{"id":"a1cac5a4-6665-482a-b84d-5b4be0e97400","arxiv_id":"2601.05031","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A 3D vorticity-stream-vector phase-field method reduces low-Reynolds deformable-body flow to two Poisson equations and produces qualitative red-blood-cell and droplet shapes.","lead":"This paper introduces a 3D simulation method for soft objects in slow viscous flows that reduces the fluid equations to two Poisson problems via vorticity and a stream vector, avoiding pressure solves. It reports qualitative recovery of red-blood-cell-like shapes and wall migration, but validation is mostly visual and no quantitative benchmarks are given.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Wall BCs fix ω and ψ to empty-channel values, so no-slip on the total velocity is not enforced.","rationale":"The manuscript's central claim rests on the solver reducing to coupled Poisson problems while correctly imposing the channel wall no-slip. The implementation described in 'Boundary conditions' sets ω and ψ at the walls to their no-object analytical values (Eqs. 12–13). For the perturbed flow, the zero-Dirichlet component ψ_p cannot ensure ∇×ψ_p = 0 at the wall, so the total velocity acquires a tangential slip at the wall. This is a modeling error, not a discretization artifact; it directly contaminates the base flow that the vesicle experiences in every reported simulation. The reader's weakest assumption was the matched-asymptotics bridge; I find that bridge may also have issues (e.g., an apparent factor error in Eq. 14–15 evaluating ∫(∂z φ0)² dz), but the boundary-condition problem is more immediately load-bearing because it undermines the numerical results themselves. The proposed test is cheap and decisive: use the released code to check wall velocities. If the wall slip is below solver tolerance, my objection would be withdrawn and the verdict could return to conditional; until then, the central claim is not established.","tokens_in":14495,"tokens_out":27642,"duration_ms":279483,"concrete_test":"Using the released GitHub code (MemPhaseFlow3D), first simulate the empty channel and verify that the base flow is reproduced exactly. Then place a vesicle off-center in a Couette cell (as in Fig. 4C) and, at several time steps, compute v = ∇×ψ on the grid points immediately adjacent to the static and moving walls. If the tangential wall velocity is not zero to solver tolerance, no-slip is violated. As a quantitative cross-check, compute the wall vorticity from the no-slip relation (e.g., ω_z = ∂v_y/∂x - ∂v_x/∂y at the wall) and compare with the imposed Dirichlet value from Eq. (13); a mismatch of order the body's perturbation would confirm the boundary condition is not correct.","verdict_should_be":"REJECT","load_bearing_attack":"Main text, 'Boundary conditions' (after Eq. 11): the no-slip condition at the channel walls is implemented by setting the vorticity ω and the vector potential ψ to their empty-channel (base-flow) expressions, Eqs. (12)–(13), with the claim that 'the BCs of the fluid are equivalent to the conditions without immersed bodies'. This is not a valid no-slip implementation for the coupled Poisson system (5b)–(5c). In this formulation v = ∇×ψ, so no-slip at a wall is a condition on the derivatives of ψ (and, through the Stokes equations, on the wall vorticity), not on the Dirichlet values of ψ itself. Decomposing ψ = ψ_base + ψ_p and ω = ω_base + ω_p, the perturbation ψ_p satisfies ∇²ψ_p = -ω_p with ψ_p = 0 on the walls; there is no constraint on ∂ψ_p/∂n, so the tangential component of ∇×ψ_p at the wall is generally nonzero. Likewise, prescribing ω_wall = ω_base fixes the wall shear to its no-object value, whereas the body's flow perturbation changes the wall shear. The resulting slip is a continuum-level error that does not vanish under grid refinement. Since all reported results—Poiseuille/Couette shapes, stress maps, and lateral migration—depend on the wall no-slip condition, the central claim that the method produces the intended confined flow is unsupported by the boundary-condition implementation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a 3D vorticity–stream vector formulation for low-Reynolds-number incompressible flow with phase-field models of deformable interfaces. The fluid equations are recast as two coupled Poisson problems for the vorticity ω and a vector potential ψ, with the interface evolved by a Cahn–Hilliard/Model-B-type equation that can incorporate Cahn–Hilliard surface tension or Canham–Helfrich bending energies. The authors demonstrate the method on vesicles and droplets in Poiseuille and Couette flows, reporting recovery of canonical red-blood-cell shapes (parachute, slipper), lateral migration, and interfacial shear stress maps. They also introduce integral observables based on deviations of ω and ψ from their empty-channel values and release the code on GitHub.","tokens_in":14887,"tokens_out":7278,"duration_ms":83271,"significance":"If the formulation is correct, it provides a comparatively simple and lightweight alternative to LBM, boundary-integral, and immersed-boundary methods for simulating deformable particles in confined low-Reynolds flows, avoiding a pressure solve and enforcing incompressibility by construction. The paper is also open about the code, which is a positive feature for reproducibility. However, the validation is entirely qualitative (shapes compared by eye, no grid-convergence or benchmark-error study), and the boundary-condition implementation raises a serious continuum-level concern that affects all confined-flow results. The sharp-interface matched-asymptotic derivation, which is load-bearing for the physical fidelity of the interface stresses, is sketched rather than rigorously established. The potential significance is real, but the current manuscript does not yet support the central claim that the method 'successfully recovers' the physical phenomenology.","major_comments":[{"comment":"The no-slip condition at the channel walls is not correctly implemented by setting ω and ψ to their empty-channel base-flow values. For the perturbation fields, ψ_p = ψ - ψ_base satisfies ∇²ψ_p = -ω_p with ψ_p = 0 on the wall. This Dirichlet condition does not constrain the normal derivative of ψ_p, so the tangential components of v_p = ∇×ψ_p at the wall are generally nonzero; e.g., at a wall y = const, v_{p,x} = ∂ψ_{p,z}/∂y - ∂ψ_{p,y}/∂z, and the first term is not forced to vanish. Similarly, prescribing ω_p = 0 at the wall fixes the wall shear to the empty-channel value, which is inconsistent with the flow perturbation caused by the body. This is a continuum-level error that does not disappear under grid refinement, and it affects every confined-flow simulation in the paper (Figures 1–4). The statement that 'the BCs of the fluid are equivalent to the conditions without immersed bodies'","section":"Results and Discussion, Figures 1–4"},{"comment":"The validation is entirely qualitative. Shapes are compared by eye to known red-blood-cell morphologies, and stress maps are presented without reference results or error measures. There is no grid-convergence study, no comparison with boundary-integral or LBM benchmarks, and no quantitative metric (e.g., Taylor deformation parameter for droplets, or a known drag/shape for a vesicle in a channel). The observables in Figure 2 are introduced as 'very reactive' but are not validated against any physical quantity. This leaves the central claim that the method 'successfully recovers' canonical shapes unsupported. Please add at least: (i) a grid-refinement study for a representative droplet/vesicle case showing convergence of shape, stress, and the proposed observables; (ii) a comparison with an established benchmark, such as small-deformation droplet theory in shear flow or published vesicle s","section":"Supplementary Materials, 'Matched asymptotic derivation', Eqs. (S9)–(S27)"},{"comment":"The sharp-interface limit of the diffuse-interface body force is load-bearing because the vorticity Poisson equation (5b) relies on -φ∇μ producing the correct surface tension and bending stress. The derivation is incomplete and contains apparent indexing errors: Eq. (S12) uses μ0 for the next-order chemical potential while earlier μ0=0 is the equilibrium condition, and Eq. (S15) yields γ = 2√2σ/3, which appears dimensionally inconsistent with the Cahn–Hilliard energy (10) where σ has dimensions of length² (the surface tension should scale as √σ). The Helfrich stress condition in Eq. (S27) is simply stated, not derived. Since the membrane simulations use the Helfrich energy, the effective boundary conditions — particularly the bending term κ(2Δ_sH + 4H(H²−K)) — are not established. I request a complete, self-contained matched-asymptotic derivation for both energies, or a clear reference t","section":"Main text Eq. (5a) and Supplementary Eq. (S38)"},{"comment":"The volume Lagrange multiplier λ_V is introduced in the phase-field equation as ∂tφ = M(∇²μ + λ_V) - v·∇φ. If λ_V is spatially uniform, ∇²λ_V=0 and the term Mλ_V does not conserve volume: integrating over a periodic or no-flux domain gives ∫∂tφ dV = Mλ_V V, not zero. If λ_V is intended to be a Lagrange multiplier that enforces the volume constraint, it must enter as a spatially varying or time-dependent field used to project the chemical potential (e.g., μ → μ - λ_V with λ_V chosen to maintain ∫φ dV), or be added inside the Laplacian in a form that produces no net source. The paper's statement that Model-B dynamics conserves volume and that λ_V only helps at high flow speeds does not resolve this inconsistency. Since volume conservation is crucial for vesicle simulations, this needs to be clarified and corrected.","section":"Main text, Results; Supplementary"}],"minor_comments":[{"comment":"Notation is inconsistent: the vector potential is denoted ψ in the main equations (3)–(5) but ξ in the observables of Figures 2 and 3. Please unify to one symbol.","section":"Abstract / Discussion"},{"comment":"Typographical errors: 'Pouiseuille' should be 'Poiseuille'; 'non-intertial' should be 'non-inertial'; 'Canh-Hilliard' should be 'Cahn–Hilliard'.","section":"Figure captions"},{"comment":"Figure 4(C): the horizontal axis label appears as 'x(x)' which is likely a typo. Also, the figure panels D and C are discussed in an order that could be clarified.","section":"Main text 'Boundary conditions'"},{"comment":"Eq. (12) and (13): the Couette boundary conditions for ω and ψ are written component-wise in a way that is ambiguous regarding which component is meant; please write them as explicit vectors with all three components, and specify the coordinate system (especially whether the moving wall is at y=0 or y=h).","section":"Supplementary Eq. (36)"},{"comment":"The inertial extension in Supplementary Eq. (36) appears to have a missing factor of ρ in the vorticity equation; as written, ρ/η ∂tω has units of 1/length² but the diffusion term ∇²ω has units of 1/(length²·time) — please check the dimensional consistency.","section":"References"},{"comment":"Some references are incomplete (e.g., [3] has a malformed title/edition), and the GitHub repository [30] should be versioned or archived (e.g., Zenodo) for reproducibility.","section":"None"}],"recommendation":"major_revision","confidential_remarks":"The boundary-condition issue is the most serious technical problem: as written, the method does not enforce no-slip at the walls, so all confined-flow results are suspect. This is fixable in principle, but the authors must re-implement the wall treatment and provide quantitative validation. The incomplete matched-asymptotic derivation and the volume-Lagrange-multiplier inconsistency further weaken the manuscript. I would consider the paper publishable only after these issues are resolved with concrete numerical evidence. The lack of quantitative benchmarking is surprising for a methods paper; the authors should compare at least one case against a well-established code or analytical solution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the 3D vorticity–stream vector generalization is a real step forward. Replacing the pressure solve by two Poisson problems and handling the interface via a phase field is clean, and the open-source code is a useful contribution. The derivation from Stokes is standard and the system (5a–c) is plausible. The paper also does something genuinely new: it defines fluctuation observables for the vorticity and stream vector and shows they track shape changes, which is a nice diagnostic.\n\nThe soft spots are real. The validation is qualitative: shapes are compared by eye, there is no grid convergence study, no error bars on the observables, and the sharp-interface asymptotics for the Helfrich case is simply stated, not derived. Those are addressable. But the boundary-condition issue is more serious. The paper fixes ω and ψ to their empty-channel values at the walls, claiming that the BCs are equivalent to the no-body case. That is not correct. For v = ∇×ψ, no-slip at a wall requires the tangential components of v to vanish, which imposes conditions on the normal derivatives of the tangential components of ψ, not just Dirichlet values. Setting ψ to the base-flow value on the wall leaves the perturbation tangential velocity arbitrary. Likewise, prescribing ω to the base-flow value fixes the wall shear to its no-object value, which is wrong when an object perturbs the flow. This is a continuum-level slip that does not vanish under grid refinement. For Poiseuille and Couette confined flows, all reported shapes and migration behavior depend on the wall no-slip condition, so this is a load-bearing flaw. It should be fixable by using the correct boundary conditions for the vector potential (e.g., zero tangential velocity via a Neumann condition on ψ_p, or by solving for the perturbation with appropriate BCs), but as written the method is not enforcing the stated physics.\n\nThere are also minor typos in Eq. (13) for the Couette boundary conditions — dimensions don't look right — and the matched-asymptotic derivation is only sketched. The self-citation to [19] is not a problem; the 2D method is clearly the basis, and they cite it honestly.\n\nBottom line: the idea is worth engaging with, and the paper deserves peer review because a competent referee could help fix the BC issue and demand quantitative convergence. Right now I wouldn't cite it for results, but I would cite it as a formulation if the BC issue is resolved.","headline":"A promising 3D phase-field/vorticity solver whose wall boundary conditions are likely under-specified; the no-slip condition is not enforced on the total velocity.","tokens_in":15312,"tokens_out":4076,"would_cite":false,"duration_ms":44300,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Three-dimensional flows with deformable membranes and droplets can be solved as two coupled Poisson equations in vorticity and a stream vector, with no pressure solve.","keywords":["vorticity–stream vector formulation","phase-field method","low-Reynolds-number flow","deformable interfaces","Canham–Helfrich energy","Cahn–Hilliard energy","vesicle and droplet simulation","lateral migration"],"falsifier":"Compute a single well-controlled case — e.g., a vesicle with a given reduced volume in unbounded shear at low Reynolds number — with this method and with an established sharp-interface (boundary-integral) solver at matching parameters. If the steady shape, tank-treading frequency (where applicable), or stress distribution differ beyond numerical error, the effective boundary conditions are not recovered. An even sharper test is to measure the pressure jump across the interface at finite ε and check it approaches 2γH + κ(2ΔsH + 4H(H²−K)) − 2λAH as ε→0.","tokens_in":14446,"feed_emoji":"🌀","tokens_out":6560,"duration_ms":67395,"temperature":0.7,"pith_summary":"This paper aims to show that low-Reynolds-number 3D flow around deformable elastic objects — vesicles, red blood cells, droplets — can be simulated without solving the pressure field. The key move is to write velocity as the curl of a stream vector and vorticity as the unknown of a Poisson equation, turning the fluid equations into coupled Poisson problems that couple to a phase-field equation for the interface. The interface can be governed by different free energies, such as Canham–Helfrich bending for membranes or Cahn–Hilliard tension for droplets. The authors demonstrate the method by recovering canonical shapes (parachute and slipper red blood cells) in Poiseuille flow, capturing non-inertial lateral migration in Couette flow with viscosity-contrast dependence, and producing membrane stress maps. If the approach holds, micron-scale biomedical flow simulations become much simpler to implement.","feed_headline":"3D membrane/droplet flow solver reduces to two Poisson equations","feed_subtitle":"A phase-field vorticity method captures red-blood-cell shapes and wall migration without a pressure solve.","key_machinery":"The stream vector ψ, the three-dimensional analogue of the 2D stream function, defined by v = ∇×ψ with the gauge ∇·ψ = 0. Together with the vorticity ω = ∇×v, the incompressible Stokes equations reduce to two Poisson problems: ∇²ω = (1/η)(∇φ)×(∇µ) and ∇²ψ = −ω, while the interface is carried by the phase-field equation ∂tφ = M(∇²µ + λ_V) − (∇×ψ)·∇φ. The identity that carries the argument is the matched-asymptotic result that the chemical-potential body force yields, at leading order, a Gibbs–Thomson relation µ₀ = γH with γ = 2√2σ/3 for Cahn–Hilliard, mapping the diffuse model onto sharp-interface Laplace-law and Helfrich bending balances.","core_discovery":"This paper establishes that the evolution of a deformable body in a Stokes flow can be written as a phase-field advection–diffusion equation for φ, a Poisson equation for the vorticity ω, and a Poisson equation for the stream vector ψ, with velocity given by v = ∇×ψ. The body force on the fluid is the phase-field term −φ∇µ, whose sharp-interface limit produces the no-slip condition, the kinematic condition, continuity of tangential velocity, and the normal stress balance with surface tension γ = 2√2σ/3 for Cahn–Hilliard interfaces, plus an added Helfrich bending term for membrane energy. The authors demonstrate that this formulation, implemented with standard finite differences, reproduces k","pith_inferences":["A systematic ε-convergence study — decreasing interface width and grid spacing together and comparing shapes and stresses to a sharp-interface boundary-integral solver — would directly test whether the effective boundary conditions hold at finite interface width; the paper does not report such a study.","The vorticity and stream-vector fluctuation observables could in principle be measured from experimental micro-PIV velocity fields, offering a lab-side fingerprint of shape transitions in flowing cells.","The formulation's Poisson structure suggests it should port readily to sparse iterative solvers and GPU accelerators; the paper does not explore performance but the equations are linear at each step, so such acceleration is a plausible extension."],"forward_implications":["The fluid part of each time step reduces to two Poisson solves (vorticity and stream vector) plus an explicit update for the phase field, so 3D vesicle and droplet simulations run on a fixed grid with standard finite differences, without pressure boundary conditions or re-meshing.","Switching between membrane and droplet physics reduces to substituting one free-energy functional for another and recomputing the chemical potential; the fluid equations are unchanged.","The squared vorticity and stream-vector fluctuation observables provide quantitative fingerprints of shape change; they plateau when the body reaches a steady, effectively rigid configuration.","The method captures non-inertial lateral migration of vesicles in Couette flow, with equilibrium position controlled by viscosity contrast and reduced volume, and computes in-plane shear stress maps on the membrane.","The same scheme extends to inertial flow by replacing the vorticity Poisson equation with a heat-type equation, so the framework can be pushed beyond the low-Reynolds regime."],"fun_headline_variants":["Vorticity-stream method simplifies 3D flow with deformable bodies","Deformable body solver: two Poisson equations instead of pressure","Phase-field vorticity captures vesicle shapes in Stokes flow","3D fluid solver for membranes: vorticity-stream and phase-field"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the diffuse-interface force used in the vorticity equation yields, in the sharp-interface limit, exactly the physical boundary conditions of an elastic membrane or droplet — correct surface tension, correct bending resistance — and the paper states rather than derives the bending case.","fun_headline_variants_meta":{"raw":{"variants":["Vorticity-stream method simplifies 3D flow with deformable bodies","Deformable body solver: two Poisson equations instead of pressure","Phase-field vorticity captures vesicle shapes in Stokes flow","3D fluid solver for membranes: vorticity-stream and phase-field"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000138,"raw_usage":{"total_tokens":992,"prompt_tokens":747,"completion_tokens":245,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":491,"completion_tokens_details":{"reasoning_tokens":184}},"tokens_in":491,"tokens_out":245,"duration_ms":3408,"temperature":1.0,"reasoning_tokens":184,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T11:47:19.105263+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute a single well-controlled case — e.g., a vesicle with a given reduced volume in unbounded shear at low Reynolds number — with this method and with an established sharp-interface (boundary-integral) solver at matching parameters. If the steady shape, tank-treading frequency (where applicable), or stress distribution differ beyond numerical error, the effective boundary conditions are not recovered. An even sharper test is to measure the pressure jump across the interface at finite ε and check it approaches 2γH + κ(2ΔsH + 4H(H²−K)) − 2λAH as ε→0.","supporting_citations":[],"review_version":1}