{"id":"8f0c146a-4ea6-4161-bd6c-ce67b16a60c8","arxiv_id":"2504.13505","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For every smooth Fano threefold of Picard rank one, slope-stable higher-rank instanton bundles exist for all charges above an explicitly computed minimal value, and no stable examples exist below it.","lead":"This paper proves that on every smooth Fano threefold with one Picard generator, stable higher-rank instanton vector bundles exist exactly once a certain integer called the charge is large enough, and it computes the smallest allowed charge. Generalists may care because it settles a basic nonemptiness question for moduli spaces that appear across algebraic geometry and gauge theory.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quadric induction rests on an unproved vanishing: Theorem 4.4 Step 3 invokes Ext^2(S,F_n^k)=0 via a nonexistent 'Lemma ??', and the stated μ-stability assumptions do not obviously imply it.","rationale":"The paper's central claim is a uniform existence and minimality theorem for all four Fano indices. The quadric case is not a corollary of the other indices; Theorem 4.4 is the only mechanism producing higher-rank instantons on Q, and its induction step depends on the unobstructedness of the auxiliary sheaves F_{n+1}^k. That unobstructedness is asserted through a vanishing whose proof is missing ('Lemma ??'). The nearest available result, Lemma 4.1, is stated for Gieseker-stable instantons, whereas the bundles entering the induction are only μ-stable and the extensions are described as Gieseker-unstable, so the implication is not automatic. This is an internal gap at a load-bearing point of the proof, rather than a mere dependence on cited results. The reader's stated weakest assumption concerns the rank-2 base cases; that is a real dependency but is partly outsourced to [BF11], [CFK24], [Fae14], and [HH85], whereas the quadric gap is fully internal and directly checkable. I do not see a demonstrated contradiction in the other cases, and the missing vanishing may well be true; the proposed test would settle it. I therefore keep the reader's CONDITIONAL verdict without moving to ACCEPT or REJECT.","tokens_in":49973,"tokens_out":34015,"duration_ms":309520,"concrete_test":"Re-derive Theorem 4.4 Step 3 without the phantom lemma: apply Hom(S,·) to 0→O(−1)→E_n^k→F_n^k→0 and use the exact sequence (16) 0→S→O^4→S^*→0 to compute Ext^2(S,F_n^k) directly, using only the induction hypotheses (μ-stability and conditions i)–iii of Theorem 4.4). In particular, test the case of a μ-stable but Gieseker-unstable E_n^k obtained as a nontrivial extension (17): compute Ext^2(S,E_n^k) from the defining exact sequence and check whether it vanishes for all k≥⌈n/3⌉; also verify Ext^3(S,O(−1))=0 explicitly. A concrete proxy is to compute, for n=2 and k=1, the sheaf F_2^1= coker(O(−1)→E_2^1) using the Kapranov monadal description of §7 on Q and check Ext^2_Q(S,F_2^1)=0.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the proof of Theorem 4.4 (quadric threefold, i_X=3), the induction step constructs F_{n+1}^k as the cokernel of a lifted section in diagram (18). To apply Lemma 2.13 and conclude that F_{n+1}^k is unobstructed, the proof needs both Ext^2(F_n^k,F_n^k)=0 and Ext^2(S,F_n^k)=0. The first is part of the induction hypothesis, but the second is justified only by the sentence 'since Ext^2(S,E_n^k)=0, see the proof of Lemma ??'. No such lemma appears in the manuscript. The closest statement, Lemma 4.1, proves Ext^2(S,E)=0 only for Gieseker-stable instantons, whereas the inductively produced bundles E_n^k are only known to be μ-stable and the extensions E_{n+1}^k are explicitly described as Gieseker-unstable. Moreover, passing from 0→O(−1)→E_n^k→F_n^k→0 to Ext^2(S,F_n^k) also requires a vanishing of Ext^3(S,O(−1)), which is not written down. If the missing vanishing fails for some μ-stable but Gieseker-unstable (n,k)-instanton, then the unobstructedness step of the quadric induction collapses, and the i_X=3 case of Main Theorem 1 is unsupported exactly in the range k≥⌈n/3⌉. This is an internal gap, not merely a reliance on external literature.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a systematic existence theory for higher-rank instanton sheaves on smooth Fano threefolds of Picard rank one. It defines (n,k)-instantons, proves lower bounds for the charge of slope-stable instantons, and constructs unobstructed μ-stable instanton bundles for all k above a rank-dependent threshold k_0^n, by induction from rank-two 't Hooft bundles. The cases i_X=4,3,2 and i_X=1 are treated separately, with the index-one case conditioned on genus and on very-ampleness of −K_X. The paper also proves stability of restrictions to general anticanonical K3 sections, discusses instantons on Fano threefolds with curvilinear Kuznetsov components, and gives a monadic description when H^3(X)=0.","tokens_in":50213,"tokens_out":12357,"duration_ms":108236,"significance":"If the gaps identified below are repaired, this would be a substantial contribution: it gives, for a large class of Fano threefolds, essentially sharp necessary and sufficient numerical conditions for the existence of slope-stable higher-rank instanton bundles, together with unobstructedness and generic splitting statements. The inductive construction from rank-two 't Hooft bundles is natural and, where carried out in detail, is convincing. The applications to K3 restrictions and to moduli spaces of sheaves on the anticanonical section are also valuable and connect with the symplectic geometry of moduli spaces.","major_comments":[{"comment":"The proof invokes the vanishing Ext^2(S,F_n^k)=0 by saying 'since Ext^2(S,E_n^k)=0, see the proof of Lemma ??', but no such lemma appears in the manuscript. This vanishing is load-bearing: it is exactly one of the two Ext^2 vanishings required to apply Lemma 2.13 and conclude that F_{n+1}^k is unobstructed. Moreover, passing from 0→O(−1)→E_n^k→F_n^k→0 to Ext^2(S,F_n^k)=0 also requires a vanishing of Ext^3(S,O(−1)), which is not stated. Without a correct proof of Ext^2(S,F_n^k)=0, the induction for the quadric threefold case of Main Theorem 1 is unsupported.","section":"§4.1, Theorem 4.4, Step 3"},{"comment":"The first paragraph of Main Theorem 1 asserts existence of unobstructed μ-stable (n,k)-instanton bundles for all n≥2−r_X and k≥k_0^n with no exception for i_X=1. However, Theorem 4.14, which is the only source for the existence statement in the index-one case, is stated only for genus g≥4 and assumes that X is not contained in a singular quadric when g=4. In particular, for g=3 or for g=4 with X contained in a singular quadric, the theorem does not prove the asserted existence; for g=4 singular quadric and n=1, k_0^1=0, while Proposition 4.10 only gives existence for k≥1. Either the statement of Main Theorem 1 must be restricted, or the exceptional cases must be proved.","section":"Main Theorem 1 vs. §4.2, Theorem 4.14"},{"comment":"The dimension computation intended to show that a general deformation of F_{n+1}^k has no global sections contains an unexplained and apparently incorrect equality: the text claims ext^1(F_{n+1}^k,F_{n+1}^k)−ext^1(F_n^k,F_n^k)−ext^1(F_n^k,O)−1 = χ(O,F_n^k) = χ(E_n^k) > 0. But for a (n,k)-instanton on P^3 one has χ(E_n^k)=n−2k, which is not positive when k≥⌈n/2⌉. The displayed equalities also appear to mix ext and χ terms without justification. This step is essential for the P^3 case of Main Theorem 1, so it must be corrected and rewritten.","section":"§3.1, Theorem 3.2, Step 5"},{"comment":"The proof of Theorem 3.8 consists of the single sentence 'The steps of the proof are equivalent to the ones we used for Theorem 3.2.' Since Theorem 3.8 is load-bearing for the index-two case of Main Theorem 1, and since the base of the induction is different here (using the smoothing result Proposition 3.5 for elliptic curves), the proof needs to be written out, or at least the analogous Step 5 dimension count and the use of Lemma 2.13 must be spelled out.","section":"§3.2, Theorem 3.8"},{"comment":"The inequality used to show that the component of M_Q(v) through [F_{n+1}^k] is not contained in Z_{v'} is justified by the formula ext^1(F_{n+1}^k,F_{n+1}^k)−ext^1(F_n^k,F_n^k)−ext^1(F_{2n}^k,S)+1 = 3k−n > 0. The object F_{2n}^k is not defined, and the equality is not derived. Since this computation is exactly what rules out the closed locus of sheaves admitting a quotient of type F_n^k, the argument is incomplete as written.","section":"§4.1, Theorem 4.4, Step 5"}],"minor_comments":[{"comment":"The abstract contains several typos ('intantons', 'such as and generic splitting'); these should be corrected.","section":"Abstract and Introduction"},{"comment":"The reference 'Lemma ??' is a broken cross-reference and must be replaced by a precise statement and proof.","section":"§4.1, Theorem 4.4, Step 3"},{"comment":"The sentence 'allows to conclude that Ext^2(F0,E)' is missing the equality '=0' and should be completed.","section":"§4.2, Lemma 4.12(a)"},{"comment":"The condition 'v0/v1 = −1/2' should presumably be 'v1/v0 = −1/2', since slope is c1/rank.","section":"§4.2, Theorem 4.14, Step 3"},{"comment":"The notation 'H 3+k, 1' for the Hilbert scheme of elliptic curves is not defined in the proof; it should be introduced explicitly.","section":"§4.2, Proposition 4.10"}],"recommendation":"major_revision","confidential_remarks":"The paper is squarely within the scope of the journal and the main ideas are promising, but the broken reference in Theorem 4.4 and the overclaim in Main Theorem 1 relative to Theorem 4.14 need to be addressed before the paper can be considered for publication. The omitted proof of Theorem 3.8 and the questionable dimension count in Theorem 3.2 Step 5 are also serious. I recommend requesting a major revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuine advance. For every smooth Fano threefold of Picard rank one, the paper determines the sharp minimal charge k0(n) for slope-stable higher-rank instanton bundles and proves existence for all k >= k0. Rank-2 existence was known piecewise; the inductive deformation argument that takes a mu-stable (n,k)-instanton, extends by O or the spinor/Mukai bundle F0, then deforms the semistable extension to a mu-stable one is the real new engine. The paper also proves stability of restriction to a general K3 section and gets Lagrangian subvarieties, plus Kuznetsov/monad descriptions. That is a solid subfield-level advance.\n\nThe good: the inductive set-up is genuinely well organized. The P3 case is written out in detail; the quadric case too, with the spinor bundle playing the role of F0. The minimality arguments (Lemmas 2.10 and 4.1, plus Lemma 4.12 for index 1) are sound. The paper is honest about what it imports: rank-2 existence from BF11, CFK24, Fae14, and the elliptic-curve smoothing from HH85. The dependence is a base, not circular reasoning.\n\nWhere it gets soft:\n- Theorem 3.8 (Del Pezzo) is not proved; the text simply says the steps are equivalent to Theorem 3.2. For a paper whose main theorem covers all indices, that is a real gap in exposition, even if the adaptation is likely straightforward.\n- In Theorem 4.4, Step 3, there is a missing cross-reference: Ext^2(S,F_n^k)=0 is asserted via a nonexistent 'Lemma ??'. I checked the stress-test concern about this. It does not land: the vanishing Ext^2(S,E_n^k)=0 follows from the proof of Lemma 4.1 using only mu-stability and H^2(E)=0, and Ext^3(S,O(-1))=0 holds by Serre duality. So the mathematics is fine, but the manuscript must fix the reference and spell out the one-line argument.\n- The index-1 theorem is conditional (g>=4, plus a singular-quadric exclusion at g=4). This is openly stated, but it means the 'all Fano threefolds' claim in the introduction is slightly stronger than what is proved.\n- The paper leans heavily on prior rank-2 results; a referee should spot-check the boundary cases in Theorem 4.8 and Proposition 4.10.\n\nNet: the central argument holds up, the result is new and significant, and the issues are fixable. This is for people working on moduli of sheaves on Fano threefolds or on mathematical instanton bundles. I would send it to a serious referee and expect minor-to-moderate revision. For a reading group, it is a good choice if you want to see the inductive deformation trick in action.","headline":"A real non-emptiness theorem for higher-rank instantons on Fano threefolds; the main ideas are sound and the missing-reference gap is fixable.","tokens_in":50829,"tokens_out":9909,"would_cite":true,"duration_ms":79683,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14J60","14F06","14F08","14D21"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper establishes the minimal charge $k_0$ for slope-stable instanton bundles on Fano threefolds and constructs such bundles at and above it.","keywords":["Instanton bundle","Fano threefold","Moduli space of instantons","Restriction of stable sheaves","Curvilinear Kuznetsov category","Monads","Charge","Slope stability"],"falsifier":"Exhibit a slope-stable $(n,k)$-instanton sheaf with $k<k_0^n$ on a smooth Fano threefold of Picard rank one with $i_X\\ge 2$, or with $i_X=1$ and genus $g\\ge 4$ outside the singular-quadric exception; the theorem asserts none exist. The companion check is to verify the base: construct or rule out an unobstructed slope-stable rank-2 instanton bundle in each charge class used in the induction.","tokens_in":49705,"feed_emoji":"🧮","tokens_out":7208,"duration_ms":61208,"temperature":0.7,"pith_summary":"The paper defines higher-rank instanton sheaves on smooth Fano threefolds with Picard group $\\mathbb{Z}$ and shows that their topological type is governed by two integers: the rank $n$ and a charge $k$. Its central result pins down the minimal charge $k_0^n$—an explicit function of the threefold's genus and of $n$—for which slope-stable instanton bundles exist, and proves that for every $k \\ge k_0^n$ the variety carries unobstructed $\\mu$-stable $(n,k)$-instanton bundles, with only a short list of index-one exceptions left open. The same bundles restrict to a general anticanonical K3 surface as $\\mu$-stable sheaves, and their images in the K3 moduli space form Lagrangian subvarieties. This makes the instanton moduli spaces on Fano threefolds as well understood, at the level of nonemptiness and generic behaviour, as on projective space.","feed_headline":"Stable instantons appear once charge passes a threshold","feed_subtitle":"Unobstructed slope-stable instanton bundles of every rank exist once the charge reaches a threshold k0.","key_machinery":"The engine is an inductive extension-and-deformation process. Start from an unobstructed $\\mu$-stable $(n,k)$-instanton bundle $E$; because $\\chi(E,F_0)<0$ for $k$ above the threshold, there is a nontrivial extension $0\\to F_0\\to E'\\to E\\to 0$ (with $F_0=\\mathcal{O}$ for even index). Lemma 2.12 pins down the unique torsion-free slope-$\\mu(E')$ quotient of $E'$, Lemma 2.13 transfers unobstructedness, and a dimension count inside the moduli space of Gieseker-semistable or simple sheaves shows that $E'$ deforms to a $\\mu$-stable bundle with the same rank and charge. The base case is supplied by rank-2 't Hooft bundles, built via Serre correspondence from locally complete intersection curves: disjoint lines for $i_X=3,4$, families of unobstructed elliptic curves for $i_X=2$, and elliptic curves or rank-zero instanton transformations for $i_X=1$.","core_discovery":"The central claim is Main Theorem 1: for a smooth Fano threefold $X$ of Picard rank one and index $i_X$, writing the Chern character of an $(n,k)$-instanton as $\\gamma(n,k)=n\\,\\mathrm{ch}(F_0)-k\\,\\mathrm{ch}(\\mathcal{O}_l(q_X-1))$, the minimal charge is $k_0^n=\\lceil n/2\\rceil$ for $i_X=4$, $k_0^n=\\lceil n/3\\rceil$ for $i_X=3$, $k_0^n=n$ for $i_X=2$, and for $i_X=1$ it is $k_0^1=0$, $k_0^n=1$ for odd genus, and $k_0^n=n$ for even genus. For all $n\\ge 2-r_X$ and $k\\ge k_0^n$ the paper constructs unobstructed $\\mu$-stable instanton bundles, with generic splitting $0^n$ for even index and $(0^n,-1^n)$ for odd index, and proves that for $n\\ge 2$ no slope-stable $(n,k)$-instanton sheaves exist below $k_0^n$, except possibly for index one with genus $\\le 3$, or genus $4$ with $X$ lying in a singular quadric. The construction is inductive: rank-2 't Hooft bundles (obtained by Serre correspondence from disjoint lines or unobstructed elliptic curves) form the base, and each higher-rank instanton arises from a non-split extension by $\\mathcal{O}$ (even index) or by the minimal instanton $F_0$ (odd index), which is then deformed to a $\\mu$-stable bundle.","pith_inferences":["The index-one boundary cases with genus $\\le 3$ or the singular-quadric genus 4 appear to be failures of the rank-2 base-case existence results, not of the inductive step; if a single unobstructed rank-2 instanton in each needed charge class were found there, the same construction should go through verbatim.","The Lagrangian subvarieties in the K3 moduli spaces are natural places to transfer enumerative invariants from the instanton components, provided the restriction map is birational onto its image; this is a concrete property one could test next.","The monadic description for $H^3(X)=0$ is likely to extend to every Fano threefold whose derived category has a full exceptional collection, with the middle term dimension $w$ read off the same cohomology table, so checking the next such threefolds would be a direct extension of Proposition 7.1."],"forward_implications":["For $\\mathbb{P}^3$, the moduli space $\\mathrm{MI}_{\\mathbb{P}^3}(n,k)$ is nonempty for all $n\\ge 2$ and $k\\ge \\lceil n/2\\rceil$, with a generically smooth component of dimension $1-n^2+4nk$.","For Del Pezzo threefolds of index 2, $\\mathrm{MI}_X(n,k)$ is nonempty for $k\\ge n$, with a generically smooth component of dimension $2kn-n^2+1$.","For the quadric threefold, $\\mathrm{MI}_X(n,k)$ is nonempty for $k\\ge \\lceil n/3\\rceil$, with a generically smooth component of dimension $1-n^2+6nk$; for index-one threefolds the corresponding dimensions are $1-n^2+2nk$ (even genus) and $1+2nk$ (odd genus).","A general member of the constructed component has stable restriction to a general anticanonical K3 surface, and the restriction map cuts out a Lagrangian subvariety of the K3 moduli space.","For Fano threefolds with $H^3(X)=0$, every Gieseker-semistable $(n,k)$-instanton is the homology of a monad $F_{-1}^{\\oplus k}\\to W\\otimes F_0\\to F_1^{\\oplus k}$, giving an explicit linear-algebra description."],"supporting_citations":[{"why":"Originates the notion of instanton bundles on $\\mathbb{P}^3$ via the twistor transform, the foundational framework being generalized.","marker":"[AHDM78]"},{"why":"Introduces instanton bundles on Fano threefolds and raises the generic splitting question that the paper addresses for higher ranks.","marker":"[Kuz12]"},{"why":"Supplies the rank-2 instanton bundle constructions and the elementary-transformation/deformation method that the higher-rank induction extends.","marker":"[Fae14]"},{"why":"Provides moduli spaces of rank-2 ACM bundles on prime Fano threefolds, used as base cases for index-one instantons via ordinary lines.","marker":"[BF11]"},{"why":"Establishes existence of rank-2 instanton bundles for charges in the range $m_g,\\dots,g+3$ via elliptic curves, forming the index-one base cases.","marker":"[CFK24]"},{"why":"Gives the smoothing results for nodal elliptic curves used to produce unobstructed elliptic curves on Del Pezzo threefolds.","marker":"[HH85]"},{"why":"Treats rank-two stable sheaves on Fano threefolds of genus nine, supporting the canonical-resolution and restriction arguments in Section 6.","marker":"[BF13]"}],"fun_headline_variants":["Stable instanton bundles exist once charge hits threshold","Charge threshold yields stable higher-rank instantons","Existence of stable instantons begins at minimal charge","Higher-rank instanton stability above a charge bound","Minimal charge for stable instantons on Fano threefolds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole construction rests on the base case: for the needed charge ranges there must exist unobstructed slope-stable rank-2 instanton bundles, a fact imported from earlier work; if such a rank-2 bundle fails to exist on some boundary Fano threefold of index one, the higher-rank induction has no starting point.","fun_headline_variants_meta":{"raw":{"variants":["Stable instanton bundles exist once charge hits threshold","Charge threshold yields stable higher-rank instantons","Existence of stable instantons begins at minimal charge","Higher-rank instanton stability above a charge bound","Minimal charge for stable instantons on Fano threefolds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000649,"raw_usage":{"total_tokens":3068,"prompt_tokens":1124,"completion_tokens":1944,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":740,"completion_tokens_details":{"reasoning_tokens":1868}},"tokens_in":740,"tokens_out":1944,"duration_ms":12282,"temperature":1.0,"reasoning_tokens":1868,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T12:07:25.792514+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a slope-stable $(n,k)$-instanton sheaf with $k<k_0^n$ on a smooth Fano threefold of Picard rank one with $i_X\\ge 2$, or with $i_X=1$ and genus $g\\ge 4$ outside the singular-quadric exception; the theorem asserts none exist. The companion check is to verify the base: construct or rule out an unobstructed slope-stable rank-2 instanton bundle in each charge class used in the induction.","supporting_citations":[],"review_version":1}