{"id":"655c6db2-1edc-434f-bb04-f2ae49d6ee2c","arxiv_id":"2506.13942","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Sessile water droplets on superhydrophobic surfaces can fail to coalesce when a neighboring merged pair's interface pushes against them, and in tight clusters this launches the droplets sideways.","lead":"Two droplets sitting on a water-repellent surface can bounce off each other instead of merging when a nearby pair of droplets merges first, and the bounce can fling them sideways. The effect appears to work for droplets down to a few hundred microns and points to a passive way to shed condensation droplets.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The kink-entrapment mechanism is not self-consistently derived: the VOF model in SI S3 omits disjoining pressure and van der Waals forces, so the claim that kink formation 'leads to' non-coalescence rests on an unvalidated assumption.","rationale":"The central empirical finding—non-coalescence and lateral propulsion in sessile droplet clusters—rests on high-speed imaging and is plausible; I see no grounds to reject the observation itself. The soft spot is the mechanistic explanation, not the phenomenon. The reader's weakest_assumption identifies exactly this issue: the VOF model omits molecular-scale film rupture. I agree with that assessment. The paper's own SI admits that molecular forces cannot be accounted for, yet the main text asserts that kink formation prevents coalescence. Because the simulation uses a single VOF marker for all droplets and no disjoining pressure, the non-coalescence outcome is effectively guaranteed by resolving the air film rather than predicted from physical rupture criteria. The missing test is to include a rupture model or a direct film-thickness measurement and compare the predicted l/d0-dependent boundary with Fig. 2. I also note a secondary textual inconsistency: the abstract claims droplets 'as small as 100 microns,' the conclusion says 200 µm, and the only sub-millimeter data shown are 300–400 µm condensation droplets. This overclaim should be corrected, but it does not change the conditional verdict, which remains appropriate.","tokens_in":14232,"tokens_out":6031,"duration_ms":68561,"concrete_test":"Re-run the SI S3 axisymmetric VOF simulation with a disjoining-pressure model added to Eq. (3.2), for example an augmented Young-Laplace term with Π(h) = -A_H/(6πh^3) for the water-air-water film (Hamaker constant A_H ≈ 10^-20 J) plus a short-range Born repulsion, or a coupled Reynolds thin-film drainage model with a critical rupture thickness near 10 nm. Apply this to the in-line geometry at l/d0 = 0.05, 0.10, and 0.14. If the model predicts film rupture for cases where experiments show bouncing, the kink-entrapment mechanism is insufficient; if the film remains above ~10 nm throughout apparent contact for the bouncing cases and thins to rupture only for the coalescing cases, the mechanism is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"SI Section S3 explicitly states that 'coalescence or non-coalescence of droplets after apparent contact depends on molecular forces, which cannot be directly accounted for in the CFD simulations,' yet the main-text conclusion is that kink formation 'leads to momentary entrapment of the air layer, leading to non-coalescence with the third droplet.' The VOF momentum equation (Eq. 3.2) contains only pressure, viscous, surface-tension (Eq. 3.6), and gravity terms; there is no disjoining pressure, van der Waals term, or rupture criterion. With d0 ≈ 100 µm and the finest mesh at 1800 cells per radius, the cell size is about 28 nm, comparable to the 10–100 nm film thicknesses at which van der Waals forces decide coalescence. The simulation therefore cannot determine whether the air film would drain to rupture; it can only show that, in a computation that never allows subgrid rupture, a kink correlates with slow drainage. Furthermore, a single VOF marker is used for all droplets, so numerical coalescence occurs only when the air film is no longer resolved; the outcome is set by grid resolution rather than by molecular physics. Only a non-coalescence event is validated (Fig. S3b), so the model does not test the actual coalescence/non-coalescence boundary. The empirical bouncing is credible, but the causal claim that the kink prevents rupture is not established by this simulation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports an experimental and numerical study of non-coalescence in sessile droplet clusters on nanotextured superhydrophobic surfaces. The authors show that when two droplets in a cluster coalesce, the evolving interface can come into apparent contact with a neighboring droplet and bounce back without merging; in staggered configurations this bounce can transfer lateral momentum and produce in-plane self-propulsion, with an energy conversion efficiency reported as high as 9%. A parameter study maps the third droplet's position (R, θ) and classifies outcomes into full coalescence, coalescence followed by reseparation, and non-coalescence. Axisymmetric Volume-of-Fluid simulations are used to argue that a kink in the interstitial air film suppresses air drainage and entraps the film, thereby preventing coalescence. The phenomenon is also reported for condensed droplets of roughly 0.3–0.4 mm diameter on a cooled superhydrophobic surface, and the authors claim applicability down to 100 µm (abstract) or 200 µm (conclusions). The paper is well organized, with careful substrate fabrication and control of wettability, high-speed imaging, and a systematic geometric survey.","tokens_in":14446,"tokens_out":12134,"duration_ms":116071,"significance":"If substantiated, the central observation is significant: it identifies a previously unreported configuration for droplet non-coalescence, requiring neither relative centroidal velocity, surfactants, high gas pressure, nor millimetric droplet sizes, and it demonstrates a passive pathway for droplet transport on superhydrophobic surfaces during condensation. The manuscript's strengths include the systematic experimental mapping over cluster geometry, high-speed imaging, reproducibility controls (vibration isolation, substrate grounding), dimensionless scaling of the results, and a falsifiable claim that the coalescence/non-coalescence thresholds in the in-line arrangement are independent of droplet size for d0 in 0.6–2 mm. The proposed mechanism, however, rests on a continuum CFD model that, by the authors' own statement in SI Section S3, cannot account for the molecular forces that decide film rupture; the causal claim therefore needs additional support. The claims about droplet size range and about efficiency also need reconciliation with the data shown.","major_comments":[{"comment":"The central causal claim—that 'kink formation leads to momentary entrapment of the air layer, leading to non-coalescence with the third droplet'—is not established by the simulations. SI Section S3 states that coalescence or non-coalescence after apparent contact depends on molecular forces 'which cannot be directly accounted for in the CFD simulations,' and Eqs. (3.1)–(3.6) indeed contain no disjoining pressure, van der Waals interaction, or rupture criterion. Because a single VOF marker is used for all droplets, numerical coalescence can occur only after the air film becomes unresolved; the finest mesh (1800 cells per radius for d0 ≈ 100 µm) corresponds to a cell size of about 28 nm, and the resolved films in Figs. 3–4 have thickness on the order of 0.01 d0, i.e., about 1 µm, well above the 10–100 nm films in which van der Waals forces decide rupture. The model is also validated only against a non-coalescence event (Fig. S3b), so it does not test the coalescence/non-coalescence boundary. The simulation therefore demonstrates a correlation between kink formation and slow drainage over the resolved part of the film's life, not a mechanism that prevents rupture. A revision should either couple the VOF solution to a thin-film drainage model with a disjoining-pressure rupture criterion, or supply an order-of-magnitude comparison between the film-drainage time and the contact-residence time; unless that is done, the causal wording in the main text and Conclusions overstates the evidence.","section":"Main text (Fig. 4 discussion) and SI Section S3"},{"comment":"The droplet-size claim is internally inconsistent and goes beyond the data shown. The abstract claims non-coalescence 'from millimeters to as small as 100 microns,' the Conclusions state 'as small as 200 µm,' and the smallest event actually presented—the condensation event in Fig. 7a—involves droplets of about 0.3–0.4 mm. The cluster experiments use d0 = 0.6–2 mm, and the numerical simulations use d0 ≈ 100 µm, so an experimental observation at 100–200 µm is not demonstrated anywhere in the manuscript. The size-independence of the in-line thresholds has been tested only over 0.6–2 mm in non-dimensional coordinates; extrapolating that trend to 100 µm is not equivalent to observing an event at that scale. The authors should either add experimental non-coalescence data at or below 200 µm or revise the abstract and conclusions to give the size range that the data actually support.","section":"Abstract; Conclusions; Fig. 7a"},{"comment":"The quantitative efficiency claims require more support. The lateral efficiency η_lateral is defined with a denominator that accounts only for the surface energy released by D1+D2, while the numerator includes the kinetic energy of all participating droplets, including the propelled droplet D3; excluding D3's surface energy from the denominator is an accounting choice that directly affects the headline value of 'as high as 9%,' and the paper should justify or discuss it. No error bars, standard deviations, or event counts are reported for η_lateral in Figs. 5e–f, and the sample sizes N in the box plots of Figs. 5a–b are not stated; consequently the regime boundaries (l/d0 ≲ 0.07 and ≳ 0.14 for the in-line arrangement), the trend with packing density, and the claim of size independence rest on unquantified sampling. The statement that the four-droplet configuration is 'higher by ~2 percentage points' appears to be based on a single event in Fig. 6. At minimum, the authors should give event counts per R–θ bin, report the spread of efficiency values across repeats, and indicate how many events define each regime boundary.","section":"Main text (efficiency definition) and Fig. 5e–f"}],"minor_comments":[{"comment":"The caption of Fig. 2 states that panel (a) triggers coalescence with D3 and panel (b) retracts without coalescence, while the main text says 'Figure 2(a) illustrates a typical non-coalescence event' and 'the coalescence proceeds to completion, as shown in Fig 2(b)'; the labels and the text should be made consistent.","section":"Fig. 2 caption vs. main text (p. 4)"},{"comment":"The reference list contains duplicates and one corrupted entry: Ref. 31 is a raw downloaded filename that duplicates Ref. 12; Refs. 3 and 17 (Jayaratne and Mason), Refs. 5 and 18 (Orme), and Refs. 15 and 20 (Tang et al.) are also duplicated. The list should be cleaned.","section":"References"},{"comment":"The phrase 'no bulk velocity prior to coming in apparent contact' should be clarified: the interface of the merged droplet approaches D3 with inertial-capillary velocity, and the distinguishing feature is the absence of centroidal droplet motion rather than the absence of approach velocity.","section":"Main text (p. 3)"},{"comment":"The axes, units, and sample sizes should be documented: the text should define the units of S in Fig. 5d and give the values of N for each configuration in Figs. 5a–b, ideally in the caption or methods.","section":"Fig. 5"},{"comment":"The axisymmetry assumption is justified by stating that the interface reaches D3 before the ripple reflects from the substrate; this should be quantified by comparing the relevant travel times, since the symmetry of the computational domain rests on this assumption.","section":"SI Section S3"},{"comment":"There are typographical inconsistencies, including 'reseperation' for 'reseparation' (S2 heading and main text) and inconsistent spelling of 'water-repellent/repellant'.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The empirical core of the paper is novel and credible, but the manuscript currently overclaims in three places: the causal attribution of non-coalescence to the kink (where the authors' own SI admits molecular forces are unmodeled), the lower size bound of the claimed range (100/200 µm versus the smallest shown event at ~0.3–0.4 mm), and the precision of the efficiency values (no error bars or event counts). All three are addressable in a revision. The duplicate reference entries also need cleaning before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this paper reports a genuinely new empirical observation. On superhydrophobic surfaces, when two sessile droplets in a cluster coalesce, the advancing interface can bounce off a neighboring droplet instead of merging with it. This happens without relative collision velocity, surfactants, charge, Leidenfrost effects, or millimeter-sized drops. The high-speed imaging is convincing, and the condensation footage makes the phenomenon plausibly relevant to dropwise condensation.\n\nThe systematic parameter sweeps over droplet position (R and theta) are a real strength, and the lateral efficiency measurement—up to about 9%, comparable to coalescence-induced jumping—is a useful quantitative handle. The paper clearly identifies a new regime: stationary sessile clusters, where the merging pair's interface acts like a soft projectile that deforms the neighbor and sometimes launches it in-plane.\n\nWhere the soft spots are, in proportion: First, the size range is overstated. The abstract claims droplets as small as 100 microns, the conclusion says 200 microns, but the smallest condensation event actually shown is 300–400 microns. The minimum diameter is not demonstrated by any experiment in the paper. That is a straightforward fix in the text but it matters because the title-level claim rests on it. Second, the mechanistic claim goes beyond what the simulation can support. The VOF model in SI S3 explicitly states that molecular forces cannot be directly accounted for, and it includes no disjoining pressure or van der Waals term. With a finest mesh cell of about 28 nm, the simulation operates exactly in the film-thickness range where molecular forces decide rupture. So the kink and stagnation zone the simulation reveals are plausible correlates of slow drainage, but the statement that kink formation 'leads to' non-coalescence is not established. The authors should either couple a thin-film rupture criterion or soften the causal language to something like 'consistent with a mechanism where the kink slows drainage.' Third, the lateral efficiencies lack error bars and the box plots show limited sample sizes; treat the 9% as a rough upper bound rather than a robust value.\n\nWho this is for: droplet dynamics, wetting, and condensation heat transfer researchers. The empirical bouncing effect is real enough to deserve a serious referee. I would send it to peer review, but with a request for major revision on the size claim, uncertainty quantification, and the mechanism framing.","headline":"Genuinely new empirical observation of non-coalescence in sessile droplet clusters, but the size claims and the kink-based mechanism need to be reined in before this is publishable.","tokens_in":15043,"tokens_out":2104,"would_cite":true,"duration_ms":23274,"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 establishes that non-coalescence is the rule, not the exception, for sessile droplet clusters on water-repellent surfaces: when two droplets coalesce, the merged interface bounces off neighbouring droplets rather than merging…","keywords":["non-coalescence","sessile droplets","superhydrophobic surfaces","air film drainage","droplet self-propulsion","dropwise condensation","droplet clusters","kink entrapment"],"falsifier":"Measure the interstitial air film thickness in real time during apparent contact between a coalescing pair and a neighbouring sessile droplet, for example by high-speed interferometry; if the film thins below the roughly 10 nm rupture threshold before the interface retracts in any event the paper classifies as non-coalescence, the kink-entrapment mechanism is falsified. Alternatively, repeat the cluster experiments in a low-pressure chamber: if non-coalescence persists when the interstitial gas density is reduced by an order of magnitude, the air film cannot be the decisive element.","tokens_in":13948,"feed_emoji":"💧","tokens_out":6299,"duration_ms":60592,"temperature":0.7,"pith_summary":"This paper reports that sessile water droplets sitting in clusters on water-repellent surfaces can come into apparent contact with each other without merging: when two droplets in a cluster coalesce, the expanding interface of the merged pair reaches a neighbouring droplet and bounces off it instead of triggering further coalescence. The authors show that this non-coalescence happens over a wide size range, from millimetric droplets down to roughly 100 microns, with no surfactants, external fields, or pre-existing droplet velocity. They trace the bounce to a thin entrapped air film whose drainage is blocked by a kink in the film profile, and show that the rejected interface can transfer enough lateral momentum to self-propel the neighbouring droplet along the surface. The finding matters because it offers a passive pathway for droplet removal and surface renewal during dropwise condensation on superhydrophobic surfaces.","feed_headline":"Coalescing droplets bounce off neighbours and launch them sideways","feed_subtitle":"The bounce turns capillary energy into sideways motion, shedding condensate droplets as small as 100 microns.","key_machinery":"The mechanism that carries the argument is kink-induced entrapment of the interstitial air film. As the interface of two coalescing droplets advances toward a neighbour, it squeezes the air between them, creating a nearly flat film region; at $t/\\tau \\approx 0.99$ the two interfaces form a kink where the film thickness is locally minimum. This kink suppresses air drainage—radial drainage velocity drops and overpressure rises in the stagnation zone behind the kink—so the film does not thin to the tens-of-nanometres scale at which van der Waals forces would trigger coalescence. During retraction the film drains again and the kink disappears. The paper arrives at this picture through axisymmetric Volume-of-Fluid simulations that resolve the air film with adaptive meshing, validated against high-speed experiments.","core_discovery":"The central claim, stated on the paper's own terms, is that non-coalescence is a general property of coalescing sessile droplet clusters on water-repellent surfaces, not a special case requiring large droplets, high collision velocities, or surfactants. When two droplets merge, the capillary ripple travelling along the merged interface brings it into apparent contact with a third droplet; whether contact leads to coalescence is decided by the initial geometric arrangement of the cluster. In most configurations the evolving interface forms a flat air film with a kink—a local minimum in film thickness—that acts as a restriction against air drainage, momentarily trapping the interstitial air and preventing film rupture. When the merged interface retracts, the neighbouring droplet, which had been deformed during contact, recoils and can be launched along the surface with a lateral efficiency up to 9% for tightly packed three-droplet clusters, increasing further when more droplets participate. The same bouncing and self-propulsion is observed in micrometric condensate droplets, indicating a new pathway for spontaneous droplet removal.","pith_inferences":["The kink-entrapment picture predicts that the non-coalescence window should shift if the interstitial gas is replaced by one of different viscosity or pressure; this is a direct, testable consequence not reported in the paper.","If confirmed at even smaller scales, cluster non-coalescence could be engineered as a surface-cleaning mechanism for condensation heat transfer, where droplet departure size is a known performance bottleneck.","The dependence of lateral efficiency on cluster packing suggests that surface textures that stabilise dense droplet clusters could convert coalescence energy into directed motion more effectively than random droplet arrangements."],"forward_implications":["Non-coalescence is the dominant outcome for most in-line cluster configurations, with coalescence confined to narrow spacing windows.","The recoil from a bounced interface can self-propel non-coalescing droplets across superhydrophobic surfaces, with lateral efficiency up to 9% for close-packed three-droplet clusters.","Lateral efficiency increases with packing density and with the number of participating droplets beyond three.","The same non-coalescence and in-plane propulsion occurs in condensate droplet clusters on nanotextured superhydrophobic surfaces, extending the phenomenon to micrometric droplets."],"supporting_citations":[{"why":"Establishes the gas-microfilm framework for when drops bounce, underpinning the interstitial-film picture used here.","marker":"[2]"},{"why":"Quantifies how squeeze forces in the air film during bouncing transfer momentum, the baseline for lateral momentum generation.","marker":"[4]"},{"why":"Shows rebounding droplet-droplet collisions on superhydrophobic surfaces, the closest prior evidence for non-coalescence in sessile-like geometries.","marker":"[12]"},{"why":"Supplies the inertial-capillary scaling and jumping-droplet measurements used to nondimensionalize the present experiments and simulations.","marker":"[21]"},{"why":"Provides the nanoscale jump-to-contact mechanism by which van der Waals forces trigger coalescence, the rupture threshold the air film must avoid.","marker":"[26]"},{"why":"Documents transitions between bouncing and coalescence in binary droplet collisions, informing the drainage-transition interpretation.","marker":"[32]"},{"why":"Describes thin air film entrapment in drop-pool impacts, the closest analogue of the flat-film/kink morphology observed here.","marker":"[33]"},{"why":"Benchmarks coalescence-induced jumping efficiency, against which the reported lateral efficiencies are compared.","marker":"[35]"}],"fun_headline_variants":["Coalescence-induced bounce launches droplets sideways","Droplet clusters bounce and self-propel without extra help","Capillary ripple makes droplets bounce and move laterally","Micron droplets bounce off neighbours and self-propel","Sessile droplets bounce and self-propel in clusters"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a continuum Volume-of-Fluid simulation without disjoining pressure or other molecular-scale forces, but with a finely resolved air film, correctly predicts whether the film ruptures; if molecular-scale rupture would occur before the retraction that the simulation shows, the kink-entrapment mechanism would not explain the observed bouncing.","fun_headline_variants_meta":{"raw":{"variants":["Coalescence-induced bounce launches droplets sideways","Droplet clusters bounce and self-propel without extra help","Capillary ripple makes droplets bounce and move laterally","Micron droplets bounce off neighbours and self-propel","Sessile droplets bounce and self-propel in clusters"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000563,"raw_usage":{"total_tokens":2716,"prompt_tokens":1031,"completion_tokens":1685,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":647,"completion_tokens_details":{"reasoning_tokens":1609}},"tokens_in":647,"tokens_out":1685,"duration_ms":24910,"temperature":1.0,"reasoning_tokens":1609,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T00:26:05.807796+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the interstitial air film thickness in real time during apparent contact between a coalescing pair and a neighbouring sessile droplet, for example by high-speed interferometry; if the film thins below the roughly 10 nm rupture threshold before the interface retracts in any event the paper classifies as non-coalescence, the kink-entrapment mechanism is falsified. Alternatively, repeat the cluster experiments in a low-pressure chamber: if non-coalescence persists when the interstitial gas density is reduced by an order of magnitude, the air film cannot be the decisive element.","supporting_citations":[{"cited_title":"& Van Den Ende, D","cited_arxiv_id":null,"evidence_quote":"Quantifies how squeeze forces in the air film during bouncing transfer momentum, the baseline for lateral momentum generation."},{"cited_title":"& Ras, R","cited_arxiv_id":null,"evidence_quote":"Shows rebounding droplet-droplet collisions on superhydrophobic surfaces, the closest prior evidence for non-coalescence in sessile-like geometries."},{"cited_title":"V., Takahashi, H., Clanet, C., Shimoyama, I","cited_arxiv_id":null,"evidence_quote":"Supplies the inertial-capillary scaling and jumping-droplet measurements used to nondimensionalize the present experiments and simulations."},{"cited_title":"& Tordjeman, P","cited_arxiv_id":null,"evidence_quote":"Provides the nanoscale jump-to-contact mechanism by which van der Waals forces trigger coalescence, the rupture threshold the air film must avoid."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents transitions between bouncing and coalescence in binary droplet collisions, informing the drainage-transition interpretation."},{"cited_title":"& Mani, A","cited_arxiv_id":null,"evidence_quote":"Describes thin air film entrapment in drop-pool impacts, the closest analogue of the flat-film/kink morphology observed here."}],"review_version":1}