{"id":"f17f2e81-c813-4aca-bd9d-7e9dee76ab35","arxiv_id":"1908.07881","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Emulsions with 95% internal phase can spontaneously organize their droplets into a random Apollonian packing, a scale-invariant space-filling structure.","lead":"A new kind of ultra-concentrated emulsion was found in which tiny oil droplets pack like the famous Apollonian gasket, filling space with ever-smaller spheres. The discovery suggests how to make scale-invariant soft materials and fractal interfaces without special equipment.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"DLS on the diluted HIPE is the load-bearing measurement: no control shows dilution preserves the 95%-volume-fraction droplet sizes, so the Apollonian exponent and the SAXS match may describe the diluted state, not the aged packing.","rationale":"The reader's verdict is CONDITIONAL, and my analysis points to the same condition. The paper has real positives: direct optical microscopy shows spherical droplets at low surfactant and high φ; the SAXS data are genuine beamline measurements; and the ORAP exponent d_f = 2.47 is consistent with independent algorithms. But the load-bearing input to the claim is the droplet-size distribution, measured only after dilution. The paper's own statement that droplets <1 μm cannot be resolved under optical microscope makes the ex-situ DLS the sole source of the small-droplet power-law tail. A dilution artifact would change d_f and would invalidate the ORAP comparison, because the simulated structure is built from the same size distribution. This is not an internal inconsistency, but an unverified assumption with a concrete falsifiable test. The mechanism section (coalescence-fragmentation) also lacks details of the Monte Carlo simulation, but that is secondary: even if the mechanism were wrong, the existence of an Apollonian-like packing could stand; conversely, if the size measurement is wrong, the existence claim falls. Therefore I keep the CONDITIONAL verdict without moving it.","tokens_in":7278,"tokens_out":5555,"duration_ms":58336,"concrete_test":"Perform confocal fluorescence microscopy on the intact aged HIPE: label the oil phase with a lipophilic dye, image thick sections, count at least 10^4 droplets, and construct the in-situ n(a) over the same diameter decade range used in Fig. 2. Compare the in-situ power-law exponent and range directly with the DLS result (d_f = 2.48–2.50). If the in-situ distribution matches, the concern is retired; if it deviates, the Apollonian assignment is an artifact of dilution.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Eq. (1) and Fig. 2 are the quantitative foundation of the paper: the droplet diameter distribution n(a) ∝ a^{-(d_f+1)} with d_f = 2.48–2.50 is what identifies the packing as Apollonian (d_f ≈ 2.47). These data, however, are Dynamic Light Scattering measurements made after the HIPE at φ = 0.95 with only 0.6% surfactant is diluted in excess continuous phase. The paper contains no control demonstrating that dilution preserves the in-situ droplet size distribution. In such a surfactant-poor, 95% internal-phase emulsion, the continuous films are thin and fragile; the act of dilution can induce coalescence (changing the large-size tail and the measured exponent) or, during sample handling, breakup (adding small droplets). Moreover, the optical microscopy in Fig. 1 cannot resolve droplets below 1 μm, so the small-droplet side of the power law is only accessible through the ex-situ DLS measurement. The SAXS analysis inherits the same assumption: P_exp(q) in Eq. (2) is the form factor of droplets with the DLS-derived size distribution, so the Fig. 4 comparison of S_exp(q) with the ORAP simulation is only as valid as the ex-situ distribution. If dilution alters droplet sizes, then neither the Apollonian exponent nor the claimed structure-factor agreement represents the actual aged HIPE, and the central claim collapses to 'diluted emulsions have a power-law distribution', which is not the paper's claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":7609,"tokens_out":3510,"duration_ms":31999,"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":[{"comment":"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.","section":"Droplet-size distributions, Fig. 2, Eq. (1)"},{"comment":"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.","section":"Relevance of ORAP model, Fig. 4"},{"comment":"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.","section":"Coalescence in Apollonian HIPEs"}],"minor_comments":[{"comment":"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.","section":"Fig. 2"},{"comment":"The phrase 'rescaled by by the Wigner-Seitz radius' contains a duplicated 'by'; it should read 'rescaled by the Wigner-Seitz radius'.","section":"Fig. 4 caption and text"},{"comment":"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).","section":"Introduction / Eq. (1)"},{"comment":"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.","section":"Relevance of ORAP model"},{"comment":"The sentence 'an latex foam production' contains a grammatical error and should read 'and latex foam production'.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The novelty is significant and the work fits the journal. The main risk is the unvalidated DLS dilution step; if the authors can supply an in-situ control or a dilution series showing invariance, the central claim will be much stronger. The SAXS comparison also needs to be framed as a consistency check unless an independent prediction is added."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe headline: this paper reports a genuinely new experimental observation—surfactant-poor HIPEs spontaneously evolve a power-law droplet-size distribution with an exponent close to the Apollonian value (df ≈ 2.47), and their SAXS structure factor matches a simulated random Apollonian packing. If it holds up, it opens a new class of scale-invariant concentrated emulsions with practical implications for templating and transport.\n\nWhat the paper does well: the protocol is simple and reproducible across three stirring speeds; the month-long aging at rest is a clean experiment; the SAXS comparison with a simulation is a sensible approach; and the authors explicitly caution that matching an ORAP does not prove the same mechanism. The Apollonian literature is well used.\n\nThe soft spots are real and need attention. The load-bearing measurement is the droplet-size distribution from a Malvern Mastersizer after diluting a 95%-volume-fraction HIPE in excess continuous phase. There is no control showing that dilution preserves the in-situ droplet sizes. In a surfactant-poor system with thin, fragile films, dilution could coalesce or fragment droplets, and the small-droplet end of the power law—essential for the exponent—is invisible to optical microscopy. If the distribution is altered, the Apollonian exponent and the SAXS-derived structure factor inherit the distortion.\n\nSecond, the SAXS validation is a consistency check, not an independent prediction: the simulated ORAP already has the Apollonian property, and the match at φ=0.92 vs. 0.95 is encouraging but not decisive. The authors say the main peak is nearly insensitive to φ, but the supporting data are not shown.\n\nThird, the claimed coalescence-fragmentation mechanism rests on a Monte Carlo simulation that is never described—no parameters, no validation, no figure. That is a substantial missing piece.\n\nFinally, the power-law fit lacks error bars and specified fit ranges, so the exponent df = 2.48–2.50 is hard to weigh.\n\nOverall, the central observation is novel and likely correct, but the Apollonian interpretation is not yet nailed down. This paper deserves a serious referee: it should go to peer review, where the dilution control and simulation details can be demanded. I would take it to our reading group as a discussion piece on what counts as evidence for self-similar packing in soft matter.\n\nBest,","headline":"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.","tokens_in":8032,"tokens_out":3009,"would_cite":false,"duration_ms":29839,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"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.","keywords":["high internal phase emulsion","Apollonian packing","polydisperse emulsion","coalescence","fractal dimension","small-angle X-ray scattering","droplet size distribution","scale invariance"],"falsifier":"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.","tokens_in":7077,"feed_emoji":"🫧","tokens_out":7060,"duration_ms":65097,"temperature":0.7,"pith_summary":"The paper reports that certain concentrated oil-in-water emulsions made with very little surfactant spontaneously organize their droplets into an Apollonian packing, a scale-invariant space-filling arrangement in which each void is occupied by the largest possible smaller sphere. At an internal-phase volume fraction of 95%, the droplets remain spherical instead of deforming into polyhedra, and their size distribution evolves by coalescence toward a power law with an exponent corresponding to a fractal dimension of about 2.47 to 2.50. Small-angle X-ray scattering from aged samples matches the structure factor of a numerically simulated random Apollonian packing, supporting the claim that the real emulsion shares the same spatial organization. The authors propose a coalescence-fragmentation mechanism in which pairs of droplets merge and then split into multiple non-overlapping spherical daughters, conserving total volume and minimizing surface area. If correct, this makes a concentrated emulsion the first experimentally realized liquid system that self-assembles into an Apollonian packing without engineering discrete droplet-size populations.","feed_headline":"Droplet packings in 95% emulsions follow Apollonian fractal rules","feed_subtitle":"At 95% oil and 0.6% surfactant, droplets self-organize into the same structure as simulated Apollonian packings.","key_machinery":"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.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Supplies the Osculatory Random Apollonian Packing algorithm used to simulate the packing and compute its structure factor.","marker":"[28]"},{"why":"Gives the standard relation connecting measured intensity, average form factor, and structure factor used to extract the experimental $S(q)$.","marker":"[31]"},{"why":"Provides an alternative Apollonian construction whose reported fractal dimension agrees with the measured exponent.","marker":"[26]"},{"why":"Reports a global-optimization Apollonian exponent range that brackets the emulsion's measured fractal dimension.","marker":"[27]"},{"why":"Establishes the fractal-geometry framework in which power-law sphere distributions fill space.","marker":"[22]"},{"why":"Provides the observable precedent of coalescence producing Apollonian packings in deposited metal drops.","marker":"[36]"},{"why":"Defines the standard surfactant-rich HIPE conditions that the present low-surfactant system contrasts with.","marker":"[12]"},{"why":"States the power-law size distribution for the densest Apollonian packing, used to identify the exponent.","marker":"[24]"}],"fun_headline_variants":["Apollonian packing emerges in 95% emulsions with spherical droplets","95% oil emulsions: droplets spherical, packing Apollonian","Surfactant-poor emulsions self-organize into Apollonian packing at 95% oil","Polydisperse emulsions at 95% oil follow Apollonian fractal rules"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Apollonian packing emerges in 95% emulsions with spherical droplets","95% oil emulsions: droplets spherical, packing Apollonian","Surfactant-poor emulsions self-organize into Apollonian packing at 95% oil","Polydisperse emulsions at 95% oil follow Apollonian fractal rules"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000774,"raw_usage":{"total_tokens":3412,"prompt_tokens":918,"completion_tokens":2494,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":534,"completion_tokens_details":{"reasoning_tokens":2410}},"tokens_in":534,"tokens_out":2494,"duration_ms":17255,"temperature":1.0,"reasoning_tokens":2410,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:53:39.014442+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[],"review_version":1}