{"id":"97e45a96-1e0d-42b0-9a0c-cd84d88f3a6a","arxiv_id":"2411.13115","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"In turbulent Taylor-Couette flow with neutrally buoyant, equally viscous droplets, interfacial tension on the stretched fore head of droplets slows the surrounding flow and is the main cause of drag enhancement.","lead":"This paper uses computer simulations to show that in turbulent Taylor-Couette flow, the surface tension of small droplets is what increases drag as more droplets are added. The finding points to a specific mechanism: stretched droplets act like tiny brakes near the rotating cylinder, and models of two-phase flows should include this effect.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Numerical coalescence and the missing high-phi resolution test leave the deformation-based mechanism unverified at the volume fractions where it is most needed.","rationale":"The reader's weakest assumption identifies the same load-bearing point: the simulations are calibrated with phi_mvf=5%, numerical coalescence is admitted, and no high-phi resolution test exists. The central claim is not a bare budget identity; it is a deformation-based mechanism, so uncertainty in droplet size and deformation propagates directly into the causal story. I see no additional fatal flaw: the angular-velocity-flux decomposition is exact, the drag enhancement is reproduced, and the authors are transparent about the limitations. A targeted coalescence-controlled and resolution-checked run at phi=40% would settle whether the mechanism survives at the highest volume fraction. Until then, CONDITIONAL is the appropriate verdict, and this concern does not require changing the reader's assessment.","tokens_in":20428,"tokens_out":9444,"duration_ms":112090,"concrete_test":"Re-run the phi=40% case with a smaller marker-group fraction (phi_mvf=1%, i.e., 40 markers) or with a short-range repulsive anti-coalescence force, and also on the finer grid from Appendix A; then recompute (i) the torque/drag, (ii) the radial J_sigma profile in Fig. 7(a), and (iii) the fore-head/rear-end azimuthal interfacial-tension PDFs in Fig. 5(b). If these quantities are unchanged and the droplet-size PDF moves toward the experimental log-normal, the mechanism is robust. If the drag enhancement or the sign of the fore-head force changes materially, the central claim must be restricted to lower volume fractions.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The mechanism in §3 is explicitly deformation-dependent: stretched droplets near the inner cylinder produce a fore-head interfacial tension opposing the flow, which slows the continuous phase and raises wall shear stress. This causal chain is only trustworthy if the simulated droplet population is faithful in size, deformation, and coalescence state. The paper's own Appendix B states that the modified multi-marker VOF method 'does not completely resolve the issue of numerical coalescence' and that at phi=40% the simulated droplet-size PDF deviates from experiment. Appendix A reports no resolution test at phi=40% because of computational cost. Because each marker group is fixed at 5% volume fraction and coalescence within a marker is not prevented, droplet sizes and near-wall droplet statistics at 30-40% are method-dependent. The standard-VOF comparison in Fig. 11 shows quantitatively different droplet/continuous velocity PDFs at 40%, confirming that the method choice matters. If numerical coalescence artificially enlarges or deforms droplets, the fore-head/rear-end asymmetry in Fig. 5(b) and the resulting J_sigma budget could reflect the numerical droplet population rather than the physical mechanism.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies turbulent Taylor–Couette flow laden with neutrally buoyant, equally viscous droplets at volume fractions up to 40%, using a modified multi-marker volume-of-fluid method with the dispersed phase split into marker groups of 5% volume fraction each. The main claim is that the observed drag enhancement originates from interfacial tension: an exact decomposition of the angular velocity flux shows that the interfacial-tension contribution J_sigma grows with phi while the turbulent and viscous contributions do not, and near-wall diagnostics indicate that stretched droplets near the inner cylinder exert a fore-head interfacial-tension force opposing the flow, which slows the continuous phase, disrupts high-speed streaks, and raises wall viscous stress. The paper reports agreement with previous experiments on global torque and presents supporting statistics: radial profiles of J_sigma, joint PDFs of viscous stress and angular velocity, phase-resolved angular velocity PDFs, and per-droplet fore-head/rear-end interfacial tension PDFs.","tokens_in":20630,"tokens_out":3426,"duration_ms":37827,"significance":"If the proposed mechanism is correct, it offers a concrete physical picture for droplet-induced drag enhancement in turbulent emulsions, connecting interfacial tension in the viscous sublayer to streak modification and wall stress. The work combines an exact flux budget with multiple independent diagnostics and includes open-source code, a comparison against experiments, and a spurious-current assessment, which are strengths. The significance is moderate: it is a mechanism-focused numerical study at a single Reynolds and Weber number, and its broader reach depends on the fidelity of the simulated droplet population at high volume fractions.","major_comments":[{"comment":"The manuscript states that the modified multi-marker VOF method 'does not completely resolve the issue of numerical coalescence' and that at phi=40% the simulated droplet-size PDF deviates from experimental results. Because the causal mechanism in §3 is explicitly deformation- and size-dependent (fore-head versus rear-end interfacial tension on stretched droplets), the mechanism is least verified exactly in the regime where the drag enhancement is largest. The standard-VOF comparison in Fig. 11 shows quantitatively different droplet and continuous phase velocity PDFs at phi=40%, confirming that the numerical method affects the droplet statistics on which the mechanism rests. The authors should either provide evidence that the remaining numerical coalescence does not change the sign or magnitude of the fore-head PDF in Fig. 5(b), or explicitly restrict the mechanism claim to volume fractions where the droplet size distribution is validated.","section":"Appendix B, Fig. 10"},{"comment":"The resolution test is conducted only for phi=0 and phi=10%, and the paper acknowledges that a resolution test at phi=40% is impractical. The proposed mechanism depends on interface curvature, droplet deformation, and interfacial tension in the viscous sublayer; the Kolmogorov-scale grid estimate does not address interface-resolution requirements such as curvature accuracy and parasitic currents at high phi. A resolution study at an intermediate high volume fraction, or at least a quantitative estimate of how the fore-head interfacial tension statistics change with grid refinement, is needed to support the deformation-based mechanism at the volume fractions where it is invoked.","section":"Appendix A"},{"comment":"The marker-group volume fraction phi_mvf=5% is a free parameter that controls the maximum droplet size and the degree of numerical coalescence, and no sensitivity study with respect to this parameter is reported. Since the fore-head/rear-end mechanism depends on droplet size and deformation, and since the choice is justified only by reproduction of the global drag from experiments, the robustness of the mechanism to different phi_mvf values should be demonstrated to rule out a method-induced artifact.","section":"§2, choice of phi_mvf"},{"comment":"The fore-head/rear-end division by the droplet center of mass is acknowledged by the authors as 'somewhat idealized.' The proposed causal chain relies on this division being physically meaningful, but the paper does not test whether the result is sensitive to the definition of the dividing surface, for example by using a different split criterion or by weighting droplets by their deformation. Such a sensitivity check would strengthen the causal interpretation beyond the exact flux budget.","section":"§3, discussion near Fig. 5"}],"minor_comments":[{"comment":"The phrase 'fore head' is written inconsistently with the standard 'forehead' and appears without hyphen in most places; please unify the spelling.","section":"§1 and Appendix B"},{"comment":"The operator 'Í' used for radial averaging is nonstandard and not defined in the main text; please define it explicitly at first use.","section":"§2, notation near Eq. (3.1)"},{"comment":"The normalization of the PDFs by the absolute value of the total azimuthal interfacial tension experienced by all droplets in the range y+<24 should be stated more clearly in the caption, since the PDFs are not conventional probability densities.","section":"Fig. 5(b) caption"},{"comment":"The sentence 'The datasets agree well with each other at phi ⩽ 30%, with a minor deviation observed at phi=40%' should be supported by an explicit error metric, as the deviation at 40% is later used to qualify the method's fidelity.","section":"§2, text after Fig. 2"},{"comment":"The statement that |u|_max/u_i < 0.02 is acceptable would be more informative with a comparison to the typical turbulent velocity fluctuations in the simulations, rather than only to the inner-cylinder velocity.","section":"Appendix C"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is appropriate for a fluids journal and the authors are transparent about the main numerical limitations. The central budget decomposition is exact and the global drag reproduction is convincing, but the deformation-based mechanism is exactly the part most sensitive to numerical coalescence and resolution at high phi. I do not see grounds for rejection, but the causal claim should be either further substantiated or appropriately tempered. The paper is somewhat incremental relative to the authors' earlier JFM Rapids publication, but it adds a specific mechanistic narrative and detailed near-wall statistics, which is a meaningful contribution if the concerns above are addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe paper you'll want to know about: it reproduces droplet-induced drag enhancement in Taylor-Couette flow with a modified multi-marker VOF method and then gives a concrete physical picture for it. The new bit is the fore-head/rear-end split: stretched droplets near the inner cylinder feel interfacial tension opposing the flow at their fore head and pulling forward at their rear end, with the fore-head effect dominating, which brakes the continuous phase, disrupts high-speed streaks, and raises wall shear stress. That mechanism is not in the earlier literature, and it is a genuinely useful hypothesis for emulsion drag modeling.\n\nThe paper does several things well. The angular velocity flux decomposition is exact and clearly explained. The result that the interfacial contribution J_sigma grows with volume fraction while the turbulent and viscous radial averages do not is robust and consistent with the authors' prior JFM Rapids. The paper is also honest: the appendices admit the multi-marker method does not fully remove numerical coalescence, that the simulated droplet-size PDF deviates from experiment at phi=40%, and that no resolution test was run at that volume fraction because of cost. That self-reporting is rare and should be credited.\n\nThe soft spots are real but not fatal. The causal chain—deformation leads to fore-head/rear-end asymmetry leads to streak disruption—rests on interpretive diagnostics. The fore-head/rear-end division along the droplet center of mass is admittedly idealized. More importantly, the same numerical-coalescence issue that the authors flag could change droplet size and deformation at 30-40%, precisely where the mechanism matters most. The comparison with standard VOF in Fig. 11 shows method dependence in the droplet velocity PDFs at 40%. So the budget identity is solid, but the deformation-based mechanism is a well-supported interpretation rather than a proven derivation. The choice of marker-group volume fraction phi_mvf=5% is calibrated to reproduce the experimental drag; that is a mild circularity, but it does not by itself manufacture the mechanism—the flux decomposition and PDFs are independent diagnostics.\n\nWho is this for? People working on emulsion rheology, multiphase turbulence, and two-phase flow modeling. The paper deserves a serious referee. The mechanistic claim should be scrutinized, especially in the high-volume-fraction regime, but the exact budget result and the clear statement of the proposed pathway are worth engaging with. I would send it to review.","headline":"Solid simulation study with an exact flux decomposition and a plausible, clearly-stated fore-head/rear-end mechanism; the high-volume-fraction causal chain rests on unverified droplet statistics, so review but with careful attention to the coalescence caveats.","tokens_in":21132,"tokens_out":1993,"would_cite":true,"duration_ms":19123,"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":"Droplet drag in Taylor–Couette flow traced to interfacial tension on stretched droplet fore heads.","keywords":["Taylor-Couette flow","turbulent drag enhancement","interfacial tension","droplet deformation","multi-marker VOF method","liquid-liquid two-phase flow","angular velocity flux","volume-of-fluid method"],"falsifier":"A direct test would be to run the same geometry and parameters with interfacial tension switched off ($\\sigma = 0$) while keeping droplets deformable by another mechanism; if drag enhancement persists, the claim that interfacial tension is the source fails. Alternatively, tracking individual droplets near the inner cylinder and measuring the azimuthal interfacial tension summed over fore heads should give a net against-flow value that grows with volume fraction; observing zero or positive fore-head forces would falsify the mechanism.","tokens_in":1717,"feed_emoji":"💧","tokens_out":1869,"duration_ms":65674,"temperature":0.7,"pith_summary":"This paper uses direct numerical simulations of liquid-liquid Taylor-Couette turbulence to identify why dispersed droplets raise the drag of the flow when they have the same density and viscosity as the carrier liquid. At Reynolds number 5200 and droplet volume fractions up to 40%, the torque on the inner cylinder rises with droplet fraction, matching earlier experiments. The paper argues that the extra drag comes from interfacial tension, not from turbulent or viscous stresses averaged over the gap: the interfacial contribution to the angular velocity flux grows with droplet fraction while the other two contributions do not. Near the inner cylinder, shear stretches droplets, and interfacial tension on the fore head (the side closer to the cylinder) pulls against the flow, so the droplet phase moves slower than the surrounding fluid and brakes it. That braking disrupts high-speed streaks, promotes low-speed streaks, and raises the viscous stress at the wall.","feed_headline":"Stretched droplets brake turbulent flow via their fore heads","feed_subtitle":"Surface tension on the side of a droplet facing the wall slows the carrier fluid and raises torque.","key_machinery":"The load-bearing tool is the angular velocity flux decomposition $J^{\\omega} = J_T^{\\omega}(r) + J_V^{\\omega}(r) + J_\\sigma^{\\omega}(r) = \\mathrm{const.}$, which splits the conserved flux into turbulent, viscous, and interfacial-tension transport. The interfacial term is $J_\\sigma^{\\omega}(r) = -\\int_{r_i}^r \\langle r^2 f_\\sigma^{\\theta} \\rangle\\,dr$, so it directly converts the azimuthal component of interfacial tension into a drag contribution. The causal mechanism is the fore-head/rear-end asymmetry: near the inner cylinder, shear stretches droplets, and interfacial tension on the fore head acts against the flow while that on the rear end acts with it; the fore-head effect dominates, producing a net hindering force that slows the continuous phase and raises wall viscous stress.","core_discovery":"The central claim is that the droplet-induced drag enhancement in this system originates from the contribution of interfacial tension, specifically the hindering of the continuous phase by the fore head of stretched droplets. In a Taylor-Couette cell at Re=5200 with neutrally buoyant, equally viscous droplets at volume fractions 0-40%, the torque required to drive the inner cylinder increases with the volume fraction, in line with experiment. Decomposing the conserved angular velocity flux into turbulent, viscous, and interfacial-tension parts shows that the interfacial part grows monotonically with volume fraction, while the radial averages of the other two stay essentially unchanged. Within the viscous and buffer layers near the inner cylinder, shear deforms droplets in the streamwise direction so that the rear end lags behind the fore head; interfacial tension then acts against the flow on the fore head and with the flow on the rear end, with the fore-head effect dominant. This braking effect lowers the angular velocity near the wall, disrupts high-speed streaks, increases the occurrence of low-speed streaks, and thereby raises viscous stress and drag.","pith_inferences":["A testable extension is to vary the Weber number at fixed Reynolds number: the fore-head braking effect should weaken as interfacial tension becomes weaker relative to inertia, giving a shear-thinning trend in effective viscosity.","If the mechanism is general, the same fore-head braking should appear in other wall-bounded shear flows with deformable droplets, including turbulent channel and pipe flows, not only Taylor-Couette geometry.","Comparing droplets with progressively higher surface tension against rigid-particle suspensions could sharpen the claim that deformability is what caps the drag enhancement.","Direct experimental measurement of droplet interface curvature near the wall could reveal the sign of the azimuthal interfacial tension on fore heads and rear ends, providing a non-simulation check on the proposed causal chain."],"forward_implications":["If the central claim is correct, models of turbulent two-phase drag must include interfacial tension in the near-wall region, and omitting it will underpredict drag at moderate volume fractions.","The linear wall law $u^+ = y^+$ breaks down in the viscous sublayer of two-phase flow, because interfacial tension acts as an additional near-wall stress rather than a simple roughness effect.","Drag enhancement should grow with droplet volume fraction at fixed Reynolds and Weber numbers, consistent with experiments up to about 30% volume fraction, with the deviation at 40% attributed to residual numerical coalescence.","The mechanism explains why deformable droplets enhance drag less than rigid particles: deformability lets droplets adjust shape and exert a weaker hindering effect on the surrounding flow.","In industrial Euler-Lagrange type predictions, the boundary-layer interfacial tension contribution must be added to correctly capture the torque increase reported here."],"supporting_citations":[{"why":"Supplies the experimental torque-versus-volume-fraction data and droplet size distribution that the simulations reproduce and extend.","marker":"Yi et al. (2021)"},{"why":"Provides the multi-marker front-capturing method that the modified multi-marker VOF approach builds on to limit numerical coalescence.","marker":"Coyajee & Boersma (2009)"},{"why":"Establishes the angular velocity flux formalism and the torque-flux relation used for the budget decomposition.","marker":"Eckhardt et al. (2007)"},{"why":"Documents how coalescence reduces interfacial area and weakens drag enhancement, used to interpret the role of numerical coalescence and the droplet-particle comparison.","marker":"de Vita et al. (2019)"},{"why":"Prior interface-resolved liquid-liquid Taylor-Couette simulation that reported advection-dominated and interface-dominated regimes and the near-wall interfacial stress behavior.","marker":"Hori et al. (2023)"},{"why":"Gives the particle-induced stress decomposition in turbulent channel flow that motivates the analogous angular velocity flux decomposition and the droplet-particle analogy.","marker":"Picano et al. (2015)"},{"why":"Previous numerical study of drag modulation by dispersed drops in Taylor-Couette flow, used for validation and for the consistent finding that interfacial tension enhances momentum transport.","marker":"Su et al. (2024b)"}],"fun_headline_variants":["Stretched droplet fore heads slow flow and boost drag","Interfacial tension at droplet fore heads raises drag","Fore-head tension on stretched droplets boosts drag","Droplet fore-head tension slows flow and ups drag"],"cache_read_input_tokens":23296,"weakest_assumption_plain":"The load-bearing assumption is that the modified multi-marker VOF method with marker-group volume fraction 5% produces droplet sizes, deformation, and coalescence behavior representative of the real experiments at all volume fractions up to 40%; the paper itself notes that numerical coalescence remains and that no resolution test was run at 40%.","fun_headline_variants_meta":{"raw":{"variants":["Stretched droplet fore heads slow flow and boost drag","Interfacial tension at droplet fore heads raises drag","Fore-head tension on stretched droplets boosts drag","Droplet fore-head tension slows flow and ups drag"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00086,"raw_usage":{"total_tokens":3797,"prompt_tokens":1073,"completion_tokens":2724,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":689,"completion_tokens_details":{"reasoning_tokens":2662}},"tokens_in":689,"tokens_out":2724,"duration_ms":23221,"temperature":1.0,"reasoning_tokens":2662,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:48:50.912713+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct test would be to run the same geometry and parameters with interfacial tension switched off ($\\sigma = 0$) while keeping droplets deformable by another mechanism; if drag enhancement persists, the claim that interfacial tension is the source fails. Alternatively, tracking individual droplets near the inner cylinder and measuring the azimuthal interfacial tension summed over fore heads should give a net against-flow value that grows with volume fraction; observing zero or positive fore-head forces would falsify the mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the multi-marker front-capturing method that the modified multi-marker VOF approach builds on to limit numerical coalescence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the angular velocity flux formalism and the torque-flux relation used for the budget decomposition."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the particle-induced stress decomposition in turbulent channel flow that motivates the analogous angular velocity flux decomposition and the droplet-particle analogy."}],"review_version":1}