{"id":"bb0933c1-23c5-45c3-8db5-b84e3d59851d","arxiv_id":"2411.13980","paper_version":1,"verdict":"REJECT","confidence":"LOW","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper claims a priori controls on the second derivatives of Leray solutions in L^{\\tilde r}_T L^r for all r >= 2 with a new relation 1/\\tilde r + 6/r = 9.","lead":"A math paper claims new bounds on the second derivatives of 3D Navier-Stokes solutions in high Lebesgue spaces, with very weak time integrability. The proof uses a modified heat equation and a new integral inequality, but the main theorem contains a range error and a suspect lemma.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 1 is false: an explicit Volterra-type solution violates the claimed rearrangement inequality, so the proof of Theorem 1 collapses.","rationale":"The paper's central claim rests on Lemma 1, a new Bihari-LaSalle type inequality. The reader flagged the monotonicity step in its proof as questionable. A concrete construction shows the lemma is not merely unproven but actually false: with ψ supported on [0,1], the comparison function K_t(t) decreases after the support of ψ, so the non-decreasing rearrangement of the solution can exceed K_t(t). This directly invalidates the proof of Lemma 1 and therefore every theorem that invokes it, including the advertised second-derivative estimate for L^r with large r. The range error in the k=1 branch of Theorem 1 (the formula r̃=r(r-4)/(4r²-13r+6) is negative for r∈(3,4)) is an additional independent defect, but the false lemma is the most load-bearing concern. Because the main results are derived from a false inequality, the appropriate verdict is REJECT; a full correction would require a genuinely different comparison argument and a careful re-derivation of all exponent ranges.","tokens_in":32688,"tokens_out":42376,"duration_ms":372413,"concrete_test":"Verify the counterexample: solve the Volterra equality with ψ=1_{[0,1]}, β=γ=1/2, a=0 on [0,2], evaluate the solution at t=2 and at t=1, and compare φ_*(2)=max φ with K_2(2)=(2(√2-1))². If φ_*(2)>K_2(2), Lemma 1 is falsified.","verdict_should_be":"REJECT","load_bearing_attack":"The new Bihari-LaSalle lemma (Lemma 1) is false. Take T=2, a≡0, β=1/2, γ=1/2, and ψ(s)=1_{[0,1]}(s). Let φ solve the equality φ(t)=∫_0^t (t-s)^{-1/2}ψ(s)φ(s)^{1/2}ds. For 0≤t≤1, φ(t)=(π²/4)t; for 1≤t≤2, φ(t)=(π/2)∫_0^1 (t-s)^{-1/2}√s ds. This φ is continuous, nonnegative, satisfies the strict form of the lemma's inequality after taking a small multiple, and its non-decreasing rearrangement at t=2 equals its maximum, φ(1)=π²/4≈2.467. But K_2(2)=(∫_0^1 (2-s)^{-1/2}ds)²=4(√2-1)²≈0.686. Thus φ_*(2)>K_2^*(2), contradicting the lemma. The proof's key claim that s↦K_s(s) is non-decreasing fails here: K_1(1)=4>K_2(2)≈0.686. Since every estimate in Theorem 1 and Corollary 2 is derived through Lemma 1, the central results are not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript claims new estimates for Leray solutions of the 3D incompressible Navier-Stokes equations, controlling u, ∇u, and ∇²u in spaces L^{~ r}_T L^r under explicit relations between ~ r and r, together with weighted singular-in-time estimates. The method is a Duhamel formula around a convected heat kernel, combined with a new Bihari-LaSalle type lemma based on nondecreasing rearrangements. The paper also interpolates these estimates with earlier results and applies them to second and first derivatives.","tokens_in":33004,"tokens_out":14394,"duration_ms":139200,"significance":"If correct, the k=2 estimates would genuinely extend known derivative controls to arbitrarily large spatial integrability exponents, and the explicit relations (for instance 1/~ r + 6/r = 9 for r ≥ 2) are concrete and falsifiable. The manuscript is self-contained in its auxiliary inequalities and introduces no fitted parameters. However, the central auxiliary lemma is not valid, and one of the main ranges in Theorem 1 is internally inconsistent; the claimed results are therefore not established.","major_comments":[{"comment":"The proof of Lemma 1 rests on the assertion, immediately after (2.23), that s ↦ K_s^β(s) is non-decreasing. This assertion is false. For T=2, a≡0, β=γ=1/2 and ψ(s)=1_{[0,1]}(s), one has K_1(1)=4 while K_2(2)=4(√2-1)^2≈0.686. If φ solves the corresponding Volterra equality, then φ(s)=π²s/4 on [0,1] and φ(t)=(π/2)∫_0^1 (t-s)^{-1/2}√s ds for t>1, which is decreasing on [1,2]; hence φ_*(2)=φ(1)=π²/4 > K_2^*(2). Multiplying φ by a constant c∈(0,1) close to 1 makes the strict inequality assumed in Lemma 1 hold while preserving the violation. Although the displayed ψ is discontinuous, replacing it by a continuous bump supported in [0,1] and close to 1 on [0,1] preserves the failure of the monotonicity claim; continuity is not what validates the step. Since this monotonicity step is the only mechanism behind Lemma 1, the subsequent applications of the lemma in Theorem 1, Corollary 2, and Theorem 5 are unsupported.","section":"Section 2.4, Lemma 1"},{"comment":"The displayed formula ~ r = r(r-4)/(4r²-13r+6) is claimed for r ∈ [3,6]. This is impossible: at r=3 the formula gives ~ r=-1, and for every r ∈ (3,4) the numerator is negative while the denominator is positive, so ~ r < 0; at r=4 the expression vanishes. A Lebesgue time exponent must be strictly positive. The proof itself obtains the formula only for r>4, see Eq. (3.35), and the following paragraph treating 3<r≤4 incorrectly concludes ~ r ≥ 0. Thus the statement of Lemma 3 and the k=1 part of Theorem 1 are invalid as written, and the interpolations in Section 4 built on the [3,6] segment need to be revisited.","section":"Theorem 1 and Lemma 3, Eq. (3.9)"},{"comment":"The passage from the mollified equation to the Leray solution is not proved. The flow in (3.11) is not uniquely defined for the non-smooth velocity field, and after defining the regularized problem the author suppresses the regularization index and states that the limit is direct, without a convergence argument for the flow-dependent kernel ^np_{t,x} or for the nonlinear term in (3.25). The available convergence L^∞_T L² ∩ L²_T H¹ does not by itself justify passing to the limit in this kernel. Consequently, even apart from Lemma 1, the estimates for ∇u and ∇²u are not rigorously established for general Leray solutions.","section":"Section 3.2.1 and Eq. (3.25)"}],"minor_comments":[{"comment":"The notation k ∈ [[0,2]] is nonstandard; it should be written k ∈ {0,1,2}.","section":"Throughout"},{"comment":"The formula for ~ r, written as “1/~ r = 1−β − γr / r”, is ambiguous; it should be parenthesized and checked against the Hardy-Littlewood-Sobolev condition (2.19).","section":"Corollary 2"},{"comment":"The text first says “If r>3, and if r ∈ (4,6)” but then treats 3 < r ≤ 4 separately; this organization should be corrected, and the sign of (3.35) on [3,4] should be checked explicitly.","section":"Section 3.2.4"},{"comment":"The captions and panels are difficult to read because several curves and labels overlap; separate panels or higher-resolution figures would improve clarity.","section":"Figures 1–3"}],"recommendation":"reject","confidential_remarks":"For the editor: the central tool of the paper, Lemma 1, has an invalid proof, and the k=1 range in Theorem 1 contains negative time exponents. These are not presentation issues but load-bearing defects in the main argument. A correct proof would require a substantially different strategy for the Bihari-LaSalle step, so I recommend rejection rather than major revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"I've read arXiv:2411.13980. The advertised extension—second-derivative controls for arbitrary large r, with 1/tilde r + 6/r = 9 for r≥2—is genuinely new and not in Constantin, Lions, or Vasseur. The author is also honest that this does not touch Prodi-Serrin. But the paper has a load-bearing flaw: Lemma 1 is false.\n\nThe proof of Lemma 1 tries to compare φ with a family K_t, and uses that s↦K_s^β(s) is non-decreasing. That monotonicity is simply wrong. Here is the counterexample: T=2, a≡0, β=γ=1/2, ψ(s)=1_{[0,1]}(s). Let φ solve the equality φ(t)=∫_0^t (t-s)^{-1/2} ψ(s) φ(s)^{1/2} ds. Then φ(t)=π²t/4 on [0,1], and for t>1 it continues by the integral over [0,1]; it is decreasing after 1. A small multiple satisfies the strict inequality in the lemma. But φ_*(2)=π²/4≈2.47, while K_2^*(2)=4(√2-1)²≈0.686. The lemma's conclusion fails. I checked the arithmetic; it holds. So every estimate in Theorem 1 and Corollary 2 that relies on Lemma 1 is unproved.\n\nSeparately, the statement of Theorem 1 has a range error: for k=1 and r∈(3,4), the formula for tilde r is negative. That is a red flag that the case split was not validated, though it is fixable.\n\nWhat is worth keeping? The Duhamel formulation around a transported heat kernel is a sensible approach, and the interpolation part in Section 4 is careful. The paper is clearly written and engages properly with the literature. But none of that saves the main theorem while the lemma fails.\n\nFor a reader: this is not something to cite for the bounds. It might be useful as a starting point for someone who wants to repair the lemma or replace it. I would not bring it to a reading group as a positive example, and I would not accept it in current form. A serious editor could still send it to a referee, because the claimed result is important and the error is non-obvious enough to deserve a written report. But the expected outcome of that review should be rejection unless the author can prove a correct version of the lemma.\n\nMy recommendation: reject in current form, but encourage a resubmission with a sound lemma.","headline":"The claimed second-derivative bounds are new, but Lemma 1 is false — an explicit counterexample — so the paper's main results are unproved.","tokens_in":33534,"tokens_out":6254,"would_cite":false,"duration_ms":57236,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q30","76D03","76D05","26D10"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves new mixed Lebesgue-space controls on Leray solutions of the 3D incompressible Navier–Stokes equations, including second derivatives in $L^{\\tilde r}_T L^r$ for arbitrarily large $r$.","keywords":["Incompressible Navier-Stokes equation","Leray solution","integral inequality","Lebesgue space estimates","second-order derivatives","Bihari-LaSalle inequality","Duhamel formula","heat kernel regularization"],"falsifier":"Test the monotonicity assertion inside Lemma 1 with elementary data: take $T=1$, $\\gamma=\\beta=1/2$, $a(t)=1$, $\\psi(s)=1+s\\sin(10s)$, form $K_t(\\tilde s) = (a(t)^{1-\\beta} + \\int_0^{\\tilde s}(t-\\sigma)^{-1+\\gamma}\\psi(\\sigma)\\,d\\sigma)^{1/(1-\\beta)}$, and compute whether $s \\mapsto \\tilde K_s^{*\\beta}(s)$ is non-decreasing as the proof requires; a single decrease at any pair $s_1<s_2$ invalidates the lemma, and with it the proof of Theorem 1.","tokens_in":32526,"feed_emoji":"🌊","tokens_out":18333,"duration_ms":137883,"temperature":0.7,"pith_summary":"This paper proves new a priori bounds for Leray solutions of the 3D incompressible Navier–Stokes equations without assuming smallness of the data. The main theorem gives explicit mixed Lebesgue-space controls for the solution, its gradient, and its Hessian: for $k=2$ and every $r\\ge 2$, the relation $1/\\tilde r + 6/r = 9$ holds, so $\\nabla^2 u \\in L^{\\tilde r}_T L^r$ with $\\tilde r = r/(9r-6)$, a regime of arbitrarily large spatial integrability that earlier second-derivative results (covering only $r<3/2$) did not reach. The proof works through a Duhamel formula around a heat equation corrected by the flow of $u$, combined with a new Bihari–LaSalle inequality that converts nonlinear time-integral bounds into Lebesgue norms. A fifth section adds weighted singular controls $\\sup_{t\\in[0,T]}\\int_0^t (t-s)^{-\\theta}\\|\\nabla^k u(s)\\|_{L^r}\\,ds < \\infty$ under the condition $\\theta < (3-kr)/(2r)$, $k\\in\\{0,1,2\\}$, $1<r<3/k$. Such quantitative controls are the currency of the Navier–Stokes regularity problem, so each new integrability range narrows the class of possible singular profiles.","feed_headline":"New estimate puts ∇²u of 3D Navier–Stokes in all L^r spaces","feed_subtitle":"Extends second-derivative controls from r<3/2 to every finite r via a Duhamel formula and Bihari–LaSalle argument.","key_machinery":"The engine is a Duhamel formula written around a heat equation whose drift is frozen along the flow $\\theta_{s,t}(x)$ of $u$: after mollification, $u$ solves $\\partial_t u + u(t,\\theta_{t,\\tau}(\\xi))\\cdot\\nabla u - \\nu\\Delta u = u^\\Delta_{[\\tau,\\xi]}\\cdot\\nabla u + \\Xi[u\\cdot\\nabla u] + Pf$, with the Gaussian kernel $\\hat p_{\\tau,\\xi}(s,t,x,y)$ centered at the frozen trajectory. Choosing the freezing point $(\\tau,\\xi)=(t,x)$ gives a representation of the vorticity $\\omega=\\nabla\\times u$ in which the nonlinear term becomes a double difference $[u(s,\\theta_{s,t}(x))-u(s,y)]^{\\otimes 2}$ against a derivative of the heat kernel. That difference is converted into the Sobolev–Slobodeckij norm $[u(s,\\cdot)]_{W^{\\gamma,2r}}$, which interpolation bounds by the known $L^\\infty_T L^2$ and $L^2_T \\dot H^1$ energy controls, leaving an integrable time singularity $(t-s)^{-1+\\gamma}$. The new Bihari–LaSalle lemma (Lemma 1) then turns the resulting inequality $\\phi(t) \\le a(t) + \\int_0^t (t-s)^{-1+\\gamma}\\psi(s)\\phi^\\beta(s)\\,ds$, $\\beta\\in[0,1)$, into $L^p$ bounds using the symmetric increasing rearrangement and the Hardy–Littlewood–Sobolev inequality.","core_discovery":"The central claim is Theorem 1: for a Leray solution $u$ of (1.2) with $u_0\\in L^2(\\mathbb{R}^3)\\cap \\dot B^{2(\\tilde r-1)/\\tilde r}_{r,\\tilde r}(\\mathbb{R}^3)$ and $f\\in L^\\infty_T L^2 \\cap L^\\infty_T \\dot B^2_{r,\\infty}(\\mathbb{R}^3)$, the norm $\\|\\nabla^k u\\|_{L^{\\tilde r}_T L^r}$ is finite whenever the indices satisfy the stated relations. The main extension is the branch $k=2$, $r\\ge 2$, where $\\tilde r = r/(9r-6)$, equivalently $1/\\tilde r + 6/r = 9$; this puts $\\nabla^2 u$ in $L^{r/(9r-6)}_T L^r$ for every finite $r$, an arbitrarily large spatial integrability that was previously out of reach, since the classical Constantin–Lions bound $2/\\tilde r + 3/r = 4$ covered only $r \\in (1,3/2)$. Theorem 5 in the same paper establishes weighted singular controls $\\int_0^T (T-t)^{-\\theta}\\|\\nabla^k u(t)\\|_{L^r}\\,dt < \\infty$ for $\\theta < (3-kr)/(2r)$, $k\\in\\{0,1,2\\}$, $1<r<3/k$. The author's aim is to establish these bounds as unconditional statements about Leray solutions; the novelty is the extension to large $r$ through the heat-flow Duhamel identity and the Bihari–LaSalle argument.","pith_inferences":["Beyond the paper: the freeze-the-flow Duhamel identity might be pushed to $k\\ge 3$, but the author notes that the Sobolev indices would force $r<1$; a genuinely new interpolation idea would be needed, so this is an obstacle explicitly acknowledged in the text rather than a proven impossibility.","Beyond the paper: the line $1/\\tilde r + 6/r = 9$ has a scaling-invariant look, and it would be natural to test sharpness by constructing near-extremal Leray profiles (self-similar or numerically optimized) that saturate the mixed-norm bound.","Beyond the paper: because each new $L^{\\tilde r}_T L^r$ control on $\\nabla^2 u$ restricts possible blow-up profiles, combining this estimate with existing regularity criteria could sharpen conditional results on the critical norm $\\|\\nabla u\\|_{L^1_T L^\\infty}$."],"forward_implications":["For every finite $r\\ge 2$, $\\|\\nabla^2 u\\|_{L^{r/(9r-6)}_T L^r} < \\infty$; in particular, as $r\\to\\infty$ the time exponent $\\tilde r$ tends to $1/9$, a quantitative bound far stronger in spatial integrability than all previous second-derivative controls.","The same framework yields $\\|\\nabla u\\|_{L^{\\tilde r}_T L^r} < \\infty$ for arbitrarily large $r$ via Sobolev embedding from the $k=2$ branch, and for $k=0$ recovers the known control $u\\in L^1_T L^\\infty$ at $r=\\infty$.","The weighted singular estimates $\\sup_{t\\in[0,T]}\\int_0^t (t-s)^{-\\theta}\\|\\nabla^k u(s)\\|_{L^r}\\,ds < \\infty$ are valid for all $\\theta < (3-kr)/(2r)$, $k\\in\\{0,1,2\\}$, $1<r<3/k$, giving a handle on the time singularities that appear in mild formulations.","Interpolating the new branches with the older $r<3/2$ controls produces a larger admissible region for $\\nabla^2 u$ and $\\nabla u$ (Figures 1–3); the equality case $(r,\\tilde r)=(4/3,4/3)$ remains the only unresolved endpoint.","These bounds do not reach the Prodi–Serrin scaling $2/\\tilde r + 3/r \\le 1+k$, so they do not by themselves imply smoothness or rule out finite-time blow-up; they tighten the constraints on any hypothetical singularity."],"supporting_citations":[{"why":"supplies the Leray weak solutions and the energy estimates (1.3)–(1.4) that every later estimate in the paper leans on.","marker":"[Ler34]"},{"why":"gives the earlier second-derivative control $2/\\tilde r + 3/r = 4$ with $r\\in(1,3/2)$, the baseline the paper's $r\\ge 2$ branch extends.","marker":"[Con90]"},{"why":"provides the same second-derivative control and a Lorentz-space $L^{4/3,\\infty}$ result used as a comparison and interpolation endpoint.","marker":"[Lio96]"},{"why":"gives local $L^r$ controls of $\\nabla^k u$ for $r<4/(k+1)$, used in Section 4 to interpolate and enlarge the admissible region.","marker":"[Vas10]"},{"why":"extends Vasseur's second-derivative estimate; its endpoint $(4/3,4/3)$ is used as the reference point for interpolation curves E–G.","marker":"[VY21]"},{"why":"states the Hardy–Littlewood–Sobolev inequality used to convert the Bihari–LaSalle output into $L^{\\tilde r}_T L^r$ bounds.","marker":"[Ste70]"},{"why":"supplies the Brezis–Mironescu Gagliardo–Nirenberg interpolation formulas used in every case analysis of Theorem 1.","marker":"[BM18]"},{"why":"defines the Sobolev–Slobodeckij norm that bounds the double-difference term $u(s,\\theta_{s,t}(x))-u(s,y)$.","marker":"[Tri83]"}],"fun_headline_variants":["Second derivative control for Navier-Stokes now holds for all r","∇²u of Navier-Stokes tamed in every Lebesgue space","All L^r bounds for second derivatives of Leray solutions","New bounds put ∇²u of 3D Navier-Stokes in any L^r","Navier-Stokes: ∇²u integrability extended to all finite r"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the new Bihari–LaSalle lemma (Lemma 1, Section 2.4) is correct; its proof assumes that the rearranged comparison function is non-decreasing in $s$ so it can be pulled out of the integral, and that monotonicity is asserted without proof.","fun_headline_variants_meta":{"raw":{"variants":["Second derivative control for Navier-Stokes now holds for all r","∇²u of Navier-Stokes tamed in every Lebesgue space","All L^r bounds for second derivatives of Leray solutions","New bounds put ∇²u of 3D Navier-Stokes in any L^r","Navier-Stokes: ∇²u integrability extended to all finite r"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000262,"raw_usage":{"total_tokens":1684,"prompt_tokens":1123,"completion_tokens":561,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":739,"completion_tokens_details":{"reasoning_tokens":460}},"tokens_in":739,"tokens_out":561,"duration_ms":44153,"temperature":1.0,"reasoning_tokens":460,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:41:11.990361+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Test the monotonicity assertion inside Lemma 1 with elementary data: take $T=1$, $\\gamma=\\beta=1/2$, $a(t)=1$, $\\psi(s)=1+s\\sin(10s)$, form $K_t(\\tilde s) = (a(t)^{1-\\beta} + \\int_0^{\\tilde s}(t-\\sigma)^{-1+\\gamma}\\psi(\\sigma)\\,d\\sigma)^{1/(1-\\beta)}$, and compute whether $s \\mapsto \\tilde K_s^{*\\beta}(s)$ is non-decreasing as the proof requires; a single decrease at any pair $s_1<s_2$ invalidates the lemma, and with it the proof of Theorem 1.","supporting_citations":[],"review_version":1}