{"id":"fa3559a5-ebb4-4e2e-90fb-e95d7dee1c41","arxiv_id":"2507.20794","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"For the minimal nonlinear thermoelasticity system on 2D and 3D tori, unique regular solutions exist globally for small initial data and locally for large data, with temperature stabilizing to a constant and the displacement splitting into persistent divergence-free oscillations and decaying…","lead":"This paper proves that a simplified nonlinear thermoelasticity model on flat two- and three-dimensional tori has unique regular solutions that exist for all time when the initial data are small, and only for a finite time in general. It also shows that over time the temperature becomes spatially constant, the curl-free part of the displacement decays, and the divergence-free part keeps oscillating like a wave.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.6 is unproved: the displayed Young-type estimate is false, so the global small-data invariance of F and Theorem 2.2 lack support as written.","rationale":"The paper's central claim is that small initial data measured by F produce global regular solutions and that the curl-free part of the displacement decays while temperature stabilizes. Everything hinges on the a priori control of F; without Lemma 3.6, the half-Galerkin sequence is only known to exist on a finite time interval and the long-time analysis has no uniform control in time. I checked the proof of Lemma 3.6 and confirmed the reader's diagnosis: the displayed estimate is not a valid Young inequality and is not true. The correct Young inequality yields a term proportional to F, not to F times the Hessian norm squared, so the invariant-sublevel mechanism is unsupported. This is a correctness risk in the argument, not a disagreement with the consensus view; the result may be true, and the exact identity in Lemma 3.2 may admit a valid proof of the same invariance, but the submitted text does not contain it. I therefore agree with the reader's identification of the weakest assumption and see no independent reason to change the conditional verdict. The asymptotic and Helmholtz arguments appear coherent conditional on global existence, so the conditional verdict is appropriate.","tokens_in":1040,"tokens_out":2066,"duration_ms":171254,"concrete_test":"Re-derive Lemma 3.6 from (3.10) using Young's inequality with a=A and b=C_2 F^{1/2}. The resulting estimate is dF/dt <= -(C_1 - eps) A^2 + (C_2^2/(4 eps)) F, not the claimed dF/dt <= C A^2 (F-D). This analytically settles that the proof as written is invalid. If the authors still claim Lemma 3.6, they must provide a different derivation, for example by estimating the sign of the exact identity in Lemma 3.2 on a family of small-data solutions.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 3.6 is the load-bearing step: it gives the sublevel invariance of F used to globalize the half-Galerkin solutions (Corollary 3.7, Lemma 4.6) and supports the global part of Theorem 2.2 and the asymptotic theorem. In the proof, after (3.10), the authors claim an inequality of the form C_2 F^{1/2} A <= eps A^2 + (C_2/eps) F A^2, where A is the L2 norm of the Hessian of sqrt(theta). This is not a consequence of Young's inequality. The correct Young step gives a second term (C_2^2/(4 eps)) F without the factor A^2. The displayed inequality is false in general: with F=1 and A small, the left side is positive and of order A, while the right side is of order A^2, so it fails for sufficiently small A. Consequently the derived differential inequality dF/dt <= C A^2 (F-D) and Corollary 3.7 do not follow. Because no alternative proof is supplied, the existence of a global regular solution for initial data satisfying (2.4) is not established as written. The exact identity in Lemma 3.2 may still imply such invariance by a more delicate argument, but the present proof does not provide it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the simplified nonlinear thermoelasticity system (1.1) on the torus T^d, d=2,3, for displacement u and positive temperature θ. It introduces the Fisher-information-type functional F(u,θ)=1/2(∫|∇u_t|²+∫(Δu)²+∫|∇θ|²/θ), proves the differential identity in Lemma 3.2, and claims an invariant-ball estimate in Lemma 3.6. The main theorems are: global existence and uniqueness for small initial data and local existence for large data (Theorem 2.2); uniform upper and lower temperature bounds (Theorem 2.3); and long-time asymptotics (Theorem 2.4) stating that the curl-free part χ of the displacement satisfies χ(t)→0 in H¹, χ_t(t)→0 in L², θ(t)→θ∞ in L², while the divergence-free part ν solves the free wave equation and persists. Section 7 states analogous results for the Lamé operator. The proof strategy extends the 1D approach of Bies–Cieślak, using the half-Galerkin scheme of Cieślak–Muha–Trifunović, maximal regularity estimates, Moser iteration for temperature bounds, and an omega-limit argument for the asymptotics.","tokens_in":26015,"tokens_out":34582,"duration_ms":302730,"significance":"If the main results are correct, this paper is a significant advance for multidimensional nonlinear thermoelasticity: it removes symmetry restrictions that were previously needed for global regular solutions and gives a precise asymptotic decomposition into an undamped divergence-free wave and a damped curl-free component. The new functional F is natural and elegant, and the paper is largely self-contained, with detailed proofs of the approximation, uniqueness, and regularity parts. Conditional on the key a priori estimate, the arguments are convincing. The Lamé extension is a useful bonus. However, the central small-data global existence hinges entirely on Lemma 3.6, whose proof contains a concrete algebraic error; until that lemma is repaired, the main theorems do not follow as written. Given that the surrounding structure is sound, the gap looks repairable, but it is load-bearing.","major_comments":[{"comment":"The proof claims dF/dt ≤ −C1 A² + C2 F^{1/2} A ≤ (ε−C1)A² + (C2/ε)F A², with A := ∥∇²θ^{1/2}∥_{L2}. The second inequality is not a consequence of Young's inequality and is false in general: for F=1, A=0.01, ε=1, the left-hand side C2A is of order A, while the right-hand side is of order A². The correct Young estimate is C2F^{1/2}A ≤ εA² + (C2²/(4ε))F, which would produce a term proportional to F and would not yield the claimed invariant-ball inequality dF/dt ≤ C A²(F−D). In addition, the first displayed inequality is not derived from Lemmas 3.5 and 3.3 as stated: Lemma 3.5 contains C2F^{1/2}∥∇θ^{1/2}∥²_{L4}, and the interpolation needed to bound this by C2F^{1/2}A is absent; the natural Gagliardo–Nirenberg plus Poincaré estimate gives ∥∇θ^{1/2}∥²_{L4} ≤ C∥∇²θ^{1/2}∥²_{L2}, i.e., A² rather than A. Since Lemma 3.6 is the only step producing the invariant ball F≤D in Corollary 3.7, which is then used in Lemma 4.6 and the global part of Theorem 2.2, and since Theorem 2.4 concerns the global solutions of Theorem 2.2, the central existence and asymptotic statements are not established as written. Please correct the exponent in the first line if it is a typo and give the complete Young argument, or supply an alternative proof of the invariant-ball estimate.","section":"Section 3, Lemma 3.6 (proof, first displayed inequality after (3.10))"}],"minor_comments":[{"comment":"The inequality is stated with an explicit constant C=1+√d/2+d/8, but no proof is provided and the cited reference [9, Lemma A.1] is not stated; please include the lemma statement or a proof adapted to the torus setting, since the constant is not obviously valid pointwise.","section":"Section 3, Lemma 3.4"},{"comment":"The Gagliardo–Nirenberg step appears to yield, after Young's inequality, a differential inequality with higher powers than those stated (F² in d=2 and F³ in d=3 rather than F^{(1+α)/2}); the finite-time blow-up conclusion is unaffected, but the displayed ODE and the definition of Tmax should be corrected or the computation shown.","section":"Section 3, Lemma 3.8"},{"comment":"The weak formulation (4.11) contains a boundary term ∫ m_t(T)φ(T), but no corresponding term at t=0; if test functions are not required to vanish at the endpoints, both boundary terms should appear, and the subsequent energy estimate should be written over (0,t) to avoid ambiguity.","section":"Section 4, Lemma 4.9"},{"comment":"In the sentence beginning 'Let us take a sequence t_n', the third sequence is printed as χ(t_n,·) again; it should be θ(t_n,·).","section":"Section 6, Theorem 6.7"},{"comment":"The small-data condition appears to have typos: it should presumably read ∥∇v0∥²_{L2} + (2ζ+λ)∥∇ div u0∥² + ζ∥curl curl u0∥² + ∥∇θ0/√θ0∥² ≤ D, consistent with the definition of F in (7.4); as printed it uses ∥v0∥² and omits the factor ζ.","section":"Section 7, Theorem 7.3, condition (7.7)"},{"comment":"The statement says the constant D depends only on µ, d and T^d, but the functional F and the system (7.1) depend on the Lamé parameters ζ and λ; the dependence on ζ and λ should be included.","section":"Section 7, Theorem 7.3"}],"recommendation":"major_revision","confidential_remarks":"The paper is well-organized and the asymptotic picture is attractive, but the error in Lemma 3.6 is central. I was close to recommending reject because Theorem 2.2(i) is the foundation of the paper, but the flaw looks like a repairable algebraic slip (the exponent A vs A²) rather than an irreparable conceptual gap; the rest of the proof structure is coherent. I would ask the authors to fix Lemma 3.6 with a complete proof and to re-verify the constants in Lemma 3.8 and the Lamé section before resubmission. The self-citations are legitimate extensions of prior work and are not excessive."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look, but not in this form. The paper's real contribution is the Fisher-information functional F and the Helmholtz splitting of the long-time dynamics: temperature goes to a constant, the curl-free part of the displacement decays, and the divergence-free part keeps oscillating. That picture is new for d=2,3 and is argued cleanly once a global solution is known. The construction machinery is also serious: half-Galerkin approximation, uniqueness in the regularity class, Moser-type L∞ bounds for temperature, and the Lamé extension. The dissipation balance (3.3) and the entropy argument in Section 6 are coherent and well referenced. The citation of the 1D papers [5,6] and [16] is honest. The problem is Lemma 3.6. As written, the proof combines Lemma 3.5 with an inequality of the form C2 F^{1/2} A ≤ εA² + (C2/ε) F A², where A = ||∇²√θ||₂. That is not Young's inequality. Young gives εA² + C₂²F/(4ε), with no A² in the second term. The displayed inequality is false: take F=1, A=0.01, ε=1. Since Lemma 3.6 is exactly what yields the invariant ball F≤D (Corollary 3.7), and that ball is what localizes the half-Galerkin solutions globally (Lemma 4.6), the global existence part of Theorem 2.2 is unsupported as written. The long-time theorem inherits the problem. This is a load-bearing flaw, not a typo. Other soft spots are minor: in the same proof, the L4 norm of ∇√θ is silently replaced by the L2 norm of ∇²√θ; this needs an interpolation argument, not Lemma 3.3. The local-in-time argument (Lemma 3.8) looks fine. Bottom line: the paper is repairable in principle. The identity in Lemma 3.2 may support a more delicate differential inequality, and the rest of the architecture is plausible. But as it stands, the central claim of global existence for small data does not follow. I would send it to referees—this is exactly the kind of paper where a referee can identify a fixable gap—but I wouldn't accept it without a corrected Lemma 3.6.","headline":"A novel Fisher-information functional and a clean asymptotic picture in 2D/3D thermoelasticity, but the global small-data theorem currently rests on a false Young-type estimate in Lemma 3.6.","tokens_in":26443,"tokens_out":3749,"would_cite":false,"duration_ms":35021,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["74A15","74H40","35M30"],"pacs":[],"model":"deepseek-v4-flash","headline":"For small data on the 2D and 3D torus, the thermoelasticity system has a unique global regular solution, and the temperature converges to a constant while only the divergence-free displacement keeps oscillating.","keywords":["thermoelasticity","global regular solutions","local existence","Fisher information","long-time asymptotics","Helmholtz decomposition","temperature stabilization","Lamé operator"],"falsifier":"Directly evaluate the displayed estimate in Lemma 3.6 with $F=1$, $\\|A\\|=0.01$, $\\varepsilon=1$, and $C_2=1$: the claimed bound $C_2F^{1/2}\\|A\\|\\le\\varepsilon\\|A\\|^2+(C_2/\\varepsilon)F\\|A\\|^2$ becomes $0.01\\le0.0002$, which is false, so the derivation of the invariant-sublevel estimate lacks support as written. A repaired argument, or a numerical simulation of (1.1) on $\\mathbb{T}^2$ with data just below the threshold $D$, would settle whether the global-existence claim holds as stated.","tokens_in":2045,"feed_emoji":"🌡️","tokens_out":2317,"duration_ms":89895,"temperature":0.7,"pith_summary":"This paper aims to prove that a simplified nonlinear thermoelasticity system on the two- and three-dimensional torus admits unique regular solutions: global in time for sufficiently small initial data and local in time for arbitrary data. The advertised long-time picture is that the temperature stabilizes to a constant determined by the initial energy, while the displacement splits into a divergence-free part that keeps oscillating according to the free wave equation and a curl-free part that decays to zero. The tool is a new Fisher-information functional whose sublevel set is forward invariant for small data, extending a one-dimensional approach to higher dimensions. The paper also states the same existence, uniqueness, and asymptotic results for the Lamé operator.","feed_headline":"Small thermoelastic data stay regular forever; temperature goes constant","feed_subtitle":"A Fisher-information functional controls the coupling, pushing temperature to a constant and damping the curl-free part of the displacement.","key_machinery":"The central object is the functional $F(u,\\theta)=\\frac12\\left(\\int|\\nabla u_t|^2+\\int(\\Delta u)^2+\\int|\\nabla\\theta|^2/\\theta\\right)$, where the last term is the Fisher information associated with the temperature. Its time derivative is computed exactly as $-\\int\\theta|\\nabla^2\\log\\theta|^2\\,dx-\\frac{\\mu}{2}\\int(|\\nabla\\theta|^2/\\theta)\\operatorname{div} u_t\\,dx$, which, together with the vector identity $\\Delta w=-\\operatorname{curl}\\operatorname{curl}w+\\nabla\\operatorname{div}w$, shows that dissipation acts only through the divergence part. The Helmholtz decomposition then decouples the divergence-free displacement, which obeys the free wave equation, from the coupled curl-free/heat system, and the paper uses the resulting invariant-sublevel estimate $\\frac{d}{dt}F\\le C\\|\\nabla^2\\theta^{1/2}\\|_{L^2}^2(F-D)$ to obtain the small-data global bound.","core_discovery":"The paper claims that on the torus $\\mathbb{T}^d$, $d=2,3$, the system $u_{tt}-\\Delta u=-\\mu\\nabla\\theta$, $\\theta_t-\\Delta\\theta=-\\mu\\theta\\operatorname{div} u_t$ has a unique global regular solution whenever the initial data satisfy the smallness condition $\\|\\nabla v_0\\|_{L^2}^2+\\|\\Delta u_0\\|_{L^2}^2+\\|\\nabla\\theta_0/\\sqrt{\\theta_0}\\|_{L^2}^2\\le D$, and a unique local solution without that condition. For every such global solution, the paper claims the temperature converges in $L^2$ to the constant $\\theta_\\infty=\\frac12\\int|\\nabla\\chi_0|^2+\\frac12\\int|\\tilde\\chi_0|^2+\\int\\theta_0$, the curl-free part $\\chi$ of the displacement converges to zero in $H^1$ with $\\chi_t\\to0$ in $L^2$, while the divergence-free part solves the homogeneous wave equation and does not decay unless it is initially zero. The same conclusions are stated for the Lamé operator.","pith_inferences":["The mechanism suggests a testable numerical signature: on a global solution, the divergence-free Fourier amplitudes of $u$ should oscillate with constant modulus, while the curl-free amplitudes should decay; a direct Helmholtz projection in simulation should reveal this dichotomy.","Because the only dissipation in the system is entropy production $\\int|\\nabla\\log\\theta|^2$, the decay of $\\chi$ likely has no exponential rate in general; extracting the sharp polynomial rate from the torus spectrum is a natural next step that the paper does not address.","The Fisher-information functional may transfer to other hyperbolic–parabolic systems in which the heat equation couples only through $\\operatorname{div} u_t$, provided the analogous invariant-sublevel estimate can be derived.","On bounded domains with boundary conditions, the Helmholtz decomposition still applies, but the wave part feels boundary damping; the paper explicitly leaves that case open, so the oscillating component may decay there."],"forward_implications":["If the main theorem is correct, initial data satisfying (2.4) never lose regularity: the sublevel set $\\{F\\le D\\}$ is forward invariant, so the solution exists for all time.","In any global solution, the divergence-free part of the displacement is a solution of the homogeneous wave equation and therefore does not relax to rest unless its initial projection is zero; persistent oscillations are generic in dimensions 2 and 3.","The temperature necessarily approaches the constant $\\theta_\\infty$ determined by the initial energy of the curl-free subsystem, and this constant value forces the curl-free part $\\chi$ to decay to zero in $H^1$.","The same regularity, uniqueness, and asymptotic statements hold for the Lamé operator with the corresponding energy and functional, so the mechanism is tied to the divergence coupling rather than to the specific form of the elastic operator.","Without the smallness condition, uniqueness and regularity still hold on a maximal time interval, but the method gives no global continuation and permits finite-time breakdown of regularity."],"supporting_citations":[{"why":"Introduces the one-dimensional Fisher-information functional whose higher-dimensional analogue is the paper's main tool.","marker":"[5]"},{"why":"Supplies the one-dimensional asymptotic-analysis strategy, including the omega-limit and entropy-monotonicity argument, that Section 6 adapts to the curl-free/heat subsystem.","marker":"[6]"},{"why":"Provides the half-Galerkin construction of global weak solutions and the approximate-problem framework that the paper uses on the torus.","marker":"[11]"},{"why":"Supplies the torus inequality $\\|\\nabla v\\|_{L^2}\\le C\\|\\Delta v\\|_{L^2}$ used in the key estimate of Lemma 3.6.","marker":"[14]"},{"why":"Provides the comparison inequality between Hessians of $\\sqrt{w}$ and Hessians of $\\log w$ used to control the Fisher-information term.","marker":"[9]"},{"why":"Is the maximal-regularity reference used to upgrade approximate-solution regularity and justify passing to the limit.","marker":"[3]"},{"why":"Supplies the Aubin–Lions compactness lemma used to pass $\\log\\theta_n$ to the limit in the construction of solutions.","marker":"[7]"},{"why":"Provides the elliptic regularity theory behind the Helmholtz decomposition estimates used for the $H^2$ bounds on $\\chi$.","marker":"[13]"}],"fun_headline_variants":["Thermoelastic regularity on tori: temperature settles, curl-free part dies","Fisher info functional tames 2D/3D thermoelastic wave equations","Wave part oscillates forever, curl-free part decays to zero","Global thermoelastic solutions: unique, regular, with constant temperature limit"],"cache_read_input_tokens":28672,"weakest_assumption_plain":"The small-data global-existence proof depends on Lemma 3.6's inequality $\\frac{d}{dt}F\\le C\\|\\nabla^2\\theta^{1/2}\\|_{L^2}^2(F-D)$, whose derivation from Lemma 3.5 uses a displayed Young-inequality step that is not valid as written; if that step cannot be fixed, the smallness condition (2.4) does not, as written, guarantee a global solution.","fun_headline_variants_meta":{"raw":{"variants":["Thermoelastic regularity on tori: temperature settles, curl-free part dies","Fisher info functional tames 2D/3D thermoelastic wave equations","Wave part oscillates forever, curl-free part decays to zero","Global thermoelastic solutions: unique, regular, with constant temperature limit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001029,"raw_usage":{"total_tokens":4324,"prompt_tokens":923,"completion_tokens":3401,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":539,"completion_tokens_details":{"reasoning_tokens":3323}},"tokens_in":539,"tokens_out":3401,"duration_ms":24884,"temperature":1.0,"reasoning_tokens":3323,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:41:48.499101+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Directly evaluate the displayed estimate in Lemma 3.6 with $F=1$, $\\|A\\|=0.01$, $\\varepsilon=1$, and $C_2=1$: the claimed bound $C_2F^{1/2}\\|A\\|\\le\\varepsilon\\|A\\|^2+(C_2/\\varepsilon)F\\|A\\|^2$ becomes $0.01\\le0.0002$, which is false, so the derivation of the invariant-sublevel estimate lacks support as written. A repaired argument, or a numerical simulation of (1.1) on $\\mathbb{T}^2$ with data just below the threshold $D$, would settle whether the global-existence claim holds as stated.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the one-dimensional Fisher-information functional whose higher-dimensional analogue is the paper's main tool."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the one-dimensional asymptotic-analysis strategy, including the omega-limit and entropy-monotonicity argument, that Section 6 adapts to the curl-free/heat subsystem."},{"cited_title":"Cieślak, B","cited_arxiv_id":null,"evidence_quote":"Provides the half-Galerkin construction of global weak solutions and the approximate-problem framework that the paper uses on the torus."},{"cited_title":"Fuest , Global solutions near homogeneous steady states in a multidimensional population model with both predator- and prey-taxis, SIAM J","cited_arxiv_id":null,"evidence_quote":"Supplies the torus inequality $\\|\\nabla v\\|_{L^2}\\le C\\|\\Delta v\\|_{L^2}$ used in the key estimate of Lemma 3.6."},{"cited_title":"Cieślak, M","cited_arxiv_id":null,"evidence_quote":"Provides the comparison inequality between Hessians of $\\sqrt{w}$ and Hessians of $\\log w$ used to control the Fisher-information term."},{"cited_title":"Amann, Linear and quasilinear parabolic problems, I","cited_arxiv_id":null,"evidence_quote":"Is the maximal-regularity reference used to upgrade approximate-solution regularity and justify passing to the limit."},{"cited_title":"Boyer, P","cited_arxiv_id":null,"evidence_quote":"Supplies the Aubin–Lions compactness lemma used to pass $\\log\\theta_n$ to the limit in the construction of solutions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the elliptic regularity theory behind the Helmholtz decomposition estimates used for the $H^2$ bounds on $\\chi$."}],"review_version":2}