{"id":"5c2bacc6-c5de-407c-bf08-8fab12d9eeff","arxiv_id":"1908.07224","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Small Besov-space initial data produce a unique global strong solution with polynomial decay for the compressible Navier-Stokes-Korteweg system in R^N, N≥3.","lead":"This paper proves that the equations describing a compressible fluid with capillary, or surface tension, effects admit a unique global strong solution whenever the initial data are small perturbations of a uniform state. The proof combines maximal Lp-Lq regularity with Lp-Lq decay estimates to control the nonlinearity for all positive times.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.1 omits q1≤4, but the L∞ decay argument at Eq. (4.14) explicitly requires q1/2≤2, so the stated parameter range is not covered for N≥5.","rationale":"The reader's weakest assumption correctly identifies the decisive gap: the proof of the L∞ decay estimates in the global fixed-point argument requires q1/2≤2, equivalently q1≤4, while Theorem 1.1 states no such condition. This is not a mere technical annoyance: for N≥5, the stated hypotheses admit q1>4, and the paper's own line before (4.14) flags the restriction. The L∞ component of the solution norm is load-bearing because the nonlinear terms are closed via (4.23), (4.29), and (4.39), and without the Lq1/2→L∞ decay on the large-time interval the fixed-point estimate N(ω,w)(∞)≤C(I+N(θ,u)(∞)^2) is not justified. The example N=5, q1=4.9, q2=245, p=3, τ=0.34 satisfies every hypothesis of Theorem 1.1, so the theorem as stated genuinely asserts more than the proof establishes. Because adding q1≤4 to the theorem would repair the statement and the remainder of the argument appears coherent, the appropriate verdict remains CONDITIONAL, which is exactly the reader's verdict; no adjustment is needed. The reliance on unpublished reference [18] is an additional support concern, but the q1≤4 omission is the more immediate, internally verifiable obstruction.","tokens_in":25102,"tokens_out":5268,"duration_ms":55111,"concrete_test":"Re-derive the bound for I1∞ and I2∞ in Section 4.2 using only estimates permitted by Theorem 4.1's hypotheses, with the admissible parameters N=5, q1=4.9, q2=245, p=3, τ=0.34. Since t−s≥1 in those integrals and q1/2>2, condition (4.6) blocks the choice (p,q)=(∞,q1/2); check whether any alternative splitting, Lq2-based estimate, or interpolation argument can produce the factor t^{-N/q1-j/2} in (4.17) and (4.18). If no such alternative is found, the theorem statement should be amended to include q1≤4.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of Theorem 1.1 is global well-posedness for all exponents satisfying 2<p<∞, q1<N<q2, 1/q1=1/q2+1/N, and 2/p+N/q2<1. The proof in Section 4.2 closes the fixed-point argument by controlling the L∞ part of the norm N(θ,u). For t>2, the Duhamel integrals I1 and I2 in (4.13) have t−s≥t/2>1, so Theorem 4.1 is invoked in the regime t≥1 with (p,q)=(∞,q1/2). The manuscript itself states immediately before (4.14) that this is done 'under the condition q1/2≤2', and Theorem 4.1's condition (4.6) restricts t≥1 to q≤2. Thus the proof requires q1≤4. Theorem 1.1 does not state this restriction. It is not automatic: for N=5, take q1=4.9, q2=245, p=3, and τ with 1/3<τ<0.354; all hypotheses of Theorem 1.1 are satisfied, but q1/2=2.45>2. For this admissible parameter set, the L∞ decay estimate used to obtain (4.23) is not available, so the theorem as stated is broader than the proof. No alternative argument is supplied in the paper that would cover q1>4. The gap is internal to the proof, not a matter of disagreement with a consensus result.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies the compressible Navier-Stokes-Korteweg system in R^N, N >= 3, and claims global well-posedness for small initial data with polynomial decay. The strategy is standard for the maximal Lp-Lq regularity approach: Section 2 establishes maximal regularity for a variable-coefficient linearized problem via R-bounded solution operators and operator-valued Fourier multipliers; Section 3 proves local well-posedness by a fixed-point argument; Section 4 proves Lp-Lq decay estimates for the constant-coefficient linearized problem and then closes a global fixed-point argument in the weighted space X_{p,q2,∞}. The central assertion is Theorem 1.1, which states that under condition (1.2), for exponents p, q1, q2 satisfying 2 < p < ∞, q1 < N < q2, 1/q1 = 1/q2 + 1/N, 2/p + N/q2 < 1, and tau in (1/p, N/q2 + 1/p), sufficiently small initial data yield a unique global solution with N(theta,u)(∞) <= L ε.","tokens_in":25452,"tokens_out":5899,"duration_ms":63638,"significance":"If the proof covers the stated parameter range, the paper would be a useful contribution: it provides global strong solutions with decay in a maximal Lp-Lq regularity setting, with initial Besov regularity independent of the dimension, and it gives quantitative nonlinear estimates. The paper is detailed and the overall architecture is coherent, with no fitted parameters and no circular dependence on the target theorem. However, the proof as written contains a load-bearing gap: the global fixed-point argument requires the additional restriction q1 <= 4, which is absent from Theorem 1.1 and is not automatic when N >= 5. The significance of the result is therefore currently conditional on either adding that restriction or supplying a new decay estimate for the missing range.","major_comments":[{"comment":"The proof of the global theorem requires q1 <= 4, but Theorem 1.1 does not state this condition. In the estimate of I^1_∞, the manuscript applies Theorem 4.1 with (p,q) = (∞, q1/2) and explicitly says this is done 'under the condition q1/2 ≤ 2' immediately before Eq. (4.14). Theorem 4.1's condition (4.6) for t >= 1 also requires q <= 2. The hypotheses of Theorem 1.1 do not imply q1 <= 4. For example, with N = 5, q1 = 4.9, q2 = 245, p = 3, and tau = 0.34, all conditions of Theorem 1.1 are satisfied, but q1/2 = 2.45 > 2. For such admissible parameters, the L∞ decay bound used to obtain (4.23) is not available, and no alternative argument is supplied. The theorem as stated is therefore broader than the proof.","section":"Section 4.2, Eq. (4.14); Theorem 1.1"},{"comment":"The missing restriction q1 <= 4 is not a cosmetic point: it is needed for the very first term in the t > 2 branch of the fixed-point argument. The linear initial-data term S(t)(rho0,u0) is estimated by Theorem 4.1 with (p,q) = (∞, q1/2), and the Duhamel terms I^1_∞ and I^2_∞ are also estimated with that choice. Since q1 > N >= 3, the alternative q = q1 is not available for the t >= 1 regime of Theorem 4.1, because it violates q <= 2. Thus, for q1 > 4 with N >= 5, the weighted L∞ part of N(theta,u) cannot be closed by the present proof. The authors should either add q1 <= 4 to Theorem 1.1 or prove a decay estimate that covers q1/2 > 2.","section":"Section 4.2, Eq. (4.23)"}],"minor_comments":[{"comment":"The base resolvent construction in the proof of Theorem 2.4 is imported from Theorem 3.1 of [18], which is cited only as a preprint. Since Theorems 2.2, 3.2, and ultimately Theorem 1.1 depend on it, the authors should either cite the published version or include the needed statement in the paper so that the dependency can be verified.","section":"References, [18]"},{"comment":"In the Lq1 estimates the display after (4.25) is labeled I^3_∞, but it should be I^3_{q1}; this is confusing because the preceding 'Estimates in L∞' paragraph uses the same symbol. Please correct the label.","section":"Section 4.2, Eqs. (4.13)-(4.28)"},{"comment":"The definition of the norm N(theta,u) has mismatched parentheses in the weighted Lp terms; the expression '‖(< s >^{ℓ_i}(theta,u)‖_{Lp((0,t),W^{3,2}_{q_i}(R^N))}' should have a consistent closing parenthesis. Please rewrite the norm with unambiguous brackets.","section":"Equation (1.3)"},{"comment":"There are numerous typographical errors, including 'In oder', 'drived', 'Golobal', 'thses', 'constrction', 'samller', and 'consraction'. A careful proofreading pass is needed.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The central issue is the q1 <= 4 gap in the proof of Theorem 1.1. If the authors add the condition q1 <= 4, the proof appears coherent for the remaining range; if the original parameter range is essential, a genuinely new decay estimate is needed. The dependence on the unpublished preprint [18] should also be resolved before final acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things first. The paper is a genuine contribution to the PDE method literature: global well-posedness for the Korteweg system in the maximal Lp-Lq setting, with polynomial decay, for small initial data in Besov spaces of dimension-independent regularity. The proof is detailed, the linear decay estimates are central and mostly self-contained, and the fixed-point framework is coherent. The second thing is that Theorem 1.1 as stated is broader than the proof. At line (4.14) the authors apply their decay estimate Theorem 4.1 with (p,q)=(∞,q1/2) and state they need q1/2≤2. Theorem 1.1 never says q1≤4, and for N≥5 the hypotheses allow q1>4. The stress-test example is valid: N=5, q1=4.9, q2=245, p=3 satisfies all stated conditions, and then the L∞ decay used to close the argument is not available. This is a load-bearing gap, though probably repairable by adding q1≤4 or reworking the decay estimates.\n\nWhat is actually new is the method: Lp-Lq maximal regularity with R-bounded solution operators applied to the Korteweg model, and the accompanying Lp-Lq decay for the linearized semigroup. Global existence for small data in critical Besov spaces was already known from Chikami-Kobayashi and Kobayashi-Tsuda, so the problem is not new, but this particular functional framework and the decay statements are. I believe that counts as a solid incremental contribution.\n\nThe soft spots beyond the exponent gap: the resolvent estimates come in part from the unpublished preprint [18]; that is not fatal but should be checked. Some lemmas are quoted from the authors' own previous papers; they are used as tools, not as restatements of the main result, so I see no circularity. The citation pattern is normal.\n\nThis paper is for PDE specialists working on maximal regularity for compressible fluids and on Lp-Lq decay semigroups. It deserves a serious referee—the argument is substantial and mostly sound. My recommendation: send it to review, but require the authors to either add q1≤4 to Theorem 1.1 or extend the proof. Do not accept as is.","headline":"A solid maximal-regularity treatment of the Korteweg model with a real gap between the stated exponents and the proof; fixable, but not as written.","tokens_in":25979,"tokens_out":4269,"would_cite":true,"duration_ms":41153,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q35","76N10","35B40"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that small initial density and velocity perturbations of the compressible Korteweg fluid model in N-dimensional space (N≥3) yield a unique global strong solution with polynomial time decay.","keywords":["Korteweg fluid model","compressible fluids with capillarity","global well-posedness","maximal Lp-Lq regularity","polynomial decay","small initial data","diffuse interface model","Besov spaces"],"falsifier":"Check Theorem 1.1 with $N=5$, $p=3$, $q_1=4.9$, $q_2=245$, and $\\tau=0.34$: all hypotheses $q_1<N<q_2$, $1/q_1=1/q_2+1/N$, $2/p+N/q_2<1$, and $\\tau\\in(1/p,N/q_2+1/p)$ are satisfied. Since $q_1/2=2.45>2$, the decay estimate (4.5) with $(p,q)=(\\infty,q_1/2)$, which the proof invokes just before (4.14) to control the Duhamel term in $L_\\infty$, is outside the allowed range (4.6) for $t\\ge1$; hence the fixed-point bound (4.10) is not justified for such data. If the theorem is true for this exponent choice, a new argument must supply the missing $L_\\infty$ control.","tokens_in":24878,"feed_emoji":"🌊","tokens_out":18628,"duration_ms":167168,"temperature":0.7,"pith_summary":"This paper proves global well-posedness for the compressible fluid model of Korteweg type, a diffuse-interface model for liquid-vapor flows that includes capillary stress in the momentum equation. The claim is that in every dimension $N \\ge 3$, sufficiently small initial perturbations of density and velocity, measured in the Besov-type space $D_{q_i,p}$ together with $L^{q_1/2}$, produce a unique solution $(\\rho,u)=(\\rho_*+\\theta,u)$ that exists for all time and lies in a maximal $L_p$-$L_q$ regularity class. The same proof yields quantitative polynomial decay: the norm $\\mathcal N(\\theta,u)(\\infty)$ is bounded by $L\\varepsilon$, with time weights such as $\\langle t\\rangle^{N/q_1}$ on the $L_\\infty$ part. A reader would care because earlier global existence results for this model required critical Besov regularity tied to the dimension or high-order Sobolev spaces, whereas here the data regularity is independent of $N$ and decay rates come out of the same fixed-point argument.","feed_headline":"Small data give unique global solutions for Korteweg fluids","feed_subtitle":"Small initial data in N dimensions (N≥3) yield a unique solution that decays polynomially in time.","key_machinery":"The central object is the linear solution operator $S(t)$ for the linearized Korteweg system (4.2), together with its decay estimate (Theorem 4.1): $\\|\\partial_x^j S(t)(f,g)\\|_{W^{1,0}_p}\\le C t^{-N(1/q-1/p)/2-j/2}\\|(f,g)\\|_{W^{1,0}_q}$ under $1<q\\le p\\le\\infty$, with the additional restriction $q\\le2\\le p$ for $t\\ge1$. The estimate is obtained by splitting Fourier space: on low frequencies the symbol's eigenvalues satisfy $\\lambda_\\pm=-(\\alpha_*+\\beta_*)|\\xi|^2/2 \\pm i\\sqrt{\\rho_*\\gamma_*}|\\xi|+O(|\\xi|^2)$ as $|\\xi|\\to0$, which gives the polynomial decay, and on high frequencies the semigroup decays exponentially. The other main ingredient is maximal $L_p$-$L_q$ regularity for the linearized problem with variable coefficients (Theorem 2.2), meaning that the time derivative and the highest spatial derivatives are bounded in $L_p$ in time with values in $L_q$ in space; this is built from $R$-bounded solution operators and an operator-valued Fourier multiplier theorem. These two tools let the Duhamel term $\\int_0^t S(t-s)(f,g)\\,ds$ be split into three time pieces and controlled entirely in terms of the weighted norm $\\mathcal N$.","core_discovery":"Under the parameter assumptions $\\mu_*>0$, $\\mu_*+\\nu_*>0$, $\\kappa_*>0$, $P'(\\rho_*)>0$, and $(\\mu_*+\\nu_*)^2/(4\\rho_*^2)\\ne \\rho_*\\kappa_*$, the paper establishes the following. For $N\\ge 3$ and exponents $2<p<\\infty$, $q_1<N<q_2$, $1/q_1=1/q_2+1/N$, $2/p+N/q_2<1$, and $\\tau\\in(1/p,\\,N/q_2+1/p)$, there is an $\\varepsilon>0$ such that any initial data $(\\rho_0,u_0)$ with $I<\\varepsilon$ admit a unique global strong solution $(\\rho,u)=(\\rho_*+\\theta,u)$ with $(\\theta,u)$ in $X_{p,q_2,\\infty}$, with the density kept in the range $\\rho_*/4\\le \\rho_*+\\theta\\le 4\\rho_*$, and with $\\mathcal N(\\theta,u)(\\infty)\\le L\\varepsilon$. The solution decays polynomially: the sup-in-time weights in $\\mathcal N$ include $\\langle t\\rangle^{N/q_1+j/2}$ for the $L_\\infty$ part, $\\langle t\\rangle^{N/(2q_1)+j/2}$ for the $L^{q_1}$ part, and $\\langle t\\rangle^{N/(2q_2)+1+j/2}$ for the $L^{q_2}$ part, so higher derivatives and lower integrability exponents pay the expected powers of $t^{-1/2}$ and $t^{-N/(2q)}$. The proof is a Banach fixed point on the Duhamel formula, with the linearized system handled by maximal $L_p$-$L_q$ regularity and its solution operator split into low-frequency (polynomial decay) and high-frequency (exponential decay) contributions.","pith_inferences":["Beyond the paper: because the two tools are modular, the same maximal-regularity-plus-decay scheme should extend to exterior domains or to Navier-Stokes-Korteweg systems with external forces, provided the linearized semigroup satisfies the analogous decay and maximal-regularity estimates.","Beyond the paper: the paper does not claim its decay rates are optimal; by optimizing $\\tau$ in the interval $(1/p,N/q_2+1/p)$ one would obtain a one-parameter family of rates, and the sharp long-time rates should be determined by the low-frequency expansion of $\\lambda_\\pm$ rather than by the fixed-point weights.","Beyond the paper: the constants in the linear estimates depend on the gap between $(\\mu_*+\\nu_*)^2/(4\\rho_*^2)$ and $\\rho_*\\kappa_*$; near that resonance the smallness threshold $\\varepsilon$ should shrink, and a separate asymptotic analysis of the coalescing eigenvalues would likely be needed there."],"forward_implications":["For every $N\\ge3$, small initial data in $D_{q_i,p}\\cap L^{q_1/2}$ give a unique global strong solution, so the local theory from Section 3 is promoted to all times under a smallness condition.","The solution decays polynomially: the $L_\\infty$ norm of $(\\theta,u)$ behaves like $t^{-N/q_1}$, and each spatial derivative of order $j$ adds a factor $t^{-j/2}$.","The maximal regularity part of the norm is weighted: $\\langle s\\rangle^{N/(2q_1)-\\tau}$ controls the highest derivatives in $L^{q_1}$ and $\\langle s\\rangle^{N/(2q_2)+1-\\tau}$ in $L^{q_2}$, giving integrable-in-time control of $\\partial_t\\theta$ and $\\partial_t u$.","The non-resonance condition $(\\mu_*+\\nu_*)^2/(4\\rho_*^2)\\ne\\rho_*\\kappa_*$ keeps the eigenvalues $\\lambda_\\pm$ distinct, so the low-frequency and high-frequency decomposition of the linearized semigroup remains valid.","The contraction argument gives an effective smallness threshold: once $L^2\\varepsilon<1$, the fixed-point map is a contraction and the solution bound $\\mathcal N(\\theta,u)(\\infty)\\le L\\varepsilon$ holds."],"supporting_citations":[{"why":"Supplies the R-bounded solution operators for the linear Korteweg system without the pressure-gradient term, the starting point for Theorem 2.4.","marker":"[18]"},{"why":"The operator-valued Fourier multiplier theorem that converts the R-bounded resolvent estimates into maximal Lp-Lq regularity for the time-dependent linear problem.","marker":"[28]"},{"why":"The real interpolation argument used to derive the semigroup estimate Theorem 2.6 from the analytic semigroup.","marker":"[21]"},{"why":"Provides the low-frequency decay estimates adapted in Theorem 4.1 for the linearized Korteweg equations.","marker":"[14]"},{"why":"Supplies the R-boundedness calculus (sums, products, Fourier multipliers) used in Lemma 2.7 to construct and invert the solution operators.","marker":"[7]"},{"why":"Provides the trace and real-interpolation lemma that yields the embedding estimates in Lemma 3.3 for the local well-posedness step.","marker":"[19]"},{"why":"The prior critical-Besov existence result that this paper's dimension-independent maximal-regularity approach extends.","marker":"[6]"}],"fun_headline_variants":["Korteweg fluids: unique global solutions for small data","Small initial data ensure global well-posedness for Korteweg","Global strong solutions for Korteweg with small data","Korteweg model: polynomial decay and unique global solutions","Unique global Korteweg solutions from small perturbations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument's closing $L_\\infty$ bound only works when $q_1\\le4$, because the linear decay estimate is applied with $q=q_1/2$ and $q\\le2$ is required for $t\\ge1$; the theorem's stated hypotheses allow $q_1>4$ for $N\\ge5$, so the proof covers fewer exponents than the statement.","fun_headline_variants_meta":{"raw":{"variants":["Korteweg fluids: unique global solutions for small data","Small initial data ensure global well-posedness for Korteweg","Global strong solutions for Korteweg with small data","Korteweg model: polynomial decay and unique global solutions","Unique global Korteweg solutions from small perturbations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000371,"raw_usage":{"total_tokens":2034,"prompt_tokens":1039,"completion_tokens":995,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":655,"completion_tokens_details":{"reasoning_tokens":920}},"tokens_in":655,"tokens_out":995,"duration_ms":7883,"temperature":1.0,"reasoning_tokens":920,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:24:42.089571+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check Theorem 1.1 with $N=5$, $p=3$, $q_1=4.9$, $q_2=245$, and $\\tau=0.34$: all hypotheses $q_1<N<q_2$, $1/q_1=1/q_2+1/N$, $2/p+N/q_2<1$, and $\\tau\\in(1/p,N/q_2+1/p)$ are satisfied. Since $q_1/2=2.45>2$, the decay estimate (4.5) with $(p,q)=(\\infty,q_1/2)$, which the proof invokes just before (4.14) to control the Duhamel term in $L_\\infty$, is outside the allowed range (4.6) for $t\\ge1$; hence the fixed-point bound (4.10) is not justified for such data. If the theorem is true for this exponent choice, a new argument must supply the missing $L_\\infty$ control.","supporting_citations":[{"cited_title":"Saito, Maximal regularityfor a compressible fluid model of Kortewe g type on general domains , preprint","cited_arxiv_id":null,"evidence_quote":"Supplies the R-bounded solution operators for the linear Korteweg system without the pressure-gradient term, the starting point for Theorem 2.4."},{"cited_title":"Weis, Operator-valued Fourier multiplier theorems and maximal Lp-regularity","cited_arxiv_id":null,"evidence_quote":"The operator-valued Fourier multiplier theorem that converts the R-bounded resolvent estimates into maximal Lp-Lq regularity for the time-dependent linear problem."},{"cited_title":"Shibata and S","cited_arxiv_id":null,"evidence_quote":"The real interpolation argument used to derive the semigroup estimate Theorem 2.6 from the analytic semigroup."},{"cited_title":"Kobayashi and Y","cited_arxiv_id":null,"evidence_quote":"Provides the low-frequency decay estimates adapted in Theorem 4.1 for the linearized Korteweg equations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the R-boundedness calculus (sums, products, Fourier multipliers) used in Lemma 2.7 to construct and invert the solution operators."},{"cited_title":"Schonbek and Y","cited_arxiv_id":null,"evidence_quote":"Provides the trace and real-interpolation lemma that yields the embedding estimates in Lemma 3.3 for the local well-posedness step."},{"cited_title":"Danchin, B","cited_arxiv_id":null,"evidence_quote":"The prior critical-Besov existence result that this paper's dimension-independent maximal-regularity approach extends."}],"review_version":1}