{"id":"f4accda8-58b6-4b5a-a9fd-ffa2f9aa7227","arxiv_id":"2411.11787","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"First proof of L1 to L∞ decay at rate t^{-3/2} for Schrödinger propagators with nontrivial short-range magnetic potentials in three dimensions.","lead":"This paper proves that solutions of the three-dimensional Schrödinger equation with a magnetic potential decay at the same rate as the free equation, with a t^{-3/2} L1 to L∞ bound. It is the first such dispersive estimate for nonzero magnetic potentials, a gap noted in prior work.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The decisive step is Proposition 9's sequence of cancellations; the proof delegates several 'entirely analogous' estimates, and a sign error there would break the t^{-3/2} bound.","rationale":"The reader's weakest_assumption is Assumption 1, but their rationale also notes that several parts are delegated as 'entirely analogous.' I find the delegated estimates in Proposition 9 and Proposition 12 to be the more load-bearing concern: they are the actual mechanism producing the resolvent bounds, and a sign or factor error there would invalidate the theorem even under Assumption 1. The H_-1 threshold condition is a scope restriction, not an internal inconsistency, and the authors explicitly flag it as possibly superfluous. My recommended verdict remains CONDITIONAL, matching the reader: the theorem is plausible and likely correct, but the proof should expand or verify the delegated cancellation identities before acceptance. The proposed concrete test would settle whether the cancellations are correct.","tokens_in":38971,"tokens_out":12889,"duration_ms":122012,"concrete_test":"Independently re-derive the cancellation identities (41)+(38), (46)+(40), and (48)+(45) in the proof of Proposition 9, and the analogous reduction of (56) in Proposition 12, using a computer algebra system or by hand. If any identity is off by a sign or a constant factor, the U(L∞) ∩ U(K*_log) bound for T1 is false and the main theorem does not follow.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The main theorem rests on Proposition 13, which needs the resolvent bounds (18)–(19) and the invertibility of I - T(λ). The core new estimate is Proposition 9, where T1(λ) is shown to lie in U(L∞) ∩ U(K*_log) via a long integral computation. The proof reduces the singular terms to expressions (31), then invokes a chain of cancellations: (41)+(38), (46)+(40), and (48)+(45). These identities are asserted after replacing derivatives by averages and are not fully derived; Proposition 12 then states that the analogous estimate for (56) is 'entirely analogous' to the delicate cancellations in Proposition 9. If any of these identities has a wrong sign or a missing factor, the bound for T1 fails, and with it the Wiener inversion step and the final t^{-3/2} decay. This is a genuine correctness risk in the central argument. By contrast, Assumption 1 (0 regular for both H and H_-1) is a clearly stated hypothesis; it restricts scope but does not threaten correctness under the theorem's assumptions.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves that for a self-adjoint magnetic Schrödinger Hamiltonian H = -Δ + U on R^3, under the assumptions A ∈ X0, V ∈ Y0 and with 0 regular for both H and H_-1 = -Δ - U, the continuous-spectrum propagator satisfies ||e^{itH} P_ac f||_{L∞} ≲ |t|^{-3/2} ||f||_{L1}. The proof follows the Beceanu–Goldberg approach: it represents the perturbed resolvent through the resolvent identity, proves bilinear estimates showing that the relevant operator T lies in a Wiener algebra Û(L∞) ∩ Û(K*_log), applies Wiener's theorem to invert I - T, and derives the decay estimate from the resulting resolvent bounds. The authors also state related wave-equation estimates and prove finiteness of negative eigenvalues in an appendix.","tokens_in":39153,"tokens_out":10541,"duration_ms":102288,"significance":"If correct, this is a substantial result: it provides the first L1 → L∞ dispersive decay estimate for Schrödinger equations with nonzero magnetic potentials, resolving an open problem quoted from [ErGoSc2]. The proof is genuinely parameter-free in the sense that no decay rate is fitted and no spectral parameter is tuned; the main hypotheses, including the threshold regularity assumption, are stated explicitly. The paper also gives credit to prior techniques and includes quite detailed integral estimates in Proposition 9. The main weakness is that some of the most delicate cancellations, and the estimates for ∂_λ T on which the final decay rate depends, are only sketched or described as 'entirely analogous' to earlier computations; these are load-bearing and need to be verified or expanded.","major_comments":[{"comment":"The central cancellations in Proposition 9 are asserted rather than fully derived. After reducing the singular terms to (31), the proof replaces ∂R A1 by its line average and then claims the cancellations (41)+(38), (46)+(40), and (48)+(45). These identities are load-bearing: they are what makes T1 belong to Û(L∞) ∩ Û(K*_log), which in turn feeds into the invertibility of I - T and the final t^{-3/2} bound. A missing factor or sign error in any of these identities would break the argument. Please provide the full endpoint evaluations, the justification for the 'spare copy' of (40), and the exact accounting of all terms in (31), (34), and (35).","section":"§3.1, Proposition 9, around (31)–(48)"},{"comment":"Proposition 12 is essential because (19), the bound ∂λR ∈ Û(L1, L∞), is used directly in the proof of Theorem 1. However, the proof repeatedly states that the required estimates are 'entirely analogous' to those for (34) in Proposition 9, including 'the delicate cancellations that take place there.' Since Proposition 9 itself is only sketched at exactly those delicate points, the verification of Proposition 12 is not complete as written. Please either write out the analogous estimates or state and prove a separate lemma that covers the ∂λ terms with all necessary cancellations.","section":"§3.1, Proposition 12, after (56) and in the T12 analysis"},{"comment":"The notation for the magnetic term is ambiguous and potentially inconsistent. The abstract defines H = -Δ + i(A∇ + ∇A) + V, while Theorem 1 states H = -Δ + U = -Δ + ∇A + V. Later, in (14), the phrase '∇A here means the composition of operators' is introduced, but this is not reflected in the statement of Theorem 1. As written, a reader could interpret ∇A as the gradient of the vector field A, which would not give a self-adjoint operator. Please introduce a consistent operator notation, for example defining U explicitly as a first-order differential operator with the chosen coefficients, and state the self-adjointness condition explicitly in Theorem 1.","section":"Theorem 1 and §2.2"},{"comment":"The proof of invertibility of I - T(λ) for λ ≠ 0 relies on the absence of embedded eigenvalues, citing [KocTat] under 'an assumption weaker than A ∈ L3, V ∈ L3/2.' Since the Hamiltonian here contains a magnetic first-order term, it is not immediately clear that the Carleman-estimate result in [KocTat] applies verbatim. Please state the precise form of the result being cited, or give a short reduction of the magnetic operator to the setting of [KocTat]; this is needed for the λ ≠ 0 part of the Wiener inversion step.","section":"§3.2, Proposition 13, λ ≠ 0 step"}],"minor_comments":[{"comment":"In (8), the norm for K_{2,log2} is written as ||f||_{K_{2,log}}, which is the same symbol used for the norm of K_{2,log} in (7); please use ||f||_{K_{2,log2}} for the log-squared space.","section":"§2.1, definitions (7)–(8)"},{"comment":"The expression 'H = -Δ + ∇A + V' should be rewritten with the operator notation defined in §2.2, or the reader is left with an apparent inconsistency with H = -Δ + i(A∇ + ∇A) + V from the abstract.","section":"§1.2, Theorem 1"},{"comment":"The statement of Proposition 11 uses A ∈ K, but the proof and the surrounding discussion sometimes write A#; please make the use of A and A# uniform in the statement and proof.","section":"§3.1, Proposition 11"},{"comment":"The use of χ_{t≥0} in the functional calculus identity is not explained; a sentence clarifying the contour/sign convention would improve readability.","section":"Corollary 14"}],"recommendation":"major_revision","confidential_remarks":"The paper is ambitious and the result, if correct, is important. My main concern is verifiability: the load-bearing cancellations in Proposition 9 and the 'entirely analogous' parts of Proposition 12 are not sufficiently detailed for a journal referee to certify the proof. I recommend requesting an expanded version of those arguments before publication. The notation issue in Theorem 1 should also be fixed. I do not see evidence of circularity or fitted parameters; the assumptions are stated honestly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline is simple: this paper proves the first L1 to L∞ dispersive bound for Schrödinger operators with a nonzero magnetic potential, at the free decay rate |t|^{-3/2}. That is a real step forward and closes an open problem quoted from Erdoğan–Goldberg–Schlag. The novelty is not just the theorem but the machinery: the Beceanu–Goldberg Wiener algebra framework has been extended to handle the first-order derivative term, which is the genuine technical obstacle.\n\nWhat the paper does well: the structure is honest and readable. The threshold assumption is stated clearly, the authors admit that the H_-1 condition may be superfluous, and they note that a more careful proof would only need V in Klog. Proposition 9, the heart of the paper, is written out at length with explicit cancellations; Proposition 13's invertibility argument uses Agmon-type bootstrapping and is careful about the λ = 0 case. The paper also honestly delegates some estimates as \"entirely analogous,\" which is normal in this area but carries real risk.\n\nThe soft spots, in proportion: first, there is a notation slip in Theorem 1 as printed: the Hamiltonian is written H = −Δ + U = −Δ + ∇A + V, missing the imaginary unit that appears everywhere else in the paper. That is confusing but fixable. Second, the central Proposition 9 relies on a chain of cancellations — (41)+(38), (46)+(40), (48)+(45) — and some steps are asserted rather than fully derived. A sign error in any of those identities would break the t^{-3/2} bound. I cannot machine-verify these integrals, and the paper would be stronger if the delegated estimates were expanded. That said, the cancellations are shown explicitly, not just claimed, and I found no internal contradiction. Third, Assumption 1 (0 regular for both H and H_-1) is a genuine restriction, but it is stated clearly and the authors flag that necessity of the H_-1 condition is open. It limits scope, not correctness.\n\nCitation pattern is fine: heavy reference to the authors' previous work is appropriate because the method is theirs, and the new step is the magnetic first-order term.\n\nWho this is for: people working on dispersive estimates, magnetic Schrödinger operators, or the Wiener algebra approach to resolvent estimates. It deserves a serious referee. The referee should spend time on Proposition 9 and check the cancellation identities carefully; if those hold, the paper is a solid contribution.","headline":"First L1-to-L∞ decay for magnetic Schrödinger propagators in 3D; the proof looks serious, but the central integrand cancellations deserve close referee scrutiny.","tokens_in":39714,"tokens_out":2337,"would_cite":true,"duration_ms":26578,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q41","35B40","47A10","47A40","47D08"],"pacs":[],"model":"deepseek-v4-flash","headline":"Magnetic Schrödinger solutions decay exactly like the free wave, at rate |t|^{-3/2} in three dimensions.","keywords":["magnetic Schrödinger equation","dispersive estimates","L1-L∞ decay","resolvent estimates","Kato spaces","Wiener theorem","algebroid","threshold regularity"],"falsifier":"Construct magnetic and electric potentials A ∈ X0, V ∈ Y0 such that 0 is a resonance or eigenvalue for H_-1 = -Δ - ∇A - V but is regular for H, and compute the L1→L∞ norm of $e^{{itH}}$P_ac; if the |t|^{-3/2} rate persists, the H_-1 condition is superfluous, and if the rate is slower or the norm grows, the condition is necessary.","tokens_in":38749,"feed_emoji":"▫️","tokens_out":3419,"duration_ms":32713,"temperature":0.7,"pith_summary":"The paper proves that, in three dimensions, the magnetic Schrödinger evolution, after projection onto the absolutely continuous spectrum, decays in time at the same L1-to-L∞ rate as the free propagator: |t|^{-3/2} times the L1 norm of the initial data. This is the first dispersive bound of this kind proved for a Hamiltonian with a nonzero magnetic potential, a case previously open. The magnetic and electric potentials may be arbitrarily large as long as they are short-range and obey the stated regularity, and the same method yields wave-equation decay estimates. The argument works by showing that a certain four-term operator built from the free resolvent and the potential is invertible in a Banach-lattice algebroid, reducing the perturbed resolvent to the free one.","feed_headline":"Magnetic quantum waves decay as fast as free ones","feed_subtitle":"First L1-to-L∞ dispersive bound for Schrödinger with a nonzero vector potential in 3D: decay at |t|^(-3/2).","key_machinery":"The central object is the algebroid U(X,Y) of operator kernels T(ρ,y,x) whose integral in ρ is a bounded operator from X to Y, together with its Fourier-transformed version Û(X,Y); composition in ρ corresponds to pointwise composition of the Fourier transforms. The proof decomposes the Kato–Birman style operator T = R0 U R0 U# into four terms T1,...,T4 and proves, in Proposition 9, that T1 ∈ Û(L∞) ∩ Û(K*_log) by integrating its kernel over ellipsoids Σ_ρ = {|x-y|+|y-z|=ρ}; there singular contributions cancel in pairs, leaving terms bounded through logarithmic Kato spaces. Wiener's theorem then converts the pointwise invertibility of I - T̂(λ) for all λ ∈ R into invertibility of I - T in the algebroid, yielding the resolvent bounds (18)-(19) that imply the L1→L∞ decay.","core_discovery":"Theorem 1 establishes the sharp dispersive estimate ||$e^{{itH}}$ P_ac f||_{L∞} ≲ |t|^{-3/2} ||f||_{L1} for every self-adjoint magnetic Schrödinger Hamiltonian H = -Δ + ∇A + V with A in X0 and V in Y0, provided 0 is neither an eigenvalue nor a resonance for H and for the sign-reversed Hamiltonian H_-1 = -Δ - U, or alternatively if the potentials are small in norm. The discovery is that the magnetic gradient term ∇A, which prevents the resolvent from acting on ordinary Kato spaces, can still be handled through a decomposition T = R0 U R0 U# into four operators, of which the hardest one, T1 = R0 ∇A R0 ∇A#, is shown to lie in the algebroid spaces Û(L∞) and Û(K*_log) via delicate cancellations in ellipsoidal coordinates. Once (I - T)^{-1} is obtained by Wiener's theorem, the perturbed resolvent inherits the free-resolvent mapping properties R(λ²) ∈ Û(K,L∞) ∩ Û(L1,K*_log) and ∂_λ R(λ²) ∈ Û(L1,L∞), which directly yields the $t^{{-3/2}}$ decay.","pith_inferences":["The four derivatives required on the magnetic potential A are likely an artifact of the proof: the structure of the estimates suggests that only ∇A ∈ K^log and ∇²A ∈ L1 should be needed, so a future refinement may weaken the hypotheses considerably.","The condition that 0 be regular for H_-1 = -Δ - U as well as for H is used only to rule out the λ = 0 case in the invertibility proof; a targeted counterexample or numerical experiment could decide whether it is genuinely necessary or merely a convenience.","The algebroid method used here could be adapted to endpoint Strichartz estimates and to threshold cases by combining it with the case-by-case analysis of zero-energy eigenstates and resonances already developed for scalar potentials."],"forward_implications":["The L1→L∞ dispersive estimate for magnetic Schrödinger equations in three dimensions is now established for arbitrarily large short-range potentials, closing a known gap in the literature.","Strichartz estimates, reversed Strichartz inequalities, and decay estimates for wave, Klein–Gordon, and related equations with short-range magnetic potentials follow from the same resolvent bounds by functional calculus.","The result shows that, apart from bound states, the magnetic Schrödinger flow spreads exactly as the free flow: the |t|^{-3/2} rate is optimal and matches the free propagator.","Together with the absence of embedded eigenvalues, the theorem implies that the only possible obstructions to free decay are threshold eigenvalues or resonances at zero energy, which are handled separately in other works."],"supporting_citations":[{"why":"Supplies the method of estimates stronger than the limiting absorption principle that the proof builds on.","marker":"[BecGol1]"},{"why":"Provides the Wiener theorem version used to pass from pointwise invertibility of I - T̂(λ) to invertibility of I - T in the algebroid.","marker":"[BecGol3]"},{"why":"Gives the absence of embedded eigenvalues in (0,∞) needed to exclude nonzero λ in the invertibility argument.","marker":"[KocTat]"},{"why":"Supplies Corollary 13 used to upgrade the vanishing of the Fourier transform on a sphere to L2 membership of the candidate eigenfunction.","marker":"[GolSch]"},{"why":"Agmon's bootstrap argument is invoked to show that the pairing ⟨U g, g⟩ is real-valued and to control eigenfunction decay.","marker":"[Agm]"},{"why":"Introduces the global Kato space K used to formulate the potential classes and the resolvent bounds.","marker":"[RodSch]"},{"why":"Provides the self-adjointness condition for H with A in the class A and the framework for more general conditions.","marker":"[IonSch]"}],"fun_headline_variants":["Magnetic potentials don't slow Schrödinger decay","Same t^{-3/2} decay for magnetic Schrödinger waves","No decay penalty for magnetic Schrödinger waves","Magnetic Schrödinger waves match free decay"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof needs 0 to be a regular point of the spectrum for both H and its sign-reversed counterpart H_-1 = -Δ - U; if 0 is an eigenvalue or resonance for either operator, the invertibility step at λ = 0 fails and the decay rate could change.","fun_headline_variants_meta":{"raw":{"variants":["Magnetic potentials don't slow Schrödinger decay","Same t^{-3/2} decay for magnetic Schrödinger waves","No decay penalty for magnetic Schrödinger waves","Magnetic Schrödinger waves match free decay"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001089,"raw_usage":{"total_tokens":4544,"prompt_tokens":935,"completion_tokens":3609,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":551,"completion_tokens_details":{"reasoning_tokens":3548}},"tokens_in":551,"tokens_out":3609,"duration_ms":26053,"temperature":1.0,"reasoning_tokens":3548,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T18:09:18.686542+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct magnetic and electric potentials A ∈ X0, V ∈ Y0 such that 0 is a resonance or eigenvalue for H_-1 = -Δ - ∇A - V but is regular for H, and compute the L1→L∞ norm of $e^{{itH}}$P_ac; if the |t|^{-3/2} rate persists, the H_-1 condition is superfluous, and if the rate is slower or the norm grows, the condition is necessary.","supporting_citations":[],"review_version":1}