{"id":"62e7d8b0-6068-4660-9ef1-c21b60395517","arxiv_id":"2502.05105","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"In a cough-machine model, viscoelastic mucosalivary fluids form larger, thinner bags that rupture into smaller expelled droplets than Newtonian fluids of similar viscosity.","lead":"Experiments and simulations show that adding a small amount of polymer to a mucus-like liquid makes the thin liquid bags that form during a simulated cough inflate larger and more uniformly, and the droplets that survive in the expelled air jet become smaller. This suggests that the viscoelasticity of airway mucus can change how many small, long-lasting aerosol droplets a cough or sneeze produces.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The single-point PDA measurement at 15 cm may exaggerate the viscoelastic droplet-size reduction if polymer solutions shift more mass into large rim-retraction droplets that fall out before the probe; this sampling bias is acknowledged but unquantified.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the droplet-size comparison is made at one downstream point that preferentially samples small droplets, and the paper does not quantify whether this bias differs between Newtonian and viscoelastic fluids. I agree with that identification. The central claim in the abstract and conclusions is that viscoelasticity reduces droplet size, and the only direct experimental evidence for that reduction is the PDA measurement. The bag-size and lifetime measurements show that viscoelastic bags are larger and live longer, and the simulations show more uniform thinning, but neither provides a measured droplet size distribution. Therefore, if the PDA sampling bias is concentration-dependent, the headline conclusion is not supported. The proposed test directly measures that dependence by comparing size distributions at an additional downstream location and at the exit. If the volume-weighted distribution at the exit already shifts to smaller sizes, the concern is resolved; if it does not, the paper's claimed reduction is an artifact of where the droplets were sampled. I do not see a need to change the reader's CONDITIONAL verdict: the mechanism is plausible and partially supported, but the quantitative droplet-size claim needs this check before the paper can be accepted at face value. I considered other candidate concerns, such as the lack of a surface-tension-matched Newtonian control and the fact that the simulations do not directly predict droplet sizes, but these are secondary to the sampling-bias issue and, in any case, would also be partially addressed by an exit-resolved droplet size measurement.","tokens_in":18700,"tokens_out":5119,"duration_ms":59348,"concrete_test":"Repeat the droplet measurements at a second downstream station, e.g., 5 cm and 25 cm on the same centerline, for Cm = 0%, 0.05%, 0.10%, and 0.20%, and add high-speed shadowgraphy at the channel exit to capture the large rim-retraction droplets and ligaments. Compute the volume-weighted size distribution and the volume fraction of droplets larger than 200 µm at each station. If the >200 µm volume fraction grows with Cm and decays between the 5 cm and 25 cm stations while the mean diameter at 15 cm falls, the reported reduction is a sampling artifact; if the exit-resolved distribution itself shifts toward smaller sizes with increasing Cm, the central claim survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that viscoelasticity reduces expelled droplet size rests on PDA size distributions measured at a single point, 15 cm downstream of the channel exit along the centerline (Section II.A). The authors explicitly note that this location preferentially samples smaller droplets because larger droplets ballistically fall out of the puff. The concern is not the existence of the bias, which is acknowledged, but whether the bias changes with polymer concentration. In the same geometry, bag rupture produces droplets larger than 200 µm from the unstable retraction of rims (Section III.A), and the viscoelastic cases visibly form filaments and ligaments; at Cm = 1.00% no droplets were detected because the liquid was expelled mostly as filaments (Section III.C). Thus, as viscoelasticity increases, a larger fraction of the expelled liquid mass may be partitioned into large droplets, ligaments, or filaments that sediment before reaching the PDA probe. If so, the measured decrease in mean diameter at 15 cm would reflect preferential removal of the large-droplet fraction rather than an intrinsic reduction in the sizes of droplets generated at rupture. The paper's own discussion of prior literature strengthens this concern: most previous studies report an increase in droplet size with polymer concentration, and the authors attribute the discrepancy to the population captured by PDA. That attribution is plausible, but it is also exactly the assumption that needs to be tested. The numerical simulations do not resolve this issue because they quantify bag thinning and uniformity (sigma-squared in Section IV.B) but do not compute a droplet size distribution; therefore the droplet-size reduction has no independent quantitative support outside the single-point PDA measurement.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript investigates how polymer-induced viscoelasticity modifies the shear-driven fragmentation of a thin liquid film in a model 'cough machine' (a 30 cm rectangular channel with a 1 mm film sheared by a pulsed 30 m/s airflow). Experiments with PEO solutions (Cm = 0.05-1.00% by mass) show that fragmentation still occurs through the bag-breakup mode, but the bags are longer, wider, and longer-lived than for a Newtonian liquid (Fig. 3). Phase-Doppler anemometry at 15 cm downstream of the channel exit reports a decrease of the mean droplet diameter from about 20 µm (pure water) to about 10 µm for Cm = 0.10-0.20%, a slight reversal at Cm = 0.50% (attributed to shear thinning), and no droplet detections at Cm = 1.00% (filament-dominated expulsion). The proposed mechanism is that viscoelasticity produces thinner and more uniformly thick bag sheets, so that rim retraction at rupture generates smaller droplets. Axisymmetric Oldroyd-B/volume-of-fluid simulations (Basilisk) of a droplet in an impulsively started airflow, run at Wi = 1 with a trial-and-error-selected elastoinertial number Π = 0.001, show more uniform bag thinning quantified by the mean-squared deviation σ̂² of the thickness profile and are presented as numerical support for this mechanism.","tokens_in":18969,"tokens_out":32597,"duration_ms":254324,"significance":"If the main claim holds, the paper provides a clean experimental demonstration in a regime (bag breakup of a sheared film in a confined geometry) where most prior work on viscoelastic atomization concerned higher-Weber-number modes and generally reported droplet-size increases with polymer concentration. The cough-machine geometry makes the result directly relevant to bioaerosol generation. Strengths: (i) the bag size and lifetime data (Fig. 3) come from direct high-speed imaging and are robust to the PDA sampling criticism; (ii) the fluid characterization (shear rheology, surface tension, capillary-thinning relaxation times; Fig. 1 and Table I) is careful and reproducible; (iii) the simulations use a well-established code (Basilisk) with the log-conformation formulation, and the σ̂² metric is a sensible quantifier of thinning uniformity; (iv) the authors are transparent about the PDA sampling bias, the trial-and-error choice of Π, and the shear-thinning complications at high Cm.","major_comments":[{"comment":"The central quantitative claim that viscoelasticity reduces the expelled droplet size rests on PDA distributions measured at a single point, 15 cm downstream of the channel exit along the centerline. As the authors acknowledge in Section II.A, this location preferentially samples droplets that remain airborne while larger droplets fall out ballistically; what is not demonstrated is that the sampled fraction is independent of polymer concentration. Section III.A reports rim-retraction droplets larger than 200 µm for all cases, and for Cm = 1.00% the expelled liquid is mostly filaments with no PDA detections (Section III.C), showing that the partition of expelled mass between the PDA-detectable and the sedimenting populations changes qualitatively with Cm. If, in the 0.05-0.20% range, increasing Cm also shifts more volume into the large-droplet or ligament fraction, the measured decrease of the mean diameter at 15 cm would reflect preferential removal of the large fraction rather than an intrinsic reduction of the droplet sizes generated at rupture. This is exactly the assumption invoked in Section III.C to reconcile the present trend with the opposite trend reported in earlier viscoelastic atomization studies; it should be tested rather than assumed. I request a quantitative assessment of the sampling bias, for instance high-speed imaging of the droplet field near the channel exit with size statistics for both populations, a mass-flux-weighted comparison, or PDA measurements at multiple downstream positions, for at least one Boger case and the Newtonian baseline.","section":"Section II.A and Section III.C; Fig. 4"},{"comment":"The numerical support for the proposed mechanism is weaker than presented because the simulation parameters are not consistent with the measured rheology. The manuscript states that Π = 0.001 was 'decided after some trial-and-error iterations to best showcase the viscoelastic response of the liquid'. More importantly, an estimate from the characterization data conflicts with that choice: using the polymer viscosity ηp = ηN − ηs (Table I) and the measured relaxation times, the Oldroyd-B modulus G = ηp/λ equals roughly 1.2-6.6 Pa over the concentration range, so Π = G/(ρlU²) ≈ 1-8 × 10⁻⁶ at U = 30 m/s, i.e., two to three orders of magnitude below the simulated value of 10⁻³. Since the paper itself notes that 'too low a value' of Π yields a Newtonian-like response, a simulation at the physically estimated Π would probably not reproduce the strongly viscoelastic thinning reported in Fig. 7b. The Wi comparison is less clear-cut (using the film timescale of Section III.B gives Wi ≈ 0.4, while the droplet radius implicit in We_d = 7.5 Γρ of Section IV.A gives Wi ≈ 10-20), but the Π mismatch is robust to the choice of length scale. The simulations should either adopt rheology-based dimensionless parameters or be explicitly presented as qualitative illustrations, with a sensitivity study over Π spanning the estimated range, and the statement of 'complete agreement with our experimental findings' should be correspondingly qualified.","section":"Section IV.B; Figs. 6 and 7"},{"comment":"The mechanism advanced in the conclusions, namely that 'thinner bags of comparatively more uniform thickness' produce smaller droplets, has no direct experimental measurement of bag thickness at rupture. The experiments connect polymer concentration to bag length, width, and lifetime (Fig. 3) and to PDA droplet size (Fig. 4), but the intermediate step, the sheet thickness, is inferred rather than measured. This inference is not forced by volume conservation, because the bags remain attached to the liquid film while inflating (Fig. 2a), so their volume is not fixed and larger Lb and Wb do not strictly imply a smaller thickness. The only direct evidence for a thinner and more uniform sheet comes from the simulations whose parameters are questioned in the preceding comment. The authors should measure or estimate the bag thickness for at least one viscoelastic and one Newtonian case (for example by transmitted-light intensity or a calibrated optical method), or explicitly state that the thickness reduction is an inference supported only qualitatively by the simulations and by the capillary-thinning filament data of Fig. 1d.","section":"Section III.B, Section III.C, and Conclusions"},{"comment":"The proposed causal chain (higher Cm → larger bags → thinner sheets → smaller droplets) is not monotonic at the upper end of the explored concentration range. At Cm = 1.00% the bags are the largest (Figs. 3a-3b), yet no droplets are detected by the PDA (Section III.C), and at Cm = 0.50% the mean droplet diameter increases relative to Cm = 0.20% (Fig. 4a) while the bags remain larger than in the dilute cases. The paper attributes both departures to shear thinning, but the shear-rate dependence in the atomization range is not measured (Fig. 1c) and no mechanistic account of the reversal is given. The central claim as stated in the abstract and the conclusions is therefore established only for the dilute Boger regime (Cm ≤ 0.20%); the manuscript should either explicitly limit its conclusions to that regime or provide a quantitative explanation of the shear-thinning crossover.","section":"Section III.B and Section III.C"}],"minor_comments":[{"comment":"The mean diameters in Fig. 4a are shown without uncertainty intervals or sample counts; please report the number of validated PDA samples per concentration and add percentile or bootstrap intervals as in Fig. 4b.","section":"Section III.C, Fig. 4a"},{"comment":"The printed formula for the bag lifetime scale, τf = Hf sqrt(Γρ/U), is dimensionally inconsistent; presumably the intended expression is τf = Hf sqrt(Γρ)/U (as in Ref. 48), and the formula should be corrected.","section":"Section III.B"},{"comment":"The color code in Fig. 7b is not decoded in the caption; please list the Wi values corresponding to each curve.","section":"Fig. 7b"},{"comment":"The attribution of the Cm = 0.50% reversal of the mean droplet diameter to shear thinning is plausible but untested, since the viscosity curves in Fig. 1c do not reach the shear-rate range of the atomization event and no shear-thinning inelastic control fluid is included; the text could note explicitly that the high-shear viscosity in the relevant range is unknown.","section":"Section III.C"},{"comment":"Because the polymeric solutions have γ = 62 mN/m while the pure-water baseline has γ = 72 mN/m, the film Weber number differs between baseline and polymeric cases (We_f ≈ 17 vs ≈ 15); a sentence acknowledging this small variation would clarify the comparison with the Newtonian data of Kant et al.","section":"Section II.B and Section V"},{"comment":"The normalization of σ̂² shown in Fig. 7b is not specified; please state whether it is normalized by the droplet radius squared or by another reference length.","section":"Section IV.A, Eq. (7)"},{"comment":"The solid line in Fig. 3d indicates Lb ~ Tb²; please state whether this is a fit to the data or the theoretical law of Ref. 48 with the constant taken from theory.","section":"Fig. 3d"},{"comment":"The panel labels of Fig. 2 are inconsistent between the caption (which lists both b-i/ii/iii and d-i/ii/iii) and the text (which refers to 'figure 2b-iii' for the arrow); please harmonize the panel labels and the arrow reference.","section":"Section III.A, Fig. 2"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the manuscript fits the journal's scope well and the experimental core (bag size, lifetime, rheology) is solid; the two main risks are the unquantified, potentially polymer-concentration-dependent PDA sampling bias and the physically inconsistent simulation parameters (Π ≈ 10⁻³ simulated vs ≈ 10⁻⁶ estimated from Table I). Neither appears unfixable: the first can be addressed with a supplementary near-exit measurement or an explicit mass-flux argument, and the second with a sensitivity study or a clear statement that the simulations are qualitative illustrations. I see no novelty-disclosure or citation-practice concerns; the authors cite and discuss the contrary atomization literature explicitly. If the PDA bias is quantified and the simulation claims are scaled back to match the parameter situation, I would support publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth engaging with. It reports a genuine new observation: in a cough-machine geometry, adding small amounts of PEO makes the atomizing bags inflate larger and produces smaller PDA-measured droplets at the exit, against the trend reported in most prior viscoelastic atomization studies.\n\nWhat is new: the shift to smaller droplets is not in the cited prior literature; previous work mostly saw larger droplets. The mechanistic argument — that viscoelasticity makes bags thinner and more uniform before rupture — is plausible and is supported by their bag-length/lifetime data and by the Oldroyd-B simulations showing lower sigma-squared at higher Wi. The paper is honest about its limitations, too.\n\nThe main soft spot is the droplet-size measurement. The PDA sits at a single point 15 cm downstream, and they acknowledge it under-samples larger droplets. If the polymer solutions partition more mass into large rim-retraction droplets that fall out before the probe, the reported reduction could be exaggerated. That concern is real and unquantified. The simulations do not fully resolve it because they do not compute droplet size distributions; they only show thinner, more uniform bags. Also, the elastoinertial number Pi is hand-tuned to reproduce the qualitative response, so the simulation support is partly circular — though the experimental bag-size trend is independent. The Cm=0.5% non-monotonicity is handled reasonably by invoking shear thinning, though it is not deeply investigated.\n\nWho this is for: anyone working on aerosol generation, respiratory droplet transport, or viscoelastic atomization. It is a credible, well-written contribution that opens a question rather than settles it. It deserves a serious referee. My recommendation: send it to review. In the revision, ask for direct sheet-thickness measurements or a quantitative droplet-size prediction from the simulations, and at least a sensitivity analysis of the PDA sampling bias.","headline":"A credible, well-executed study showing viscoelasticity can shift expelled droplet sizes down in a cough-machine geometry, but the droplet-size evidence rests on a single-point PDA measurement and the simulation support is partly tuned; worth reviewing, not yet conclusive.","tokens_in":19540,"tokens_out":2611,"would_cite":true,"duration_ms":25308,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["76A10","76T10","76D45"],"pacs":[],"model":"deepseek-v4-flash","headline":"Adding polymer elasticity to the airway fluid shifts cough droplets from around 20 µm to around 10 µm by making the pre-rupture bags thinner.","keywords":["viscoelasticity","respiratory droplets","bag breakup","cough machine","mucosalivary fluid","droplet size distribution","Oldroyd-B","sheet atomization"],"falsifier":"Measure the complete droplet-size distribution, including the large-droplet fraction, at the channel exit and at several downstream positions for water and for the polymer solutions; if the mean diameter including all droplets does not decrease with polymer concentration, the central claim fails. Alternatively, a three-dimensional simulation of viscoelastic bag rupture that resolves hole growth without mesh-limited breakup and yields equal or larger droplets at nonzero Weissenberg number would refute the thickness-controlled mechanism.","tokens_in":18521,"feed_emoji":"💧","tokens_out":8229,"duration_ms":84976,"temperature":0.7,"pith_summary":"This paper tries to establish that the viscoelasticity of the fluid lining the airways is a controlling factor in the sizes of droplets expelled during coughing and sneezing. In a channel that shears a thin liquid film with a cough-like blast of air at 30 m/s, polymer-laden mucosalivary mimics break apart through the same bag-inflation route as ordinary water, but their bags inflate longer and become thinner before bursting. Because the droplets are shed by the unstable retraction of the bag rims, the thinner bags yield smaller expelled droplets, with the mean diameter falling from roughly 20 µm for water to about 10 µm at polymer fractions of 0.1–0.2%. If this is right, the elastic, memory-bearing character of mucus, not just its viscosity, must be included in predictions of how far respiratory aerosol travels and in mitigation strategies aimed at the airway lining fluid.","feed_headline":"Stretchy airway mucus makes cough droplets smaller","feed_subtitle":"Thin bags burst into fine, long-floating droplets, so mucus elasticity may shape infection spread.","key_machinery":"The central object is the inflated liquid bag: a thin sheet of liquid stretched between a rim and the still-intact film by aerodynamic pressure, which eventually pops through nucleated holes whose rims retract and shed droplets. Bag size just before rupture sets sheet thickness, and sheet thickness sets the droplet size produced by rim retraction; the paper uses the bag-inflation scaling $L\\sim T^2$ to connect longer bag lifetime to larger bag size. Viscoelasticity enters through the Oldroyd-B constitutive model with polymer relaxation time $\\lambda$ and elastoinertial number $\\Pi$, and the uniformity of thinning is quantified by the mean-squared deviation $\\hat{\\sigma}^2$ of the local sheet thickness about its minimum: larger Weissenberg number lowers $\\hat{\\sigma}^2$, meaning bags thin more evenly and survive longer. This chain of objects—bag, hole, Taylor-Culick rim, thickness variance—carries the argument.","core_discovery":"The paper claims that viscoelastic liquids and Newtonian liquids fragment by the same bag-mediated route in this shear-driven setup, and that the only difference—larger, longer-lived bags in the viscoelastic case—is enough to shift the droplet-size distribution. The proposed mechanism is a chain: polymer relaxation stretches the bag in a more spatially uniform way, so the bag reaches a greater size before the first hole appears; a larger bag of the same liquid volume is a thinner sheet; and when holes open and the bounding rims retract in the Taylor-Culick fashion, thinner sheets emit smaller droplets. The experiments support each link by showing that bag length, width, and lifetime grow with polymer concentration, that the sheet-thickness profile becomes more uniform in Oldroyd-B simulations at higher Weissenberg number, and that the measured mean droplet diameter decreases with polymer concentration before rising again at concentrations where shear thinning appears. The paper presents this as the mechanism by which mucosalivary viscoelasticity controls the expelled aerosol size distribution.","pith_inferences":["The droplet statistics come from a single measurement point 15 cm downstream that preferentially captures small airborne droplets; the size reduction could be partly an artifact if viscoelasticity causes more large droplets to fall out before reaching that point, a bias the paper acknowledges but does not quantify.","A natural extension is to measure the full size distribution at multiple downstream stations or at the channel exit, which would separate true size reduction from differential settling of large droplets.","The same chain—elastic stretching, thinner bags, smaller rim-retraction droplets—should apply to other sheared viscoelastic films, for example sea-spray generation or industrial atomization of polymer solutions, where analogous large-bag structures are observed.","Real mucus contains mucin networks, salts, and surfactants, so dilute polyethylene-oxide solutions isolate only the elasticity; how mucin-specific rheology, including strain hardening and heterogeneity, shifts the droplet spectrum remains open."],"forward_implications":["Mucus viscoelasticity should be treated as a first-order input to respiratory aerosol size predictions, alongside surface tension and viscosity.","A modest elastic response can produce droplets near 10 µm at low Ohnesorge number, matching sizes that Newtonian fluids reach only at much higher viscosity.","The non-monotonic behavior at higher polymer concentration suggests that shear thinning or other high-concentration rheology can partially undo the elastic size reduction.","Mitigation strategies that modify airway-lining fluid properties would be expected to change the airborne small-droplet fraction rather than only the total amount of expelled liquid.","Bag-mediated atomization models for Newtonian sheets should be extended to include a relaxation time to capture elongated bag shapes and downstream droplet sizes."],"supporting_citations":[{"why":"Establishes the cough-machine geometry, the Newtonian bag-mediated atomization route, and the viscosity baseline against which the present viscoelastic results are compared.","marker":"[36]"},{"why":"Supplies the bag-inflation scaling $L\\sim T^2$ and the single-drop fragmentation framework connecting bag size to droplet size.","marker":"[48]"},{"why":"Provides the fragmentation-versus-cohesion picture of bag breakup, hole nucleation, and rim retraction used to interpret the experiments.","marker":"[50]"},{"why":"Fixes the Weber-number regime classification identifying bag breakup for $11\\lesssim We\\lesssim 18$, supporting the analogy to droplet aerobreakup.","marker":"[56]"},{"why":"Supplies the axisymmetric numerical setup and Newtonian parameters used as the baseline for the viscoelastic bag-thinning simulations.","marker":"[39]"},{"why":"Defines the class of constant-viscosity elastic liquids, which isolates the elastic effect in the dilute polymer solutions.","marker":"[88]"},{"why":"Provides the exponential filament-thinning measurement used to extract relaxation times and the elastocapillary framework for viscoelastic sheet thinning.","marker":"[78]"},{"why":"Shows how viscoelastic filaments retract more slowly, delaying pinch-off and supporting the longer, thinner bag morphology seen here.","marker":"[92]"},{"why":"Links droplet size to the local thickness of the liquid structures, the protoblob argument that converts thinner sheets into smaller droplets.","marker":"[77]"}],"fun_headline_variants":["Stretchy mucus makes cough droplets tinier","Viscoelastic mucus spawns finer cough spray","Bag-bursting mucus shrinks aerosol droplets","Mucus elasticity trims expelled droplet size"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the single-point droplet measurement 15 cm downstream samples Newtonian and viscoelastic fluids with the same bias, so that the recorded shift toward smaller mean diameters reflects a real change in atomization rather than a change in how many large droplets fall out of the puff before they are counted.","fun_headline_variants_meta":{"raw":{"variants":["Stretchy mucus makes cough droplets tinier","Viscoelastic mucus spawns finer cough spray","Bag-bursting mucus shrinks aerosol droplets","Mucus elasticity trims expelled droplet size"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000517,"raw_usage":{"total_tokens":2506,"prompt_tokens":943,"completion_tokens":1563,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":559,"completion_tokens_details":{"reasoning_tokens":1505}},"tokens_in":559,"tokens_out":1563,"duration_ms":14365,"temperature":1.0,"reasoning_tokens":1505,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T20:13:55.188963+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the complete droplet-size distribution, including the large-droplet fraction, at the channel exit and at several downstream positions for water and for the polymer solutions; if the mean diameter including all droplets does not decrease with polymer concentration, the central claim fails. Alternatively, a three-dimensional simulation of viscoelastic bag rupture that resolves hole growth without mesh-limited breakup and yields equal or larger droplets at nonzero Weissenberg number would refute the thickness-controlled mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Fixes the Weber-number regime classification identifying bag breakup for $11\\lesssim We\\lesssim 18$, supporting the analogy to droplet aerobreakup."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the class of constant-viscosity elastic liquids, which isolates the elastic effect in the dilute polymer solutions."},{"cited_title":"Elastocapillary Worthington jets","cited_arxiv_id":"2207.07928","evidence_quote":"Provides the exponential filament-thinning measurement used to extract relaxation times and the elastocapillary framework for viscoelastic sheet thinning."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows how viscoelastic filaments retract more slowly, delaying pinch-off and supporting the longer, thinner bag morphology seen here."},{"cited_title":"Keshavarz, E","cited_arxiv_id":null,"evidence_quote":"Links droplet size to the local thickness of the liquid structures, the protoblob argument that converts thinner sheets into smaller droplets."}],"review_version":1}