{"id":"bbeacda0-dd86-4e8c-98d9-512be751ccb4","arxiv_id":"2505.07009","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For odd n ≤ 4m-1, the L^p range of the wave operators for (-Δ)^m+V is 1<p<∞ for regular or subcritical resonances, and shrinks to 1<p<p_k for stronger resonances, with p_k = n/(n-2m+k+k_c-1); the range is shown sharp.","lead":"The paper proves sharp ranges of exponents p for which the wave operators of higher order Schrödinger operators are bounded on L^p, covering all odd dimensions up to 4m-1 and all threshold resonance types. The result completes a program on L^p bounds for wave operators and yields dispersive decay estimates for the perturbed propagator.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central low-energy expansion of (M±(λ))^{-1} (Theorem 2.5) is imported from the authors' preprint [4] and only outlined in Appendix A; all Lp-range thresholds in Theorems 1.5 and 1.7 flow from it, so the headline result is conditional on that expansion being correct.","rationale":"The reader's weakest_assumption is exactly the one I consider load-bearing. The paper's main positive contribution is a detailed stationary-phase/kernel analysis in Sections 3-4, which appears internally coherent: phases, cancellation from Q_j, dyadic Hilbert-transform localization, and the loose radial estimate in (3.42) are repairable. However, none of that analysis can start without the inverse-resolvent expansion Theorem 2.5, and the appendix does not prove it: it refers to [4] for the spectral properties of D and for the eigenvalue-case factorization. Since the final p-ranges are sensitive to the exact powers λ^{2m-n-i-j} and to the remainder exponents, an error in those powers would propagate directly into Theorem 1.5 and Theorem 1.7. This is not a disagreement with consensus; it is a verification gap. Because there is no formal verification and the key expansion is imported, CONDITIONAL is the right verdict, and I see no reason to change it. I also note the paper's own statements support this reading: Appendix A says 'we outline the processes' and cites [4] for Lemma 2.6. I credit the paper for making the reduction from resolvent expansion to Lp bounds quite explicit, including the unboundedness direction via Kato-Jensen-type lower bounds; if Theorem 2.5 holds, the proofs are plausible and detailed enough to be checked. The concrete test above would either close the gap or locate the failure.","tokens_in":62517,"tokens_out":32013,"duration_ms":299544,"concrete_test":"Verify Theorem 2.5 independently for the smallest new cases: take m=3,n=1 and m=2,n=5 (and n=3,m=2 as a benchmark), choose a compactly supported generic V, compute the finite-dimensional operators Q_j, D00, D11, d in Appendix A explicitly, and check (i) the stated positivity/negativity and invertibility of D, and (ii) by numerical evaluation of M±(λ) = U+vR0^±(λ^{2m})v on the Q_j subspaces for λ=10^{-2},10^{-3},10^{-4}, that the expansion (2.36) matches the computed inverse to the predicted order with remainders satisfying (2.37)-(2.38). Alternatively, compare term-by-term with the full proof in [4, Theorem 2.7]; if the exact λ-powers of any Γ_{i,j} differ, recompute the p_k ranges in Propositions 3.5 and 3.7.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The entire low-energy analysis and the unboundedness proof rest on Theorem 2.5: the expansion (2.36) with powers λ^{2m-n-i-j}, the projections Q_j from (2.21)-(2.31), and the remainder bounds (2.37)-(2.38). Every estimate in Section 3 (Lemmas 3.3, 3.8, 3.11 and Propositions 3.5, 3.7) begins from this expansion via (3.1)-(3.5), and Section 5 uses the same expansion to isolate the principal term in (5.3)-(5.6). Appendix A is explicitly only an outline: the crucial invertibility and signature properties of D (D00 strictly positive, D11 strictly negative, d invertible and rigidly negative definite on ⊕_{j∈J''_k} Q_j L^2) are quoted from Lemma 2.6 of [4], and the eigenvalue-case invertibility of Q vG_{4m-n}v Q is similarly imported. If any one of these facts fails, or if the remainder Γ has weaker λ-powers than (2.37)-(2.38), the p-thresholds in Theorem 1.5(iii)-(iv) and the optimality claims in Theorem 1.7 would shift. I found no independent internal error in Sections 3-5; the only blemish I noticed is a repairable typo in the radial estimate (3.42) (a missing power of r∼2^h), which can be fixed by keeping the support factor and does not undermine the uniform bound. Thus the main claim is coherent but not self-contained: its correctness is exactly as secure as Theorem 2.5.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proves sharp L^p-boundedness ranges for the wave operators W_±(H;(-Δ)^m) for H=(-Δ)^m+V on R^n with odd 1≤n≤4m-1, under explicit decay and regularity assumptions on V and under a classification of zero-energy resonances. The main results are Theorems 1.5 and 1.7: for regular zero or resonance types k≤k_c the wave operators are bounded on L^p for all 1<p<∞; for k_c<k≤m_n the range is 1<p<n/(n-2m+k+k_c-1); and for zero being an eigenvalue the range is 1<p<2n/(n-1). Unboundedness is proved for p beyond these thresholds. The proof splits the stationary representation into low- and high-energy parts. The low-energy part is controlled by combining the asymptotic expansion of (M_±(λ))^{-1} near λ=0 (Theorem 2.5) with detailed kernel estimates for Q_j v R_±^0 (Lemmas 2.7 and 2.10) and oscillatory-integral arguments; the high-energy part is handled via existing results of Erdogan-Green and resolvent estimates. Unboundedness is reduced to optimal time-decay estimates for e^{itH}P_ac(H).","tokens_in":62857,"tokens_out":9831,"duration_ms":103373,"significance":"If the low-energy expansion is fully justified, the paper settles a natural open problem: the L^p-boundedness of wave operators for higher-order Schrödinger operators in all low odd dimensions with all zero-energy resonance types, and it gives explicit, parameter-free thresholds that agree with the known m=1,2 cases. The proof has a clear architecture, contains many detailed estimates, and the main claims are falsifiable and sharp in the stated range. The paper also correctly gives credit to the dependency of the resolvent expansion on the authors' preprint [4]; this dependency is the main source of risk rather than an internal inconsistency.","major_comments":[{"comment":"The central low-energy machine is Theorem 2.5, the expansion (2.36) of (M_±(λ))^{-1} with remainder bounds (2.37)-(2.38). The proof in Appendix A is explicitly only an outline: the decisive facts that D_00 is strictly positive, D_11 is strictly negative, d is invertible and rigidly negative definite on ⊕_{j∈J_k''} Q_j L^2, and that Q v G_{4m-n} v Q is invertible in the eigenvalue case are all quoted from Lemma 2.6 of [4]. Every low-energy kernel estimate in Section 3 (Lemmas 3.3, 3.8, 3.11 and Propositions 3.5, 3.7) and the extraction of the principal term in Section 5, equations (5.3)-(5.6), starts from (2.36). Thus the main theorems are exactly as secure as Theorem 2.5. The manuscript should either include a complete proof of Theorem 2.5, including Lemma 2.6 of [4], or state the main results as conditional on that preprint and explain why this external foundation is acceptable for the journal.","section":"Section 2.2 and Appendix A"},{"comment":"The if-and-only-if characterization of p-wave resonances is asserted without proof: a non-zero distributional solution φ of Hφ=0 in ∩_{s<-1/2} L^2_s \\setminus L^2 exists if and only if ψ=Uvφ belongs to Q_{(2m-n+1)/2}L^2. This equivalence is used in Case II to guarantee Q_{(2m-n+1)/2}≠0 and hence that the principal term in (5.6) does not vanish. Since Theorem 1.7(ii) is stated precisely for the eigenvalue case with such a distributional solution, this equivalence is load-bearing and needs either a proof or a precise reference to a result establishing it.","section":"Section 5, Remark 5.2"}],"minor_comments":[{"comment":"The displayed radial estimate contains a repairable typo: after applying the L^p-boundedness of the truncated Hilbert transform, the final integral should have the weight r^{n-1}, not r^{(n-1)/2}. With the correction the estimate is uniform in s,h and the conclusion of Proposition 3.13 is unaffected.","section":"Section 3.2.2, equation (3.42)"},{"comment":"The sentence 'W_+f=W_-f' is not literally correct; the proof treats W_- and then W_+ must be handled by the analogous stationary formula or by complex conjugation of the kernels. This does not affect the argument, but the wording should be corrected.","section":"Section 1.4"},{"comment":"The appendix says 'we outline the processes' and refers to [4] for several key algebraic facts; in addition to the major concern above, the notation M^±_{i,j} and the statement of the Neumann-series remainder Γ^±(λ) would benefit from a slightly fuller explanation of why the displayed remainder bounds are uniform over the finite index set J_k.","section":"Appendix A"}],"recommendation":"major_revision","confidential_remarks":"The main risk is external dependency: Theorem 2.5 is the engine of the paper and is imported from the authors' preprint [4]. The editor may wish to verify that [4] is in a complete, citable form before accepting the current paper, or require the authors to make the proof of Theorem 2.5 fully self-contained. This is not a novelty concern, but it is a correctness risk that should be resolved in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this paper closes the odd-dimensional zero-resonance problem for L^p wave-operator bounds of (−Δ)^m + V, including the previously open n ≤ 2m−1 cases with m ≥ 3, and it also proves matching unboundedness, so the p-ranges are genuinely sharp. Second, the whole edifice rests on Theorem 2.5, the asymptotic expansion of (M±(λ))^{-1} near zero: that theorem is imported from the same group's preprint [4] and only sketched in Appendix A. If the expansion is correct, this is a significant and careful advance. If not, the thresholds shift. Everything else I read hangs together.\n\nOn the positive side: this is a real derivation, not a fitting exercise. The stationary representation, the low/high energy split, and the reduction to kernel estimates are done in detail. Lemmas 2.7 and 2.10 are hard and proved in Section 6; the good/bad splitting in Section 3 is intricate and mostly convincing. The sharpness argument via optimal weighted decay is elegant and new. The results agree with the known m = 1, 2 cases, and the citation pattern is honest—[4] is a dependency, not padding.\n\nMy main concern is self-containedness. Theorem 2.5 is load-bearing, and the appendix explicitly says it only outlines the proof; the invertibility and signature properties of D are quoted from Lemma 2.6 of [4]. So a referee has to verify [4] as well as the present paper. I did not find a flaw in Sections 3–5, and I checked enough of the oscillatory integral arguments to believe they are sound. The only blemish I noticed is a small typo in (3.42): the radial change of variables loses a factor of 2^h (or a support factor), but it is repairable and does not affect the uniform bound. Assumption 1.3 is fairly strong decay, but it is in line with the literature, and the absence of embedded positive eigenvalues is standard and needed for the M±(λ) inversion.\n\nWho is this for? Analysts working on dispersive estimates for higher-order Schrödinger operators; they will use it as a reference. The paper deserves a serious referee, not a desk rejection. If I were the editor, I would send it out with a specific request: check Theorem 2.5, either by verifying [4] or by asking the authors to include the full proof. I would not accept as-is on the basis of the appendix alone.","headline":"A serious, detailed proof of sharp L^p wave-operator bounds for all zero-resonance types in the remaining odd low-dimensional higher-order Schrödinger cases, conditional on a resolvent expansion imported from the authors' own preprint [4].","tokens_in":63455,"tokens_out":2546,"would_cite":true,"duration_ms":27952,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35P25","47A40","35J30"],"pacs":[],"model":"deepseek-v4-flash","headline":"Zero-energy resonances determine exactly the L^p range of wave operators for higher-order Schrödinger operators in low odd dimensions.","keywords":["L^p-boundedness","wave operator","zero resonance","higher-order Schrödinger operator","zero-energy singularity","resolvent expansion","dispersive estimates","scattering theory"],"falsifier":"Test the predicted threshold in the case $n=3$, $m=2$, where a second-kind resonance is claimed to give boundedness exactly for $1<p<3$ and unboundedness for all $p>3$: construct an explicit potential $V$ satisfying Assumption 1.3 with a known second-kind zero resonance and check whether $\\|W_+(H;(-\\Delta)^2)\\|_{L^p\\to L^p}$ stays finite for some $p>3$; a single such example would refute Theorem 1.7.","tokens_in":62294,"feed_emoji":"🎯","tokens_out":13734,"duration_ms":122747,"temperature":0.7,"pith_summary":"This paper establishes sharp $L^p$-boundedness for the wave operators $W_\\pm(H;(-\\Delta)^m)$ of the higher-order Schrödinger operator $H=(-\\Delta)^m+V$ on $\\mathbb{R}^n$, for odd dimensions $1\\le n\\le 4m-1$ and $m\\ge2$, with real-valued decaying potentials $V$. If zero energy is regular or a resonance of kind $k\\le k_c$, the wave operators are bounded on $L^p$ for every $1<p<\\infty$; if zero is a resonance of kind $k_c<k\\le m_n$, boundedness holds for $1<p<p_k$ with $p_k=\\frac{n}{n-2m+k+k_c-1}$, and if zero is an eigenvalue, for $1<p<\\frac{2n}{n-1}$. The same thresholds are proved to be sharp: $W_\\pm$ are unbounded for every $p$ above $p_k$, with the eigenvalue statement requiring a $p$-wave resonance. A direct consequence is matching $L^p$-$L^{p'}$ decay estimates for $e^{itH}P_{ac}(H)$, so dispersive applications inherit the same $p$-restrictions. The proof splits the wave operator into low- and high-energy parts, expands the low-energy resolvent inverse near zero, and reduces the kernel estimates to oscillatory integrals whose singular term is a Hilbert transform.","feed_headline":"Zero-energy resonances shrink the L^p range of wave operators","feed_subtitle":"In odd dimensions n≤4m-1, each zero-resonance type k fixes the exact L^p threshold for W±((−Δ)^m+V).","key_machinery":"The carrying object is the operator $M_\\pm(\\lambda)=U+vR_0^\\pm(\\lambda^{2m})v$ on $L^2$, where $v=\\sqrt{|V|}$ and $U=\\operatorname{sgn}V$; the symmetric second resolvent identity writes the perturbed resolvent as $R^\\pm(\\lambda^{2m})V=R_0^\\pm(\\lambda^{2m})v(M_\\pm(\\lambda))^{-1}v$. The proof pivots on the asymptotic expansion of $(M_\\pm(\\lambda))^{-1}$ as $\\lambda\\to0^+$, a sum over resonance projections $Q_j$ with powers $\\lambda^{2m-n-i-j}$, together with the cancellation relations $Q_j(x^\\alpha v)=0$ for $|\\alpha|\\le\\max\\{0,\\lfloor j+1/2\\rfloor\\}-1$. These cancellations turn the low-energy wave-operator kernel into oscillatory integrals with phases $|x|\\pm|y|$; the sharp $p$-threshold appears when the remaining singular kernel is a weighted truncated Hilbert transform. Sharpness comes from a complementary lower bound for $|\\langle e^{itH}P_{ac}(H)\\psi,\\psi\\rangle|$, a Kato–Jensen-type time-decay estimate whose decay exponent matches the free dispersive rate exactly at $p=p_k$.","core_discovery":"On the paper's own terms, the central discovery is that in odd dimensions $1\\le n\\le 4m-1$ the $L^p$ mapping properties of the wave operators $W_\\pm(H;(-\\Delta)^m)$ are controlled entirely by the kind of zero-energy singularity of $H$, through a single critical index $k_c$. With regular zero counted as $k=0$ and a zero eigenvalue as $k=m_n+1$, Theorem 1.5 gives full-range boundedness $1<p<\\infty$ for $k\\le k_c$, the finite range $1<p<\\frac{n}{n-2m+k+k_c-1}$ for $k_c<k\\le m_n$, and $1<p<\\frac{2n}{n-1}$ in the eigenvalue case. Theorem 1.7 shows these ranges are sharp: for $p$ beyond the threshold the operators are unbounded, so the low-energy singularity itself is the obstruction. The paper also derives the corresponding $L^p$-$L^{p'}$ decay rates for the perturbed Schrödinger group and verifies that the results reproduce the known $m=1,2$ cases.","pith_inferences":["The paper leaves $p=p_k$ open; the classical Hilbert transform analogy suggests a weak-type $(p_k,p_k)$ bound may survive where strong boundedness fails, so endpoint dispersive applications could still be possible in a restricted sense.","The parity of $n$ enters the free-resolvent coefficients, so even-dimensional thresholds likely involve logarithmic corrections or shifted critical indices; the paper's advertised sequel for even dimensions should reveal whether the formula $p_k$ persists in that setting.","One testable consequence is that the leading angular term $\\sum_{|\\alpha|=|\\beta|=\\vartheta(j)} y^\\alpha Q_j(vz^\\beta)(x)/|y|^{(n-1)/2+\\vartheta(j)}$ controls the threshold: potentials for which this term vanishes could have a better effective $p$-range than the generic formula, suggesting the resonance type alone may not be the full story for every potential."],"forward_implications":["The dispersive estimate $\\|e^{itH}P_{ac}(H)\\|_{L^p\\to L^{p'}}\\lesssim |t|^{-(n/m)(1/p-1/2)}$ holds exactly for the endpoint ranges $p'$ listed in Corollary 1.10, so any nonlinear application of the perturbed group inherits the same resonance-dependent restriction.","In dimensions $2m+1\\le n\\le4m-1$, $k_c=0$, so the mere presence of any zero resonance or eigenvalue shrinks the range from the full $1<p<\\infty$ (or $1\\le p\\le\\infty$ in the regular case) to a finite interval.","The endpoint $p=p_k$ is not settled by the paper; endpoint boundedness or weak-type results at $p=p_k$ would be needed for applications requiring the optimal decay rate at the threshold.","The sharpness proof links $L^p$-unboundedness to optimality of Kato–Jensen time decay, so these two phenomena are equivalent up to the paper's assumptions: improving one improves the other.","The results unify the known $m=1,2$ cases and place the fourth-order line $n=1,3,5,7$ and its resonance classifications into a single formula."],"supporting_citations":[{"why":"Supplies the asymptotic expansion of $(M_\\pm(\\lambda))^{-1}$ stated as Theorem 2.5; the whole low-energy kernel analysis starts from this expansion.","marker":"[4]"},{"why":"Provides the free-resolvent expansions, the zero-resonance classification for $n>2m$, and the high-energy resolvent bounds used in Sections 2 and 4.","marker":"[15]"},{"why":"Gives the unified threshold resolvent-inversion method adapted in Appendix A to obtain $(M_\\pm(\\lambda))^{-1}$.","marker":"[30]"},{"why":"Establishes the high-energy part $W_H$ for the regular case with $n>2m$, used as the baseline in Section 4.","marker":"[10]"},{"why":"Provides the stationary representation of wave operators and the symmetric second resolvent identity that produces the low/high-energy split.","marker":"[33]"},{"why":"Supplies the free dispersive estimate that converts wave-operator boundedness into $L^p$-$L^{p'}$ decay and is used in the sharpness contradiction.","marker":"[55]"}],"fun_headline_variants":["Zero-energy resonances determine wave operator L^p bounds","Each zero-resonance kind sets the exact L^p threshold","Critical index k_c governs L^p boundedness of wave operators","Wave operator L^p range shrinks with stronger zero singularities","Sharp L^p bounds for wave operators from zero-resonance type"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's low-energy conclusions stand on the asymptotic expansion of $(M_\\pm(\\lambda))^{-1}$ near zero energy that is quoted from the earlier preprint [4] and only summarized in Appendix A; if that expansion is incomplete in any term, the stated $p$-thresholds would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Zero-energy resonances determine wave operator L^p bounds","Each zero-resonance kind sets the exact L^p threshold","Critical index k_c governs L^p boundedness of wave operators","Wave operator L^p range shrinks with stronger zero singularities","Sharp L^p bounds for wave operators from zero-resonance type"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000939,"raw_usage":{"total_tokens":4248,"prompt_tokens":1415,"completion_tokens":2833,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":1031,"completion_tokens_details":{"reasoning_tokens":2743}},"tokens_in":1031,"tokens_out":2833,"duration_ms":20322,"temperature":1.0,"reasoning_tokens":2743,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:27:11.261121+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Test the predicted threshold in the case $n=3$, $m=2$, where a second-kind resonance is claimed to give boundedness exactly for $1<p<3$ and unboundedness for all $p>3$: construct an explicit potential $V$ satisfying Assumption 1.3 with a known second-kind zero resonance and check whether $\\|W_+(H;(-\\Delta)^2)\\|_{L^p\\to L^p}$ stays finite for some $p>3$; a single such example would refute Theorem 1.7.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the free-resolvent expansions, the zero-resonance classification for $n>2m$, and the high-energy resolvent bounds used in Sections 2 and 4."},{"cited_title":"Jensen and G","cited_arxiv_id":null,"evidence_quote":"Gives the unified threshold resolvent-inversion method adapted in Appendix A to obtain $(M_\\pm(\\lambda))^{-1}$."},{"cited_title":"Erdo˘ gan and W.R","cited_arxiv_id":null,"evidence_quote":"Establishes the high-energy part $W_H$ for the regular case with $n>2m$, used as the baseline in Section 4."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the stationary representation of wave operators and the symmetric second resolvent identity that produces the low/high-energy split."},{"cited_title":"Zheng, X","cited_arxiv_id":null,"evidence_quote":"Supplies the free dispersive estimate that converts wave-operator boundedness into $L^p$-$L^{p'}$ decay and is used in the sharpness contradiction."}],"review_version":1}