{"id":"134e063a-2442-4a4b-a1c2-bde2e9738ea3","arxiv_id":"2501.17001","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"At high Galilei numbers, near-wall bubbles either escape after one or two bounces via a Magnus-like lift generated during collision, or, if path-unstable and moderately deformed, become trapped in a periodic near-wall zigzag.","lead":"Simulations of bubbles rising next to a vertical wall reveal two high-inertia behaviors: small bubbles escape after one or two bounces, while moderately deformed bubbles can get trapped and zigzag along the wall. This helps explain why millimeter bubbles in water sometimes stick to walls instead of drifting away, a question relevant to bubble columns and flotation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The BTE escape threshold and the NWZ trapping boundary rest on collision dynamics computed at the minimum grid resolution, where the grid-refinement test shows a 19% shift in the final escape position.","rationale":"I read the paper as a numerical exploration of two new near-wall bubble regimes, BTE and NWZ, with mechanistic explanations supported by flow-field analysis and comparison with experiments. The qualitative mechanisms are physically plausible: the BTE mechanism is anchored in the observed flow reversal in the gap and the resulting surface rotation, and the NWZ mechanism is supported by the measured correlation between reduced rise speed, transient de-oblateness, and larger departing frontal area, plus the existence of NWZ without collisions. The authors are also transparent about the lubrication-film under-resolution and include a grid-refinement study. My stress-test focuses on the weakest load-bearing point, which is identical to the reader's weakest assumption: the quantitative thresholds (max|Omega_z| = 1.8 for BTE, and the Ga-Bo boundaries of BTE and NWZ) are extracted from simulations whose collision dynamics are not converged. Figure 17 shows a 19% change in the final escape position between the two finest gap resolutions, so the collision physics is demonstrably resolution-sensitive, and the spinning rate that controls escape is generated during those collisions. This does not invalidate the qualitative story, but it means the specific regime boundaries and the critical spinning-rate threshold should not be taken as quantitatively reliable without a finer-grid or subgrid-lubrication check. The reader's conditional verdict is therefore appropriate; my analysis does not change it. I recommend no change to the verdict: the paper should be accepted with the condition that the resolution sensitivity of the collision dynamics be addressed before the regime map is used quantitatively.","tokens_in":28124,"tokens_out":4017,"duration_ms":46630,"concrete_test":"Re-run two cases with an increased gap resolution, e.g., Delta_min = 1/256 (or 1/512) whenever delta < 0.15, keeping all other numerical parameters identical: (i) a BTE boundary case such as (Bo, Ga) = (0.05, 60) or (0.05, 70), and (ii) an NWZ case with collisions such as (Bo, Ga) = (0.25, 90). Record whether the bubble escapes or remains trapped, the maximum value of |Omega_z| during collisions, the final rest position X_b, and the maximum frontal-area asymmetry in the NWZ case. If the bouncing-to-BTE transition shifts by more than one Ga increment, or if max(|Omega_z|) changes by more than 10%, the claimed critical spinning rate and the BTE/NWZ boundaries in Figure 2(a) are not robust. If the (0.25, 90) case changes from NWZ to WMA, the trapping mechanism is resolution-sensitive.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claims are the BTE escape mechanism (escape when the surface spinning rate exceeds a critical value near 1.8) and the NWZ trapping mechanism (transient oblateness reduction makes the departing frontal area larger, so transverse drag and added mass oppose the repulsive wake-wall force). Both mechanisms are activated during very close bubble-wall approaches, precisely the regime where the authors state the flow resolution is insufficient. Section 2 reports that even Delta_min = 1/136 does not properly resolve the lubrication film when the dimensionless gap delta drops below Delta_min, and such events are labelled 'bubble-wall collisions'. Appendix A, Figure 17 shows that for the BTE case (Bo, Ga) = (0.05, 90), the final wall-normal rest position shifts from about 6.75R at Delta_min = 1/68 to about 5.5R at Delta_min = 1/136, i.e. a 19% change, and the departure velocity peak also changes. This demonstrates that the post-collision dynamics are not grid-converged at the production resolution. Since the spinning rate Omega_z that drives the Magnus-like repulsive force is generated during these under-resolved collisions, the claimed threshold max(|Omega_z|) = 1.8 used to separate periodic bouncing from BTE (Section 4, Figure 8) may be resolution-dependent, and the BTE region in the phase diagram (Figure 2a) could shift with further refinement. Similarly, the NWZ trapping mechanism relies on the transient rise-speed drop and the resulting oblateness reduction during near-wall approaches; if the gap flow is under-resolved, the computed deformation and hence the frontal-area asymmetry documented in Section 5.3 may be quantitatively unreliable. The paper partially mitigates this by observing NWZ in cases without collisions (e.g., Bo = 1, Ga = 70), which indicates that the trapping mechanism is not purely a grid artifact.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports three-dimensional numerical simulations of a single deformable bubble rising near a vertical wall in the range 30 < Ga <= 90 and 0.02 <= Bo <= 2, extending the authors' Part 1 to highly inertial regimes. It identifies three new types of motion: bouncing-tumbling-escaping (BTE) for low Bond numbers and high Galilei numbers, in which the bubble escapes after one or two wall collisions because strong interfacial rotation generates a Magnus-like repulsive force; near-wall zigzagging (NWZ) for intermediate Bond numbers at high Ga, in which a transient reduction of bubble oblateness during close wall approaches increases the frontal area during departure and thereby enhances transverse drag and added-mass resistance; and wavy migration away (WMA) for path-unstable bubbles that depart from the wall while zigzagging or spiralling. The regimes are organized in a phase diagram, and simulation results are compared against experiments in water and silicone oils, with good agreement on Strouhal numbers and path characteristics.","tokens_in":28436,"tokens_out":7075,"duration_ms":66356,"significance":"If the proposed mechanisms are correct, the paper resolves a striking contradiction between moderate-Reynolds-number observations, where path-unstable bubbles migrate away from the wall, and high-Reynolds-number experiments in water, where bubbles are trapped near the wall. The study is valuable for its systematic parameter sweep, its quantitative comparisons with independent experiments (de Vries 2001; Jeong & Park 2015; Estepa-Cantero et al. 2024), and its explicit discussion of numerical limitations. The critical spin rate and the Magnus lift coefficient are falsifiable phenomenological predictions, although they are extracted from the same simulations used to construct the regime map. The work is a solid basis for low-order models of near-wall bubble motion, provided the resolution sensitivity of the collision phase is addressed.","major_comments":[{"comment":"The production resolution is not demonstrated to be converged for the collision phase that drives both new mechanisms. The paper states that even Delta_min = 1/136 is insufficient to resolve the lubrication film when the dimensionless gap delta drops below Delta_min, and Figure 17 shows that for (Bo, Ga) = (0.05, 90) the final escape position changes from about 6.75R to 5.5R when Delta_min is refined from 1/68 to 1/136, a 19% shift. Since the interfacial spin Omega_z in the BTE scenario is generated during these under-resolved collisions, and the NWZ trapping depends on wake-wall interaction at minimal gap, the critical value max(|Omega_z|) = 1.8 and the phase boundaries in Figure 2(a) may be resolution-dependent. The authors should either demonstrate convergence at a finer resolution (e.g., Delta_min = 1/256) for representative BTE and NWZ cases, or quantify the uncertainty of the thresholds arising from this numerical limitation.","section":"Section 2 / Appendix A, Figure 17"},{"comment":"The threshold max(|Omega_z|) = 1.8 is obtained from simulations with different initial separations: X0 = 2 for most cases, but X0 = 3.5 for 40 <= Ga <= 70. Appendix B shows that X0 affects the number of bounces and the impact velocity in BTE cases. Because the maximum spin rate is expected to depend on the collision kinematics, the threshold should be shown to be robust to X0, or the dataset used to infer it should be made homogeneous in X0. Without this, the quoted critical value may reflect an artefact of the initialization protocol rather than a universal feature of the BTE transition.","section":"Section 4, Figure 8 and Appendix B"},{"comment":"The NWZ trapping mechanism is inferred from kinematic quantities (surface area Sigma and frontal area S_perp) rather than from a direct assessment of the transverse forces. The claim that increased transverse drag and added mass are what counteract the repulsive wake-wall force would be more convincing if the authors provided a force budget over one zigzag period, at least for one NWZ case, or a quantitative estimate of the two resistive contributions during the departing stage. As presented, the correlation between area asymmetry and trapping is suggestive but does not close the force balance.","section":"Section 5.3"}],"minor_comments":[{"comment":"The legend uses the same open-square symbol for four different regimes (periodic bouncing with collisions, periodic bouncing without collisions, damped bouncing, and migration away from the wall). Please use distinct symbols or clarify the entries.","section":"Figure 2(a) caption"},{"comment":"The estimate C_L^Omega = 0.5 +/- 0.07 should be reported with a wider uncertainty, since it depends on the assumed added-mass coefficient C_I = 1/2 and on neglecting the trailing-vortex contribution, both of which are acknowledged as uncertain in the same paragraph.","section":"Section 4, force-balance paragraph"},{"comment":"The symbol Delta_min is used both for the dimensional minimum cell size and for its dimensionless counterpart (e.g., 'Delta_min = 1/68' and 'Delta_min/R'). Please distinguish the two, for instance by writing Delta_min/R explicitly wherever dimensionless values are used.","section":"Section 2, notation"},{"comment":"The reference for Estepa-Cantero et al. (2024) is listed with the same title as Cano-Lozano et al. (2016); please verify the title and bibliographic data.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong candidate for publication in JFM if the grid-convergence concern is addressed: the central qualitative picture (BTE, NWZ, WMA) is supported by comparison with independent experiments, but the quantitative thresholds and the two proposed mechanisms rely on the under-resolved collision phase. I would ask the editor to require either a finer-resolution test for representative cases or an explicit uncertainty statement, and to check the internal consistency of the X0 values used in the Figure 8 dataset."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two genuinely new things here, and both look physically plausible. The BTE escape mechanism—collision-driven rotation around the bubble producing a Magnus-like lift—explains de Vries's observation that bubbles escape only after one or two bounces. The NWZ trapping mechanism, where a transient drop in oblateness raises the transverse drag and added mass during departure, explains the wall trapping in water experiments. Those are not just labels; the flow-field analysis supports them, and the regime map anchors them against independent experiments.  The paper also does the right kind of honest scientific work: it states without hedging that the lubrication film is not resolved below a dimensionless gap of about 1/136, it labels those events 'collisions,' and it shows a grid test that reveals real sensitivity. Appendix A shows the final BTE separation moving from 6.75R at Δmin=1/68 to about 5.5R at Δmin=1/136. The production runs actually use adaptive refinement to 1/136 when the gap is small, and the adaptive result matches the full 1/136 run, so the stress-test note's phrase 'computed at the minimum grid resolution' is misleading. The legitimate concern is that there is no test at 1/272, so we don't know whether 1/136 is converged; the spin threshold of about 1.8 and the Magnus coefficient around 0.5 come from these runs and could shift with further refinement.  The NWZ mechanism is less exposed to this worry because it occurs in at least one case without any collision, so it isn't purely a lubrication artifact. Its quantitative boundary, however, could move. The force-balance estimate of C_L^Ω is admittedly crude—added-mass coefficient assumed 1/2, Bernoulli interaction coefficient taken from a potential-flow formula, trailing vortices neglected—but the authors say that plainly and the conclusion only needs order-one accuracy.  Overall, the central qualitative claims hold up. The mechanisms are new, mechanistic, and consistent with the experiments they cite. The soft spot is quantitative: the collision-driven thresholds are not grid-converged. I'd send this to a serious referee and ask for a finer-grid convergence test of the collision phase and a statement of how the spin threshold would change. It's a solid contribution. I'd cite it and I'd bring it to the reading group.","headline":"Two genuinely new mechanisms for high-Re bubble-wall interactions, both worth taking seriously despite unresolved grid-convergence questions in the near-wall collision phase.","tokens_in":28996,"tokens_out":2979,"would_cite":true,"duration_ms":28558,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["76D05","76T10"],"pacs":[],"model":"deepseek-v4-flash","headline":"What decides if a rising bubble escapes a wall or sticks to it","keywords":["bubble dynamics","path instability","wake dynamics","Magnus-like force","near-wall zigzag","Galilei number","Bond number","numerical simulation"],"falsifier":"Measure the interface spinning rate around a bubble during a wall collision (for example, with particle image velocimetry in a liquid with Ga ≈ 70 and Bo ≈ 0.05): if the bubble escapes from the wall even though the maximum |Ω_z| stays below about 1.8, the Magnus-based escape criterion is wrong. Alternatively, a direct measurement of the bubble's frontal area during the approaching and departing halves of a near-wall zigzag that shows no increase during departure would contradict the proposed trapping mechanism.","tokens_in":27900,"feed_emoji":"🫧","tokens_out":6594,"duration_ms":55272,"temperature":0.7,"pith_summary":"This paper reports numerical simulations of deformable gas bubbles rising next to a vertical wall in highly inertial liquids. It shows that when the Galilei number (the buoyancy-to-viscous force ratio) is large enough, a bubble that bounces against the wall acquires a strong rotational flow at its surface, generating a Magnus-like force that pushes it away; the bubble then escapes after one or two bounces. In a separate regime, bubbles that would zigzag in an unbounded fluid can be trapped near the wall: when the gap gets very thin, a sudden drop in rise speed temporarily reduces the bubble's oblateness, enlarging its frontal area and making it harder to leave than to approach, so the bubble keeps executing a near-wall zigzag. The results reconcile earlier experiments in water where some bubbles escape and others remain trapped, and identify the mechanisms that decide between these fates.","feed_headline":"What decides if a rising bubble escapes a wall or sticks to it","feed_subtitle":"Simulations trace escape to a Magnus-like spin; trapping to a transient flattening that raises drag.","key_machinery":"The argument rests on two quantitative objects. (1) The interface spinning rate $\\boldsymbol{\\Omega}(t) = \\frac{3}{2V_s}\\int_{V_s} \\frac{\\mathbf{r} \\times \\mathbf{u}}{|\\mathbf{r}|^2}\\, dV_s$, which measures the average fluid rotation over the bubble surface; the paper shows escape sets in once the maximum of $|\\Omega_z|$ exceeds about 1.8 during a bounce, and uses a force balance to estimate the Magnus lift coefficient $C_L^\\Omega \\approx 0.5 \\pm 0.07$. (2) The frontal-area scaling $S_\\perp \\propto \\chi^{-1/3}$ of an oblate spheroid, which connects the transient reduction of the aspect ratio $\\chi$ during wall approach to an increase of transverse drag and virtual-mass forces during departure, the ingredient that traps the bubble in the NWZ regime.","core_discovery":"The central discovery is a pair of inertial mechanisms that determine the near-wall fate of bubbles at Reynolds numbers of order 100–1000. In the absence of path instability, once the Galilei number exceeds a critical value, a bubble that collides with the wall experiences an abrupt flow reversal in the gap, which generates a strong rotational motion around the interface (the interface spinning rate $\\Omega_z$ reaches values of order unity). This rotation produces a repulsive Magnus-like lift force $\\mathbf{F}_L^M \\propto \\boldsymbol{\\Omega} \\times \\mathbf{V}$ that acts for a long time after the bounce, allowing the bubble to escape the wall region after one or two bounces — the bouncing-tumbling-escaping (BTE) scenario. For bubbles beyond the path-instability threshold, the paper identifies a competing wall-ward trapping mechanism: when the bubble approaches the wall, its rise speed drops sharply, transiently reducing its oblateness; because the frontal area opposing transverse motion scales as $\\chi^{-1/3}$, the bubble presents a larger area when departing than when returning, so transverse drag and added-mass forces are larger during departure, counteracting the repulsive wake–wall interaction and confining the bubble to a periodic near-wall zigzag (NWZ). A third scenario, wavy migration away (WMA), occurs when these effects are too weak, and the bubble drifts away while zigzagging or spiralling.","pith_inferences":["One could test the spin criterion directly in experiments by using particle image velocimetry around a bubble during a wall bounce and correlating the measured surface vorticity with the escape outcome.","The trapping mechanism suggests that surfactants, which alter shape oscillations and surface mobility, could shift the Bond-number window for near-wall trapping; this is a natural extension to contaminated systems.","Since the BTE mechanism relies on the attractive Bernoulli force to build up impact velocity, changes in initial bubble-wall separation should affect escape only when the impact velocity is too low, as the paper's appendix confirms; one might therefore expect the critical Galilei number for escape to depend weakly on release distance.","The long memory of the initial separation in NWZ regimes at high Ga implies that short experimental columns may not reach the developed trapped state; results from such experiments should be interpreted with caution."],"forward_implications":["If the spin-based escape criterion is correct, the maximum interface spinning rate during a wall collision can be used as a predictor for whether a given bubble will escape or keep bouncing.","The estimated Magnus lift coefficient (about 0.5 ± 0.07) provides a quantitative input for low-order models of near-wall bubble motion.","The NWZ trapping mechanism implies that bubbles with larger zigzag amplitude are more likely to be trapped, which explains the existence of a finite Bond-number window for trapping.","The wall does not trigger path instability; it only selects the oscillation plane, so the threshold Bond number from unbounded-flow stability analysis applies in wall-bounded configurations.","Wall proximity can increase the Strouhal number of zigzagging by up to about 20% at Bo ≥ 0.25, so near-wall oscillations are faster than their unbounded counterparts."],"supporting_citations":[{"why":"Part 1 of this study; supplies the lower-Ga baseline regimes, numerical setup, and validation that this work extends.","marker":"Shi, Zhang & Magnaudet (2024)"},{"why":"Provides the neutral curve of path instability in an unbounded fluid, used to classify bubbles as path-stable or path-unstable.","marker":"Bonnefis et al. (2024)"},{"why":"Gives the transverse-force model and interaction coefficient C_P used to estimate the Bernoulli attractive force in the BTE force balance.","marker":"Takemura & Magnaudet (2003)"},{"why":"Experimental observation of bouncing and escape of bubbles in water; supplies the (Bo, Ga) point compared with the BTE scenario.","marker":"de Vries (2001)"},{"why":"Experimental demonstration of near-wall trapping at high Reynolds number in water, which the NWZ regime reproduces and explains.","marker":"Jeong & Park (2015)"},{"why":"Experimental WMA trajectories used to validate the present simulations in the path-unstable, migration-away regime.","marker":"Estepa-Cantero et al. (2024)"},{"why":"High-Reynolds-number drag prediction for spherical bubbles used in the force balance that yields the Magnus lift coefficient.","marker":"Moore (1963)"},{"why":"Characterizes the '4R' vortex shedding mode used to interpret the wake structure of zigzagging bubbles and its interaction with the wall.","marker":"Horowitz & Williamson (2010)"}],"fun_headline_variants":["Bubble escape from walls hinges on spin versus flattening","Magnus force frees bubbles, drag traps them near walls","Rising bubbles bounce off or stick due to inertial tricks","Simulations: why bubbles either escape or hug the wall"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the under-resolved lubrication film at gaps smaller than about 1/136 of the bubble radius does not change the qualitative dynamics; the paper itself notes that spin generation and rebound details during 'collisions' remain sensitive to grid resolution.","fun_headline_variants_meta":{"raw":{"variants":["Bubble escape from walls hinges on spin versus flattening","Magnus force frees bubbles, drag traps them near walls","Rising bubbles bounce off or stick due to inertial tricks","Simulations: why bubbles either escape or hug the wall"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000256,"raw_usage":{"total_tokens":1663,"prompt_tokens":1120,"completion_tokens":543,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":736,"completion_tokens_details":{"reasoning_tokens":475}},"tokens_in":736,"tokens_out":543,"duration_ms":6022,"temperature":1.0,"reasoning_tokens":475,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T05:13:44.556391+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the interface spinning rate around a bubble during a wall collision (for example, with particle image velocimetry in a liquid with Ga ≈ 70 and Bo ≈ 0.05): if the bubble escapes from the wall even though the maximum |Ω_z| stays below about 1.8, the Magnus-based escape criterion is wrong. Alternatively, a direct measurement of the bubble's frontal area during the approaching and departing halves of a near-wall zigzag that shows no increase during departure would contradict the proposed trapping mechanism.","supporting_citations":[{"cited_title":", Sierra-Ausin, J","cited_arxiv_id":null,"evidence_quote":"Provides the neutral curve of path instability in an unbounded fluid, used to classify bubbles as path-stable or path-unstable."},{"cited_title":"& Magnaudet, J","cited_arxiv_id":null,"evidence_quote":"Gives the transverse-force model and interaction coefficient C_P used to estimate the Bernoulli attractive force in the BTE force balance."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Experimental observation of bouncing and escape of bubbles in water; supplies the (Bo, Ga) point compared with the BTE scenario."},{"cited_title":"& Park, H","cited_arxiv_id":null,"evidence_quote":"Experimental demonstration of near-wall trapping at high Reynolds number in water, which the NWZ regime reproduces and explains."},{"cited_title":", Mart\\'inez-Baz\\'an, C","cited_arxiv_id":null,"evidence_quote":"Experimental WMA trajectories used to validate the present simulations in the path-unstable, migration-away regime."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"High-Reynolds-number drag prediction for spherical bubbles used in the force balance that yields the Magnus lift coefficient."},{"cited_title":"& Williamson, C","cited_arxiv_id":null,"evidence_quote":"Characterizes the '4R' vortex shedding mode used to interpret the wake structure of zigzagging bubbles and its interaction with the wall."}],"review_version":1}