{"id":"95067a3a-5626-485e-b09c-833b3558624e","arxiv_id":"2607.21315","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Droplet trains from jet breakup experimentally confirm the 2/3 scaling law of one-dimensional ballistic aggregation, with a single length scale ℓ=τv̄²/σ_v collapsing mass growth, despite air drag flipping velocity-correlation signs.","lead":"A falling liquid jet breaks into a train of droplets that merge as they fly; this paper shows their growth follows a universal 2/3-power law controlled by one length scale. The result gives the first clean experimental look at a classic model of ballistic aggregation and suggests why common spray-size models fail.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Experimental confirmation is not decisive: air drag is asserted, not modeled, to preserve the 2/3 exponent and tail shape.","rationale":"The reader's weakest assumption directly identifies the same load-bearing concern: the experimental confirmation relies on the unverified claim that air drag modifies only the prefactor. I agree this is the critical soft spot. The paper has independent strengths: event-driven solvers give an exact treatment of the ideal model, the code and data are available, and the simulation collapse (Fig. 1c) supports the single length scale for the no-drag model. The numerical validation of Frachebourg's distribution is also valuable. However, the experimental system is not drag-free, and the paper provides no quantitative check that drag preserves the exponent. Since the paper's headline claim is experimental confirmation of a universal scaling, this missing drag model is load-bearing. The reader's conditional verdict already requests exactly this; my read does not change that verdict, so UNCHANGED is appropriate.","tokens_in":22035,"tokens_out":4302,"duration_ms":52869,"concrete_test":"Run the event-driven solver augmented with a size-dependent drag force calibrated to the experimental conditions (e.g., Stokes drag with the drag coefficient for water droplets in air over the parameter ranges in Table I). Measure ⟨m⟩/m0 versus z/ℓ up to z/ℓ ≈ 100 for the five experimental flow conditions, and fit the slope. If the fitted exponent deviates from 2/3 by more than the reported ±0.08, the experimental collapse does not uniquely confirm pure ballistic scaling; if the exponent remains 2/3 within errors, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that experiments confirm the 2/3 scaling and the single length scale ℓ rests on the assumption that air drag changes the prefactor but not the exponent or the large-mass tail. The paper states this directly: “We attribute this discrepancy to (i) the effect of air drag…” and later “Despite different correlation behaviors, they both follow a similar scaling…”. However, no quantitative drag model or drag-inclusive simulation is provided. This matters because drag is size-dependent (deceleration ~1/r), so it selectively removes small droplets and increases the average mass. The same mechanism could plausibly accelerate the growth of larger droplets, thereby changing the effective exponent within the probed z/ℓ < 100 window. The experimental fits give 0.67 ± 0.08, but with limited z positions and no drag model, a drag-induced crossover could mimic 2/3. The paper’s claim that a 1D ordering constraint “protects” bulk characteristics is an expectation, not a demonstrated result. Since drag flips the sign of velocity correlations and narrows the mass distribution relative to the ballistic solver, the dynamics are substantially altered; it is not established that the growth exponent survives. Thus the experimental confirmation of the central claim is not decisive: the observed slope could be an effective exponent from drag-modified aggregation rather than the pure ballistic exponent.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper argues that a train of droplets produced by Rayleigh-Plateau breakup of a liquid jet is a physical realization of one-dimensional ballistic aggregation. The authors derive a single length scale, ℓ = τ v̄² / σ_v, and predict ⟨m⟩/m₀ ∼ (z/ℓ)^{2/3} (Eq. 1). They support this with event-driven simulations of both the canonical bidirectional model and the unidirectional droplet-train model, showing a collapse of the local density and average mass growth. Experiments on water jets of three diameters and five velocities are reported, with average-mass growth curves that roughly collapse when plotted against z/ℓ and individual fitted slopes averaging 0.67 ± 0.08. The paper further uses the large-mass tail of the mass distribution to distinguish ballistic aggregation from Smoluchowski-type generalized-gamma predictions, and it reports neighbor velocity/mass correlation coefficients from simulation and experiment. The central conclusion is that a single length scale governs the process and that drag, while inverting velocity correlations and changing the prefactor, does not change the growth exponent or the shape of the large-mass tail.","tokens_in":22404,"tokens_out":5101,"duration_ms":61029,"significance":"If the claims hold, this is an important contribution: it provides the first experimental realization of 1D ballistic aggregation; it identifies a concrete, measurable length scale ℓ for droplet trains; it gives the first numerical validation of Frachebourg's scaling function; and it demonstrates that mean-field Smoluchowski descriptions fail when collision dynamics carry directional memory. The paper is strengthened by the release of the event-driven solver code and processed experimental data, which improves reproducibility. The simulation part is convincing and the identification of ℓ is well grounded. The experimental part, however, is less decisive because the system is drag-contaminated and the drag robustness of the exponent is asserted rather than demonstrated.","major_comments":[{"comment":"The abstract claims that experiments confirm Eq. (1), but the experimental confirmation rests on an unmodeled assumption: that air drag changes the prefactor but not the exponent. The text states this explicitly ('We attribute this discrepancy to (i) the effect of air drag...') but provides no drag-inclusive simulation or quantitative drag model. Since drag deceleration scales inversely with droplet radius, it preferentially removes small droplets and changes collision rates; within the probed window z/ℓ < 100, the fitted slopes 0.67 ± 0.08, with substantial scatter and few z positions per condition, could be an effective exponent produced by drag. I request a drag-inclusive event-driven simulation (e.g., Stokes drag with mass-dependent deceleration) or a quantitative argument showing that the exponent is stable over the probed window. If this cannot be provided, the 'experiments confirm","section":"Fig. 2 and discussion of air drag"},{"comment":"The large-mass-tail discrimination between ballistic aggregation and Smoluchowski-type models is presented as strong evidence, but the statistical support is thin. Eq. (3) is fitted with a fixed λ = 10 and a free amplitude, while Eq. (4) is fitted with ν = 65 in the text and ν = 51 in the caption — an unresolved inconsistency. With ν as a free parameter, a generalized gamma can mimic a stretched/super-exponential tail over a finite interval; one fitted curve does not establish that the mean-field form fails. Please provide a systematic fit scan over both λ and ν, report confidence intervals and goodness-of-fit statistics, and show that the best generalized gamma is outside the data's confidence band for the large-mass tail. Without this, the statement 'mean-field models fail to capture the essential physics' is stronger than the evidence.","section":"Eq. (3), Eq. (4), Fig. 3"},{"comment":"The paper states that a '1D ordering constraint protects bulk characteristics' such as the scaling exponent and tail shape, despite drag flipping the signs of C_mm and C_vv and making C_mv non-antisymmetric. This is an expectation, not a demonstrated result. The sign flips show that the dynamics are substantially altered, yet no mechanism or simulation connects these correlation changes to an unchanged growth exponent. A drag-inclusive simulation that reproduces the experimental correlations and recovers the same exponent would directly test this. As written, the robustness of the exponent to drag is an ad-hoc assumption rather than a finding.","section":"Fig. 4 and concluding discussion"}],"minor_comments":[{"comment":"In the version I received, the normalization of the Frachebourg distribution is not typeset correctly; the denominator of the integral is missing from the displayed equation. Please correct the formula.","section":"Equation (2)"},{"comment":"There are internal inconsistencies in the figure: the text gives ν = 65 for the generalized gamma while the caption gives ν = 51. Also, some equation cross-references appear as 'Eq. (X)', 'Eq. (Y)', 'Eq. (Z)' in figure captions; these placeholders must be resolved in the final version.","section":"Fig. 3 caption/text"},{"comment":"The extracted manuscript contains repeated blocks of figures and captions (e.g., Fig. 3 and Fig. 5 captions are duplicated in the text). The final version should be cleaned to avoid confusion.","section":"General presentation"},{"comment":"The word 'confirm' in the abstract is too strong given the drag caveats and the scatter of the experimental slopes. 'Support' or 'are consistent with' would better match the evidence, unless the requested drag-inclusive analysis is added.","section":"Abstract/Conclusion wording"}],"recommendation":"major_revision","confidential_remarks":"The core theoretical and numerical contribution is sound and publishable, and the open-source code release is a real plus. The main problem is that the experimental part is asked to carry more weight than it can support without a drag-inclusive model or simulation. If the authors add such a simulation and tighten the tail-fitting statistics, the paper would be a strong candidate for publication. The inconsistencies in figure cross-references should also be resolved before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper deserves a serious referee, but the experimental confirmation is softer than the abstract suggests. The new things are real: first experimental realization of 1D ballistic aggregation in a droplet train, first numerical check of Frachebourg's scaling function, and a serviceable one-parameter collapse via ℓ = τv̄²/σ_v. The mapping between unidirectional and bidirectional systems is clean, and the event-driven solvers are public. That part holds up.\n\nWhat I'm less sold on: the drag story. Air drag is invoked to explain the factor-of-two prefactor and the narrowed mass distribution, and to claim that the 2/3 exponent and the tail shape survive. But there's no drag-inclusive simulation and no quantitative drag model in the letter. With slopes averaging 0.67±0.08 over a limited z/ℓ window, the experiments would look the same if drag were producing an effective exponent that mimics 2/3. The '1D ordering constraint protects bulk characteristics' line is an assertion, not a demonstrated result. I don't think that breaks the main claim — the simulation and scaling argument stand on their own — but it means the experimental validation is not decisive on its own.\n\nThe tail-fit evidence is the weakest section. Eq. (3) uses λ=10 as a fit parameter plus a free amplitude, Eq. (4) uses a fitted ν, and the contrast between them is a stretch; both are flexible enough that the comparison doesn't strongly discriminate the two models. The text mentions ν=65 while the figure says ν=51, and the figures have unfinished references (Eq. (X), (Y), (Z)). These are minor but should be caught before publication.\n\nBottom line: the central scaling and the length-scale identification are new and useful, and the paper is honest about its errors and shares code and data. It's a good paper for a reading group and a genuine candidate for peer review — but the authors should either produce a drag model or soften the claim that drag leaves the bulk exponent untouched.","headline":"First experimental staging of 1D ballistic aggregation, with a clean length-scale collapse; the drag story is asserted rather than modeled, so the experimental confirmation is suggestive, not decisive.","tokens_in":22840,"tokens_out":2108,"would_cite":true,"duration_ms":22783,"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":"A droplet train from a liquid jet realizes one-dimensional ballistic aggregation, governed by a single length scale, with mean mass growing as (z/l)^(2/3).","keywords":["ballistic aggregation","droplet trains","liquid jet breakup","scaling law","coalescence","velocity correlations","mass distribution tail","one-dimensional aggregation"],"falsifier":"Measure <m>/m0 and the mass distribution at significantly larger rescaled distances (z/l > 100) or at low ambient air pressure; if the fitted exponent deviates outside roughly 0.6 to 0.75, or if a gamma-type tail fits better than the Frachebourg-derived expression under reduced drag, the claimed universality of the ballistic scaling would be refuted.","tokens_in":1416,"feed_emoji":"💧","tokens_out":2079,"duration_ms":51737,"temperature":0.7,"pith_summary":"This paper claims the first experimental realization of one-dimensional ballistic aggregation, using a train of droplets formed by the breakup of a liquid jet. Scaling arguments identify a single length scale l = tau v^2 / sigma_v such that the average droplet mass grows as (z/l)^(2/3) downstream, and the paper shows this collapse holds across jet diameters from 32 to 180 micrometers and speeds from 2 to 120 m/s. Simulations match the predicted scaling, and the number-weighted mass distribution's large-mass tail matches the ballistic-aggregation form rather than a Smoluchowski-type gamma distribution. Air drag inverts the sign of neighbor velocity correlations without changing the 2/3 exponent or the tail shape. The paper concludes that mean-field descriptions fail when collisions carry directional memory, as demonstrated by these jet-generated sprays.","feed_headline":"Droplet trains confirm 2/3 ballistic aggregation law","feed_subtitle":"Mass growth in a jet's droplet train follows (z/l)^(2/3) across five jet conditions, one length scale doing the work.","key_machinery":"The central identity is the scaling law <m>/m0 ~ (sigma_v z / (tau v_bar^2))^(2/3) = (z/l)^(2/3), where l = tau v_bar^2 / sigma_v is the length scale governing the process. It is derived by a CPY-type scaling argument: a clump of mass N m0 sweeps distance a0 N over time t, and momentum conservation plus the central limit theorem gives its typical momentum fluctuation delta P ~ sqrt(N) m0 sigma_v, an argument valid in one dimension because the geometric ordering prevents the correlations that break the mean-field assumption in higher dimensions. The paper also uses the equivalence between the unidirectional (spatial) and bidirectional (temporal) formulations, and the Frachebourg distribution","core_discovery":"We show that a train of droplets formed by jet breakup is a physical realization of one-dimensional ballistic aggregation, with downstream distance playing the role of time in the comoving frame. The system is statistically equivalent to the canonical bidirectional ballistic aggregate under the mapping t = z/v_bar, and the dynamics is ruled by a single length scale l = tau v_bar^2 / sigma_v. The average mass obeys <m>/m0 ~ (z/l)^(2/3), confirmed experimentally with an average fitted exponent of 0.67 plus or minus 0.08 and by event-driven simulations showing collapse of the local density. Even when air drag makes the velocity correlations negative and flips the signs of mass-mass and velocity","pith_inferences":["If the scaling survives outside the probed z/l < 100 window, the length scale l could serve as a universal collapse parameter for other strongly directional spray and train-like systems, including inkjet printing and drug-delivery sprays, where similar coalescence dynamics occur.","The inversion of correlation signs induced by drag suggests a testable staging: at sufficiently low ambient pressure (as the paper mentions), experimental correlations should approach the ballistic-solver predictions while the 2/3 growth law and tail shape should remain, cleanly separating drag effects from intrinsic ballistic correlations.","The demonstrated sensitivity of the large-mass tail to the coalescence mechanism implies that high-precision droplet-sizing methods, such as the volume-integration technique used here, could be used to infer the physical origin of coalescence in industrial sprays where mean-field fits are currently standard.","One might extend the unidirectional/bidirectional equivalence to non-identical initial velocity distributions (e.g., skewed or heavy-tailed) to test whether the collapse in l and the 2/3 exponent hold beyond the near-normal distributions considered."],"forward_implications":["The observed collapse of <m>/m0 against z/l across five jet conditions establishes l as the controlling parameter for droplet-train coalescence, meaning measurement or prediction of l suffices to predict average mass growth in such systems.","The 2/3 exponent and the large-mass tail shape are preserved even when air drag reverses the sign of neighbor correlations, implying that the 1D ordering constraint protects these bulk characteristics beyond the ideal ballistic model.","The failure of the generalized-gamma (Smoluchowski) fit and success of the Frachebourg-derived tail provide a sharp experimental discriminator for coalescence mechanisms in sprays: if velocity correlations matter, ballistic scaling should appear in the mass tail.","The first numerical validation of Frachebourg's scaling function for one-dimensional ballistic aggregation offers a benchmark for future solvers and for testing other initial velocity distributions.","The framework suggests that directional memory is essential in jet sprays; mean-field models should only be trusted when collision velocities are effectively randomized."],"fun_headline_variants":["Droplet trains confirm one length scale rules ballistic aggregation","Single length scale governs droplet train mass growth","Jet droplets validate 2/3 power law for ballistic aggregation","One length scale dictates droplet aggregation in jet trains"],"cache_read_input_tokens":24192,"weakest_assumption_plain":"The experimental confirmation rests on the premise that air drag alters only the prefactor, not the 2/3 exponent or the large-mass tail shape, without a quantitative drag model or drag-inclusive simulation in the paper.","fun_headline_variants_meta":{"raw":{"variants":["Droplet trains confirm one length scale rules ballistic aggregation","Single length scale governs droplet train mass growth","Jet droplets validate 2/3 power law for ballistic aggregation","One length scale dictates droplet aggregation in jet trains"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000449,"raw_usage":{"total_tokens":2062,"prompt_tokens":667,"completion_tokens":1395,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":411,"completion_tokens_details":{"reasoning_tokens":1332}},"tokens_in":411,"tokens_out":1395,"duration_ms":10936,"temperature":1.0,"reasoning_tokens":1332,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T07:48:37.509639+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure <m>/m0 and the mass distribution at significantly larger rescaled distances (z/l > 100) or at low ambient air pressure; if the fitted exponent deviates outside roughly 0.6 to 0.75, or if a gamma-type tail fits better than the Frachebourg-derived expression under reduced drag, the claimed universality of the ballistic scaling would be refuted.","supporting_citations":[],"review_version":1}