{"id":"83dc961b-a8ed-428a-9451-efc8c5515b60","arxiv_id":"2504.17932","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Gallery wave constructions give counterexamples to Strichartz estimates for the acoustic wave equation of physical vacuum compressible Euler, indicating a derivative loss.","lead":"Small perturbations of a gas with a vacuum boundary obey a wave equation whose 'whispering gallery' solutions hug the boundary and reflect many times. The authors build explicit such waves to show that Strichartz estimates lose derivatives, which suggests the known low-regularity well-posedness threshold is optimal.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section 7 exponent algebra overstates the growth of ∇²ψ_j by d/(2r) for finite r, so Theorem 1.5 as stated fails for finite r.","rationale":"The reader's weakest_assumption (the uniform lower bound for B in Lemma 4.2) is a genuine gap, but it is likely repairable: B(μ,z) is continuous, B(μ,0)>0, and the μ-range is a fixed compact set, so a small positive strip can be justified by taking ε and [a,b] appropriately. The Section 7 exponent error is more load-bearing because it makes a stated main theorem quantitatively false for finite r, not merely insufficiently justified. The reader's rationale does flag this same error, so the verdict CONDITIONAL remains appropriate, but the weakest assumption should be the exponent algebra rather than Lemma 4.2. A revision should restate Theorem 1.5 with the corrected exponent and address the H^{2s} normalization issue in the proof.","tokens_in":907,"tokens_out":852,"duration_ms":218209,"concrete_test":"Recompute Section 7 exponents for d=3, κ=1, (q,r)=(4,4), γ=7/8, using Proposition 6.1. The corrected growth of ‖∇²ψ_j‖ is 2^{2j·9/4}; the paper claims 2^{2j·21/8}. Evaluate the limit with s=0 and α=5/2: if the paper's exponent were correct, 2^{−2jα}‖∇²ψ_j‖ would blow up, while the corrected computation gives decay to zero. This settles that the stated α-range in Theorem 1.5 is invalid for finite r.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central counterexample construction in Section 7 contains a concrete algebraic error. Proposition 6.1 gives ‖U_j‖_{L^r_x} ≃ (2^{2j})^{d−1−d/r}, and after normalization ψ_j = U_j / 2^{2j(d/2 − 1/(2κ))} one obtains ‖∇²ψ_j‖_{L^q_t L^r_x([0,1]×Ω)} ≃ (2^{2j})^{d+1−d/r−d/2+1/(2κ)}. The paper identifies this exponent with 1/q+γ+1/(2κ)+1 using the wave-admissible relation 1/q+d/(2r)=d/2−γ. But substituting d/r = d−2γ−2/q gives the exponent 1+2γ+2/q−d/2+1/(2κ), which is smaller by d/(2r) whenever r<∞. For example, d=3, κ=1, (q,r)=(4,4), γ=7/8: the paper's claimed exponent is 21/8, while the corrected exponent is 9/4. With s=0 and α=5/2, the claimed divergence range would require 2^{−2jα}‖∇²ψ_j‖ to blow up, but the correct computation gives decay to zero. Thus Theorem 1.5, as stated with the full range α < 1/q+γ+1/(2κ)+1−s, is quantitatively false for every finite r. The construction may still produce divergence for the smaller corrected range, so the overall phenomenon is not destroyed, but the stated exponent needs correction. A second issue is the normalization: the proof asserts ‖(∂tψ_j, ∇xψ_j)‖_{H^{2s}} ≈ 2^{2js} while the data have unit H-norm; this ratio does not imply the required sup ≤ 1 after the same normalization, so the H^{2s} normalization in Theorem 1.5 needs to be rechecked as well.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the irrotational compressible Euler equations in a physical-vacuum free-boundary setting. It derives a velocity-potential formulation in which the linearized equation reduces, in the model case (r = x_d, v = 0), to the acoustic wave equation ∂_t^2 ψ − κ x_d Δψ − ∂_d ψ = 0 on the upper half-space. The authors analyze the associated bicharacteristics, construct whispering-gallery modes from Laguerre-type eigenfunctions, prove Strichartz-type estimates for these modes (Theorem 1.9), and then build frequency-localized wave packets U_j to claim counterexamples showing that Strichartz estimates with the expected scaling must lose derivatives (Theorems 1.5 and 1.7). The paper explicitly frames the optimality of the Ifrim–Tataru low-regularity threshold as a suggestion, not a rigorous conclusion.","tokens_in":28657,"tokens_out":29174,"duration_ms":248909,"significance":"The construction is concrete and self-contained: it produces explicit exact solutions of the linearized equation and computes their norms directly, so it does not assume the Strichartz estimates it seeks to disprove. The positive estimate for gallery modes via stationary phase and TT* is a useful contribution in its own right, and the connection between multiply reflecting geodesics and derivative loss is natural. However, the central counterexample theorems contain a quantitative exponent error and a normalization error; once these are corrected, the range of derivative loss is smaller than stated, so Theorems 1.5 and 1.7 need revision. The underlying wave-packet construction appears salvageable.","major_comments":[{"comment":"The identity asserted in the first display of Section 7 is false. From Proposition 6.1, one has ‖U_j(t,·)‖_{L^r_x} ≃ (2^{2j})^{d−1−d/r}, so with ψ_j = U_j / 2^{2j(d/2 − 1/(2κ))} the true exponent of ‖ψ_j‖_{L^q_t L^r_x} is d−1−d/r+1/(2κ)−d/2, not 1/q+γ+1/(2κ)−1. Using the wave Strichartz relation 1/q + d/(2r) = d/2 − γ, one computes 1/q+γ+1/(2κ)−1 = d/2 − d/(2r) + 1/(2κ) − 1, which exceeds the true exponent by d/(2r). Consequently the displayed equivalence for ‖∇²_xψ_j‖ is also off by 2^{2j·d/(2r)}, and the divergence range in Theorem 1.5 should involve 1/q+γ+1/(2κ)+1−d/(2r) rather than 1/q+γ+1/(2κ)+1. The error is quantitative: for d=3, κ=1, (q,r)=(4,4), γ=7/8 and s=0, the paper's range allows α=5/2, but the corrected exponent gives 2^{−2jα}‖∇²ψ_j‖ → 0, so the asserted divergence fails.","section":"Section 7, exponent computations after the definition of ψ_j"},{"comment":"The sequence ψ_j does not satisfy the normalization hypothesis of Theorem 1.5 for s>0. The proof shows ‖(∂_tψ_j^0, ∇_xψ_j^0)‖_H ≈ 1 and then states that Bernstein's inequality gives ‖(∂_tψ_j^0, ∇_xψ_j^0)‖_{H^{2s}} ≈ 2^{2js}. For s>0 this grows with j, so the hypothesis sup_j ‖(∂_tψ_j^0, ∇_xψ_j^0)‖_{H^{2s}} ≤ 1 is false. To obtain data with H^{2s}-norm O(1), one must multiply by 2^{−2js}; after this renormalization the L^q_t L^r_x norm of ∇²ψ_j acquires the additional factor 2^{−2js}, so the correct divergence threshold in Theorem 1.5 is α < 1/q+γ+1/(2κ)+1−d/(2r)−s, with the correction from the previous comment, rather than the threshold stated in the theorem. The same normalization defect affects Theorem 1.7, where the proof's displayed range α < 1/(2q)−γ+2−s and the theorem's stated range α < 1/(2q)−γ−s must also be reconciled.","section":"Section 7, normalization of the H^{2s} norm before Theorem 1.5"}],"minor_comments":[{"comment":"The phrase 'Euler Strcihartz triple' contains a typo and should read 'Euler Strichartz triple'.","section":"Definition 1.6"},{"comment":"The phrase 'Let (q,r,γ) be wave-admissible' should be 'wave Strichartz triple', since Definition 1.2 introduces γ through the equality in item (2), not through the wave-admissibility inequality alone.","section":"Theorem 1.5 statement"},{"comment":"The statement of Proposition 5.1 omits the factor (2^{2j})^{(3(d−1)/4)(1/2−1/r)} that appears in its proof and in Remark 4.3; the statement should agree with the proof.","section":"Proposition 5.1 and Remark 4.3"},{"comment":"The step 'B(μ,0)>0. Thus, there exists an interval [a,b]...' should specify that [a,b] is chosen in a neighborhood of s=0 where B(μ,·) is uniformly positive by continuity, since B is a Laguerre-type function and may have zeros away from 0.","section":"Lemma 4.2, proof of the norm equivalence"},{"comment":"The displayed H-norm bound in Proposition 6.1 writes 2^{2j(d/2 − 1/κ)}, while the proof and the later normalization in Section 7 use 2^{2j(d/2 − 1/(2κ))}; these should be made consistent.","section":"Proposition 6.1 statement"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely to become a solid contribution after revision: the wave-packet construction is explicit, and the positive gallery-mode estimate is a concrete step. The main obstacle is that Theorem 1.5, as stated, is quantitatively false because of the Section 7 exponent error, and the H^{2s} normalization in both counterexample theorems needs to be redone. I would ask the authors to restate the theorems with the corrected ranges and to make the heuristic nature of the Ifrim–Tataru optimality claim explicit, as they already partly do."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know. The gallery-wave construction is real and new: explicit Laguerre-type modes for the degenerate acoustic operator (1.13), a positive Strichartz estimate for those modes (Theorem 1.9), and a wave-packet counterexample that really does exhibit derivative loss. The heuristics about multiply reflecting geodesics and the connection to Ifrim-Tataru's threshold are worth taking seriously, at least in the (2,∞)/τ²≲|ξ′| regime. But the main counterexample theorem as stated is not supported by the paper's own computation.\n\nIn Section 7 they equate 2^{2j(d+1−d/r+1/(2κ)−d/2)} with 2^{2j(1/q+γ+1/(2κ)+1)} using 1/q+d/(2r)=d/2−γ. Substituting d/r=d−2γ−2/q gives an exponent smaller by d/(2r). So for finite r, the divergence range in Theorem 1.5 is too large; for α between the true exponent and the claimed one, 2^{−2jα}‖∇²ψ_j‖ actually decays. The (2,∞) case is fine since d/(2r)=0. There's also a normalization problem: the theorem requires sup_j‖(ψ_j¹,∇ψ_j⁰)‖_{H^{2s}}≤1, but their own Bernstein estimate gives ≈2^{2js}, which is unbounded for s>0. That affects Theorems 1.5 and 1.7 alike. This looks fixable by rescaling and restating the admissible α range, but as written the statements are not correct.\n\nThe uniform lower bound on B(μ,s) in Lemma 4.2 is asserted rather than proved; I think that one is minor, because B is continuous and B(μ,0)>0 on a compact μ-range, so compactness gives a strip. Still, it should be written down.\n\nWho gets value: anyone working on physical vacuum Euler or Strichartz estimates for degenerate wave equations. The gallery-mode construction and the counterexample idea are a useful starting point, and the paper is honest about the literature. But the main theorem needs a major revision before the precise ranges can be trusted. Recommendation: send it to peer review with instructions to verify the exponent algebra and the rescaling; desk rejection would throw away a genuinely new construction.","headline":"Genuinely new gallery-wave construction for the physical-vacuum acoustic operator, but the main counterexample theorem has a concrete exponent error and a normalization gap; worth a serious referee with major revision.","tokens_in":29148,"tokens_out":6619,"would_cite":false,"duration_ms":58519,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q75","35L10","35Q35","35P05","35L81"],"pacs":[],"model":"deepseek-v4-flash","headline":"Gallery waves break Strichartz bounds for the vacuum Euler equation","keywords":["compressible Euler equations","physical vacuum","free boundary problem","Strichartz estimates","gallery modes","whispering gallery waves","derivative loss","acoustic wave equation"],"falsifier":"Evaluate the explicit profile $B(\\mu,s)=e^{-s}L_{(\\mu/\\kappa-1/\\kappa)/2}^{1/\\kappa-1}(2s)$ numerically on a strip $s\\in[a,b]$ with $\\mu$ close to $1$: a single zero there would invalidate the norm equivalence in Lemma 4.2. Alternatively, for one admissible triple and a few large $j$, compute the ratio $\\|\\nabla^2\\psi_j\\|_{L^q_tL^r_x([0,1]\\times\\Omega)}/\\|(\\psi_j^1,\\nabla_x\\psi_j^0)\\|_{H^{2s}}$; if the ratio stays bounded as $j$ grows, the claimed derivative loss is absent.","tokens_in":27989,"feed_emoji":"🌊","tokens_out":12370,"duration_ms":109277,"temperature":0.7,"pith_summary":"This paper asks what dispersive space-time estimates, known as Strichartz estimates, are possible for the linearized acoustic wave equation satisfied by the velocity potential of an irrotational compressible Euler gas in a physical vacuum, in the model case of the upper half-space with density variable $r=x_d$ and zero background velocity. Its central claim is that the natural Strichartz estimates fail: for every admissible pair of exponents $(q,r)$ and any regularity $s$ below a critical line, there are frequency-localized solutions with normalized initial data in $H^{2s}$ whose second derivatives have $L^q_tL^r_x$ norms over the time interval $[0,1]$ that diverge as the frequency tends to infinity. The obstruction comes from gallery waves, solutions highly concentrated in tangential frequency that propagate along acoustic geodesics reflecting repeatedly off the vacuum boundary and that spend a positive fraction of their time in a thin boundary layer. If correct, this means any Strichartz estimate for this equation must lose derivatives, and it suggests that the low-regularity well-posedness threshold established in [40] may be optimal in the frequency regime $\\tau^2\\lesssim|\\xi'|$. The paper presents what it identifies as the first counterexamples of this kind for the irrotational compressible Euler equations in a physical vacuum.","feed_headline":"Gallery waves break Strichartz bounds for the vacuum Euler equation","feed_subtitle":"Waves bouncing off the vacuum boundary force a derivative loss, so the known regularity threshold may be optimal.","key_machinery":"The central object is the family of whispering gallery modes $B(\\mu,|\\xi'|x_d)$, explicit eigenfunctions of the transverse operator $(\\kappa x_d\\partial_d^2+\\partial_d)B=(\\kappa x_d|\\xi'|^2-\\mu|\\xi'|)B$; they are built from generalized Laguerre polynomials and hypergeometric functions, and the choice of solution is forced by the weighted energy space of the physical vacuum, giving $B(\\mu,0)>0$. Superposing these modes with tangential frequencies in a shell near $2^{2j}$ and adding the time oscillation $e^{it2^j}$ yields exact solutions $U_j(t,x)=2^{2j(d-1)}U_0(2^jt,2^{2j}x)$ of the wave equation. The argument then runs through two norm transfers: Lemma 4.2 converts $L^r$ norms of the mode into $L^r$ norms of its tangential profile $\\phi$ via $(2^{2j})^{-1/r}\\|u\\|_{L^r_x}\\approx\\|\\phi\\|_{L^r_{x'}}$, and the bicharacteristic computation for the Hamiltonian $H=\\kappa x_d\\xi_d^2+\\kappa x_d|\\xi'|^2-\\tau^2$ shows that geodesics hit the boundary about $2^j$ times in $[0,1]$ while lingering near $x_d\\approx 2^{-2j}$ for a positive fraction of the time. A stationary-phase dispersive estimate for the reduced half-dimensional equation $\\partial_t^2\\phi+\\mu|\\nabla_{x'}|\\phi=0$, followed by the $TT^*$ argument supplied in Lemma 5.3, gives the positive Strichartz bound for individual gallery modes in Theorem 1.9; the wave-packet superposition then turns this into the divergence, because each packet's $L^r_x$ norm scales like $(2^{2j})^{d-1-d/r}$ while its $H$-norm scales like $2^{2j(d/2-1/(2\\kappa))}$.","core_discovery":"The paper proves, on its own terms, that the equation $\\partial_t^2\\psi-\\kappa x_d\\Delta\\psi-\\partial_d\\psi=0$ on $\\{x_d>0\\}$ does not admit the Strichartz estimates one would expect from the classical wave equation. Theorem 1.5 shows that for every wave-admissible triple $(q,r,\\gamma)$ and every $s<1/q+\\gamma+1/(2\\kappa)+1$, a sequence of solutions localized at tangential frequency $2^{2j}$ and time frequency $2^j$ satisfies $\\sup_j\\|(\\psi_j^1,\\nabla_x\\psi_j^0)\\|_{H^{2s}}\\leq 1$ while $\\|\\nabla^2\\psi_j\\|_{L^q_tL^r_x([0,1]\\times\\Omega)}$ diverges at the rate $2^{2j\\alpha}$ for any $\\alpha$ below the gap. Theorem 1.7 gives the analogous statement for Euler-admissible triples, with the scaling line $2s<1/q-2\\gamma$. At the endpoint $(q,r)=(2,\\infty)$, the conclusion is that controlling $\\|\\nabla^2\\psi\\|_{L^2_tL^\\infty_x}$ would require $s>k_0+1/2$, namely more derivatives than the threshold $2k>2k_0+1$ of the Eulerian well-posedness theory in [40]. In other words, the estimates fail with a loss of derivatives, and the mechanism is a geometric one: the acoustic metric supports multiply reflecting geodesics that concentrate wave packets at the boundary for a positive proportion of time.","pith_inferences":["A natural nonlinear extension is to ask whether the multiply reflecting acoustic geodesics survive under the full Euler flow; if the gallery-wave concentration is stable, the regularity barrier would be dynamical rather than a linearization artifact.","The proof is restricted to the frequency regime $\\tau^2\\lesssim|\\xi'|$; away from it the geodesic geometry changes, so the threshold in [40] might still be improvable there.","Because the modes are explicit Laguerre-type functions, the derivative loss is directly checkable by computation for moderate $j$, which could make the failure visible before a full proof is read.","The same two-scale wave-packet construction could be adapted to other vacuum decay rates or curved free boundaries; whether the loss persists should be governed by whether the reflecting geodesics remain periodic."],"forward_implications":["No classical Strichartz estimate can hold for $\\partial_t^2\\psi-\\kappa x_d\\Delta\\psi-\\partial_d\\psi=0$ without a loss of derivatives; the loss is measured by the gap between $s$ and $1/q+\\gamma+1/(2\\kappa)+1$ (or between $2s$ and $1/q-2\\gamma$ for Euler-admissible triples).","The low-regularity well-posedness threshold of the Eulerian theory in [40] cannot be improved by Strichartz-type estimates in the frequency regime $\\tau^2\\lesssim|\\xi'|$, because the endpoint control of $\\nabla^2\\psi$ in $L^2_tL^\\infty_x$ would require more derivatives than that threshold.","Individual gallery modes do satisfy a Strichartz bound with a quantified loss (Theorem 1.9); the divergence in Theorems 1.5 and 1.7 is caused by coherent superpositions of many modes, not by a single mode.","The wave-packet construction gives explicit frequency-localized solutions that can be used to test any proposed dispersive estimate for the vacuum acoustic equation, including nonlinear and higher-dimensional variants."],"supporting_citations":[{"why":"Defines the weighted $H^{2k}$ energy spaces, the critical exponent $2k_0=d+1+1/\\kappa$, and the well-posedness threshold $2k>2k_0+1$ that the counterexamples are built to test.","marker":"[40]"},{"why":"Supplies the gallery-wave counterexample strategy for strictly convex domains, including the dispersive-to-Strichartz lemma reused here as Lemma 5.3.","marker":"[41]"},{"why":"Supplies the stationary-phase asymptotic lemma used to derive the dispersive decay for the reduced half-dimensional equation.","marker":"[74]"}],"fun_headline_variants":["Whispering gallery waves break Euler bounds","Gallery reflections doom Strichartz estimates","Vacuum boundary waves defeat Strichartz estimates","Derivative loss for Strichartz in vacuum Euler","Whispering gallery modes break Strichartz"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction assumes that the explicit gallery profile $B(\\mu,s)$ stays bounded away from zero on a fixed strip of normal heights $s\\in[a,b]$ for all frequencies $\\mu$ near $1$; the paper verifies $B(\\mu,0)>0$ and continuity but does not prove this uniform strip bound.","fun_headline_variants_meta":{"raw":{"variants":["Whispering gallery waves break Euler bounds","Gallery reflections doom Strichartz estimates","Vacuum boundary waves defeat Strichartz estimates","Derivative loss for Strichartz in vacuum Euler","Whispering gallery modes break Strichartz"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000333,"raw_usage":{"total_tokens":1947,"prompt_tokens":1137,"completion_tokens":810,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":753,"completion_tokens_details":{"reasoning_tokens":741}},"tokens_in":753,"tokens_out":810,"duration_ms":6882,"temperature":1.0,"reasoning_tokens":741,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:33:06.342941+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the explicit profile $B(\\mu,s)=e^{-s}L_{(\\mu/\\kappa-1/\\kappa)/2}^{1/\\kappa-1}(2s)$ numerically on a strip $s\\in[a,b]$ with $\\mu$ close to $1$: a single zero there would invalidate the norm equivalence in Lemma 4.2. Alternatively, for one admissible triple and a few large $j$, compute the ratio $\\|\\nabla^2\\psi_j\\|_{L^q_tL^r_x([0,1]\\times\\Omega)}/\\|(\\psi_j^1,\\nabla_x\\psi_j^0)\\|_{H^{2s}}$; if the ratio stays bounded as $j$ grows, the claimed derivative loss is absent.","supporting_citations":[{"cited_title":"The compressible Euler equation s in a physical vacuum: A com- prehensive Eulerian approach","cited_arxiv_id":null,"evidence_quote":"Defines the weighted $H^{2k}$ energy spaces, the critical exponent $2k_0=d+1+1/\\kappa$, and the well-posedness threshold $2k>2k_0+1$ that the counterexamples are built to test."},{"cited_title":"Counterexamples to Strichartz estimates for the wave equation in domains","cited_arxiv_id":null,"evidence_quote":"Supplies the gallery-wave counterexample strategy for strictly convex domains, including the dispersive-to-Strichartz lemma reused here as Lemma 5.3."},{"cited_title":"Lecture notes 8 for 247b","cited_arxiv_id":null,"evidence_quote":"Supplies the stationary-phase asymptotic lemma used to derive the dispersive decay for the reduced half-dimensional equation."}],"review_version":1}