betaFeedback

WorksMathematics

Twisted Fourier-Mukai partners of genus 1 curves

Abstract

We study twisted derived equivalence between genus 1 curves over an arbitrarly perfect field.

Introduction

In this paper we study twisted Fourier–Mukai partners of genus 1 curves over a perfect field k\mathbb{k}.

We recall that a twisted variety over k\mathbb{k} is given by a pair (X,α)(X,\alpha), with XX a smooth projective geometrically connected variety over k\mathbb{k} and α∈Br⁡(X)\alpha\in\operatorname{Br}(X) an element of its Brauer group. Given a twisted variety, there is a Gm\mathbb{G}_{m}-gerbe π:X→X\pi:\mathcal{X}\to X with class α\alpha, unique up to isomorphism. We write Db(X,α)=Db(X)(1)D^{b}(X,\alpha)=D^{b}(\mathcal{X})^{(1)} for the derived category of 1-twisted sheaves. We say that two varieties XX and YY are twisted derived equivalent if there are two Brauer classes α∈Br⁡(X)\alpha\in\operatorname{Br}(X) and β∈Br⁡(Y)\beta\in\operatorname{Br}(Y) and an exact equivalence

Φ:Db(X,α)→Db(Y,β).\Phi:D^{b}(X,\alpha)\to D^{b}(Y,\beta).

We say that two varieties XX and YY are derived isogenous if there are a finite sequence of varieties

X≃X0,X1,…,Xn≃YX\simeq X_{0},X_{1},\ldots,X_{n}\simeq Y

and Brauer classes βi∈Br⁡(Xi)\beta_{i}\in\operatorname{Br}(X_{i}) and αi+1∈Br⁡(Xi+1)\alpha_{i+1}\in\operatorname{Br}(X_{i+1}) such that Db(Xi,βi)≃Db(Xi+1,αi+1)D^{b}(X_{i},\beta_{i})\simeq D^{b}(X_{i+1},\alpha_{i+1}) for 0≤i<n0\leq i<n. Investigating the hidden relationships between twisted derived equivalent or derived isogenous varieties is a question that has raised, both recently and less recently, much interest (see for example [2 LLT25, 6, 5]).

In this paper we completely characterize derived isogenous genus 1 curves, generalizing the results of [1] to the twisted setting.

Main Theorem. Let k\mathbb{k} be a perfect field and let C,C′C,C' be genus 1 curves over k\mathbb{k}. The following are equivalent:

  1. CC and C′C' are derived isogenous;

  2. the Jacobians J(C)J(C) and J(C′)J(C') are isomorphic over k\mathbb{k};

  3. CC and C′C' are connected by a chain (of length two) of twisted derived equivalences passing through J(C)J(C).

Ramachandran and Rosenberg in [14] show that a genus 1 curve CC with Jacobian JJ satisfies Db(C)≃Db(J,αC)D^{b}(C)\simeq D^{b}(J,\alpha_{C}) for a Brauer class αC\alpha_{C} determined by the class of CC in H1(k,J)H^{1}(\mathbb{k},J).

Hence any two genus 1 curves with isomorphic Jacobians are derived isogenous. From this, the implications (2) ⇒\Rightarrow (3) ⇒\Rightarrow (1) in the above statement follow easily. Our main input is the proof that derived isogenous genus 1 curves have the same Jacobian, which will be a consequence of the following:

Theorem. If CC and C′C' are twisted derived equivalent genus 1 curves, then there is an isomorphism between their Jacobians J(C)≃J(C′)J(C) \simeq J(C').

Our strategy is to leverage the main construction of [8], where, starting from a twisted variety (X,α)(X,\alpha), the author provides a group algebraic space O(X,α)\mathrm{O}(X,\alpha) which is a twisted derived invariant. More precisely, given a twisted equivalence Φ:Db(X,α)→Db(Y,β)\Phi: D^{\mathrm{b}}(X,\alpha) \to D^{\mathrm{b}}(Y,\beta), Olsson constructs an isomorphism of group algebraic spaces

RΦ:O(X,α)⟶O(Y,β).R_{\Phi} : \mathrm{O}(X,\alpha) \longrightarrow\mathrm{O}(Y,\beta).

When XX and YY are genus 1 curves, using that the classes α\alpha and β\beta trivialize when passing to the algebraic closure of kk, we construct, in § 3, isomorphisms

θτ:J(X)×J(X)^⟶O(X,α),θτ′:J(Y)×J(Y)^⟶O(Y,β)\theta_{\tau} : J(X) \times\widehat{J(X)} \longrightarrow\mathrm{O}(X,\alpha), \qquad\theta_{\tau'} : J(Y) \times\widehat{J(Y)} \longrightarrow\mathrm{O}(Y,\beta)

such that the composition θτ′−1∘RΦ∘θτ\theta_{\tau'}^{-1} \circ R_{\Phi} \circ\theta_{\tau} is symplectic in the sense of [11] (see also Subsection 2.1 for more details). We can then conclude thanks to Proposition 2.2, which shows that a symplectic isomorphism E×E^⟶E′×E′^E \times\widehat{E} \longrightarrow E' \times\widehat{E'}, where EE and E′E' are elliptic curves, yields an isomorphism E≃E′E \simeq E'.

Notation. We work over a perfect field kk with algebraic closure k‾\overline{k} and absolute Galois group Gk=Gal⁡(k‾/k)G_k = \operatorname{Gal}(\overline{k}/k). Given XX a reduced scheme over kk we denote by X‾\overline{X} its base extension over k‾\overline{k}. For γ∈Gk\gamma\in G_k and an object or morphism ZZ defined over kk we write γZ{}^{\gamma}Z for its Galois conjugate.

A genus 11 curve is a smooth projective geometrically connected curve of genus 1;1; its Jacobian J(C)=Pic⁡C0J(C) = \operatorname{Pic}^{0}_{C} is an elliptic curve.

Given an abelian variety AA over kk, we denote by A^\widehat{A} its dual, and we identify A^^=A\widehat{\widehat{A}} = A. Given a homomorphism f:A→Bf : A \to B of abelian varieties over kk, we denote by f^:B^→A^\widehat{f} : \widehat{B} \to\widehat{A} its dual. For an integer mm we denote by [m]:A→A[m] : A \to A the multiplication by mm. For an elliptic curve EE we write λE:E→E^\lambda_E : E \to\widehat{E} for its canonical principal polarization; recall that λ^E=λE\widehat{\lambda}_E = \lambda_E.

Acknowledgments. The second and third authors would like to thanks D. Rydh for valuable mathematics conversation and M. Olsson for a very nice e-mail exchange. The second author was partially supported by the grant number VR2023-03837 of the Vetenskapsrådet and by the grant number 2023.0395 of the Knut and Alice Wallenberg foundation.

AI Usage. Claude Opus 5.5 was used to check for correand polish the exposition.

Preliminary results and Background Material

Abelian varieties and Symplectic Isomorphisms

Let AA and BB be abelian varieties over kk. A morphism f:A×A^→B×B^f : A \times\widehat{A} \to B \times\widehat{B} can be represented by a matrix of morphisms:

f=(xyzw),f = \begin{pmatrix} x & y \\ z & w \end{pmatrix},

with x:A→Bx : A \to B, y:A^→By : \widehat{A} \to B, z:A→B^z : A \to\widehat{B} and w:A^→B^w : \widehat{A} \to\widehat{B}. We recall the following definition.

Definition 2.1. An isomorphism f:A×A^→B×B^f : A \times\widehat{A} \to B \times\widehat{B} is said to be symplectic (or isometric) if

f−1=(w^−y^−z^x^).f^{-1} = \begin{pmatrix} \widehat{w} & -\widehat{y} \\ -\widehat{z} & \widehat{x} \end{pmatrix}.

The fact that two abelian varieties AA and BB are tied by a symplectic isomorphism A×A^≃B×B^A \times\widehat{A} \simeq B \times\widehat{B} is very strong: over an algebraically closed field of characteristic not 2, this is equivalent to AA and BB being derived equivalent ([Orl02, Pol96, Pol03]). Since for elliptic curves derived equivalence implies isomorphism ([1]), we directly get that elliptic curves tied by a symplectic isomorphism are isomorphic over (most) algebraically closed fields. The next result says that the same is true over any field.

Proposition 2.2. Let kk be a field and let E1E_1, E2E_2 be elliptic curves over kk. If there is a symplectic isomorphism f:E1×E^1→E2×E^2f:E_1 \times\widehat{E}_1 \to E_2 \times\widehat{E}_2 over kk, then E1E_1 and E2E_2 are isomorphic over kk.

Proof. Write ff in matrix form, that is f=(xyzw)f = \begin{pmatrix} x & y \\ z & w \end{pmatrix}, and set a:=deg⁡xa := \deg x, b:=deg⁡yb := \deg y, c:=deg⁡zc := \deg z and d:=deg⁡wd := \deg w. Using the principal polarizations λEi:Ei→E^i\lambda_{E_i}: E_i \to\widehat{E}_i we identify E^i\widehat{E}_i with EiE_i. Then x,y,z,wx,y,z,w become elements of Hom⁡k(E1,E2)\operatorname{Hom}_k(E_1,E_2), and the dual of a homomorphism u:E1→E2u:E_1 \to E_2 becomes its dual isogeny λE1−1∘u^∘λE2:E2→E1\lambda_{E_1}^{-1} \circ\widehat{u} \circ\lambda_{E_2}: E_2 \to E_1, which we still denote by u^\widehat{u}; recall that u^∘u=[deg⁡u]\widehat{u} \circ u = [\deg u] and u∘u^=[deg⁡u]u \circ\widehat{u} = [\deg u].

Since ff is symplectic and the isomorphisms λEi\lambda_{E_i} are symmetric, f−1f^{-1} can be expressed as in (2.1). Comparing the entries of the first row of f∘f−1=Id⁡E2×E^2f \circ f^{-1} = \operatorname{Id}_{E_2 \times\widehat{E}_2} gives

x∘w^−y∘z^=1E2,y∘x^=x∘y^.(1)x \circ\widehat{w} - y \circ\widehat{z} = 1_{E_2}, \qquad y \circ\widehat{x} = x \circ\widehat{y}. \tag*{(1)}

If x=0x = 0, the first equation gives −y∘z^=1E2-y \circ\widehat{z} = 1_{E_2}, so yy is an isomorphism and E1E_1 and E2E_2 are isomorphic. Similarly, if z=0z = 0, then x∘w^=1E2x \circ\widehat{w} = 1_{E_2} and xx is an isomorphism. Therefore we can assume that x,z≠0x,z \ne0, and thus a,c≠0a,c \ne0.

As the dual of y∘x^y \circ\widehat{x} is x∘y^x \circ\widehat{y}, the second equation in (2.2) implies that y∘x^y \circ\widehat{x} is self-dual, so it must be multiplication by an integer nn. Indeed, the map u↦u^u \mapsto\widehat{u} is the Rosati involution of the canonical principal polarization, whose fixed points in End⁡(E2)\operatorname{End}(E_2) are the integers (see e.g. [Sil09, III.6.2 and §III.9]).

By comparing degrees, we have n2=abn^2 = ab; in particular abab is a perfect square. Now we compose both sides of the second equation of (2.2) on the right with xx. Using that x^∘x=[a]\widehat{x} \circ x = [a], we get

ay=nx.(2)ay = nx. \tag*{(2)}

Now we turn our attention to the second row of f∘f−1f \circ f^{-1}, which gives the equations

z∘w^=w∘z^,w∘x^−z∘y^=1E2.(3)z \circ\widehat{w} = w \circ\widehat{z}, \qquad w \circ\widehat{x} - z \circ\widehat{y} = 1_{E_2}. \tag*{(3)}

As before, z∘w^z \circ\widehat{w} is self-dual, hence multiplication by an integer mm with m2=cdm^2 = cd. Composing on the right with zz, and using z^∘z=[c]\widehat{z} \circ z = [c], we get

cw=mz.(4)cw = mz. \tag*{(4)}

By (2.3) and (2.5) we have y=naxy = \frac{n}{a}x in Hom⁡(E1,E2)⊗Q\operatorname{Hom}(E_1,E_2) \otimes\mathbb{Q}, and, taking duals, w^=mcz^\widehat{w} = \frac{m}{c}\widehat{z} in Hom⁡(E2,E1)⊗Q\operatorname{Hom}(E_2,E_1) \otimes\mathbb{Q}. Substituting into the first equation of (2.2), we get the following equalities in End⁡(E2)⊗Q\operatorname{End}(E_2) \otimes\mathbb{Q}:

1E2=x∘(mcz^)−nax∘z^=(mc−na)x∘z^.(5)\begin{aligned} 1_{E_2} &= x \circ\left(\frac{m}{c}\widehat{z}\right) - \frac{n}{a}x \circ\widehat{z} \\ &= \left(\frac{m}{c} - \frac{n}{a}\right)x \circ\widehat{z}. \tag*{(5)} \end{aligned}

This means that x∘z^x \circ\widehat{z} is a rational multiple hh of the identity. Comparing degrees, we get h2=ach^2 = ac, and therefore hh is an integer. Now observe that, from (2.6), we get

1=hmc−hna.1 = \frac{hm}{c} - \frac{hn}{a}.

Put p:=hm/cp := hm/c and q:=hn/aq := hn/a. These are rational numbers with

p2=h2m2c2=ac⋅cdc2=ad,q2=h2n2a2=ac⋅aba2=bc,p^2 = \frac{h^2m^2}{c^2} = \frac{ac \cdot cd}{c^2} = ad,\qquad q^2 = \frac{h^2n^2}{a^2} = \frac{ac \cdot ab}{a^2} = bc,

so pp and qq are integers. Since p−q=1p-q=1, the integers pp and qq are relatively prime. Therefore the same is true for their squares p2=adp^2=ad and q2=bcq^2=bc. Any common divisor of aa and bb divides both adad and bcbc. Thus we deduce that aa and bb are coprime. As ab=n2ab=n^2 is a perfect square, aa and bb are themselves perfect squares. Write a=r2a=r^2 with rr a positive integer. Since r2=ar^2=a divides n2=abn^2=ab, the integer rr divides nn. Put s:=n/r∈Zs:=n/r\in\mathbb{Z}, so that s2=bs^2=b. Then (2) reads r2y=rsxr^2y=rsx. Since Hom⁡k(E1,E2)\operatorname{Hom}_k(E_1,E_2) is torsion free and r≠0r\ne0, we get

ry=sx.ry=sx.

Moreover rr and ss are coprime, because r2=ar^2=a and s2=bs^2=b are. Choose u,v∈Zu,v\in\mathbb{Z} with ur+vs=1ur+vs=1. Then

x=(ur+vs)x=urx+vsx=urx+vry=r(ux+vy).x=(ur+vs)x=urx+vsx=urx+vry=r(ux+vy).

Again comparing degrees we have

a=deg⁡x=r2deg⁡(ux+vy)=adeg⁡(ux+vy).a=\deg x=r^2\deg(ux+vy)=a\deg(ux+vy).

Therefore deg⁡(ux+vy)=1\deg(ux+vy)=1 because a≠0a\ne0. Hence ux+vy:E1→E2ux+vy:E_1\to E_2 is an isomorphism defined over kk, which concludes the proof. □

We conclude by remarking that a variant of the proof of Proposition 2.2 works for principally polarized abelian varieties AA and BB over kk, with polarizations λA\lambda_A and λB\lambda_B, whose only endomorphisms fixed by the Rosati involution are the integers, i.e. with Hom⁡ksym(A,A^)=ZλA\operatorname{Hom}^{\mathrm{sym}}_k(A,\widehat{A})=\mathbb{Z}\lambda_A and Hom⁡ksym(B,B^)=ZλB\operatorname{Hom}^{\mathrm{sym}}_k(B,\widehat{B})=\mathbb{Z}\lambda_B. In that case the self-dual composites y∘x^y\circ\widehat{x} and x^∘x\widehat{x}\circ x are integers by hypothesis, the second one nonzero when x≠0x\ne0 by positivity of the Rosati involution. This shows that xx and yy are proportional, and one concludes by comparing degrees in the first equation of (1), using dim⁡A=dim⁡B\dim A=\dim B.

Olsson groups

In this paragraph we review the main construction in [8]. Before beginning, we observe that the introduction of [8] is phrased over an algebraically closed field, but the constructions and proofs of the results we use are carried out over an arbitrary field; see [8 3.1 and §5].

Let XX be a smooth projective geometrically connected variety over kk and let π:X→X\pi:\mathcal{X}\to X be a Gm\mathbb{G}_m-gerbe with class α∈Br⁡(X)\alpha\in\operatorname{Br}(X). As in the introduction, we write Db(X,α)=Db(X)D^b(X,\alpha)=D^b(\mathcal{X}) for the derived category of 1-twisted sheaves. Let AutX\mathcal{A}ut_{\mathcal{X}} be the stack which to any kk-scheme TT associates the groupoid of isomorphisms XT→XT\mathcal{X}_T\to\mathcal{X}_T inducing the identity on the stabilizer groups Gm\mathbb{G}_m. By [8], AutX\mathcal{A}ut_{\mathcal{X}} is a Gm\mathbb{G}_m-gerbe over a group algebraic space Aut⁡X\operatorname{Aut}_{\mathcal{X}}, locally of finite type over kk. We set

O⁡(X,α):=Aut⁡X0,\operatorname{O}(X,\alpha):=\operatorname{Aut}_{\mathcal{X}}^0,

the identity component, and we call it the Olsson group of (X,α)(X,\alpha). Up to isomorphism, it only depends on α\alpha, since gerbes associated to the same class are isomorphic.

By [8 Theorem 1.1(2)] there is an exact sequence of group algebraic spaces

1⟶Pic⁡X→ιXAut⁡X→rXAut⁡X;1\longrightarrow\operatorname{Pic}_{\mathcal{X}}\xrightarrow{\iota_{\mathcal{X}}}\operatorname{Aut}_{\mathcal{X}}\xrightarrow{r_{\mathcal{X}}}\operatorname{Aut}_X;
Equation not extracted
(6)

more precisely, when the set is nonempty, the automorphisms of XT\mathcal{X}_T lying over a given automorphism of XTX_T form a torsor under Pic⁡X(T)\operatorname{Pic}_{\mathcal{X}}(T) [8 3.5 and Remark 3.6].

The map ιX\iota_{\mathcal{X}} sends a line bundle L\mathcal{L} to the automorphism tLt_{\mathcal{L}} of X\mathcal{X} over id⁡X\operatorname{id}_X given by twisting by L\mathcal{L}. Explicitly, if ξ\xi is an object of X\mathcal{X} over a TT-point xx of XX, then tL(ξ)=ξ⊗π∗Lt_{\mathcal{L}}(\xi)=\xi\otimes\pi^*\mathcal{L}, where ⊗\otimes denotes the action of Gm\mathbb{G}_m-torsors on X\mathcal{X} given by its banding. (Depending on conventions, [8 3.5] may use L−1\mathcal{L}^{-1} instead of L\mathcal{L}; this plays no role below.) The map rXr_{\mathcal{X}} sends σ\sigma to the automorphism σˉ\bar{\sigma} of XX that it induces. Since XX is the rigidification of X\mathcal{X}, every morphism from X\mathcal{X} to an algebraic space factors uniquely through π\pi. So σˉ\bar{\sigma} is the unique automorphism of XX with π∘σ≃σˉ∘π\pi\circ\sigma\simeq\bar{\sigma} \circ\pi.

Restricting to identity components we obtain an exact sequence

0→KX,α→O(X,α)→AutX0,KX,α:=PicX∩O(X,α).(7)0 \to K_{X,\alpha} \to\mathrm{O}(X,\alpha) \to\mathrm{Aut}^{0}_{X}, \qquad K_{X,\alpha} := \mathrm{Pic}_{X} \cap\mathrm{O}(X,\alpha). \tag*{(7)}

Since PicX0\mathrm{Pic}^{0}_{X} is connected, PicX0⊆KX,α\mathrm{Pic}^{0}_{X} \subseteq K_{X,\alpha}, but equality need not hold. On the other hand, O(X,α)\mathrm{O}(X,\alpha) is of finite type, being the identity component of a group algebraic space locally of finite type over k\mathsf{k}. So KX,αK_{X,\alpha} has finitely many connected components, and its image in NS(X‾)=Pic(X‾)/Pic0(X‾)\mathrm{NS}(\overline{X}) = \mathrm{Pic}(\overline{X})/\mathrm{Pic}^{0}(\overline{X}) is a finite group. Therefore

PicX0⊆KX,α⊆PicXτ.\mathrm{Pic}^{0}_{X} \subseteq K_{X,\alpha} \subseteq\mathrm{Pic}^{\tau}_{X}.

In particular KX,α=PicX0K_{X,\alpha} = \mathrm{Pic}^{0}_{X} whenever NS(X‾)\mathrm{NS}(\overline{X}) is torsion free, for instance when XX is a curve or an abelian variety.

Olsson group for trivial Brauer classes

Let X=BGm,X\mathcal{X} = B\mathbb{G}_{m,X} be the trivial gerbe. Then

Db(X)→∼Db(BGm,X)(1)=Db(X,α),M↦M⊠χ.D^{b}(X) \xrightarrow{\sim} D^{b}(B\mathbb{G}_{m,X})^{(1)} = D^{b}(X,\alpha), \qquad\mathcal{M} \mapsto\mathcal{M} \boxtimes\chi.

where χ\chi is the universal invertible sheaf on BGmB\mathbb{G}_{m} (its pullback along the TT-point of BGmB\mathbb{G}_{m} given by a line bundle N\mathcal{N} on TT is N\mathcal{N}). The stabilizers act on χ\chi through the standard character. By [8 Example 1.2] there is an isomorphism of stacks AutX×PicX→AutBGm,X\mathrm{Aut}_{X} \times\mathcal{P}\mathrm{ic}_{X} \to\mathcal{A}\mathrm{ut}_{B\mathbb{G}_{m,X}} sending a pair (a,M)(a,\mathcal{M}), with a∈AutX(T)a \in\mathrm{Aut}_{X}(T) and M∈PicX(T)\mathcal{M} \in\mathcal{P}\mathrm{ic}_{X}(T), to the automorphism σ(a,M)\sigma_{(a,\mathcal{M})} which sends a TT-point (x,N)(x,\mathcal{N}) of X×BGmX \times B\mathbb{G}_{m} to (a(x),N⊗x∗M)(a(x),\mathcal{N} \otimes x^{*}\mathcal{M}). Under this isomorphism the composition law reads

σ(a,M)∘σ(a′,M′)=σ(aa′,M′⊗a′∗M).(8)\sigma_{(a,\mathcal{M})} \circ\sigma_{(a',\mathcal{M}')} = \sigma_{(aa',\mathcal{M}' \otimes a'^{*}\mathcal{M})}. \tag*{(8)}

so that AutBGm,X≃AutX⋉PicX\mathrm{Aut}_{B\mathbb{G}_{m,X}} \simeq\mathrm{Aut}_{X} \ltimes\mathrm{Pic}_{X}. If PicX0\mathrm{Pic}^{0}_{X} is an abelian variety, then AutX0\mathrm{Aut}^{0}_{X} acts trivially on PicX0\mathrm{Pic}^{0}_{X} (a connected group scheme acts trivially on an abelian variety by group automorphisms), and therefore

O(X,0)≃AutX0×PicX0\mathrm{O}(X,0) \simeq\mathrm{Aut}^{0}_{X} \times\mathrm{Pic}^{0}_{X}

as group algebraic spaces over k\mathsf{k}. In any case, the identity component of AutX×PicX\mathrm{Aut}_{X} \times\mathrm{Pic}_{X} has underlying space AutX0×PicX0\mathrm{Aut}^{0}_{X} \times\mathrm{Pic}^{0}_{X}. So, for the trivial gerbe the kernel in (7) is KX,0=PicX0K_{X,0} = \mathrm{Pic}^{0}_{X}, and the map to AutX0\mathrm{Aut}^{0}_{X} is surjective. We will need the following two lemmas. The first one describes the action of O(X,0)(T)\mathrm{O}(X,0)(T) on Db(XT)D^{b}(X_{T}).

Lemma 2.3. Let (a,M)∈AutX(T)×PicX(T)(a,\mathcal{M}) \in\mathrm{Aut}_{X}(T) \times\mathcal{P}\mathrm{ic}_{X}(T) and σ=σ(a,M)\sigma= \sigma_{(a,\mathcal{M})}. Under the identification Db(BGm,XT)(1)≃Db(XT)D^{b}(B\mathbb{G}_{m,X_{T}})^{(1)} \simeq D^{b}(X_{T}), the functor σ∗\sigma^{*} is F↦a∗F⊗M\mathcal{F} \mapsto a^{*}\mathcal{F} \otimes\mathcal{M}.

Proof. For a=ida = \mathrm{id} this is simply [8 Proposition 6.1] in the case of the trivial gerbe. For the general case, let pp and qq be the projections from XT×BGmX_{T} \times B\mathbb{G}_{m} to first and second factor, respectively. By definition, σ\sigma sends a TT-point (x,N)(x,\mathcal{N}) to (a(x),N⊗x∗M)(a(x),\mathcal{N} \otimes x^{*}\mathcal{M}). Hence p∘σ=a∘pp \circ\sigma= a \circ p. Moreover, the pullback of (q∘σ)∗χ(q \circ\sigma)^{*}\chi along (x,N)(x,\mathcal{N}) is N⊗x∗M\mathcal{N} \otimes x^{*}\mathcal{M}, which is the pullback of q∗χ⊗p∗Mq^{*}\chi\otimes p^{*}\mathcal{M} along the same point; since this holds functorially in (x,N)(x,\mathcal{N}), we get (q∘σ)∗χ≃q∗χ⊗p∗M(q \circ\sigma)^{*}\chi\simeq q^{*}\chi\otimes p^{*}\mathcal{M}. Therefore

σ∗(F⊠χ)=(p∘σ)∗F⊗(q∘σ)∗χ≃p∗a∗F⊗q∗χ⊗p∗M=(a∗F⊗M)⊠χ.\sigma^{*}(\mathcal{F} \boxtimes\chi) = (p \circ\sigma)^{*}\mathcal{F} \otimes(q \circ\sigma)^{*}\chi\simeq p^{*}a^{*}\mathcal{F} \otimes q^{*}\chi\otimes p^{*}\mathcal{M} = (a^{*}\mathcal{F} \otimes\mathcal{M}) \boxtimes\chi.

The second lemma is about the kernel of the action on k\mathsf{k}-valued points when k\mathsf{k} is an algebraically closed field.

Lemma 2.4. Assume that k\mathsf{k} is algebraically closed. If (a,M)∈AutX(k)×PicX(k)(a,\mathcal{M}) \in\mathrm{Aut}_{X}(\mathsf{k}) \times\mathcal{P}\mathrm{ic}_{X}(\mathsf{k}) and a∗(−)⊗M≃ida^{*}(-) \otimes\mathcal{M} \simeq\mathrm{id} as functors on Db(X)D^{b}(X), then a=idXa = \mathrm{id}_{X} and M≃OX\mathcal{M} \simeq\mathcal{O}_{X}.

Proof. This is a special case of [8 Remark 1.8], which deduces the faithfulness of the action of AutX(k)\mathcal{A}ut_{\mathcal{X}}(\mathsf{k}) on Db(X,α)D^b(X,\alpha) from the uniqueness of Fourier–Mukai kernels [3 Theorem 1.1]. Directly: evaluating on the skyscraper sheaves k(x)\mathsf{k}(x) gives a−1(x)=xa^{-1}(x)=x for all closed points xx, so a=id⁡Xa=\operatorname{id}_X, and evaluating on OX\mathcal{O}_X gives M≃OX\mathcal{M}\simeq\mathcal{O}_X.

Twisted Rouquier isomorphisms

The main result of [8] is that the Olsson group is a twisted derived invariant. Let YY be another smooth projective geometrically connected variety over k\mathsf{k}, let Y→Y\mathcal{Y}\to Y be a Gm\mathbb{G}_{\mathrm{m}}-gerbe with class β∈Br⁡(Y)\beta\in\operatorname{Br}(Y), and let Φ:Db(X,α)→Db(Y,β)\Phi:D^b(X,\alpha)\to D^b(Y,\beta) be an exact equivalence. By [3 Theorem 1.1], Φ\Phi is of Fourier–Mukai type, and by [8 Theorems 1.7 and 1.9] it induces an isomorphism of group algebraic spaces

RΦ:O(X,α)⟶O(Y,β),R_\Phi:\mathcal{O}(X,\alpha)\longrightarrow\mathcal{O}(Y,\beta),

which we call the twisted Rouquier isomorphism associated to $\Phi.

Let an automorphism σ\sigma of a gerbe act on the twisted derived category by σ∗\sigma^*, as in [8 Example 1.5]. Then, on k‾\overline{\mathsf{k}}-points, RΦR_\Phi is given by conjugation [8 Remark 7.2]:

Φk‾∘σ∗∘Φk‾−1≃RΦ(σ)∗.(9)\Phi_{\overline{\mathsf{k}}}\circ\sigma^*\circ\Phi_{\overline{\mathsf{k}}}^{-1}\simeq R_\Phi(\sigma)^*. \tag*{(9)}

When α\alpha and β\beta are trivial and O(X,0)\mathcal{O}(X,0) and O(Y,0)\mathcal{O}(Y,0) are reduced, then the twisted Rouquier isomorphism is the isomorphism constructed by Rouquier in [13]. In particular, when XX and YY are abelian varieties with an equivalence Φ:Db(X)→Db(Y)\Phi:D^b(X)\to D^b(Y), we have O(A,0)=X×X^\mathcal{O}(A,0)=X\times\widehat{X} and O(B,0)=Y×Y^\mathcal{O}(B,0)=Y\times\widehat{Y}, up to some negligible conventions discussed in Remark 2.5. Then the twisted Rouquier isomorphism is exactly the isomorphism fΦf_\Phi constructed by Orlov [Orl02]. In particular, RΦR_\Phi is symplectic by [Orl02, Corollary 2.13].

Remark 2.5 (Conventions). The comparison above depends on two conventions: whether an automorphism aa acts on Db(X)D^b(X) through a∗a^* or through a∗=(a−1)∗a_*=(a^{-1})^*, and whether a line bundle L\mathcal{L} acts through −⊗L-\otimes\mathcal{L} or through −⊗L−1-\otimes\mathcal{L}^{-1} (see the parenthetical remark after (2.7)). These choices do not affect symplecticity. Indeed, let AA and BB abelian varieties with Φ:Db(A)≃Db(B)\Phi:D^b(A)\simeq D^b(B) an equivalence between their untwisted derived categories. The isomorphism fΦ:A×A^→B×B^f_\Phi:A\times\widehat{A}\to B\times\widehat{B} constructed by Orlov, is such that, over k‾\overline{\mathsf{k}}, we have fΦ(a,ξ)=(b,η)f_\Phi(a,\xi)=(b,\eta) if, and only if

Φ∘(ta∗(−)⊗Pξ)∘Φ−1≃tb∗(−)⊗Pη,(10)\Phi\circ(t_a^*(-)\otimes\mathcal{P}_\xi)\circ\Phi^{-1}\simeq t_b^*(-)\otimes\mathcal{P}_\eta, \tag*{(10)}

where tat_a denotes the translation by aa and Pξ\mathcal{P}_\xi denotes the line bundle in Pic⁡0(A)\operatorname{Pic}^0(A) corresponding to ξ∈A^(k)\xi\in\widehat{A}(\mathsf{k}). Now, let f′f' be the isomorphism characterized by the analogue of (2.11) in which ta∗t_a^* and tb∗t_b^* are replaced by ta∗=t−at_a^*=t_{-a} and tb∗=t−bt_b^*=t_{-b}. Then f′(a,ξ)=(b,η)f'(a,\xi)=(b,\eta) if and only if f(−a,ξ)=(−b,η)f(-a,\xi)=(-b,\eta), that is,

f′=ϵB∘f∘ϵA,where ϵA:=(−1001):(a,ξ)⟼(−a,ξ).f'=\epsilon_B\circ f\circ\epsilon_A,\qquad\text{where }\epsilon_A:=\begin{pmatrix}-1&0\\0&1\end{pmatrix}:(a,\xi)\longmapsto(-a,\xi).

Similarly, replacing Pξ\mathcal{P}_\xi and Pη\mathcal{P}_\eta by Pξ−1=P−ξ\mathcal{P}_\xi^{-1}=\mathcal{P}_{-\xi} and Pη−1=P−η\mathcal{P}_\eta^{-1}=\mathcal{P}_{-\eta} replaces ff by ϵB′∘f∘ϵA′\epsilon'_B\circ f\circ\epsilon'_A, where ϵA′:=−ϵA\epsilon'_A:=-\epsilon_A; which is again ϵB∘f∘ϵA\epsilon_B\circ f\circ\epsilon_A (and changing both conventions leaves ff unchanged). In matrix form f′′f'' is represented as follows

ϵB∘(xyzw)∘ϵA=(x−y−zw).\epsilon_B\circ\begin{pmatrix}x&y\\z&w\end{pmatrix}\circ\epsilon_A=\begin{pmatrix}x&-y\\-z&w\end{pmatrix}.

If ff is symplectic, then by (2.1) the inverse of this isomorphism is

ϵA∘f−1∘ϵB=ϵA∘(w^−y^−z^x^)∘ϵB=(w^y^z^x^),\epsilon_A\circ f^{-1}\circ\epsilon_B=\epsilon_A\circ\begin{pmatrix}\widehat{w}&-\widehat{y}\\-\widehat{z}&\widehat{x}\end{pmatrix}\circ\epsilon_B=\begin{pmatrix}\widehat{w}&\widehat{y}\\\widehat{z}&\widehat{x}\end{pmatrix},

which is exactly the matrix that Definition 2.1 prescribes for (x−y−zw)\begin{pmatrix} x & -y \\ -z & w \end{pmatrix}, because −u^=−u^\widehat{-u}=-\hat{u}. Hence eB∘f∘eAe_B \circ f \circ e_A is symplectic whenever ff is, and we may ignore these conventions. □

We now describe two twisted Rouquier isomorphisms that play a central role in § 3.

Example 2.6 (Conjugation by an isomorphism of gerbes). Let X\mathcal{X} and X′\mathcal{X}' be Gm\mathbb{G}_m-gerbes over XX, and let τ:X→X′\tau:\mathcal{X}\to\mathcal{X}' be an isomorphism of Gm\mathbb{G}_m-gerbes over XX, that is, an isomorphism of stacks which commutes with the projections to XX and induces the identity on stabilizer groups. Then τ∗:Db(X′)(1)→Db(X)(1)\tau^*:D^b(\mathcal{X}')^{(1)}\to D^b(\mathcal{X})^{(1)} is an equivalence and thus induces an isomorphism Rτ∗:O(X,β)→O(X,α)R_{\tau_*}:\mathcal{O}(X,\beta)\to\mathcal{O}(X,\alpha), where α\alpha and β\beta are the Brauer classes associated to X→X\mathcal{X}\to X and X′→X\mathcal{X}'\to X (of course α=β\alpha=\beta, since the two gerbes are isomorphic). By (9) and the faithfulness of the action [8 Remark 1.8], we see that Rτ∗R_{\tau_*} is the isomorphism induced after rigidification by the isomorphism of group stacks AutX′→AutX\mathcal{A}ut_{\mathcal{X}'}\to\mathcal{A}ut_{\mathcal{X}} given by conjugation by τ\tau: σ↦τ−1∘σ∘τ\sigma\mapsto\tau^{-1}\circ\sigma\circ\tau.

To lighten the notation we will, in the sequel, denote Rτ∗R_{\tau_*} simply by RτR_\tau.

We see that RτR_\tau behaves well with respect to the exact sequences (6) for X′\mathcal{X}' and X\mathcal{X}, i.e. the diagram

1⟶\longrightarrowPic⁡X\operatorname{Pic}_{X}→ιX′\xrightarrow{\iota_{\mathcal{X}'}}AutX′\mathcal{A}ut_{\mathcal{X}'}→rX′\xrightarrow{r_{\mathcal{X}'}}AutX\mathcal{A}ut_X
∥\Vert↓Rτ\downarrow R_\tau∥\Vert
1⟶\longrightarrowPic⁡X\operatorname{Pic}_{X}→ιX\xrightarrow{\iota_{\mathcal{X}}}AutX\mathcal{A}ut_{\mathcal{X}}→rX\xrightarrow{r_{\mathcal{X}}}AutX\mathcal{A}ut_X

Table 2.12.

commutes. Let π:X→X\pi:\mathcal{X}\to X and π′:X′→X\pi':\mathcal{X}'\to X be the structure maps. Saying that τ\tau is an isomorphism of gerbes over XX means that π′∘τ≃π\pi'\circ\tau\simeq\pi. Hence, for every automorphism σ\sigma of X′\mathcal{X}',

π∘(τ−1στ)≃π′∘σ∘τ≃σˉ∘π′∘τ≃σˉ∘π.\pi\circ(\tau^{-1}\sigma\tau)\simeq\pi'\circ\sigma\circ\tau\simeq\bar{\sigma}\circ\pi'\circ\tau\simeq\bar{\sigma}\circ\pi.

By the uniqueness in the definition of rr, this gives rX(τ−1∘σ∘τ)=σˉ=rX′(σ)r_{\mathcal{X}}(\tau^{-1}\circ\sigma\circ\tau)=\bar{\sigma}=r_{\mathcal{X}'}(\sigma), so the right square commutes. For the left square, observe that τ\tau induces the identity on stabilizers, so it is compatible with the action of Gm\mathbb{G}_m-torsors: τ(ξ⊗P)≃τ(ξ)⊗P\tau(\xi\otimes P)\simeq\tau(\xi)\otimes P, functorially in the object ξ\xi and in the Gm\mathbb{G}_m-torsor PP. Taking P=x∗LP=x^*\mathcal{L} gives τ∘tL≃tL∘τ\tau\circ t_{\mathcal{L}}\simeq t_{\mathcal{L}}\circ\tau, that is, Rτ(ιX′(L))=ιX(L)R_\tau(\iota_{\mathcal{X}'}(\mathcal{L}))=\iota_{\mathcal{X}}(\mathcal{L}).

The case of main interest for us will be when, over kˉ\bar{k}, the gerbe Xkˉ→X‾\mathcal{X}_{\bar{k}}\to\overline{X} is associated to the trivial Brauer class and τ:Xkˉ→BGm,X‾\tau:\mathcal{X}_{\bar{k}}\to B\mathbb{G}_{m,\overline{X}} is a trivialization. Applying the above over kˉ\bar{k} and passing to identity components, by Subsection 2.2.1 the top row of (2.12) restricts to the split exact sequence 0→Pic⁡X0→O(X,0)kˉ→Aut⁡X0→00\to\operatorname{Pic}^{0}_{X}\to\mathcal{O}(X,0)_{\bar{k}}\to\operatorname{Aut}^{0}_{X}\to0. We therefore obtain an isomorphism of group algebraic spaces Rτ:O(X,0)kˉ→O(X,α)kˉR_\tau:\mathcal{O}(X,0)_{\bar{k}}\to\mathcal{O}(X,\alpha)_{\bar{k}} fitting in the commutative diagram

0⟶\longrightarrowPic⁡X0\operatorname{Pic}^{0}_{X}⟶\longrightarrowO(X,0)kˉ\mathcal{O}(X,0)_{\bar{k}}⟶\longrightarrowAut⁡X0\operatorname{Aut}^{0}_{X}⟶\longrightarrow0
∥\Vert↓Rτ\downarrow R_\tau∥\Vert
0⟶\longrightarrowPic⁡X0\operatorname{Pic}^{0}_{X}⟶\longrightarrowO(X,α)kˉ\mathcal{O}(X,\alpha)_{\bar{k}}⟶\longrightarrowAut⁡X0\operatorname{Aut}^{0}_{X}⟶\longrightarrow0

Table 2.13.

The bottom row is exact, because the top row is and RτR_\tau is an isomorphism. When Pic⁡X0\operatorname{Pic}^{0}_{X} is an abelian variety, O(X,0)=Aut⁡X0×Pic⁡X0\mathcal{O}(X,0)=\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X}.

Example 2.7 (The automorphisms Ad⁡(Q)\operatorname{Ad}(\mathcal{Q})). Let Q\mathcal{Q} be a line bundle on XX. Then σ(id⁡,Q)\sigma_{(\operatorname{id},\mathcal{Q})} is an automorphism of the trivial gerbe BGm,XB\mathbb{G}_{m,X}, with inverse σ(id⁡,Q)−1=σ(id⁡,Q−1)\sigma_{(\operatorname{id},\mathcal{Q})}^{-1}=\sigma_{(\operatorname{id},\mathcal{Q}^{-1})} by (8). Applying Example 2.6 to τ=σ(id⁡,Q−1)\tau=\sigma_{(\operatorname{id},\mathcal{Q}^{-1})}, we obtain the automorphism

Ad⁡(Q):=Rσ(id⁡,Q−1):ρ⟼σ(id⁡,Q)∘ρ∘σ(id⁡,Q)−1\operatorname{Ad}(\mathcal{Q}):=R_{\sigma_{(\operatorname{id},\mathcal{Q}^{-1})}}:\rho\longmapsto\sigma_{(\operatorname{id},\mathcal{Q})}\circ\rho\circ\sigma_{(\operatorname{id},\mathcal{Q})}^{-1}

of Aut⁡BGm,X\operatorname{Aut}_{\mathcal{B}\mathcal{G}_{m,X}}, that is, conjugation by σ(id⁡,Q)\sigma_{(\operatorname{id},\mathcal{Q})}. By Example 2.6 and Lemma 2.3, it is the twisted Rouquier isomorphism of the autoequivalence σ(id⁡,Q−1)∗=−⊗Q−1\sigma^{*}_{(\operatorname{id},\mathcal{Q}^{-1})}=-\otimes\mathcal{Q}^{-1} of Db(X)D^{b}(X). Note that Ad⁡(Q)\operatorname{Ad}(\mathcal{Q}) is not an automorphism of BGm,X\mathcal{B}\mathcal{G}_{m,X} itself: it acts on automorphisms of BGm,X\mathcal{B}\mathcal{G}_{m,X}. It preserves the identity component O⁡(X,0)\operatorname{O}(X,0), and we denote the induced automorphism again by Ad⁡(Q)\operatorname{Ad}(\mathcal{Q}).

By (2.9) we have Ad⁡(Q⊗Q′)=Ad⁡(Q)∘Ad⁡(Q′)\operatorname{Ad}(\mathcal{Q}\otimes\mathcal{Q}')=\operatorname{Ad}(\mathcal{Q})\circ\operatorname{Ad}(\mathcal{Q}'), and

Ad⁡(Q)(σ(a,L))=σ(a,L⊗a∗Q⊗Q−1)(11)\operatorname{Ad}(\mathcal{Q})(\sigma_{(a,\mathcal{L})})=\sigma_{(a,\mathcal{L}\otimes a^{*}\mathcal{Q}\otimes\mathcal{Q}^{-1})} \tag*{(11)}

Assume now that Pic⁡X0\operatorname{Pic}^{0}_{X} is an abelian variety, so that O⁡(X,0)=Aut⁡X0×Pic⁡X0\operatorname{O}(X,0)=\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X}. Then Ad⁡(Q)\operatorname{Ad}(\mathcal{Q}) acts on it by (a,L)↦(a,L⊗a∗Q⊗Q−1)(a,\mathcal{L})\mapsto(a,\mathcal{L}\otimes a^{*}\mathcal{Q}\otimes\mathcal{Q}^{-1}). If Q∈Pic⁡0(X)\mathcal{Q}\in\operatorname{Pic}^{0}(X), this automorphism is the identity: the morphism a↦a∗Q⊗Q−1a\mapsto a^{*}\mathcal{Q}\otimes\mathcal{Q}^{-1} from Aut⁡X0\operatorname{Aut}^{0}_{X} to Pic⁡X0\operatorname{Pic}^{0}_{X} is trivial, because Aut⁡X0\operatorname{Aut}^{0}_{X} acts trivially on the abelian variety Pic⁡X0\operatorname{Pic}^{0}_{X}. Consequently, on Aut⁡X0×Pic⁡X0\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X} the automorphism Ad⁡(Q)\operatorname{Ad}(\mathcal{Q}) only depends on the class of Q\mathcal{Q} in NS⁡(X)\operatorname{NS}(X). In Section 3 we use this over k‾\overline{k}.

Olsson groups for Algebraic Brauer classes

The main result of this section is the following

Proposition 3.1. If CC is a genus 11 curve and α∈Br⁡(C)\alpha\in\operatorname{Br}(C), then

O⁡(C,α)≃J(C)×J^(C).\operatorname{O}(C,\alpha)\simeq J(C)\times\widehat{J}(C).

We say that α∈Br⁡(X)\alpha\in\operatorname{Br}(X) is algebraic when it sits in the kernel of the map Br⁡(X)→Br⁡(X‾)\operatorname{Br}(X)\to\operatorname{Br}(\overline{X}). Note that every class α∈Br⁡(C)\alpha\in\operatorname{Br}(C) becomes trivial over k‾\overline{k}: since C‾\overline{C} is a smooth projective curve over an algebraically closed field, Br⁡(C‾)=0\operatorname{Br}(\overline{C})=0 by Tsen’s theorem. Thus every α∈Br⁡(C)\alpha\in\operatorname{Br}(C) is algebraic.

We will put ourselves in a much more general situation and work with XX, a smooth geometrically connected projective variety over a perfect field kk such that Pic⁡X0\operatorname{Pic}^{0}_{X} is an abelian variety. The idea is that, when α\alpha is algebraic, over k‾\overline{k} there is an isomorphism

O⁡(X,α)k‾≃(Pic⁡X0×Aut⁡X0)k‾,\operatorname{O}(X,\alpha)_{\overline{k}}\simeq(\operatorname{Pic}^{0}_{X}\times\operatorname{Aut}^{0}_{X})_{\overline{k}},

since α‾\overline{\alpha} is trivial by hypothesis and Pic⁡X0\operatorname{Pic}^{0}_{X} is an abelian variety (see Subsection 2.2.1). So we want to define a morphism Aut⁡X0×Pic⁡X0→O⁡(X,α)\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X}\to\operatorname{O}(X,\alpha) over kk whose base change to k‾\overline{k} is the isomorphism RτR_{\tau} of Example 2.6.

If α\alpha is algebraic, then Xk‾\mathcal{X}_{\overline{k}} is a trivial gerbe, so there is an isomorphism of Gm\mathbb{G}_{m}-gerbes

τ:Xk‾→∼BGm,X‾,\tau:\mathcal{X}_{\overline{k}}\xrightarrow{\sim}\mathcal{B}\mathcal{G}_{m,\overline{X}},

which, by a limit argument, is defined over some finite Galois extension k′/kk'/k. Let γ∈Gk\gamma\in G_{k}. The projections Xk‾→X‾\mathcal{X}_{\overline{k}}\to\overline{X} and BGm,X‾→X‾\mathcal{B}\mathcal{G}_{m,\overline{X}}\to\overline{X} are base changes of morphisms defined over kk, hence Galois equivariant. So the conjugate γτ{}^{\gamma}\tau again commutes with the projections to X‾\overline{X} and induces the identity on stabilizers. Consider the composite

ργ:=γτ∘τ−1:BGm,X‾→BGm,X‾.\rho_{\gamma}:={}^{\gamma}\tau\circ\tau^{-1}:\mathcal{B}\mathcal{G}_{m,\overline{X}}\to\mathcal{B}\mathcal{G}_{m,\overline{X}}.

Clearly, it induces the identity on stabilizers, so it is an object of AutBGm,X‾(k‾)\mathcal{A}ut_{\mathcal{B}\mathcal{G}_{m,\overline{X}}}(\overline{k}). In addition, it commutes with the projection to X‾\overline{X}, so its image under the map AutBGm,X‾→Aut⁡X‾\mathcal{A}ut_{\mathcal{B}\mathcal{G}_{m,\overline{X}}}\to\operatorname{Aut}_{\overline{X}} of [8 Theorem 1.1(2)] is id⁡X‾\operatorname{id}_{\overline{X}}. Hence, by [8 Example 1.2] (see Subsection 2.2.1), it is of the form σ(id⁡,Mγ)\sigma_{(\operatorname{id},\mathcal{M}_{\gamma})} for a line bundle Mγ\mathcal{M}_{\gamma} on X‾\overline{X}. By Lemma 2.4 this line bundle is unique up to isomorphism.

Lemma 3.2. In the notation above we have the following:

(1) for γ∈Gal⁡(k‾/k′)\gamma\in\operatorname{Gal}(\overline{k}/k'), Mγ≃OX‾\mathcal{M}_{\gamma}\simeq\mathcal{O}_{\overline{X}};

(2) for all γ,δ∈Gk\gamma,\delta\in G_{k}, Mγδ≃Mγ⊗γMδ\mathcal{M}_{\gamma\delta}\simeq\mathcal{M}_{\gamma}\otimes{}^{\gamma}\mathcal{M}_{\delta};

(3) every other trivialization of Xk‾\mathcal{X}_{\overline{k}} is of the form τ′=σ(id,Q)∘τ\tau'=\sigma_{(\mathrm{id},Q)}\circ\tau with Q∈Pic⁡(X‾)Q\in\operatorname{Pic}(\overline{X}). In addition, if we denote by Mγ′\mathcal{M}_{\gamma}' the line bundle associated to γτ′∘τ′−1\gamma\tau'\circ\tau'^{-1} then

Mγ′≃Mγ⊗γQ⊗Q−1.\mathcal{M}_{\gamma}'\simeq\mathcal{M}_{\gamma}\otimes{}^{\gamma}Q\otimes Q^{-1}.

Proof. The first claim is clear since τ\tau is defined over kk, and thus left fixed by conjugation by γ\gamma. We turn our attention to (2). Since the isomorphism Aut⁡X×Pic⁡X≃AutGm,X\operatorname{Aut}_{X}\times\operatorname{Pic}_{X}\simeq\mathcal{A}ut_{\mathbb{G}_{m},X} is defined over kk, we have that

γδτ∘τ−1=γ(δτ)∘τ−1=γ(δτ∘τ−1)∘γτ∘τ−1=σ(id,γMδ)∘σ(id,Mγ).\begin{aligned} \gamma\delta_{\tau}\circ\tau^{-1}=\gamma(\delta_{\tau})\circ\tau^{-1} \\ &=\gamma(\delta_{\tau}\circ\tau^{-1})\circ\gamma\tau\circ\tau^{-1} \\ &=\sigma_{(\mathrm{id},{}^{\gamma}\mathcal{M}_{\delta})}\circ\sigma_{(\mathrm{id},\mathcal{M}_{\gamma})}. \end{aligned}

The result follows from (8).

For (3), observe that the composite τ′∘τ−1\tau'\circ\tau^{-1} is an automorphism of Gm,X‾\mathbb{G}_{m,\overline{X}} over the identity inducing the identity on stabilizers. Thus

τ′∘τ−1=σ(id,Q)\tau'\circ\tau^{-1}=\sigma_{(\mathrm{id},Q)}

for some line bundle QQ. Hence τ′\tau' has the stated form. Moreover

γτ′∘τ′−1=σ(id,γQ)∘σ(id,Mγ)∘σ(id,Q−1)=σ(id,Mγ⊗γQ⊗Q−1),\gamma\tau'\circ\tau'^{-1}=\sigma_{(\mathrm{id},{}^{\gamma}Q)}\circ\sigma_{(\mathrm{id},\mathcal{M}_{\gamma})}\circ\sigma_{(\mathrm{id},Q^{-1})}=\sigma_{(\mathrm{id},\mathcal{M}_{\gamma}\otimes{}^{\gamma}Q\otimes Q^{-1})},

where the last step is (2.9), and the statement is proven.

By Lemma 3.2(2), the map γ↦[Mγ]∈NS⁡(X‾)\gamma\mapsto[\mathcal{M}_{\gamma}]\in\operatorname{NS}(\overline{X}) is a 1-cocycle, so it defines a class in H1(k,NS⁡(X‾))\mathrm{H}^{1}(k,\operatorname{NS}(\overline{X})). By Lemma 3.2(3), replacing τ\tau by another trivialization σ(id,Q)∘τ\sigma_{(\mathrm{id},Q)}\circ\tau changes this cocycle by γ↦[γQ⊗Q−1]=γ[Q]−[Q]\gamma\mapsto[{}^{\gamma}Q\otimes Q^{-1}]=\gamma[Q]-[Q], which is a coboundary; hence the class does not depend on the choice of τ\tau. It does not depend on the gerbe X\mathcal{X} either, but only on α\alpha. In fact, given another gerbe X′\mathcal{X}' representing α\alpha and an isomorphism i:X′→Xi:\mathcal{X}'\to\mathcal{X} over XX, we have

γ(τ∘i)∘(τ∘i)−1=γτ∘τ−1.\gamma(\tau\circ i)\circ(\tau\circ i)^{-1}=\gamma\tau\circ\tau^{-1}.

Thus we can define the following

Definition 3.3. For α∈Br⁡(X)\alpha\in\operatorname{Br}(X) algebraic we define the class

ob⁡(α):=[γ↦Mγ mod Pic⁡0(X‾)]∈H1(k,NS⁡(X‾)).\operatorname{ob}(\alpha):=[\gamma\mapsto\mathcal{M}_{\gamma}\bmod\operatorname{Pic}^{0}(\overline{X})]\in\mathrm{H}^{1}(k,\operatorname{NS}(\overline{X})).

Remark 3.4. If NS⁡(X‾)\operatorname{NS}(\overline{X}) is torsion free with trivial Galois action, then

H1(k,NS⁡(X‾))=0,\mathrm{H}^{1}(k,\operatorname{NS}(\overline{X}))=0,

hence ob⁡(α)=0\operatorname{ob}(\alpha)=0. This applies to genus 1 curves, for which NS⁡(C‾)=Z\operatorname{NS}(\overline{C})=\mathbb{Z} via the degree.

From now on, until the end of this section, we work under the following assumptions.

Setting 3.5. Let XX be a smooth projective geometrically connected variety over kk such that Pic⁡X0\operatorname{Pic}^{0}_{X} is an abelian variety, and let α∈Br⁡(X)\alpha\in\operatorname{Br}(X) be algebraic.

Let τ:Xk‾→Gm,X‾\tau:\mathcal{X}_{\overline{k}}\to\mathbb{G}_{m,\overline{X}} be a trivialization. By Example 2.6, conjugation by τ\tau gives an isomorphism of group algebraic spaces over k‾\overline{k}

Rτ:(Aut⁡X0×Pic⁡X0)k‾⟶O(X,α)k‾,R_{\tau}:(\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X})_{\overline{k}}\longrightarrow\mathcal{O}(X,\alpha)_{\overline{k}},

fitting in the commutative diagram (2.13) with exact rows. Here we use that O(X,0)≃Aut⁡X0×Pic⁡X0\mathcal{O}(X,0)\simeq\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X}, because Pic⁡X0\operatorname{Pic}^{0}_{X} is an abelian variety (Subsection 2.2.1). For a line bundle QQ on X‾\overline{X} we also use the automorphism Ad⁡(Q)\operatorname{Ad}(Q) of (Aut⁡X0×Pic⁡X0)k‾(\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X})_{\overline{k}} of Example 2.7. It is the identity if Q∈Pic⁡0(X‾)Q\in\operatorname{Pic}^{0}(\overline{X}), and in general it only depends on the class of QQ in NS⁡(X‾)\operatorname{NS}(\overline{X}).

Lemma 3.6. Under the assumption in Setting 3.5, we have the following:

  1. If ob⁡(α)=0\operatorname{ob}(\alpha)=0, there is a trivialization τ:Xk‾→BGm,X‾\tau:\mathcal{X}_{\overline{k}}\to BG_{\mathrm{m},\overline{X}} with Mγ∈Pic⁡0(X‾)\mathcal{M}_{\gamma}\in\operatorname{Pic}^{0}(\overline{X}) for all γ∈Gk\gamma\in G_k.

  2. Let τ\tau and τ′\tau' be two trivializations of Xk‾\mathcal{X}_{\overline{k}}, and write τ′=σ(id⁡,Q)∘τ\tau'=\sigma_{(\operatorname{id},Q)}\circ\tau with Q∈Pic⁡(X‾)Q\in\operatorname{Pic}(\overline{X}) as in Lemma 3.2(3). Then Rτ′=Rτ∘Ad⁡(Q)−1R_{\tau'}=R_{\tau}\circ\operatorname{Ad}(Q)^{-1}. If moreover ob⁡(α)=0\operatorname{ob}(\alpha)=0 and both τ\tau and τ′\tau' are as in (1), then the class of QQ in NS⁡(X‾)\operatorname{NS}(\overline{X}) is Galois invariant.

Proof. (1) Let τ0\tau_0 be any trivialization, with associated line bundles Mγ0\mathcal{M}^{0}_{\gamma}. Since ob⁡(α)=0\operatorname{ob}(\alpha)=0, there is ν∈NS⁡(X‾)\nu\in\operatorname{NS}(\overline{X}) such that the class of Mγ0\mathcal{M}^{0}_{\gamma} is γν−ν\gamma\nu-\nu for all γ∈Gk\gamma\in G_k. Let Q∈Pic⁡(X‾)Q\in\operatorname{Pic}(\overline{X}) be a lift of −ν-\nu and put τ=σ(id⁡,Q)∘τ0\tau=\sigma_{(\operatorname{id},Q)}\circ\tau_0. Denote the associated line bundles by Mγ\mathcal{M}_{\gamma}. By Lemma 3.2(3), the class of Mγ\mathcal{M}_{\gamma} in NS⁡(X‾)\operatorname{NS}(\overline{X}) is (γν−ν)−(γν−ν)=0(\gamma\nu-\nu)-(\gamma\nu-\nu)=0, that is, Mγ∈Pic⁡0(X‾)\mathcal{M}_{\gamma}\in\operatorname{Pic}^{0}(\overline{X}).

(2) For every ρ\rho we have

Rτ′(ρ)=τ−1∘σ(id⁡,Q)−1∘ρ∘σ(id⁡,Q)∘τ=Rτ(Ad⁡(Q)−1(ρ)).R_{\tau'}(\rho)=\tau^{-1}\circ\sigma_{(\operatorname{id},Q)}^{-1}\circ\rho\circ\sigma_{(\operatorname{id},Q)}\circ\tau=R_{\tau}\bigl(\operatorname{Ad}(Q)^{-1}(\rho)\bigr).

If both τ\tau and τ′\tau' are as in (1), Lemma 3.2(3) gives γQ⊗Q−1≃Mγ′⊗Mγ−1∈Pic⁡0(X‾)\gamma Q\otimes Q^{-1}\simeq\mathcal{M}'_{\gamma}\otimes\mathcal{M}_{\gamma}^{-1}\in\operatorname{Pic}^{0}(\overline{X}) for all γ∈Gk\gamma\in G_k, so the class of QQ in NS⁡(X‾)\operatorname{NS}(\overline{X}) is Galois invariant. □

The next result shows that O⁡(X,α)≃Aut⁡X0×Pic⁡X0\operatorname{O}(X,\alpha)\simeq\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X} when the obstruction ob⁡(α)\operatorname{ob}(\alpha) vanishes. Proposition 3.1 then follows immediately from Remark 3.4 and the classical facts that, for a genus 1 curve CC, we have Pic⁡C0≃J(C)\operatorname{Pic}^{0}_{C}\simeq J(C) and Aut⁡C0≃J(C)\operatorname{Aut}^{0}_{C}\simeq J(C).

Proposition 3.7. Under the hypothesis of Setting 3.5, assume in addition that ob⁡(α)=0\operatorname{ob}(\alpha)=0, and let τ:Xk‾→BGm,X‾\tau:\mathcal{X}_{\overline{k}}\to BG_{\mathrm{m},\overline{X}} be a trivialization as in Lemma 3.6(1). Then RτR_{\tau} descends to an isomorphism of group algebraic spaces over kk

θτ:Aut⁡X0×Pic⁡X0→∼O⁡(X,α),\theta_{\tau}:\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X}\xrightarrow{\sim}\operatorname{O}(X,\alpha),

*compatible with the inclusions of Pic⁡X0\operatorname{Pic}^{0}_{X} and with the maps to Aut⁡X0\operatorname{Aut}^{0}_{X}; that is, θτ\theta_{\tau} fits in the analog of (2.13) over kk. In particular, in (2.8) the map O⁡(X,α)→Aut⁡X0\operatorname{O}(X,\alpha)\to\operatorname{Aut}^{0}_{X} is surjective and KX,α=Pic⁡X0K_{X,\alpha}=\operatorname{Pic}^{0}_{X}, so that O⁡(X,α)\operatorname{O}(X,\alpha) is a split extension of Aut⁡X0\operatorname{Aut}^{0}_{X} by Pic⁡X0\operatorname{Pic}^{0}_{X}.

Proof. Once we know that RτR_{\tau} descends to an isomorphism θτ\theta_{\tau} of group algebraic spaces over kk, we are done. In fact, the commutativity of the analog of (2.13) over kk can be checked over k‾\overline{k}, where it is (2.13) itself. The surjectivity of O⁡(X,α)→Aut⁡X0\operatorname{O}(X,\alpha)\to\operatorname{Aut}^{0}_{X} and the equality KX,α=Pic⁡X0K_{X,\alpha}=\operatorname{Pic}^{0}_{X} can also be checked over k‾\overline{k}, where they follow from the exactness of the bottom row of (2.13). Finally, θτ\theta_{\tau} itself gives the splitting.

To construct θτ\theta_{\tau}, we use Galois descent. We first observe that the trivialization τ\tau is defined over some finite Galois extension k′k' of kk, and thus so is RτR_{\tau}.

Now, for γ∈Gk\gamma\in G_k, we have γ(Rτ)=Rγτ\gamma(R_{\tau})=R_{\gamma\tau}. In addition, by definition of Mγ\mathcal{M}_{\gamma} we have

γτ=ργ∘τ=σ(id⁡,Mγ)∘τ.\gamma\tau=\rho_{\gamma}\circ\tau=\sigma_{(\operatorname{id},\mathcal{M}_{\gamma})}\circ\tau.

Hence γ(Rτ)=Rτ∘Ad⁡(Mγ)−1\gamma(R_{\tau})=R_{\tau}\circ\operatorname{Ad}(\mathcal{M}_{\gamma})^{-1} by Lemma 3.6(2).

Since τ\tau is as in Lemma 3.6(1), we have Mγ∈Pic⁡0(X‾)\mathcal{M}_{\gamma}\in\operatorname{Pic}^{0}(\overline{X}) for all γ∈Gk\gamma\in G_k. Hence, as observed after (2.14), the automorphism Ad⁡(Mγ)\operatorname{Ad}(\mathcal{M}_{\gamma}) of (Aut⁡X0×Pic⁡X0)k‾(\operatorname{Aut}^{0}_{X}\times\operatorname{Pic}^{0}_{X})_{\overline{k}} is the identity. Therefore RτR_{\tau} is invariant under the action of GkG_k, which factors through Gal⁡(k′/k)\operatorname{Gal}(k'/k). So RτR_{\tau} descends to a morphism θτ\theta_{\tau} over kk. The same applies to its inverse, and thus θτ\theta_{\tau} is an isomorphism. It is a homomorphism of group algebraic spaces, since this can be checked over k‾\overline{k}, where it holds for RτR_{\tau}. □

We now specialize to the case relevant for the main result: torsors under abelian varieties. To this aim, we let XX be a torsor under an abelian variety AA over kk. Our main example will be a genus 1 curve, which is a torsor under its Jacobian. The action of AA on XX by translations identifies Aut⁡X0\operatorname{Aut}^{0}_{X} with AA [4 Lemma A.3]. Moreover, Pic⁡X0\operatorname{Pic}^{0}_{X} is canonically isomorphic to Pic⁡A0=A^\operatorname{Pic}^{0}_{A}=\widehat{A} [12 XIII, 1.1] (see also [14 §4]).

In particular Pic⁡X0\operatorname{Pic}^{0}_{X} is an abelian variety. If α∈Br⁡(X)\alpha\in\operatorname{Br}(X) is an algebraic class, we are under the assumptions of Setting 3.5. Thus, if ob⁡(α)=0\operatorname{ob}(\alpha)=0, Proposition 3.7 gives isomorphisms

θτ:A×A^→∼O⁡(X,α).\theta_{\tau}: A \times\widehat{A} \xrightarrow{\sim} \operatorname{O}(X,\alpha).

Proof of the main theorem

In this section we prove the Main Theorem from the introduction.

To this aim, let CC and C′C^{\prime} be two genus 1 curves over k\mathbb{k} with Jacobians JJ and J′J^{\prime} respectively. If (2) hold, then they are torsors over the same abelian variety JJ, and there are two classes α\alpha and α′\alpha^{\prime} in H⁡1(k,J)\operatorname{H}^{1}(\mathbb{k},J) such that [C]=α[C]=\alpha and [C′]=α′[C^{\prime}]=\alpha^{\prime}. Regarding them as Brauer classes on JJ via Br⁡(J)≃Br⁡(k)⊕H⁡1(k,J)\operatorname{Br}(J)\simeq\operatorname{Br}(\mathbb{k})\oplus\operatorname{H}^{1}(\mathbb{k},J), by [14] we have twisted equivalences

D⁡b(C)≃D⁡b(J,α),D⁡b(J,α′)≃D⁡b(C′).\operatorname{D}^{b}(C)\simeq\operatorname{D}^{b}(J,\alpha), \qquad\operatorname{D}^{b}(J,\alpha^{\prime})\simeq\operatorname{D}^{b}(C^{\prime}).

We deduce that (3) holds, which implies directly (1). To conclude the proof we need to show that two derived isogenous genus 1 curves have isomorphic Jacobians. Suppose that CC and C′C^{\prime} are derived isogenous. Then there is a sequence of varieties C=X0,…,Xn=C′C=X_{0},\ldots,X_{n}=C^{\prime} and with twisted equivalences

D⁡b(Xi,βi)≃D⁡b(Xi+1,αi+1).\operatorname{D}^{b}(X_{i},\beta_{i})\simeq\operatorname{D}^{b}(X_{i+1},\alpha_{i+1}).

Since twisted derived equivalence preserve dimension and Kodaira dimension (or order of the canonical bundle), we can show by induction that XiX_{i} are curves of genus 1. Therefore, the last implication of the main theorem will be a consequence of Proposition 2.2 and the following result.

Proposition 4.1. Let CC and C′C^{\prime} be as before, and let α∈Br⁡(C)\alpha\in\operatorname{Br}(C) and β∈Br⁡(C′)\beta\in\operatorname{Br}(C^{\prime}). If there is an exact equivalence D⁡b(C,α)≃D⁡b(C′,β)\operatorname{D}^{b}(C,\alpha)\simeq\operatorname{D}^{b}(C^{\prime},\beta), then there is a symplectic isomorphism J×J^≃J′×J′^J\times\widehat{J}\simeq J^{\prime}\times\widehat{J^{\prime}} defined over k\mathbb{k}.

Proof. Let Φ:D⁡b(C,α)→D⁡b(C′,β)\Phi:\operatorname{D}^{b}(C,\alpha)\to\operatorname{D}^{b}(C^{\prime},\beta) be an equivalence, and let RΦ:O⁡(C,α)→O⁡(C′,β)R_{\Phi}:\operatorname{O}(C,\alpha)\to\operatorname{O}(C^{\prime},\beta) be its twisted Rouquier isomorphism (Subsection 2.2.2). Let X→C\mathcal{X}\to C and Y→C′\mathcal{Y}\to C^{\prime} be the gerbes associated to α\alpha and β\beta, choose trivializations τ:Xk‾→BGm,C‾\tau:\mathcal{X}_{\overline{\mathbb{k}}}\to B\mathbb{G}_{m,\overline{C}} and τ′:Yk‾→BGm,C′‾\tau^{\prime}:\mathcal{Y}_{\overline{\mathbb{k}}}\to B\mathbb{G}_{m,\overline{C^{\prime}}} as in Lemma 3.6(1), and put

F:=θτ′−1∘RΦ∘θτ:J×J^⟶J′×J′^.F:=\theta_{\tau^{\prime}}^{-1}\circ R_{\Phi}\circ\theta_{\tau}:J\times\widehat{J}\longrightarrow J^{\prime}\times\widehat{J^{\prime}}.

We want to show that FF is symplectic. This can be checked after base change to k‾\overline{\mathbb{k}}. By construction, (θτ)k‾(\theta_{\tau})_{\overline{\mathbb{k}}} is simply RτR_{\tau} and similarly (θτ′)k‾=Rτ′(\theta_{\tau^{\prime}})_{\overline{\mathbb{k}}}=R_{\tau^{\prime}}. In addition, the formation of RΦR_{\Phi} commutes with extension of the base field (see [8 7.1–7.6]). Thus we have

Fk‾=Rτ′−1∘RΦk‾∘Rτ.F_{\overline{\mathbb{k}}}=R_{\tau^{\prime}}^{-1}\circ R_{\Phi_{\overline{\mathbb{k}}}}\circ R_{\tau}.

Recall (see Example 2.6) that the trivialization τ\tau induces an equivalence τ∗:D⁡b(C‾)=D⁡b(BGm,C‾)(1)→D⁡b(C‾,α‾)\tau^{*}:\operatorname{D}^{b}(\overline{C})=\operatorname{D}^{b}(B\mathbb{G}_{m,\overline{C}})^{(1)}\to\operatorname{D}^{b}(\overline{C},\overline{\alpha}), and RτR_{\tau} is simply the twisted Rouquier isomorphism associated to τ∗\tau^{*}. Similarly Rτ′∗=Rτ′R_{\tau^{\prime*}}=R_{\tau^{\prime}}. If we consider the equivalence

Φ0:=(τ′∗)−1∘Φk‾∘τ∗:D⁡b(C‾)→D⁡b(C′‾),\Phi_{0}:=(\tau^{\prime*})^{-1}\circ\Phi_{\overline{\mathbb{k}}}\circ\tau^{*}:\operatorname{D}^{b}(\overline{C})\to\operatorname{D}^{b}(\overline{C^{\prime}}),

since the construction of twisted Rouquier isomorphisms is functorial, we have

RΦ0=R(τ′∗)−1∘RΦk‾∘Rτ∗=Rτ′−1∘RΦk‾∘Rτ=Fk‾.R_{\Phi_{0}}=R_{(\tau^{\prime*})^{-1}}\circ R_{\Phi_{\overline{\mathbb{k}}}}\circ R_{\tau^{*}}=R_{\tau^{\prime}}^{-1}\circ R_{\Phi_{\overline{\mathbb{k}}}}\circ R_{\tau}=F_{\overline{\mathbb{k}}}.

Now it is immediate to see that Fk‾F_{\overline{\mathbb{k}}} is symplectic, as it is the (untwisted) Rouquier isomorphism associated to an equivalence between abelian varieties and so symplectic by the work of Orlov [Orl02] (see Subsection 2.2.2).

Department of Mathematics, Indian Institute of Technology Gandhinagar, Palaj Campus, Gandhi-Nagar, Gujarat, India

Email address: [email protected]

Department of Mathematics, Stockholm University, Kraftriket Hus 6, Stockholm, Sweden

Email address: [email protected]

Institut Mathématiques de Jussieu-Paris Rive Gauche, Université Paris Cité, 8 Place Aurélie Némours, Paris, France

Email address: [email protected]

References

  1. [1]Benjamin Antieau, Daniel Krashen, and Matthew Ward. Derived categories of torsors for abelian schemes. Advances in Mathematics, 306:1–23, 2017.
  2. [2]Nicolas Addington and Andrew Wray. Twisted Fourier–Mukai partners of Enriques surfaces. arXiv preprint arXiv:1803.03250, 2018.DOI
  3. [3]Alberto Canonaco and Paolo Stellari. Twisted Fourier–Mukai functors. Adv. Math., 212(2):484–503, 2007.
  4. [4]Katrina Honigs. Derived equivalence, Albanese varieties, and the zeta functions of 3-dimensional varieties. Proc. Amer. Math. Soc., 146(3):1005–1013, 2018. With an appendix by Jeffrey D. Achter, Sebastian Casalaina-Martin, Katrina Honigs, and Charles Vial.arxiv.org/abs/1604.00042
  5. [5]Tyler Lane. Twisted derived equivalences between abelian varieties, 2026. arXiv:2601.16871.arxiv.org/abs/2601.16871
  6. [6]Ruoxi Li. Point objects and derived equivalences of twisted derived categories of abelian varieties, 2025. arXiv:2511.19282.arxiv.org/abs/2511.19282
  7. [8]Martin Olsson. Twisted derived categories and Rouquier functors. International Mathematics Research Notices, 2025(18):rnaf290, 2025.arxiv.org/abs/2501.05599
  8. [11]Alexander Polishchuk. Abelian Varieties, Theta Functions and the Fourier Transform. Cambridge University Press, April 2003.DOI
  9. [12]Michel Raynaud. Faisceaux amples sur les schémas en groupes et les espaces homogènes, volume 119 of Lecture Notes in Mathematics. Springer-Verlag, Berlin-New York, 1970.DOI
  10. [13]Raphaël Rouquier. Automorphismes, graduations et catégories triangulées. J. Inst. Math. Jussieu, 10(3):713–751, 2011.DOI
  11. [14]Niranjan Ramachandran and Jonathan Rosenberg. Derived categories of curves of genus one and torsors over abelian varieties. Math. Res. Lett., 32:293–318, 2025. arXiv:2212.14497.DOI

Paper details

vlnc:W_jzv7qhmm
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Claude Opus 5.5

View statement
Looking for related papers…
Contents