{"id":"0523122f-5049-486f-815a-29fee8301f67","arxiv_id":"2502.03151","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The operator (1+L_A)^{-γ/2} e^{it√L_A} is bounded on Lp(R^2) for all 1|1/p-1/2|, with norm growing like (1+t)^γ.","lead":"The paper proves sharp Lp bounds for the 2D wave equation with a scaling-critical magnetic potential. The result completes the Lp theory for these magnetic wave propagators and improves the previously known regularity range to the optimal one.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proof of Theorem 1.1 depends on the Gaussian heat-kernel bound (2.21) from [6], which is cited but neither reproved nor have its hypotheses restated; if it fails for the full class A in W^{1,∞}(S^1), Lemma 2.2 and the Section 4 multiplier reduction collapse.","rationale":"I read the full manuscript in good faith and checked the main structural steps: the functional calculus of Section 2.2, the kernel construction of Proposition 3.1, the pointwise bounds of Proposition 3.2, the L^p argument in Section 3.3, and the Bessel-asymptotic reduction in Section 4. I found no internal contradiction in the kernel estimates: the triple Bessel integral identities are applied in their stated ranges, the Poisson summation factors are consistent with the angular eigenfunctions, and the final L^p bounds follow from the displayed integral estimates (3.32)-(3.33). The Stein interpolation for ell ≤ 1/4 is sketched very briefly, but it is a standard analytic-family interpolation between the trivial L^2 point and the proved region ell > 1/4, and I do not see a concrete obstacle to it. The dependence on the Gaussian heat-kernel bound (2.21) is, however, genuinely load-bearing: Lemma 2.2 and all of Section 4 rest on it, and the current paper only cites the prior result without reproducing the hypotheses or proof. This is exactly the reader's identified weakest assumption, and I agree with that assessment. If [6] indeed proves the general uniform Gaussian bound, the central argument holds; otherwise the proof of Theorem 1.1 is incomplete. I recommend keeping the reader's ACCEPT verdict because the concern is external and the reader has already priced it in as a moderate-confidence risk; the proposed check would settle it definitively.","tokens_in":25727,"tokens_out":39061,"duration_ms":330477,"concrete_test":"Consult [6, Prop. 3.1 and 3.2] and verify three points: (i) the class of potentials treated there is exactly A in W^{1,infty}(S^1) with A(x̂) · x̂ = 0, not only the Aharonov-Bohm example (1.4); (ii) the estimate (2.21) holds for every t>0 and all x,y in R^2 with an implicit constant independent of t, x, and y; (iii) the proof does not secretly require |x|, |y| bounded away from 0 or t bounded away from 0. As a numerical cross-check, for the Aharonov-Bohm potential with alpha in (0,1), compute the heat kernel via the eigenfunction expansion of Section 2.2 on a grid of t in [10^{-3}, 10^3] and radii r1, r2 in [10^{-3}, 10^3] with |x-y| varying, and compare the ratio |e^{-tL_A}(x,y)| / [t^{-1} exp(-|x-y|^2/(4t))] to a uniform constant. If the ratio is unbounded or the general-A statement fails, Lemma 2.2 and the proof of Theorem 1.1 have a genuine gap.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The entire spectral-multiplier machinery of Section 2.3 is derived from the Gaussian upper bound (2.21), |e^{-tL_A}(x,y)| ≲ t^{-1} exp(-|x-y|^2/(4t)), quoted from the authors' earlier paper [6, Prop. 3.1, 3.2]. Lemma 2.2 uses this bound to obtain the imaginary-power estimates (2.25)-(2.26), the Mikhlin-Hörmander multiplier theorem (2.28), and the decaying-multiplier estimate (2.30). In Section 4, every reduction step for Theorem 1.1 invokes one of these consequences: the factorization through ((1+t^2L_A)/(1+L_A))^ell uses (2.28), the control of the remainder M_ell uses (2.28), and the control of the Bessel remainder N(ell, t sqrt(L_A)) uses (2.30). Thus, if (2.21) does not hold uniformly for all t>0 and all x,y in R^2 for the general class A in W^{1,infty}(S^1) satisfying (1.3), then Theorem 1.1 is not established. The operator L_A has an inverse-square-type singularity on each angular mode, so the Gaussian bound is a strong, dimension-sensitive statement; the paper gives neither a proof nor a statement of the precise hypotheses under which [6] establishes it, nor does it indicate whether the proof is restricted to the Aharonov-Bohm potential (1.4). This is the single most load-bearing external input, and it is not independently verified here.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies L^p estimates for the wave equation associated with the scaling-critical magnetic Schrödinger operator L_A = (i∇ + A(x̂)/|x|)^2 on R^2, where A ∈ W^{1,∞}(S^1; R^2) satisfies the transversality condition A·x̂ = 0. The main result, Theorem 1.1, states that for γ > |1/p - 1/2| the operator (1+L_A)^{-γ/2} e^{it√L_A} is bounded on L^p(R^2) with norm O((1+t)^γ), matching the sharp Euclidean threshold. The proof constructs the kernel of an analytic family f_{w,t}(L_A) of Bessel multipliers, proves pointwise kernel estimates (Propositions 3.1 and 3.2), and then uses these bounds together with spectral multiplier theorems derived from a Gaussian heat kernel estimate to control the half-wave multiplier. The paper also proves Theorem 1.2, an L^p bound for the sine propagator sin(t√L_A)/√L_A, by quoting kernel estimates from the authors' earlier work. The central novelty is the explicit kernel construction and the pointwise estimates for the analytic family, which are carried out in detail in Section 3.","tokens_in":26063,"tokens_out":28599,"duration_ms":219766,"significance":"If correct, Theorem 1.1 is a sharp result: it shows that for a class of singular magnetic potentials with critical scaling, the L^p regularity threshold for the wave propagator is the same as for the Euclidean Laplacian. The proof is genuinely two-dimensional and avoids the parametrix methods used for cones, instead deriving exact kernel formulas through Bessel-function identities and Poisson summation. The paper is honest about its external inputs: the Gaussian heat kernel bound (2.21) from [6] and spectral multiplier results from [29] are cited rather than reproved. The manuscript contains no fitted parameters and the main new estimates are explicit and checkable. The result is likely to be of interest to researchers in harmonic analysis, spectral multipliers, and dispersive equations with singular potentials.","major_comments":[],"minor_comments":[{"comment":"The Gaussian heat kernel bound (2.21) is the sole input for Lemma 2.2 and hence for the spectral multiplier reductions in Section 4. The paper cites [6, Prop. 3.1, 3.2] but does not restate the hypotheses under which it holds. Please add a precise statement that (2.21) holds for the class A∈W^{1,∞}(S^1) satisfying (1.3) and indicate whether the proof in [6] covers the general class or only the Aharonov-Bohm potential (1.4).","section":"Section 2.3, Eq. (2.21)"},{"comment":"The Stein interpolation step for ℓ≤1/4 is only sketched. Please specify the analytic family of operators, the endpoint estimates at A=(1/2,0) and on the boundary ℓ=1/4, and the resulting interpolation inequality that yields (1.9) for the remaining range.","section":"Section 4, paragraph on Ω1"},{"comment":"The sentence 'Combining with (4.1), this yields Theorem 1.1 for region Ω3' is imprecise: applying (4.1) with ℓ=3/8 gives growth (1+t)^{3/4}, which is not bounded by (1+t)^ℓ for all ℓ∈[1/2,3/4). Please clarify that one chooses an auxiliary ℓ0∈(1/4,1/2) with 2ℓ0≤ℓ (or uses the Ω1 interpolation for intermediate ℓ).","section":"Section 4, reduction to (4.1)"},{"comment":"The equality in (3.19) is not literally correct for complex w because cosh(β2)-cosh τ is negative for τ>β2; the displayed identity should involve an absolute value (or a phase factor) on the right, since only the modulus is used in the subsequent estimates.","section":"Section 3.2, Eq. (3.19)"},{"comment":"The identity cosh(τ)-cos(θ̄+π)=sinh^2(τ/2)+sin^2((θ̄+π)/2) is missing a factor of 2 on the right-hand side; the correct identity is cosh τ - cos φ = 2sinh^2(τ/2)+2sin^2(φ/2). The omission does not affect the bounds, but the formula should be corrected.","section":"Section 3.2, Eq. (3.24)"},{"comment":"Reference [13] lists 'Josaroop' while the cited author's name is 'Jotsaroop'; reference [42] similarly misprints the author's name. Please correct these.","section":"References"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript's central claim is sound and the proof is quite detailed. The main external dependence is the Gaussian heat kernel bound (2.21) from the authors' earlier paper [6]; the paper would be strengthened by including the exact statement of that bound and its hypotheses. The result is a nice contribution and fits the journal's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper proves the right sharp Lp bound for the 2D magnetic wave propagator, and the proof is built on an explicit kernel construction rather than black-box multiplier arguments. I think it deserves a serious referee.\n\nWhat is new: Theorem 1.1 is the first sharp range gamma>|1/p-1/2| for this class of scaling-critical magnetic operators. The improvement over gamma>2|1/p-1/2| is not cosmetic; it matches the Euclidean wave propagator threshold. The main work is the construction of the kernel of the analytic family f_{w,t}(L_A), using Macdonald identities and Poisson summation. Proposition 3.2 gives explicit pointwise bounds for the geometric and diffractive terms, and the cancellations in the diffractive term are checked rather than merely asserted. That is the paper's real contribution.\n\nWhat I would want checked: the proof rests on the Gaussian heat kernel bound (2.21) quoted from [6]. That is a strong statement for this operator class, and it is genuinely load-bearing: Lemma 2.2 and the whole Section 4 reduction go through it. But it is a published result in [6] for the same operators, so quoting it is legitimate. The stress tester's worry would only land if [6] had proved it under narrower hypotheses, say only for the Aharonov-Bohm potential, and nothing in this text suggests that. A referee should verify that the statement in [6] covers A in W^{1,infty}(S^1) with the transversality condition. I'd also ask the authors to expand the Stein interpolation step for ell <= 1/4; it is sketched in a few lines in Section 4. The proof is organized around ell in (1/4,1/2), and the reduction from the general gamma>|1/p-1/2| to that range is compressed. I don't see a circularity problem: [6] is independent of Theorem 1.1. The sine-propagator theorem, Theorem 1.2, is essentially a repackaging of [6, Props 4.1-4.2], as the authors acknowledge; that reduces the novelty of that part, but the main theorem carries the paper.\n\nBottom line: good paper, no fatal flaw visible. The kernel estimates are explicit and the proof is formally grounded. For people working on spectral multipliers, wave equations with singular potentials, and magnetic Schrodinger operators, this is worth careful refereeing. I would send it to review.","headline":"The paper likely proves the sharp Lp range for the 2D magnetic wave propagator, and the kernel-construction proof deserves a serious referee, with the Gaussian heat-kernel input from [6] and the sketched interpolation step flagged for checking.","tokens_in":26628,"tokens_out":3163,"would_cite":true,"duration_ms":29534,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35L05","42B15","35P05"],"pacs":[],"model":"deepseek-v4-flash","headline":"For magnetic waves in 2D, the Euclidean L^p smoothing threshold survives.","keywords":["Lp-estimates","scaling-critical magnetic field","Aharonov-Bohm potential","wave equation","magnetic Schrödinger operator","spectral multipliers","Bessel functions"],"falsifier":"Take the Aharonov-Bohm potential (1.4) with a non-integral flux \\$\\alpha$ and numerically evaluate the claimed kernel bound (3.13) at t just above |x-y| for 1/2<\\operatorname{Re}w<1; any violation of the pointwise inequality would refute Theorem 1.3 and hence Theorem 1.1. Alternatively, check whether the norm inequality in Theorem 1.1 remains true at the endpoint \\gamma=|1/p-1/2| by testing specific p and t; the paper leaves this endpoint open, so a concrete counterexample there would settle the sharpness question.","tokens_in":25514,"feed_emoji":"🌊","tokens_out":6179,"duration_ms":54724,"temperature":0.7,"pith_summary":"The paper proves that the magnetic wave propagator $e^{{it\\sqrt{\\mathcal{L}}$_\\mathbf{A}}}, after smoothing by (1+\\mathcal{L}_\\mathbf{A})^{-\\gamma/2}, is bounded on L^p(\\mathbb{R}^2) whenever \\gamma>|1/p-1/2|, exactly the threshold for the ordinary Laplacian in two dimensions. The operator \\mathcal{L}_\\mathbf{A} is a Schr\\\"odinger operator with a scaling-critical magnetic potential, such as the Aharonov-Bohm field, which is singular at the origin. If the result is right, the magnetic singularity does not force any extra Sobolev regularity for the L^p wave flow, and the growth in time is only the mild factor (1+t)^\\gamma. As a corollary, the sine propagator \\sin(t\\sqrt{\\mathcal{L}_\\mathbf{A}})/\\sqrt{\\mathcal{L}_\\mathbf{A}} is L^p bounded at each fixed time with linear-in-t constant for every 1\\le p\\le\\infty.","feed_headline":"Critical magnetic waves match Euclidean Lp smoothing","feed_subtitle":"In 2D, a scaling-critical magnetic potential does not worsen the Sobolev regularity needed for Lp bounds on the wave flow.","key_machinery":"The machine is the analytic operator family f_{w,t}(\\mathcal{L}_\\mathbf{A}) = (\\pi/2)^{1/2}(t\\sqrt{\\mathcal{L}_\\mathbf{A}})^{w-1}J_{1-w}(t\\sqrt{\\mathcal{L}_\\mathbf{A}}), with w=\\epsilon+iy and 1/2<\\epsilon<1. Its kernel is computed explicitly using the Macdonald triple-Bessel integral, the angular eigenfunctions of \\mathcal{L}_\\mathbf{A}, and Poisson summation; the result is a geometric term G plus a diffractive term D satisfying the pointwise bounds |G|\\lesssim $t^{{2(\\operatorname{Re}}$w-1)}($t^{2}$-|x-y|^2)^{-\\operatorname{Re}w} and |D|\\lesssim $t^{{2(\\operatorname{Re}}$w-1)}($t^{2}$-(r_1+r_2)^2)^{-\\operatorname{Re}w}. These bounds are what convert the spectral multiplier problem into a convolution estimate; the family is designed so that the wave multiplier \\psi(s)$s^{{-2\\ell}}$$e^{{is}}$ can be written as a linear combination of such Bessel multipliers plus a well-controlled remainder, using Bessel asymptotics and the identity relating \\cos(s-\\pi(\\ell+i)) and \\cos(s-\\pi\\ell) to $e^{{is}}$.","core_discovery":"The central claim, Theorem 1.1, is that for every \\gamma>0 and 1r_1+r_2, with pointwise bounds (3.13)-(3.14). These kernel bounds transfer to L^p estimates by Young's inequality, and the main theorem follows by decomposing the multiplier m(\\ell,s)=(1+$s^{2}$)^{-\\ell}$e^{{is}}$ into a Bessel term controlled by the family and a remainder controlled by standard spectral multipliers.","pith_inferences":["I would expect the same threshold to hold for Klein-Gordon propagators e^{it\\sqrt{\\mathcal{L}_\\mathbf{A}+1}} with appropriate smoothing, since the kernel construction here is spectral rather than purely hyperbolic; the paper cites earlier Klein-Gordon Strichartz work for the same operator, but does not state the L^p smoothness analogue.","The proof's angular-mode summation suggests that the Aharonov-Bohm flux \\alpha only enters through phase factors e^{\\pm i\\alpha(\\theta_1-\\theta_2)} and indicator functions, so the L^p bounds should be uniform in \\alpha over compact intervals; this uniformity is not explicitly claimed.","A numerical check of the pointwise kernel at t near |x-y| could test the endpoint sharpness: if (3.13) fails at \\operatorname{Re}w=1/2, the endpoint \\gamma=|1/p-1/2| is genuinely excluded."],"forward_implications":["If Theorem 1.1 holds, the L^p regularity threshold for the wave equation with a scaling-critical magnetic potential in \\mathbb{R}^2 is identical to the Euclidean threshold |1/p-1/2|, up to the endpoint.","The fixed-time sine propagator is L^p bounded for every 1\\le p\\le\\infty with constant C|t|, matching the Euclidean behavior in two space dimensions.","The explicit kernel of f_{w,t}(\\mathcal{L}_\\mathbf{A}) gives pointwise control of the magnetic wave propagator, which can be used to prove further dispersive and spectral multiplier estimates for \\mathcal{L}_\\mathbf{A}.","The time-growth factor (1+t)^\\gamma in the main estimate is uniform in t>0 and is the natural analogue of the Euclidean result, not a worsened power forced by the singularity."],"supporting_citations":[{"why":"Supplies the Gaussian heat-kernel upper bound (2.21) and the sine-propagator kernel used in Theorem 1.2 and Lemma 5.1.","marker":"[6]"},{"why":"Provides the method of approximating the half-wave multiplier by the analytic Bessel family f_{w,t} and the spectral multiplier decomposition in Proposition 4.1.","marker":"[15]"},{"why":"Gives the triple Bessel integral formula (2.18), the basis for the explicit kernel of f_{w,t}(\\mathcal{L}_\\mathbf{A}).","marker":"[17]"},{"why":"Supplies the sharp Euclidean L^p range for wave propagators and the Legendre-function integral identities used in the kernel construction.","marker":"[18]"},{"why":"Provides the Bessel function asymptotics and integral representations used to split the wave multiplier into Bessel and remainder terms.","marker":"[44]"},{"why":"Provides the spectral multiplier and imaginary-power theorems quoted in Lemma 2.2, which control the remainder terms.","marker":"[29]"}],"fun_headline_variants":["Magnetic field no penalty for 2D wave Lp smoothing","Scaling-critical magnetic potential preserves wave smoothing","2D wave smoothing unaffected by critical magnetic field","Sharp Lp estimates for wave equation with magnetic field"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof imports a Gaussian upper bound for the heat kernel of \\mathcal{L}_\\mathbf{A} from an earlier paper; if that bound were false, the spectral multiplier estimates in Lemma 2.2, and with them the proof of Theorem 1.1, would fall apart.","fun_headline_variants_meta":{"raw":{"variants":["Magnetic field no penalty for 2D wave Lp smoothing","Scaling-critical magnetic potential preserves wave smoothing","2D wave smoothing unaffected by critical magnetic field","Sharp Lp estimates for wave equation with magnetic field"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000433,"raw_usage":{"total_tokens":2227,"prompt_tokens":988,"completion_tokens":1239,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":604,"completion_tokens_details":{"reasoning_tokens":1176}},"tokens_in":604,"tokens_out":1239,"duration_ms":9116,"temperature":1.0,"reasoning_tokens":1176,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T05:46:43.185723+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the Aharonov-Bohm potential (1.4) with a non-integral flux \\$\\alpha$ and numerically evaluate the claimed kernel bound (3.13) at t just above |x-y| for 1/2<\\operatorname{Re}w<1; any violation of the pointwise inequality would refute Theorem 1.3 and hence Theorem 1.1. Alternatively, check whether the norm inequality in Theorem 1.1 remains true at the endpoint \\gamma=|1/p-1/2| by testing specific p and t; the paper leaves this endpoint open, so a concrete counterexample there would settle the sharpness question.","supporting_citations":[{"cited_title":"Fanelli, J","cited_arxiv_id":null,"evidence_quote":"Supplies the Gaussian heat-kernel upper bound (2.21) and the sine-propagator kernel used in Theorem 1.2 and Lemma 5.1."},{"cited_title":"Li, Estimations Lp de l’´ equation des ondes sur les va ri´ et´ es ` a singularit´ e conique,Math","cited_arxiv_id":null,"evidence_quote":"Provides the method of approximating the half-wave multiplier by the analytic Bessel family f_{w,t} and the spectral multiplier decomposition in Proposition 4.1."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the triple Bessel integral formula (2.18), the basis for the explicit kernel of f_{w,t}(\\mathcal{L}_\\mathbf{A})."},{"cited_title":"Miyachi, On some estimates for the wave equation in Lp and H p, Journal of the Faculty of Science, the University of Tokyo","cited_arxiv_id":null,"evidence_quote":"Supplies the sharp Euclidean L^p range for wave propagators and the Legendre-function integral identities used in the kernel construction."},{"cited_title":"Taylor, Partial Differential Equations, vol","cited_arxiv_id":null,"evidence_quote":"Provides the Bessel function asymptotics and integral representations used to split the wave multiplier into Bessel and remainder terms."},{"cited_title":"I n: dans la serie London Math","cited_arxiv_id":null,"evidence_quote":"Provides the spectral multiplier and imaginary-power theorems quoted in Lemma 2.2, which control the remainder terms."}],"review_version":1}