 wt% and still remain stable enough to be studied, because the polydisperse packing eliminates the need for surfactant films to resist droplet deformation.
- The droplet-size distribution spontaneously evolves to a power law with the Apollonian exponent regardless of the initial shear rate (200 to 1000 rpm), so the final structure is an attractor of the ageing process rather than a mixing artefact.
- A concentrated polydisperse emulsion can have a nearly featureless structure factor with a low main peak around 1.1 to 1.2, clearly distinguishing it from translationally ordered monodisperse HIPEs.
- Emulsion templating can use these Apollonian emulsions to produce ultra-dense or ultra-porous solids, and the fractal oil-water interfaces provide test systems for models of thermal and electrical conduction in fractal media.
- Because the droplets stay spherical at 95% volume fraction, the process attains higher internal-phase ratios than would be possible by translational repetition of identical spherical droplets.
Reading between the lines
- Extension: the proposed coalescence-fragmentation rule could be tested directly with time-resolved confocal microscopy, which should show a single coalescence event immediately followed by multiple fission events in the crowded emulsion.
- Extension: the same geometrical constraint set, volume conservation plus spherical daughters, may apply to other coalescing dispersions beyond emulsions, such as metal films or sintering particles, so the Apollonian exponent might be a generic attractor for coalescence-driven coarsening.
- Extension: the power-law distribution implies a huge population of very small droplets, so measurements with a finite resolution cut-off will bias the measured fractal dimension; in-situ scattering or microscopy across a wider size range could reveal whether the smallest droplets truly follow the same power law.
- Extension: if the dilution step used for dynamic light scattering preserves the in-situ droplet-size distribution, then routine size measurements could screen surfactant-poor formulations for Apollonian packing; if dilution does not preserve it, the claimed exponent would need to be re-measured in the concentrated state.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript reports an experimental study of surfactant-poor high internal-phase emulsions (HIPEs) at internal volume fraction φ = 0.95. It claims that aged HIPEs contain spherical droplets whose size distribution obeys a power law n(a) ∝ a^{-(d_f+1)} with d_f = 2.48–2.50, close to the Apollonian exponent, and that the SAXS structure factor matches a numerically simulated random Apollonian packing (ORAP) at φ = 0.92. The authors propose a coalescence-fragmentation mechanism with volume and sphericity conservation as the origin of the Apollonian arrangement. The central claim is the existence of self-organized Apollonian droplet packings in these emulsions.
Significance. If correct, the result would be significant: it would demonstrate that a concentrated emulsion can spontaneously evolve into a scale-invariant, space-filling droplet packing, a phenomenon that has generally been regarded as practically inaccessible. The paper's strengths are the simple and reproducible emulsification protocol, the quantitative link between the measured power-law exponent and the known random Apollonian packing exponent, and the explicit caution that the SAXS comparison does not imply that the emulsion evolved by the same mechanism as ORAP. However, the support is currently incomplete because the key measurements rely on an unvalidated dilution step and on a simulation whose input already contains the Apollonian property.
major comments (3)
- [Droplet-size distributions, Fig. 2, Eq. (1)] The power-law exponent d_f = 2.48–2.50, which is the main quantitative evidence for Apollonian packing, is measured by dynamic light scattering after diluting the HIPE in excess continuous phase. The paper provides no control demonstrating that dilution preserves the in-situ droplet size distribution; in a surfactant-poor HIPE at φ = 0.95, dilution can plausibly trigger coalescence or breakup, and the optical microscope images in Fig. 1 cannot resolve droplets below 1 μm. Because the SAXS analysis in Eq. (2) also uses the DLS-derived size distribution to compute P_exp(q), the dilution uncertainty propagates into the structure-factor comparison. Please provide an in-situ control or an explicit quantitative argument that the measured distribution is unchanged by dilution.
- [Relevance of ORAP model, Fig. 4] The comparison between S_exp(q) and S_sim(q) is a consistency check rather than an independent test, because the simulated ORAP is generated with 'the same droplet-diameter distribution (therefore the same d_f and the same ratio a_min/a_max)' as the experimental system, which already encodes an Apollonian-like exponent. The agreement therefore does not independently confirm Apollonian ordering. Additionally, the simulation is at φ = 0.92 while the experiments are at φ = 0.95; the statement that S_sim(q) is almost insensitive to φ is only checked numerically for 0.84 ≤ φ ≤ 0.92. Please provide a test that does not input the measured distribution, or state explicitly what feature of S_sim(q) is independent of the input distribution.
- [Coalescence in Apollonian HIPEs] The Monte Carlo simulation of the proposed coalescence-fragmentation mechanism is described only in qualitative terms ('by allowing pairs of coalescing spheres to fission into multiple non-overlapping daughter spheres'). No algorithm, rule for choosing daughter spheres, volume-fraction range, or statistical analysis is given, so the reader cannot assess whether the claimed Apollonian exponent is an emergent outcome or an imposed constraint. This is load-bearing for the mechanistic claim, which is a major part of the paper's novelty.
minor comments (5)
- [Fig. 2] The fit to Eq. (1) is shown only as a guide line; please report the fit range in droplet diameters, the number of data points, and error bars or a goodness-of-fit measure.
- [Fig. 4 caption and text] The phrase 'rescaled by by the Wigner-Seitz radius' contains a duplicated 'by'; it should read 'rescaled by the Wigner-Seitz radius'.
- [Introduction / Eq. (1)] The quantity d_f is called the fractal dimension of interfaces, but Eq. (1) defines it through the exponent of the diameter distribution; please clarify the relationship (e.g., via the cumulative distribution and the space-filling condition).
- [Relevance of ORAP model] The numerical check that S_sim(q) is insensitive to φ for 0.84 ≤ φ ≤ 0.92 is not shown; please provide the data or a figure supporting this statement.
- [Abstract and Introduction] The sentence 'an latex foam production' contains a grammatical error and should read 'and latex foam production'.
Circularity Check
No significant circularity: the measured power-law exponent and the SAXS comparison are independent of the Apollonian model's definition.
full rationale
The paper's derivation chain is: (i) measure the droplet-size distribution by DLS and fit n(a) ∝ a^{-(d_f+1)} with d_f = 2.48–2.50; (ii) note that external literature values for Random Apollonian Packing give the same power law with d_f ≈ 2.45–2.52; (iii) simulate an ORAP and compare its structure factor with the experimentally extracted S_exp(q). No step reduces to its own input by construction. The size-distribution exponent is an independent measurement, not fitted to the Apollonian model. The ORAP simulation is built with the measured size distribution, but the spatial arrangement entering S_sim(q) is not encoded by that distribution alone, so the SAXS comparison is a genuine test of the packing arrangement rather than a self-definition. The Monte-Carlo coalescence–fragmentation simulation is not shown in detail, but it is a proposed mechanism rather than the basis of the main observation, and the paper explicitly cautions that ORAP does not allow growth or fragmentation, so the mechanism is not equated with the simulation. Self-citations (refs 33–35, used as a counterexample for S_max) are not load-bearing for the central claim. The dilution-sensitive DLS measurement is a legitimate experimental concern, but it is a correctness/robustness issue, not circularity: the paper does not define the Apollonian claim in terms of the diluted-state measurement. Overall, the core identification of Apollonian behavior rests on an external, parameter-free comparison with known Apollonian exponents and on a structure-factor match that is not predetermined by the fitted size distribution. No significant circularity is present.
Assumptions & free parameters
free parameters (3)
- Power-law exponent d_f =
2.48-2.50 (measured)
- Simulation volume fraction φ_sim =
0.92
- Size range ratio a_min/a_max =
matched between simulation and experiment
assumptions (4)
- domain assumption Dilution of the HIPE in excess continuous phase does not alter the droplet size distribution.
- domain assumption The Vrij equation I(q) = (N/V) P(q) S(q) applies to these concentrated polydisperse emulsions.
- ad hoc to paper Coalescence-fragmentation conserves total volume and sphericity and locally maximizes droplet radii.
- ad hoc to paper The Monte Carlo simulation of coalescence-fragmentation produces a power-law distribution with the Apollonian exponent.
Cite this review
Pith. "Pith review of Apollonian Packing in Polydisperse Emulsions." pith.science (2026). https://pith.science/paper/B4GGJREG
@misc{pith2026190807881,
author = {Pith},
title = {Pith review of: Apollonian Packing in Polydisperse Emulsions},
year = {2026},
howpublished = {\url{https://pith.science/paper/B4GGJREG}},
note = {Machine review of arXiv:1908.07881}
}
read the original abstract
We have discovered the existence of polydisperse High Internal-Phase-Ratio Emulsions (HIPE) in which the internal-phase droplets, present at 95% volume fraction, remain spherical and organize themselves in the available space according to Apollonian packing rules. These polydisperse HIPE are formed during emulsification of surfactant-poor compositions of oil-surfactant-water two-phase systems. Their droplet size-distributions evolve spontaneously towards power laws with the Apollonian exponent. Small-Angle X-Ray Scattering performed on aged HIPEs demonstrated that the droplet packing structure coincided with that of a numerically simulated Random Apollonian Packing. We argue that these peculiar, space-filling assemblies are a result of coalescence and fragmentation processes obeying simple geometrical rules of conserving total volume and minimizing surface area.
Reference graph
Works this paper leans on
-
[1]
1 K. J. Lissant, J. Colloid Interface Sci., 1966, 22, 462–468. 2 K. J. Lissant, J. Soc. Cosmet. Chem., 1970, 21, 141–154. 3 H. M. Princen, Langmuir, 1988, 4, 164–169. 4 H. M. Princen and P. Levinson, J. Colloid Interface Sci., 1987, 120, 172–175. 5 S. U. Pickering, J. Chem. Soc., Trans., 1907, 91, 2001–2021. 6 N. R. Cameron and D. C. Sherrington, Adv. Pol...
work page 1966
-
[8]
30 P. S. Dodds and J. S. Weitz, Phys. Rev. E, 2002, 65, 056108. 31 A. Vrij, J. Chem. Phys., 1979, 71, 3267–3270. 32 T. Narayanan, M. Sztucki, P. Van Vaerenbergh, J. Léonardon, J. Gorini, L. Claustre, F. Sever, J. Morse and P. Boesecke, J. Appl. Crystallogr., 2018, 51, 1511–1524. 33 Y. Chevalier, C. Pichot, C. Graillat, M. Joanicot, K. Wong, J. Maquet, P. ...
work page 2002
-
[10]
9 United States Patent, 5633220, 1994,
work page 1994
- [16]
-
[19]
8 United States Patent, US3977472A, 1976,
work page 1976
-
[1820]
15 H. M. Princen and A. D. Kiss, Langmuir, 1987, 3, 36–41. 16 R. A. Leonard and R. Lemlich, AIChE J., 1965, 11, 18–25. 17 T. G. Mason, M. D. Lacasse, G. S. Grest, D. Levine, J. Bibette and D. A. Weitz, Phys. Rev. E, 1997, 56, 3150–3166. 18 T. G. Mason, J. Bibette and D. A. Weitz, Phys. Rev. Lett., 1995, 75, 2051–2054. 19 H. . Princen, J. Colloid Interface...
work page 1987
-
[1983]
23 S. S. Manna and H. J. Herrmann, J. Phys. A. Math. Gen., 1991, 24, L481–L490. 24 S. D. S. Reis, N. A. M. Araújo, J. S. Andrade and H. J. Herrmann, Europhys. Lett., 2012, 97, 18004. 25 M. Borkovec, W. De Paris and R. Peikert, Fractals, 1994, 2, 521–526. 26 S. V. Anishchik and N. N. Medvedev, Phys. Rev. Lett., 1995, 75, 4314–4317. 27 A. Amirjanov and K. S...
work page 1991
-
[2010]
12 H. M. Princen, J. Colloid Interface Sci., 1985, 105, 150–171. 13 M. P. Aronson and M. F. Petko, J. Colloid Interface Sci., 1993, 159, 134–149. 14 A. Viard, Fouret and Pierhugue, Le cuisinier royal: ou l’Art de faire la cuisine, la patisserie et tout ce qui concerne l’office, pour toutes les fortunes, J.-N. Barba, Paris, 10th edn.,
work page 1985
Show all 9 references
-
[8150]
38 W. Wei, J. Cai, X. Hu and Q. Han, Geophys. Res. Lett., 2015, 42, 4833–4840
2015
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.