{"id":"110b786c-b99c-44d8-900c-c2034a174ab4","arxiv_id":"2607.10676","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"L2 density patches for 2D inhomogeneous Navier-Stokes are uniquely well-posed and the velocity is log-Lipschitz, so the interface keeps Hausdorff dimension 1 forever.","lead":"This paper proves uniqueness of energy-class solutions for 2D density-patch Navier-Stokes with vacuum and L2 velocity, plus a log-Lipschitz bound on velocity. The patch boundary therefore stays a continuous curve of Hausdorff dimension 1 for all time, answering Lions' question at the natural energy level.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The central claims (uniqueness of immediately strong solutions at L2 energy level for Lipschitz density patches, and the log-Lipschitz estimate implying preservation of Hausdorff dimension 1) rest on a carefully closed relative-energy argument and an atomic decomposition that extends Chemin–Lerner. The only potentially delicate step is the quantitative control of trajectory distortion under the weak integrability √t \nabla u ∈ L2t L∞x; that step is proved in Proposition 2.4 by elementary estimates already available from the energy class, and the subsequent localized L4 bounds (Lemma 2.1 + Remark 2.2) and duality formula (Proposition 2.7) follow directly. The atomic construction of the modulus of continuity in Section 4 is likewise self-contained once the parabolic decay estimates from [23] are granted. No hidden assumption, circularity, or gap that would undermine the strongest claim was found. The reader's ACCEPT / HIGH / low-risk assessment is therefore left unchanged.","tokens_in":21886,"tokens_out":636,"duration_ms":8246,"concrete_test":"Independently re-derive the Lipschitz bound (2.5) for \nabla X2(t,s) from the energy functionals A1 and A2 alone (without invoking external references beyond the scale-invariant inequality already quoted), then verify that the constant in |Ys(t,x)-x|≤C√t remains finite and depends only on Cu1,Cu2,T,D; if the constant blows up for some admissible energy class the uniqueness argument would fail.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption (parabolic displacement |X(t,x)-x|≤C√t and its propagation to composed flows Ys under only √t \nabla u ∈ L^{2}t L∞x, Prop. 2.4) is correctly identified as load-bearing for the localized L4 estimates that close the relative-energy Gronwall. However, the manuscript supplies a self-contained proof: the bound \nabla X2(t,s) ≲ (t/s)1/2 follows from Cauchy–Schwarz on \tau∥\nabla u2∥L∞^{2} ∈ L1 (itself from the A1/A2 energies via the scale-invariant inequality of [17]), and the triangle inequality with the individual displacements then yields |Ys(t,x)-x|≲√t. The same control feeds the duality representation of δ\rho and the atomic log-Lipschitz argument of Theorem B. No internal inconsistency or missing estimate appears; dependence on the concurrent existence result [23] is standard sequential practice and does not create circularity.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper studies the 2D inhomogeneous incompressible Navier–Stokes system with vacuum for initial data consisting of a Lipschitz density patch ρ₀=1_D and a divergence-free velocity u₀∈L²_σ(ℝ²). It introduces the class of immediately strong solutions (Definition 1.1) controlled by the time-weighted energies A₀–A₃ and proves uniqueness of such solutions (Theorem A). Combined with the existence theory of the concurrent work [23], this yields global well-posedness at the natural energy level. The second main result (Theorem B) establishes a log-Lipschitz estimate |u(t,x)-u(t,y)|≤γ(t)|x-y|(-log|x-y|)^{1-η} for η∈(0,1/2) and γ∈L¹_loc, both for strictly positive density and for the vacuum-patch case. As a consequence the associated flow lies in L^∞_t C^{1-ε}_x and the Hausdorff dimension of the patch boundary remains equal to 1 for all time (Corollary 1.3). The uniqueness argument proceeds via a relative-energy inequality closed by localized Gagliardo–Nirenberg estimates that exploit the parabolic displacement |X(t,x)-x|≲√t and its propagation to composed flows under the weaker integrability √t∇u∈L²_t L^∞_x (Proposition 2.4). The log-Lipschitz estimate is obtained from an atomic decomposition of L² data together with previously established parabolic decay estimates for the linearized system.","tokens_in":22084,"tokens_out":843,"duration_ms":6915,"significance":"The result closes a long-standing gap in the density-patch problem of Lions by establishing uniqueness at the pure energy level L², thereby completing the well-posedness theory for vacuum patches in two dimensions. The extension of the classical Chemin–Lerner log-Lipschitz estimate to the inhomogeneous setting (with or without vacuum) is of independent interest and yields a sharp geometric conclusion: the Hausdorff dimension of the free boundary is preserved even though Lipschitz regularity of the interface is lost. The technical core—localized L⁴ estimates on transported patches that rely only on parabolic displacement and the scale-invariant bound √t∇u∈L²L^∞—is cleanly executed and appears reusable in other free-boundary or vacuum problems. Dependence on the concurrent existence paper [23] is standard sequential practice and does not diminish the novelty of the uniqueness and regularity statements proved here.","major_comments":[],"minor_comments":[{"comment":"Page 1, abstract and title: the arXiv identifier is written 2607.10676 while the manuscript header carries the same number; consistency with the final journal citation should be checked once the paper is accepted.","section":null},{"comment":"Definition 1.1 and the subsequent energy functionals A_i: a short remark clarifying that the constant C_u is allowed to depend on the initial energy and on the Lipschitz character of D would help the reader track the dependence throughout the estimates.","section":null},{"comment":"Proposition 2.4: the elementary inequality a√x≤½x+a²/2 is used without citation; a one-line reference or a brief verification would improve readability.","section":null},{"comment":"Section 4, Step C: the convention 1/∞=0 is stated but the case q=∞ is treated separately a few lines later; a single unified statement would avoid a minor notational hiccup.","section":null},{"comment":"References [23] and [24] are listed as arXiv preprints; once they appear in print the bibliographic data should be updated.","section":null}],"recommendation":"accept","confidential_remarks":"The paper is a natural and high-quality sequel to the authors’ concurrent existence work [23]. The uniqueness and log-Lipschitz statements are independent and self-contained once the parabolic displacement from [23] is granted. I see no circularity or priority issue that would require editorial intervention. The manuscript is ready for acceptance after the minor polishing listed above."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This paper closes uniqueness for immediately strong solutions of the 2D inhomogeneous NS system with vacuum when the density is a Lipschitz patch and the velocity is only L2. Combined with the concurrent existence result, that gives global well-posedness at the natural energy level—something that previously needed H1 or Ḃ^{0}_{2,1}. The second half extends Chemin-Lerner log-Lipschitz regularity to the vacuum setting by an atomic decomposition rather than frequency localization, so the flow stays in L∞t C^{1-ε}x and the free boundary keeps Hausdorff dimension 1.\n\nWhat works well is the relative-energy argument. The author localizes the Gagliardo-Nirenberg estimates to the transported patches using only the parabolic displacement |X(t,x)-x|≲√t, then propagates that control to the composed flows Ys under the weaker integrability √t ∇u ∈ L^{2}t L∞x (Proposition 2.4). That estimate is self-contained: Cauchy-Schwarz on the weighted L∞ bound coming from the A1/A2 energies plus the scale-invariant inequality of Hao-Shao-Wei-Zhang gives the Lipschitz constant of the intermediate flow, and the triangle inequality finishes it. The same control feeds the duality representation of δρ and closes the Gronwall. The atomic construction of the modulus is clean and does not rely on Fourier tools, so it may travel to other settings.\n\nThe soft spots are minor and proportional. Existence and the atom-decay estimates are imported from the concurrent arXiv [23]; that is sequential practice, not circularity. The class of “immediately strong solutions” is defined by the four time-weighted energies, which is natural but slightly non-standard. The paper does not rule out cusp formation or infinite perimeter, and it is strictly two-dimensional. None of these undermine the main claims.\n\nMath and citations look solid; the argument structure has no load-bearing gaps. This is for people working on free-boundary NS or low-regularity uniqueness. It deserves a serious referee and I would cite the uniqueness and the log-Lipschitz statement.","headline":"Solid uniqueness at pure L2 energy for Lions' density-patch problem, plus a clean log-Lipschitz extension that preserves Hausdorff dimension 1.","tokens_in":22689,"tokens_out":535,"would_cite":true,"duration_ms":5455,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["76D03","76D05","46E35","35Q30"],"pacs":[],"model":"grok-4.5","headline":"L2 density patches for 2D Navier-Stokes with vacuum are unique, and their boundaries keep Hausdorff dimension 1 forever.","keywords":["inhomogeneous Navier-Stokes","density patch","vacuum","L2 well-posedness","log-Lipschitz regularity","Hausdorff dimension","relative energy","immediately strong solutions"],"falsifier":"Exhibit two distinct immediately strong solutions that start from the same Lipschitz patch and the same L2 velocity, or produce a single such solution whose free boundary has Hausdorff dimension strictly larger than 1 at some positive time.","tokens_in":22766,"feed_emoji":"🌊","tokens_out":668,"duration_ms":6693,"temperature":0.7,"pith_summary":"The paper settles Lions' density-patch problem at the natural energy level for the two-dimensional inhomogeneous incompressible Navier-Stokes equations with vacuum. When the initial density is the indicator of a bounded Lipschitz domain and the initial velocity is merely square-integrable and divergence-free, there is at most one solution in the class of immediately strong solutions; combined with prior existence, this yields global well-posedness. The same solutions satisfy a logarithmic Lipschitz estimate on the velocity, so the associated flow map is Hölder continuous of every order strictly less than 1. Consequently the evolving free boundary remains a continuous curve of Hausdorff dimension exactly one for all time, preserving the dimension of the initial interface even though genuine Lipschitz regularity may fail.","feed_headline":"Density patches stay 1-dimensional for all time at L2 energy","feed_subtitle":"Uniqueness at natural energy plus log-Lipschitz velocity keep free-boundary dimension exactly one","key_machinery":"Relative-energy comparison of two immediately strong solutions, closed by localized L4 estimates on the transported patches that rest on the parabolic displacement |X(t,x)-x|≲√t and its propagation to composed flows under the weaker integrability √t ∇u∈L^{2}_t L^∞_x, together with an atomic decomposition of L2 that yields the log-Lipschitz modulus via parabolic decay of each atom.","core_discovery":"For a Lipschitz density patch and L2 divergence-free initial velocity there is at most one immediately strong solution of the 2D inhomogeneous Navier-Stokes system with vacuum, and that velocity obeys the log-Lipschitz bound that forces the flow into L^∞_t C^{1-ε}_x for every ε∈(0,1), so the Hausdorff dimension of the patch boundary stays equal to 1 for all positive times.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["L2 uniqueness for 2D density patches with vacuum","Log-Lipschitz velocity preserves patch boundary dimension 1","Global L2 well-posedness of inhomogeneous NS density patches","Flow stays C^{1-ε} so free boundary remains Hausdorff dim 1","Energy-level uniqueness keeps density patches one-dimensional"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The quantitative control that particle paths can wander at most order square-root of time, and that this bound survives composition of two different flows, is indispensable; without it the localized estimates that close uniqueness and the duality argument for the density difference both fail.","fun_headline_variants_meta":{"raw":{"variants":["L2 uniqueness for 2D density patches with vacuum","Log-Lipschitz velocity preserves patch boundary dimension 1","Global L2 well-posedness of inhomogeneous NS density patches","Flow stays C^{1-ε} so free boundary remains Hausdorff dim 1","Energy-level uniqueness keeps density patches one-dimensional"]},"model":"grok-4.5","effort":"low","cost_usd":0.003902,"raw_usage":{"total_tokens":1163,"prompt_tokens":721,"num_sources_used":0,"completion_tokens":70,"cost_in_usd_ticks":39020000,"prompt_tokens_details":{"text_tokens":721,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":372,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":721,"tokens_out":70,"duration_ms":3772,"temperature":1.0,"reasoning_tokens":372,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-14T10:02:26.931816+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Exhibit two distinct immediately strong solutions that start from the same Lipschitz patch and the same L2 velocity, or produce a single such solution whose free boundary has Hausdorff dimension strictly larger than 1 at some positive time.","supporting_citations":[],"review_version":1}