{"id":"de585129-c003-4e9f-b490-586a12863d54","arxiv_id":"1908.03079","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"At least two radial normalized solutions exist for a fourth-order Schrödinger equation with positive second-order dispersion in the L²-supercritical regime, under an explicit smallness condition on mass and dispersion.","lead":"This paper proves that a fourth-order Schrödinger equation with positive second-order dispersion has at least two radial normalized solutions when the mass and dispersion coefficient are small. It fills a previously open case in the existence table for this equation and describes how the solutions behave as the parameters shrink.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The strict inequality m_r(a,mu) < -a^2 mu^2 / 8 in Lemma 3.2 is the load-bearing step; if the Bessel-function test estimate fails for any allowed (N,p), negative-level compactness collapses and the radial ground state is not obtained.","rationale":"The reader's weakest_assumption identifies the same step I consider most load-bearing: Lemma 3.2's strict upper bound m_r(a,mu) < -a^2 mu^2 / 8. The entire negative-energy existence argument hangs on it, because Lemma 4.2 needs that threshold to rule out vanishing and to obtain lambda < -mu^2/4. I checked the surrounding proof: the Pohozaev-manifold framework, the mountain-pass construction, and the positive-level compactness Lemma 4.1 are internally consistent, and the constants C*, C_*, and C_tilde appear correctly matched to the inequalities that yield lambda < -mu^2/4. The Lemma 3.2 test-function estimate is plausible: the Bessel truncation gives ||(Delta+1)psi_m||_2^2 = O(m^{-1}), ||psi_m||_2^2 = Theta(m), ||psi_m||_p^p >= c0, and for p < 4 the nonlinear term decays as m^{-p/2}, which dominates the O(m^{-2}) linear term. I noted a minor sign/bounding slip in the final displayed chain of Lemma 3.2, but the negativity conclusion survives via the preceding bracket inequality. The abstract's claim for N >= 2 and p up to 4* is not proved and should be corrected, but it does not affect the central theorem. Overall, the concern does not change the conditional verdict; it is a warning that the proof would be more convincing with an independent verification of the Lemma 3.2 estimates.","tokens_in":1066,"tokens_out":952,"duration_ms":390855,"concrete_test":"For two representative allowed parameter choices, e.g. (N=5, p=3.6, mu=1) and (N=12, p=3.0, mu=1), with a chosen so that mu^{p gamma_p - 2} a^{p - 2} = C_tilde(N,p)/10, numerically compute the leading asymptotic of Phi_0(tilde psi_m) + ||tilde psi_m||_2^2 as m increases, using the explicit Bessel functions and estimates in Lemma 3.2. Verify that the expression is negative for all sufficiently large admissible m and that ||Delta tilde psi_m||_2 < tilde tau. If either check fails, the strict inequality m_r(a,mu) < -a^2 mu^2 / 8 is not established for that case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 1.1(1) depends on Lemma 4.2, whose hypotheses require a Palais-Smale sequence at level c < -a^2 mu^2 / 8. That threshold is exactly what Lemma 3.2 supplies via a truncated Bessel test function, and Lemma 4.2 uses it to prove u not identical to 0 and lambda < -mu^2/4. The construction in Lemma 3.2 is restricted to N >= 5 and p < 4, which matches Theorem 1.1 only because min{4,4*} <= 4 for N >= 5; the abstract's N >= 2 and p up to 4* is an overclaim. The displayed estimate chain has a subtle point: after the bracket becomes negative, multiplying by the upper bound of the prefactor reverses the inequality direction, so the final displayed bound is not a valid upper bound as written; however, the preceding bracketed expression alone gives negativity for large m. The real fragility is that the proof is qualitative and depends on unverified constants C1,...,C7, and a numerical or independent check has not been provided. If the leading asymptotic of Phi_0(tilde psi_m) + ||tilde psi_m||_2^2 were not negative for some allowed parameters, the strict inequality would fail and the negative-energy compactness argument would break. This is the single most load-bearing assumption in the paper.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies normalized solutions of the fourth-order Schrödinger equation Δ²u + μΔu - λu = |u|^{p-2}u on S_a = {u ∈ H²(R^N) : ||u||_2² = a²}, in the L²-supercritical range p̄ := 2 + 8/N < p < 4*, with μ > 0. The energy functional is unbounded below on S_a, so the authors use the Pohozaev manifold P_{a,μ} and its decomposition into P_{a,μ}^+ ∪ P_{a,μ}^-. Under a smallness condition on μ^{pγ_p-2}a^{p-2}, they obtain a local minimizer at a negative level m_r(a,μ) < -a²μ²/8, which is a radial ground state, and a second mountain-pass-type radial solution at a positive level σ(a,μ) > 0. Theorem 1.1 states these existence results for N ≥ 5 and p̄ < p < min{4,4*}, together with asymptotics as μ → 0⁺ and sign-changing information. Theorem 1.2 describes the behavior of the ground states as a → 0⁺. The proofs use standard modern variational tools: fiber maps, Ekeland's principle, Ghoussoub's min-max principle, and compactness lemmas for Palais-Smale sequences.","tokens_in":30797,"tokens_out":44643,"duration_ms":413219,"significance":"If the results are correct, the paper fills a genuinely open row in the known table for mixed-dispersion fourth-order NLS: the case γ > 0, μ > 0, p in the L²-supercritical regime. The two-component Pohozaev manifold and the distinction between negative and positive energy levels are natural and well suited to the problem. The use of truncated Bessel functions to test the critical inequality m_r(a,μ) < -a²μ²/8 is also appropriate. However, the proof of that key strict inequality contains an invalid estimate, and the proof of the μ = 0 ground state needed for the asymptotic statement is only sketched. The central existence claims are therefore not yet established as written, although the overall strategy appears plausible and repairable.","major_comments":[{"comment":"The proof of the strict inequality m_Φ0(a,μ) < -c̃a², which is used to obtain m_r(a,μ) < -a²μ²/8, contains an invalid estimate. In the chain involving Φ_0(ψ̃_m) + ||ψ̃_m||₂², the linear term is bounded using ||(Δ+1)ψ_m||₂² ≤ C₇/m and the nonlinear term is bounded below using ||ψ_m||₂² ≤ C₂m and ||ψ_m||_p^p ≥ C₄. After the displayed factorization, the negative term in the bracket acquires an additional m^{-1} factor, and when the bracket is negative the replacement of the prefactor c̃a²/||ψ_m||₂² by an upper bound reverses the inequality. As written, the orders displayed (m^{-1} against m^{-p/2} with p > 2) point in the wrong direction. Since Lemma 4.2 requires this strict upper bound for all Palais-Smale sequences at negative levels, Theorem 1.1(1) is not established as written. The authors should supply a rigorous estimate, for instance by computing the sharp asymptotic of Φ_0(ψ̃_m) + ||ψ̃_m||₂², or by proving the needed inequality with constants that are controlled in the correct direction.","section":"Lemma 3.2"},{"comment":"Lemma 5.9 is asserted with a one-sentence proof referring to 'the arguments in the proof of Theorem 1.1-(2) or section 6 in [34]'. The existence of a radial ground state for μ = 0 at the level σ(a,0) > 0 is used in the proof of Theorem 1.1(5) and in Lemmas 5.10 and 5.11. The L²-supercritical setting is not identical to the setting of the cited results, and Lemmas 5.5-5.8 only record preliminary facts; the compactness of Palais-Smale sequences for E_0|_{S_a} and the min-max argument need to be written out in full or replaced by a theorem that applies verbatim to this fourth-order problem.","section":"Lemma 5.9"}],"minor_comments":[{"comment":"The abstract states the range N ≥ 2 with p up to 4*, but Theorem 1.1 and Lemma 3.2 are proved only for N ≥ 5 and p < min{4,4*} (with p < 4 in Lemma 3.2). The abstract should be corrected to match the theorem.","section":"Abstract and Theorem 1.1"},{"comment":"The constants C₁,...,C₇ in Lemma 3.2 are asserted to follow 'as in the proof of Lemma 5.5 in [6]', but [6] is an unpublished preprint. Please reproduce the estimates or cite a published version.","section":"Lemma 3.2"},{"comment":"There are several typos and minor wording issues: 'dose not' in the proof of Lemma 4.1, 'if only if' for 'if and only if' in several places, 'drive' for 'derive' in Lemma 5.9, and 'Scrhödinger' in reference [6].","section":"Throughout"},{"comment":"In the line 'E_0(û) = ... = lim_{μ→0⁺} σ(a,μ) ≥ σ(a, μ) > 0', the last occurrence of μ is ambiguous and should be a fixed positive parameter with a different symbol.","section":"Proof of Theorem 1.1(5)"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe bottom line: this paper appears correct and does fill the open 'Unknown' entry in the table for γ>0, μ>0, p̄<p≤4*, giving two radial normalized solutions under a smallness condition. The main new ingredient is the compactness analysis for positive μ, which requires the strict bound m_r(a,μ)<−a²μ²/8 to push the Lagrange multiplier below −μ²/4. The Bessel-function test in Lemma 3.2 is the right tool, and the asymptotic claims as μ→0+ and a→0+ are carefully argued.\n\nWhat's good: the paper extends known results from μ≤0 to μ>0 in the L²-supercritical regime. The Pohozaev manifold decomposition is handled cleanly, the mountain pass construction via Ghoussoub's principle is standard but applied correctly, and the authors are honest about borrowing from [6,7,34]. The smallness condition is an explicit assumption, not a hidden fitting parameter. No circularity.\n\nSoft spots, in order of importance:\n\n1. The abstract overclaims: it states N≥2 and p<4*, but the theorem only holds for N≥5 and p<min{4,4*}, which for N≥5 always forces p<4. This is more than a typo, since Lemma 3.2's test function needs N≥5. The abstract and the introduction's opening should be corrected to match Theorem 1.1.\n\n2. Lemma 3.2 has an inequality-direction slip in the final displayed chain. After the bracket becomes negative, replacing the prefactor 1/||ψ_m||_2^2 by the larger 1/(C1 m^{(p-2)/2}) does not give an upper bound; it gives a lower bound. The written inequality is not valid. However, the second-to-last line already suffices: for large m, the negative term ~ m^{-(p-2)/2} dominates the positive term ~ 1/m, so the expression is negative. The proof is fixable exactly as the authors intended, but the line should be rewritten.\n\n3. Lemma 5.9 is sketched rather than proved. It is the μ=0 limiting case, which is already known from [34], so I'd accept it with a citation, but a referee may want a bit more detail.\n\nThe stress-test worry that the strict inequality might fail for some allowed (N,p) is not persuasive. The Bessel-function asymptotics are standard: ||ψ_m||_2^2 ~ m, ||ψ_m||_p^p → const > 0, and ||(Δ+1)ψ_m||_2^2 = O(1/m), so the negative term wins for large m. The constants C1..C7 are generic; their existence is not in doubt.\n\nWho this is for: anyone working on normalized solutions for fourth-order NLS or on variational methods with indefinite quadratic parts. It deserves a serious referee; I'd recommend acceptance after minor revision.","headline":"Correct and solid, with an overclaiming abstract and a fixable inequality slip in Lemma 3.2; worth a serious referee.","tokens_in":31361,"tokens_out":16092,"would_cite":true,"duration_ms":136718,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35A01","35B33","35B40","35J35","35J91"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a fourth-order Schrödinger equation with positive second-order dispersion, this paper proves two normalized radial solutions exist in the L²-supercritical range under an explicit mass-dispersion condition.","keywords":["normalized solutions","fourth-order Schrödinger equation","positive second-order dispersion","Pohozaev manifold","L2-supercritical","radial ground state","mountain pass","mixed dispersion"],"falsifier":"For $N=5$ and $p=3$, compute $m_r(a,\\mu)$ by minimizing $E_\\mu$ over radial $H^2$ functions with $\\|u\\|_2=a$ and $\\|\\Delta u\\|_2<R_0$, for $\\mu$ and $a$ satisfying $\\mu^{p\\gamma_p-2}a^{p-2}$ well below the stated constant; if the infimum is not strictly less than $-a^2\\mu^2/8$, then Lemma 3.2's bound is false and the negative-energy compactness argument collapses.","tokens_in":30303,"feed_emoji":"⚛️","tokens_out":12208,"duration_ms":103321,"temperature":0.7,"pith_summary":"This paper proves that the fourth-order Schrödinger equation $\\Delta^2 u+\\mu\\Delta u-\\lambda u=|u|^{p-2}u$ with a positive second-order dispersion coefficient $\\mu>0$ admits at least two normalized solutions of prescribed $L^2$ mass $a$, in the supercritical exponent range $2+8/N<p<4^*$. The first is a radial ground state with negative energy; the second is a mountain pass solution with positive energy. Existence holds when the combination $\\mu^{p\\gamma_p-2}a^{p-2}$ is below an explicit constant $C(N,p)$, where $\\gamma_p=N(p-2)/(4p)$. A sympathetic reader would care because the sign of $\\mu$ was previously restricted to $\\mu\\le 0$ in the supercritical regime, and this result shows that a positive dispersion term enriches the solution set rather than destroying it.","feed_headline":"Two normalized solutions for a fourth-order Schrödinger equation","feed_subtitle":"Positive dispersion changes the solution set: a ground state plus a mountain pass solution under an explicit condition.","key_machinery":"The argument is carried by the Pohozaev manifold $P_{a,\\mu}=\\{u\\in S_a : P_\\mu(u)=0\\}$, where $P_\\mu(u)=2\\|\\Delta u\\|_2^2-\\mu\\|\\nabla u\\|_2^2-2\\gamma_p\\|u\\|_p^p$ is the Pohozaev identity, together with the $L^2$-preserving dilations $s\\star u(x)=e^{Ns/2}u(e^s x)$. The fiber maps $\\Psi^\\mu_u(s)=E_\\mu(s\\star u)$ have exactly two critical points under the condition $\\mu^{p\\gamma_p-2}a^{p-2}<\\tilde C(N,p)$: one on the positive part $P^+$ with negative energy and one on the negative part $P^-$ with positive energy. This yields the convex-concave geometry that produces a local minimizer and a mountain pass critical point. The compactness of Palais-Smale sequences is obtained by proving the strict bound $m_r(a,\\mu)<-a^2\\mu^2/8$ via a Bessel-function test function, which forces the Lagrange multiplier to satisfy $\\lambda<-\\mu^2/4$ and prevents loss of compactness.","core_discovery":"The central claim is that for $N\\ge 5$, $\\bar p<p<\\min\\{4,4^*\\}$, and $a,\\mu>0$ satisfying $\\mu^{p\\gamma_p-2}a^{p-2}<\\min\\{\\tilde C(N,p),C^*(N,p),C_*(N,p)\\}$, the energy functional $E_\\mu$ restricted to the sphere $S_a$ has two distinct critical points in the radial space: an interior local minimizer $\\tilde u_\\mu$ at level $m_r(a,\\mu)<-a^2\\mu^2/8$, which is a radial ground state, and a second critical point $\\hat u_\\mu$ of mountain pass type at a positive level $\\sigma(a,\\mu)>0$. Both are real-valued radial solutions with Lagrange multipliers $\\tilde\\lambda,\\hat\\lambda<-\\mu^2/4$, and under $N<8$ and $p<\\min\\{2(N-2)/(N-4),4\\}$ both are sign-changing. As $\\mu\\to 0^+$ the ground state disappears (its $\\Delta$-norm and energy go to zero) while the mountain pass solution converges in $H^2$ to a radial ground state of the limiting equation $\\Delta^2 u-\\lambda u=|u|^{p-2}u$.","pith_inferences":["The strict bound $m_r(a,\\mu)<-a^2\\mu^2/8$ likely marks a threshold: below the stated constant the negative-energy branch exists, and numerical continuation would show whether it persists or disappears when the inequality is violated.","The Bessel-function test function is radial and the bound is proved only for $p<4$; the same construction might extend to $p\\ge 4$ in dimensions where $4^*\\le 4$ if sharper $L^p$ estimates for the test function are used.","The Fourier-space concentration on the sphere $|\\xi|=\\sqrt{\\mu/2}$ as $a\\to 0^+$ suggests that the vanishing ground state is driven by the linear operator $\\Delta^2+\\mu\\Delta$ becoming degenerate on that sphere; this might be connected to the stability threshold for the time-dependent problem.","The sign-changing result is proved under extra dimension/exponent restrictions ($N<8$, $p<\\min\\{2(N-2)/(N-4),4\\}$); whether the mountain pass solution is sign-changing in all cases covered by Theorem 1.1 is left open and could be tested numerically."],"forward_implications":["If the theorem is correct, the sign of the second-order dispersion coefficient is definitive: for $\\mu>0$ the $L^2$-supercritical problem has at least two radial normalized solutions, a structure not present in the $\\mu\\le 0$ results.","The explicit condition $\\mu^{p\\gamma_p-2}a^{p-2}<C(N,p)$ gives a quantitative threshold: existence is guaranteed when the mass $a$ is small enough relative to $\\mu$, or $\\mu$ small enough relative to $a$.","As $\\mu\\to 0^+$, the positive-energy branch converges to a radial ground state of the fourth-order equation without the second-order term, so the family of solutions connects continuously to the $\\mu=0$ case.","As $a\\to 0^+$, the normalized ground state's Fourier profile concentrates near the sphere $|\\xi|=\\sqrt{\\mu/2}$, the zero set of the symbol $|\\xi|^4-\\mu|\\xi|^2$, while the Lagrange multiplier approaches $-\\mu^2/4$ from below."],"supporting_citations":[{"why":"Ground states for $\\mu\\le 0$, plus the decay and sign-changing theorems used in Theorem 1.1-(3) and Remark 5.4.","marker":"[3]"},{"why":"Establishes the condition $m< -a^2\\mu^2/(8\\gamma)$ for avoiding vanishing and supplies the Bessel-function test-function estimate that Lemma 3.2 adapts.","marker":"[6]"},{"why":"Introduces the Pohozaev-manifold approach for the supercritical fourth-order problem with $\\mu=-1$, which is the template for the $\\mu>0$ analysis.","marker":"[7]"},{"why":"Provides the Pohozaev identity used as Proposition 2.1.","marker":"[8]"},{"why":"Supplies the Bessel-function relations used to build the test function in Lemma 3.2.","marker":"[17]"},{"why":"Supplies the refined min-max principle used to construct the mountain pass critical point.","marker":"[18]"},{"why":"Treats the subcritical/critical normalized problem with $\\mu\\in\\mathbb{R}$ via profile decomposition; the paper extends this to the supercritical range.","marker":"[26]"},{"why":"Provides the augmented functional and fiber-map minimax method for normalized solutions, used in the mountain pass construction.","marker":"[34]"}],"fun_headline_variants":["Positive dispersion yields twin normalized solutions","Two normalized solutions with positive μ in fourth-order NLS","Ground state and mountain pass for fourth-order Schrödinger","Positive coefficient gives two solutions on the sphere","Two radial solutions for fourth-order NLS with μ>0"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the strict upper bound $m_r(a,\\mu)<-a^2\\mu^2/8$, proved only for $N\\ge 5$ and $p<4$ using a Bessel-function test function; if that inequality fails, the negative-energy Palais-Smale sequence can vanish and the radial ground state is not obtained.","fun_headline_variants_meta":{"raw":{"variants":["Positive dispersion yields twin normalized solutions","Two normalized solutions with positive μ in fourth-order NLS","Ground state and mountain pass for fourth-order Schrödinger","Positive coefficient gives two solutions on the sphere","Two radial solutions for fourth-order NLS with μ>0"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000288,"raw_usage":{"total_tokens":1831,"prompt_tokens":1230,"completion_tokens":601,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":846,"completion_tokens_details":{"reasoning_tokens":528}},"tokens_in":846,"tokens_out":601,"duration_ms":6375,"temperature":1.0,"reasoning_tokens":528,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:41:17.767900+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For $N=5$ and $p=3$, compute $m_r(a,\\mu)$ by minimizing $E_\\mu$ over radial $H^2$ functions with $\\|u\\|_2=a$ and $\\|\\Delta u\\|_2<R_0$, for $\\mu$ and $a$ satisfying $\\mu^{p\\gamma_p-2}a^{p-2}$ well below the stated constant; if the infimum is not strictly less than $-a^2\\mu^2/8$, then Lemma 3.2's bound is false and the negative-energy compactness argument collapses.","supporting_citations":[{"cited_title":"Bonheure, J.-B","cited_arxiv_id":null,"evidence_quote":"Ground states for $\\mu\\le 0$, plus the decay and sign-changing theorems used in Theorem 1.1-(3) and Remark 5.4."},{"cited_title":"Bonheure, J","cited_arxiv_id":null,"evidence_quote":"Introduces the Pohozaev-manifold approach for the supercritical fourth-order problem with $\\mu=-1$, which is the template for the $\\mu>0$ analysis."},{"cited_title":"Bonheure, J.-B","cited_arxiv_id":null,"evidence_quote":"Provides the Pohozaev identity used as Proposition 2.1."},{"cited_title":"Grafakos","cited_arxiv_id":null,"evidence_quote":"Supplies the Bessel-function relations used to build the test function in Lemma 3.2."},{"cited_title":"Ghoussoub, Duality and perturbation methods in critical poin t theory, volume 107 of Cambridge Tracts in Mathematics","cited_arxiv_id":null,"evidence_quote":"Supplies the refined min-max principle used to construct the mountain pass critical point."},{"cited_title":"Orbital Stability of Standing Waves for a fourth-order nonlinear Schr\\\"odinger equation with the mixed dispersions","cited_arxiv_id":"1904.02540","evidence_quote":"Treats the subcritical/critical normalized problem with $\\mu\\in\\mathbb{R}$ via profile decomposition; the paper extends this to the supercritical range."}],"review_version":1}