A twisted derived category of hyper-K\"ahler varieties of $K3^{[n]}$-type

Ruxuan Zhang

REVIEW 2 major objections 4 minor 17 references

A twisted derived category of hyper-K\"ahler varieties of $K3^{[n]}$-type

T0 review · 2 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The Markman-Mukai lattice controls a natural twisted derived category of K3^[n]-type hyper-Kähler varieties under a primitivity condition.

desk verdict A serious, genuinely new twisted derived Torelli theorem for K3^[n]-type hyper-Kähler varieties under a primitive-span condition, with one apparent proof gap that dissolves on close reading. read the letter →

arxiv 2502.02143 v3 pith:G2C4AMA5 submitted 2025-02-04 math.AG

Ruxuan Zhang This is my paper · ORCID

classification math.AG MSC 14F0814J2814J4214F22

keywords $K3^{[n]}$-typehyper-KählervarietiesMarkman-MukailatticetwistedderivedcategoriesBrauerclassesmodulispacesofstablesheaveson$K3$surfacesequivalenceHodgeisometryprojectivelyhyperholomorphicbundles

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes that every hyper-Kähler variety of $K3^{[n]}$-type carries a canonical Brauer class, and that its twisted derived category is determined by the oriented Markman-Mukai lattice. The main theorem proves this when the span of the relevant lattice vectors is a primitive sublattice. As a consequence, for a projective $K3$ surface and a primitive Mukai vector $w$ with $w^2=2n-2\ge 2$, the twisted derived category of the moduli space $M_w$ is equivalent to the untwisted derived category of the Hilbert scheme $S^{[n]}$. In the fine case the Brauer classes vanish, answering a question about derived equivalence of fine moduli spaces of stable sheaves on $K3$ surfaces.

What carries the argument

The oriented Markman-Mukai lattice $(L(X),v)$ carries the argument: $L(X)$ extends $H^2(X,\mathbb{Z})$ by a rank-one primitive class $v$ with $v^2=2n-2$, and choosing $v$ fixes an orientation. Associated to $v$ is a canonical class $\theta_v\in H^2(X,\mu_{2n-2})$, defined through auxiliary classes $\delta_v$ with $\delta_v^2=2-2n$ and divisibility $2n-2$; the Brauer class of $\theta_v$ is the obstruction to $X$ being a fine moduli space. The paper also introduces the twisted extended Mukai lattice $\widetilde{H}(X,\delta_v/(2n-2),\mathbb{Z})\cong H^2(X,\mathbb{Z})\oplus U$, and uses integral isometries called Eichler transvections to move the orientation vector into a convenient position. The decisive criterion, Theorem 3.8, states that an orientation-preserving Hodge isometry between twisted extended Mukai lattices that sends a $\delta$-class to a $\delta$-class is realized by a twisted derived equivalence; this is powered by projectively hyperholomorphic bundles and by the twisted D-equivalence theorem for birational $K3^{[n]}$-type varieties.

What would settle it

Compute Hochschild homology for a pair satisfying the hypotheses of Theorem 1.4 but with a non-primitive span. Since derived equivalences preserve Hochschild homology, a pair for which these vector spaces differ would show the claimed equivalence fails.

Watch

Extended reading notes

Core claim

The central claim is Conjecture 1.2: if $X$ and $Y$ are hyper-Kähler varieties of $K3^{[n]}$-type and there is a Hodge isometry $\phi:(L(X),v)\to (L(Y),w)$ between oriented Markman-Mukai lattices, then $D^b(X,[n\theta_v])\cong D^b(Y,[\epsilon n\theta_w])$, where $\epsilon=\pm1$ according as $\phi$ preserves or reverses orientation. Here $L(X)$ is the Markman-Mukai lattice, an extension of $H^2(X,\mathbb{Z})$ by a primitive generator $v$ of square $2n-2$, and $\theta_v$ is the canonical class in $H^2(X,\mu_{2n-2})$ determined by that orientation. The paper proves the conjecture when $\operatorname{Span}(\phi(v),w)\subset L(Y)$ is a primitive lattice embedding. The proof builds a Hodge isometry of twisted extended Mukai lattices and then uses a projectively hyperholomorphic bundle construction together with the twisted D-equivalence theorem to upgrade this isometry to a twisted derived equivalence. The main application is that for a projective $K3$ surface $S$ and a primitive class $w$ with $w^2=2n-2\ge 2$, one has $D^b(M_w,[n\theta_w])\cong D^b(M_w,[-n\theta_w])\cong D^b(S^{[n]})$.

Load-bearing premise

The proof relies, without reproving, on the twisted D-equivalence theorem and Markman's projectively hyperholomorphic bundle theorem; if either fails in the range used here, the derived equivalences in Theorem 1.4 and Corollary 1.6 do not follow.

Editorial extensions

If this is right

Theorem 1.4 gives a twisted derived Torelli statement under the primitivity hypothesis: a Hodge isometry of oriented Markman-Mukai lattices forces an equivalence of the corresponding twisted derived categories.
For any projective $K3$ surface and primitive Mukai vector $w$ with $w^2=2n-2\ge 2$, the twisted derived category of $M_w$ is derived equivalent to the Hilbert scheme $S^{[n]}$.
When $\operatorname{div}(w)=1$, the Brauer class vanishes and $D^b(M_w)\cong D^b(M_v)$ for any fine $v$ with the same square; in particular, any two fine moduli spaces of stable sheaves on a $K3$ surface of the same dimension are derived equivalent.
The criterion in Theorem 3.8 provides a reusable template for lifting Hodge isometries of twisted extended Mukai lattices to twisted derived equivalences for other $K3^{[n]}$-type pairs.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If Conjecture 1.2 holds without the primitivity hypothesis, then the full twisted derived Torelli theorem for $K3^{[n]}$-type would follow, suggesting that the conjectural $K3$ category associated to a hyper-Kähler variety is itself determined by the oriented Markman-Mukai lattice.
The same lattice-governed mechanism may extend to moduli spaces of Bridgeland-stable objects on arbitrary $K3$ categories, with the canonical Brauer class controlling which twists appear when the moduli space is not fine.
The boundary of the theorem could be mapped by testing explicit pairs with non-primitive spans, for instance low-dimensional examples where the twisted derived equivalence can be checked by direct Fourier–Mukai kernels.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Here's my take. The paper does something real: it formulates a natural conjecture that the twisted derived category of a K3^[n]-type hyper-Kähler variety is controlled by its oriented Markman-Mukai lattice, and proves it under a primitive-span condition. The main theorem, Theorem 1.4, is not a restatement of known results. The canonical Brauer class θ_v and the criterion in Theorem 3.8 are new, and the application to fine moduli spaces (Corollary 1.6) answers a question of Huybrechts in the stated cases. The proof strategy is honest: it builds on Markman's projectively hyperholomorphic bundles and the twisted D-equivalence theorem of [16], and it does not hide that dependency.

The soft spots are real but not disqualifying. Theorem 3.8 is a long argument with several 'direct computation' steps, and the decisive inputs — [16], Markman's theorems, Yoshioka's numerical condition — are imported rather than reproved. That is acceptable when the inputs are published and independent, but a referee should check the range of applicability, especially condition (12). The paper also has the occasional typo (Step 5 of Theorem 3.8 says 'concluded by Theorem 3.8' where it should refer to Theorem 3.2), and the compressed computations in Section 3.5 deserve scrutiny. The full conjecture remains open; the theorem covers a special but natural case.

One concern I saw in the stress-test does not hold up. The claim that Step 2 of Theorem 3.8 needs a γ with pairing 1 with η_Y(β) and that this may not exist because η_Y(β) might be non-primitive is not compatible with the text: β is explicitly taken to be a primitive class in H^2(Y,Z), and the marking η_Y is an isometry. Since η_Y(β) lies in Λ_K3, primitivity in the full lattice implies primitivity in Λ_K3, so the unimodularity of Λ_K3 gives the required γ. The worry evaporates on a careful reading.

The citation pattern is fine. [16] is a coauthored paper, but it is a published independent theorem, and the paper does not derive it from Conjecture 1.2. There is no fitted data or circular reasoning.

This is a paper for people working on derived categories and moduli of sheaves on hyper-Kähler varieties. It deserves a serious referee. I would send it out and ask for verification of the external inputs and the omitted computations. My own verdict is close to accept, with moderate confidence.

Referee Report

2 major / 4 minor

Summary. The paper proposes Conjecture 1.2: for hyper-Kähler varieties X and Y of K3^[n]-type, an oriented Hodge isometry between their Markman-Mukai lattices (L(X),v) and (L(Y),w) should induce an equivalence of twisted derived categories D^b(X,[nθ_v]) and D^b(Y,[ϵnθ_w]). It introduces a canonical Brauer class θ_v, a twisted extended Mukai lattice, and proves a lattice-level Hodge isometry (Theorem 1.3). The main new theorem, Theorem 1.4, asserts that Conjecture 1.2 holds when Span(ϕ(v),w) is a primitive sublattice of L(Y); this is deduced from a derived Torelli-type criterion (Theorem 3.8) built on Markman's projectively hyperholomorphic bundles and the twisted D-equivalence theorem of [16]. Sections 4 and 5 apply Theorem 1.4 to moduli spaces of stable objects on K3 surfaces, yielding Theorem 1.5 (twisted derived equivalences between M_w and S^[n]) and Corollary 1.6 (untwisted derived equivalences for fine moduli spaces of the same dimension).

Significance. If the proof were fully correct, the paper would provide substantial evidence for Conjecture 1.2 and would answer a question of Huybrechts on derived equivalences between fine moduli spaces of stable sheaves on K3 surfaces. The formulation of the natural Brauer class θ_v and the twisted extended Mukai lattice is elegant, the lattice-theoretic computations in Section 2 are explicit, and the paper introduces no fitted parameters. These are genuine strengths. However, the central derivation is conditional on substantial external inputs ([12], [16], [17]) and, more seriously, on a step in the proof of Theorem 3.8 that is not justified as written. The claims are therefore not fully established in the present form.

major comments (2)

[Section 3.5, Theorem 3.8, Step 2] The proof chooses an integral vector γ ∈ Λ_K3 satisfying ⟨η_Y(β),γ⟩ = 1, and then defines v_k using the B-field shift e_{kγ}. The displayed formula for s_k = s + km + rk^2γ^2/2 is valid only when ⟨η_Y(β),γ⟩ = 1, because the definition of e_B in Section 2.3 contains the term ⟨c,B⟩ = km⟨η_Y(β),γ⟩ in the f-coefficient. Existence of such γ is equivalent to primitivity of η_Y(β) in the even unimodular lattice Λ_K3. The hypotheses of Theorem 3.8 imply only that ψ(f) = re + mβ + sf is primitive, i.e. gcd(r, s, m·div(β)) = 1, and that ψ(δ_v) = δ_w gives ⟨δ_w, β⟩ = 0; neither condition forces η_Y(β) to be primitive. The numerical conditions are compatible, for instance, with β = 2h for a primitive class h with h ⊥ δ_w, for which no such γ exists. Without γ, the construction of the K3 surface S, the coprime shift k, and the fine moduli space M in Step 3 collapses. Since Theorem 1.4 invokes Theorem 3.8 directly, this is a load-bearing gap in the proof of the main theorem.
[Section 3.2, Theorem 3.1] Theorem 3.1 is stated as a slight generalization of the main result of [16], but its proof is condensed to a reference to Markman's decomposition of parallel transport operators into prime exceptional reflections and birational maps, with no derivation that prime exceptional reflections act by twisted derived equivalences with the stated B-field. Theorem 3.1 is used in Cases 1 and 3 of Theorem 3.8, so this is not a purely cosmetic omission. The proof should be expanded, or the exact statement in [16] that implies the theorem should be quoted.

minor comments (4)

[Section 3.5, end of Step 5] The sentence 'This case is now concluded by Theorem 3.8' is self-referential; it should read 'by Theorem 3.2'.
[Throughout] The notation for inner products is inconsistent: some pairings are written with periods (e.g., ⟨e.f⟩, ⟨δ_v.v⟩) and others with commas (e.g., ⟨η_Y(β),γ⟩). Please unify the notation.
[Section 4.1, Lemma 4.1] The assertion that x,y ∈ Z follows immediately from primitivity of v, w, and H is not immediate, since the factor 1/k can divide numerators. Please add a sentence explaining why k divides y as well as x.
[Introduction, abstract] The abstract says 'a natural twisted derived category of any hyper-Kähler variety of K3^[n]-type is controlled by its Markman-Mukai lattice'; this is the conjecture, not an established fact, and the wording should make the conditional status clearer.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: Theorem 1.4 is reduced to independent external theorems and explicit lattice constructions; the author-overlapping citation [16] is independent support, not a circular input.

full rationale

No circular step is present in the claimed derivation chain. Theorem 1.4 is obtained from Theorem 3.8, whose proof invokes Theorem 3.2, assembled from Markman's projectively hyperholomorphic bundle theorem [12], Kapustka–Kapustka [9], and the twisted D-equivalence theorem [16]. The only author-overlapping citation is [16] (Maulik–Shen–Yin–Zhang), but it is an independent published theorem with its own proof; its assumptions do not include Conjecture 1.2, and the present paper does not use Conjecture 1.2 to prove it. Markman's [12] and Yoshioka's [17] are likewise external. The lattice-theoretic evidence (Theorem 1.3, Section 2) is an explicit isometry construction with Eichler transvections and δ-classes; no parameter is fitted to the desired equivalence, and no equation is imposed to equal the conclusion. In Theorem 1.4, the primitive-embedding hypothesis is used to construct a δ-class δ_w with ψ(δ_v)=δ_w; this is shown in the text and is not a restatement of the desired derived equivalence. The skeptical issue about the existence of γ with ⟨η_Y(β),γ⟩=1 in Theorem 3.8 Step 2 is a possible proof gap about primitivity, not a circular reduction of the conclusion to its inputs. Overall, the derivation chain reduces to prior independent results and an explicit lattice argument, so it is not circular.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

No empirical constants or fitted parameters appear; all choices, such as δ-classes and the integer t, are constructed in the proof and do not affect the statements. No speculative entities are introduced: the canonical class θ_v and oriented lattice are constructed from established lattice theory, not postulated.

assumptions (8)

domain assumption Markman's construction of the Markman-Mukai lattice L(X) with H^2(X,Z) ⊂ L(X), its Hodge structure, and the parallel transport criterion.
Invoked throughout Section 2.1 and in Lemma 3.6 to pass from integral Hodge isometries preserving δ-classes to parallel transport operators.
domain assumption Markman's projectively hyperholomorphic bundle theorem: existence of universal bundle U^[n] and rational Hodge isometry F_{U^[n]}.
Used in Theorem 3.2 and Example 3.5 to construct the model Hilbert schemes M^[n], S^[n] and the associated Hodge isometries.
domain assumption Twisted D-equivalence theorem for birational and parallel-transport related hyper-Kähler varieties of K3^[n]-type.
The main engine of the paper: gives the initial twisted derived equivalences in Subsection 3.1 and Theorem 3.1, and is cited in Theorem 3.2. [16] is a published, independent theorem; the present author is a coauthor, but the current paper does not use Conjecture 1.2 to prove it.
domain assumption Yoshioka's criterion: the numerical condition (12) implies the moduli space consists of stable vector bundles.
Used in Theorem 3.2 setup and in Case 2 Step 3 to ensure M is a moduli space of vector bundles.
standard math Dirichlet's theorem on primes in arithmetic progressions.
Used in Lemma 4.2 to choose infinitely many t with k'+r't prime and coprime to r^2-1.
standard math Witt's extension theorem for isometries of primitive sublattices of unimodular lattices.
Used in the proof of Theorem 1.4 to extend the isometry g' of Span(ϕ(v),w) to an isometry g ∈ O(L(Y)).
standard math Surjectivity of the period map and global Torelli theorem for K3 surfaces and for hyper-Kähler varieties of K3^[n]-type.
Used in Theorem 3.8 Case 2 Step 2 to realize lattice vectors as Picard classes and to identify M with M_{v0} via Hodge isometries.
standard math Lefschetz (1,1)-theorem: rational (1,1) classes on a projective complex manifold are rational multiples of integral classes in the Néron-Severi group.
Used in Lemma 3.3 to conclude [mβ/r - B] = 0 in Br(Y) from orthogonality to H^{2,0} and H^{0,2}.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A twisted derived category of hyper-K\"ahler varieties of $K3^{[n]}$-type." pith.science (2026). https://pith.science/paper/G2C4AMA5

@misc{pith2026250202143,
  author       = {Pith},
  title        = {Pith review of: A twisted derived category of hyper-K\"ahler varieties of $K3^[n]$-type},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G2C4AMA5}},
  note         = {Machine review of arXiv:2502.02143}
}

abstract

We conjecture that a natural twisted derived category of any hyper-K\"ahler variety of $K3^{[n]}$-type is controlled by its Markman-Mukai lattice. We prove the conjecture under numerical constraints, and our proof relies heavily on Markman's projectively hyperholomorphic bundle and a recently proven twisted version of the D-equivalence conjecture. In particular, we prove that any two fine moduli spaces of stable sheaves on a $K3$ surface are derived equivalent if they have the same dimension.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 13 canonical work pages

[16]

The D-equivalence conjecture for hyper- k¨ ahler varieties via hyperholomorphic bundles.Invent

Davesh Maulik, Junliang Shen, Qizheng Yin, and Ruxuan Zhang. The D-equivalence conjecture for hyper- k¨ ahler varieties via hyperholomorphic bundles.Invent. Math. , 2025

work page 2025
[12]

Rational Hodge isometries of hyper-K¨ ahler varieties ofK3[n] type are algebraic

Eyal Markman. Rational Hodge isometries of hyper-K¨ ahler varieties ofK3[n] type are algebraic. Compos. Math., 160(6):1261–1303, 2024

work page 2024
[17]

Stability and the Fourier-Mukai transform

K¯ ota Yoshioka. Stability and the Fourier-Mukai transform. II.Compos. Math., 145(1):112–142, 2009. Fudan University Email address : rxzhang18@fudan.edu.cn

work page 2009
[1]

Mori cones of holomorphic symplectic varieties of K3 type

Arend Bayer, Brendan Hassett, and Yuri Tschinkel. Mori cones of holomorphic symplectic varieties of K3 type. Ann. Sci. ´Ec. Norm. Sup´ er. (4), 48(4):941–950, 2015

work page 2015
[2]

Derived categories of hyper-K¨ ahler manifolds: extended Mukai vector and integral structure

Thorsten Beckmann. Derived categories of hyper-K¨ ahler manifolds: extended Mukai vector and integral structure. Compos. Math., 159(1):109–152, 2023

work page 2023
[3]

Derived categories of Fano varieties of lines

Alessio Bottini and Daniel Huybrechts. Derived categories of Fano varieties of lines. arXiv preprint arXiv:2501.03534, 2024

arXiv 2024
[4]

Every rational Hodge isometry between two K3 surfaces is algebraic

Nikolay Buskin. Every rational Hodge isometry between two K3 surfaces is algebraic. J. Reine Angew. Math., 755:127–150, 2019

work page 2019
[5]

Nonfine moduli spaces of sheaves on K3 surfaces

Andrei C˘ ald˘ araru. Nonfine moduli spaces of sheaves on K3 surfaces. Int. Math. Res. Not. , pages 1027– 1056, 2002

work page 2002

Show all 17 references

[6]

Gritsenko, K

V. Gritsenko, K. Hulek, and G. K. Sankaran. Abelianisation of orthogonal groups and the fundamental group of modular varieties. J. Algebra, 322(2):463–478, 2009

2009
[7]

Lectures on K3 surfaces, volume 158 of Cambridge Studies in Advanced Mathematics

Daniel Huybrechts. Lectures on K3 surfaces, volume 158 of Cambridge Studies in Advanced Mathematics . Cambridge University Press, Cambridge, 2016

2016
[8]

D. G. James. On Witt’s theorem for unimodular quadratic forms. Pacific J. Math. , 26:303–316, 1968

1968
[9]

Constructions of derived equivalent hyper-K¨ ahler ahler four- folds

Grzegorz Kapustka and Micha l Kapustka. Constructions of derived equivalent hyper-K¨ ahler ahler four- folds. arXiv preprint arXiv:2312.14543 , 2023

2023 arXiv
[10]

A survey of Torelli and monodromy results for holomorphic-symplectic varieties

Eyal Markman. A survey of Torelli and monodromy results for holomorphic-symplectic varieties. In Com- plex and differential geometry , volume 8 of Springer Proc. Math. , pages 257–322. Springer, Heidelberg, 2011

2011
[11]

The Beauville-Bogomolov class as a characteristic class

Eyal Markman. The Beauville-Bogomolov class as a characteristic class. J. Algebraic Geom. , 29(2):199– 245, 2020

2020
[13]

Stable vector bundles on a hyper-K¨ ahler manifold with a rank 1 obstruction map are modular

Eyal Markman. Stable vector bundles on a hyper-K¨ ahler manifold with a rank 1 obstruction map are modular. Kyoto J. Math. , 64(3):635–742, 2024

2024
[14]

Integral transforms and deformations of k3 surfaces

Eyal Markman and Sukhendu Mehrotra. Integral transforms and deformations of k3 surfaces. arXiv preprint arXiv:1507.03108, 2015

2015 arXiv
[15]

Obstruction classes for moduli spaces of sheaves and lagrangian fibrations

Dominique Mattei and Reinder Meinsma. Obstruction classes for moduli spaces of sheaves and lagrangian fibrations. arXiv preprint arXiv:2404.16652 , 2024

2024 arXiv

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.

[1] [16]

The D-equivalence conjecture for hyper- k¨ ahler varieties via hyperholomorphic bundles.Invent

Davesh Maulik, Junliang Shen, Qizheng Yin, and Ruxuan Zhang. The D-equivalence conjecture for hyper- k¨ ahler varieties via hyperholomorphic bundles.Invent. Math. , 2025

work page 2025

[2] [12]

Rational Hodge isometries of hyper-K¨ ahler varieties ofK3[n] type are algebraic

Eyal Markman. Rational Hodge isometries of hyper-K¨ ahler varieties ofK3[n] type are algebraic. Compos. Math., 160(6):1261–1303, 2024

work page 2024

[3] [17]

Stability and the Fourier-Mukai transform

K¯ ota Yoshioka. Stability and the Fourier-Mukai transform. II.Compos. Math., 145(1):112–142, 2009. Fudan University Email address : rxzhang18@fudan.edu.cn

work page 2009

[4] [1]

Mori cones of holomorphic symplectic varieties of K3 type

Arend Bayer, Brendan Hassett, and Yuri Tschinkel. Mori cones of holomorphic symplectic varieties of K3 type. Ann. Sci. ´Ec. Norm. Sup´ er. (4), 48(4):941–950, 2015

work page 2015

[5] [2]

Derived categories of hyper-K¨ ahler manifolds: extended Mukai vector and integral structure

Thorsten Beckmann. Derived categories of hyper-K¨ ahler manifolds: extended Mukai vector and integral structure. Compos. Math., 159(1):109–152, 2023

work page 2023

[6] [3]

Derived categories of Fano varieties of lines

Alessio Bottini and Daniel Huybrechts. Derived categories of Fano varieties of lines. arXiv preprint arXiv:2501.03534, 2024

arXiv 2024

[7] [4]

Every rational Hodge isometry between two K3 surfaces is algebraic

Nikolay Buskin. Every rational Hodge isometry between two K3 surfaces is algebraic. J. Reine Angew. Math., 755:127–150, 2019

work page 2019

[8] [5]

Nonfine moduli spaces of sheaves on K3 surfaces

Andrei C˘ ald˘ araru. Nonfine moduli spaces of sheaves on K3 surfaces. Int. Math. Res. Not. , pages 1027– 1056, 2002

work page 2002