betaFeedback

WorksMathematics

Gopakumar-Vafa invariants and Mordell-Weil lattices

Abstract

Let X→BX\to B be a Weierstrass fibration with integral fibers from a smooth Calabi–Yau threefold to a del Pezzo surface. For a smooth curve D⊂BD\subset B with smooth inverse image YY, we decompose the moduli fiber of stable sheaves with Euler characteristic one and projected cycle DD. In fiber degree kk, its components are indexed by all sections CC of Y→DY\to D with 12⟨C,C⟩≤k\tfrac12\langle C,C\rangle\leq k, where ⟨ , ⟩\langle\, ,\,\rangle is the Mordell–Weil height pairing. For infinitesimally rigid sections we compute the GV contributions, recovering the E8E_8 theta product over a (−1)(-1)-curve. We formulate the corresponding conjecture for families of sections.

Introduction

Huang, Katz, Klemm, and Wang (HKKW) compute refined BPS numbers for compact elliptic Calabi–Yau threefolds using Wilson loops [25 §§3.3–3.4 and 5.1]. In their calculation, distinct sections can contribute to the same curve class [25 §§5.1.4 and 5.1.7]. Fix a smooth curve DD in the base. Which sections over DD occur in a given fiber degree, and what does each contribute to the Gopakumar–Vafa (GV) invariants after fibers are added? Since the projected cycle is DD with multiplicity one, each support has exactly one horizontal component, of degree one over DD. Theorem 3.1 describes the moduli fiber as the section varies, and Theorem 1.1 computes the ambient GV contribution of each infinitesimally rigid section.

In Gromov–Witten (GW) theory, such contributions enter the series of virtual counts of stable maps through the predicted GV multicover formula

∑β>0g≥0Ng,βGWλ2g−2Qβ=∑β>0,h≥0m≥1nh,βm(2sin⁡(mλ2))2h−2Qmβ,(1)\sum_{\substack{\beta>0\\g\geq0}} N^{\mathrm{GW}}_{g,\beta}\lambda^{2g-2}Q^\beta = \sum_{\substack{\beta>0,h\geq0\\m\geq1}} \frac{n_{h,\beta}}{m}\left(2\sin\left(\frac{m\lambda}{2}\right)\right)^{2h-2}Q^{m\beta}, \tag*{(1)}

where β>0\beta>0 runs over nonzero effective curve classes, Ng,βGWN^{\mathrm{GW}}_{g,\beta} is the connected genus-gg GW invariant, nh,βn_{h,\beta} is the genus-hh GV number in class β\beta, λ\lambda is the genus variable, and the formal symbols satisfy QβQγ=Qβ+γQ^\beta Q^\gamma=Q^{\beta+\gamma} [15 Part II, equation (3.2)]. The agreement of these numbers with those defined by vanishing cycles below is conjectural in general [38].

For an elliptic threefold, one can group the terms of this series by base and fiber degrees. Alim–Scheidegger and Klemm–Manschot–Wotschke studied the modular structure and holomorphic anomaly equations that arise in this way [1, 34]. At fixed base degree, Huang, Katz, and Klemm predicted that the all-genus partition function is expressed in meromorphic Jacobi forms with prescribed poles. In their formula the fiber parameter is the modular variable and the string coupling is the elliptic variable [23 equations (1.1)–(1.2)]. Their later work extends the calculation to refined invariants and examines the dependence on complex structure and orientation [24 §§3.5–3.7].

The Jacobi transformation has a monodromy interpretation. Schimmann computed the action of relative Fourier–Mukai monodromy on brane charges [46 §3], and Plioline and Schimmann relate this action to the transformation of the fixed-base partition function [43 §4]. On the enumerative side, Oberdieck and Pixton prove the quasi-Jacobi and holomorphic anomaly equations for GW invariants of rational elliptic surfaces [40]. In Pandharipande–Thomas (PT) theory one takes virtual counts of stable pairs OX→F\mathcal{O}_X \to\mathcal{F} with F\mathcal{F} pure of dimension one and zero-dimensional cokernel [42]. Here Oberdieck and Shen prove the elliptic transformation law for reduced base classes [41 Theorem 1].

HKKW, on the other hand, fix the fiber degree and allow the base degree to vary. Over P2\mathbb{P}^2 they propose a finite expansion of the compact partition function in local Wilson loops [25 §3.4.1, equation (3.58)], whose values are computed in [33, 26, 52]. For this expansion, a section and its added fibers count toward the same degree. Our first task is therefore to determine the degree of the section itself. Appendix A formulates the conjecture for families motivated by HKKW’s expansion.

We work over C\mathbb{C}. Let BB be a smooth projective del Pezzo surface, so that −KB-K_B is ample, and let π:X→B\pi: X \to B be a Weierstrass fibration with smooth projective total space and zero section σ:B→X\sigma: B \to X. Locally over BB, the fibration is defined in Weierstrass form by

y2=x3+ax+b,a∈H0(B,KB−4),b∈H0(B,KB−6).y^2 = x^3 + ax + b,\qquad a \in H^0(B,K_B^{-4}),\qquad b \in H^0(B,K_B^{-6}).

We assume that the discriminant divisor 4a3+27b2=04a^3 + 27b^2 = 0 is reduced with only ordinary cusps at the transverse intersections of a=0a = 0 and b=0b = 0, and that all fibers of π\pi are integral. We further assume that XX has trivial canonical bundle, H1(X,OX)=0H^1(X,\mathcal{O}_X) = 0, H2(X,Z)H_2(X,\mathbb{Z}) is torsion-free, and the Picard and curve groups decompose integrally as

Pic⁡(X)=Z[S]⊕π∗Pic⁡(B),H2(X,Z)=σ∗H2(B,Z)⊕Zf,(2)\operatorname{Pic}(X) = \mathbb{Z}[S] \oplus\pi^* \operatorname{Pic}(B),\qquad H_2(X,\mathbb{Z}) = \sigma_* H_2(B,\mathbb{Z}) \oplus\mathbb{Z}f, \tag*{(2)}

where S:=σ(B)S := \sigma(B) and ff is the class of an elliptic fiber. Since the fibers are integral, Wazir’s Shioda–Tate formula gives rank⁡MW⁡(X/B)=rank⁡Pic⁡(X)−rank⁡Pic⁡(B)−1\operatorname{rank}\operatorname{MW}(X/B) = \operatorname{rank}\operatorname{Pic}(X) - \operatorname{rank}\operatorname{Pic}(B) - 1 for the group MW⁡(X/B)\operatorname{MW}(X/B) of rational sections [53 Theorem 3.3]. The first integral hypothesis in (1.2) therefore forces the global Mordell–Weil rank to be zero.

To obtain sections of positive height, we restrict to a smooth connected curve D⊂BD \subset B whose preimage Y:=π−1(D)Y := \pi^{-1}(D) is smooth. Set

p:=π∣Y:Y→D,i:Y↪X,O:=σ(D),g:=g(D),χY:=χ(OY)=−KB⋅D>0.(3)p := \pi|_Y : Y \to D,\qquad i : Y \hookrightarrow X,\qquad O := \sigma(D),\qquad g := g(D),\qquad\chi_Y := \chi(\mathcal{O}_Y) = -K_B \cdot D > 0. \tag*{(3)}

For x∈Dx \in D, write fx=p−1(x)f_x = p^{-1}(x) for the scheme-theoretic fiber. Its class is FF on YY and f=i∗Ff = i_*F on XX. By a section C⊂YC \subset Y of pp, we mean the image of a morphism D→YD \to Y right inverse to pp. Under fiberwise addition on the generic fiber, the sections form the Mordell–Weil group MW⁡(Y/D)\operatorname{MW}(Y/D) with identity OO. We write ⟨ , ⟩\langle\ ,\ \rangle for the height pairing on this group and call ⟨C,C⟩\langle C,C\rangle the height of CC [49 Definition 8.5]. Since every fiber of pp is integral, Shioda’s formula [49 Theorem 8.6] gives

r(C):=12⟨C,C⟩=χY+C⋅O,(4)r(C) := \frac{1}{2}\langle C,C\rangle= \chi_Y + C \cdot O, \tag*{(4)}

and in particular r(O)=0r(O) = 0. By Proposition 2.1, r(C)r(C) is the fiber degree of CC in XX. § 2 compares this degree with the Jacobi index for a global section and distinguishes it from the discrete charge in Schimannek’s GV series.

To record the nn added fibers and their multiplicities, let Y[n]Y^{[n]} denote the Hilbert scheme of length-nn zero-dimensional subschemes of YY, let D(n):=Sym⁡nDD^{(n)} := \operatorname{Sym}^n D be the nn-th symmetric product of DD (with D(0)=D(0)D^{(0)} = D^{(0)} a single reduced point), and let Pic⁡d(D)\operatorname{Pic}^d(D) be the Picard variety of degree-dd line bundles on DD. Pushing the zero-cycle [Z]=∑z∈Supp⁡(Z)length⁡(OZ,z)[z][Z] = \sum_{z \in\operatorname{Supp}(Z)} \operatorname{length}(\mathcal{O}_{Z,z})[z] forward along pp defines the morphism

ρn:Y[n]⟶D(n),Z⟼p∗[Z].\rho_n : Y^{[n]} \longrightarrow D^{(n)},\qquad Z \longmapsto p_*[Z].

We now pass from these cycles to the sheaves they support. For an effective class β∈H2(X,Z)\beta\in H_2(X,\mathbb{Z}), let MβM_\beta be Simpson’s projective moduli scheme [50 Theorem 1.21] of Gieseker stable pure one-dimensional sheaves F\mathcal{F} on XX with class β\beta and χ(F)=1\chi(\mathcal{F}) = 1. Purity means that every nonzero subsheaf has one-dimensional support, and we write [F][\mathcal{F}] for the fundamental cycle weighted by generic lengths. Since χ(F)=1\chi(\mathcal{F}) = 1, stability is independent of the ample polarization and is equivalent to requiring χ(F′)≤0\chi(\mathcal{F}') \leq0 for every proper nonzero subsheaf F′⊂F\mathcal{F}' \subset\mathcal{F}. In particular, semistability coincides with stability [38 Remark 3.2]. Sending a sheaf to its support cycle defines the Hilbert–Chow morphism

hβ:(Mβ)red⟶Chow⁡β(X),F⟼[F],h_\beta: (M_\beta)_{\mathrm{red}} \longrightarrow\operatorname{Chow}_\beta(X),\qquad\mathcal{F} \longmapsto[\mathcal{F}],

from the underlying reduced scheme (Mβ)red(M_\beta)_{\mathrm{red}} to the Chow variety Chow⁡β(X)\operatorname{Chow}_\beta(X) of effective one-cycles of class β\beta [45 Corollary 7.15][38 Remark 3.6].

The role of hβh_\beta in the GV count comes from the BPS interpretation in terms of spins [15, 31]. Hosono, Saito, and Takahashi use the relative Lefschetz action for the support morphism to define the invariants [22]. Kiem and Li introduce vanishing cycles into this construction [32]. We use Maulik and Toda’s definition, which extracts the GV numbers from the direct image of vanishing cycles under hβh_\beta [38 Definition 3.7].

To specify the vanishing cycles, fix a trivialization of KXK_X. Then MβM_\beta carries a natural dd-critical structure [5 Corollary 3.19][38 Theorem 3.3] recording local presentations as critical loci of regular functions on smooth varieties [29]. An orientation is a line bundle on (Mβ)red(M_\beta)_{\mathrm{red}} whose square is the virtual canonical line bundle [29 Definition 2.31], and it determines a global perverse sheaf of vanishing cycles ϕMβ\phi_{M_\beta} by Brav, Bussi, Dupont, Joyce, and Szendrői [6 Theorem 6.9].

Following [38 Definition A.4], an orientation on a proper open and closed subscheme M⊂MβM \subset M_{\beta} is Calabi–Yau with respect to hβh_{\beta} if its line bundle is pulled back from the intermediate variety in the Stein factorization of hβ∣Mredh_{\beta}|_{M_{\mathrm{red}}}. For such an orientation, the GV contributions nhMn^{M}_{h} are defined by

∑a∈Zχ(pHaRhβ∗ϕM)ya=∑h≥0nhM(y+2+y−1)h.(5)\sum_{a\in\mathbb{Z}}\chi\left({}^{p}\mathcal{H}^{a}Rh_{\beta*}\phi_{M}\right)y^{a}=\sum_{h\geq0}n^{M}_{h}\left(y+2+y^{-1}\right)^{h}. \tag*{(5)}

where pHa{}^{p}\mathcal{H}^{a} denotes middle perverse cohomology and χ(K):=∑b∈Z(−1)bdim⁡Hb(Chow⁡β(X),K)\chi(\mathcal{K}):=\sum_{b\in\mathbb{Z}}(-1)^{b}\dim\mathbb{H}^{b}(\operatorname{Chow}_{\beta}(X),\mathcal{K}). When MM is smooth and endowed with its canonical orientation KMK_{M}, one has ϕM=QM[dim⁡M]\phi_{M}=\mathbb{Q}_{M}[\dim M]. The following theorem identifies such a component for each infinitesimally rigid section and each nn.

Theorem 1.1. Let C⊂YC\subset Y be a section of p:Y→Dp:Y\to D whose normal bundle NC/XN_{C/X} satisfies H0(C,NC/X)=0H^{0}(C,N_{C/X})=0. For every integer n≥0n\geq0, the locus of stable sheaves with support cycle

[C]+∑x∈Dmx[fx],mx≥0,∑x∈Dmx=n,(6)[C]+\sum_{x\in D}m_{x}[f_{x}],\qquad m_{x}\geq0,\qquad\sum_{x\in D}m_{x}=n, \tag*{(6)}

forms a smooth projective connected component MC,nM_{C,n} of Mσ∗[D]+(r(C)+n)fM_{\sigma_{*}[D]+(r(C)+n)f}, equipped with an isomorphism

MC,n≅Pic⁡g+r(C)(D)×Y[n].(7)M_{C,n}\cong\operatorname{Pic}^{g+r(C)}(D)\times Y^{[n]}. \tag*{(7)}

Under this identification, the support morphism factors through the projection onto Y[n]Y^{[n]} and sends (N,Z)(N,Z) to [C]+∑x∈Dmx[fx][C]+\sum_{x\in D}m_{x}[f_{x}], where ∑x∈Dmx[x]=ρn(Z)\sum_{x\in D}m_{x}[x]=\rho_{n}(Z). Moreover, the canonical orientation on MC,nM_{C,n} is Calabi–Yau with respect to the support morphism. Let nh,nCn^{C}_{h,n} be the genus-hh GV contribution of MC,nM_{C,n} in (1.5), and put w:=y+2+y−1w:=y+2+y^{-1}. With qq recording the number of added fibers, we have

∑n,h≥0nh,nCwhqn=wg∏m≥1(1+yqm)2g−2(1+y−1qm)2g−2(1−qm)12χY+4g−4.(8)\sum_{n,h\geq0}n^{C}_{h,n}w^{h}q^{n} = w^{g}\prod_{m\geq1} \frac{(1+yq^{m})^{2g-2}(1+y^{-1}q^{m})^{2g-2}} {(1-q^{m})^{12\chi_{Y}+4g-4}}. \tag*{(8)}

In particular, nh,nC=0n^{C}_{h,n}=0 unless g≤h≤g+ng\leq h\leq g+n, and

∑h≥0nh,0Cwh=wg,∑h≥0nh,1Cwh=12χYwg+(2g−2)wg+1.(9)\sum_{h\geq0}n^{C}_{h,0}w^{h}=w^{g},\qquad \sum_{h\geq0}n^{C}_{h,1}w^{h}=12\chi_{Y}w^{g}+(2g-2)w^{g+1}. \tag*{(9)}

For a fixed pair (D,C)(D,C), the section starts in fiber degree r(C)r(C), so its contribution with nn added fibers occurs in degree r(C)+nr(C)+n. In (8) the Picard variety contributes wgw^{g}, while the maps ρn\rho_{n} record the added fibers in D(n)D^{(n)} and give the infinite product. In this sense (8) is the pointwise version of Conjecture A.1, in which (D,C)(D,C) varies.

For a rational elliptic surface, the fixed-section product already appears in the relative Lefschetz calculation of Hosono, Saito, and Takahashi [22 Theorem 4.7]. In the compact threefold one must also determine when this surface calculation gives a contribution to the ambient GV invariant. Infinitesimal rigidity makes the fixed-section locus a smooth component of the threefold moduli space, on which we determine the orientation. Over a single line in P2\mathbb{P}^{2}, Example 5.3 gives six rigid sections in the same ambient class. For very general coefficients in this example, Corollary 5.4 computes the contribution of every nonzero section in the rank-two subgroup they generate.

There is a related local calculation. When all singular fibers of Y→DY\to D are nodal, Bryan and Kool compute the connected unweighted Donaldson–Thomas series of the total space of KYK_{Y} [8 Corollary 2]. Setting their variable p=−yp=-y gives (−w)g−1(-w)^{g-1} times the product in (8).

Finally, when DD is a (−1)(-1)-curve, every section is infinitesimally rigid and ∣D∣|D| is a point. We can therefore sum the fixed-section contributions to obtain the complete GV series in classes σ∗[D]+kf\sigma_{*}[D]+kf. § 6 identifies the resulting E8E_{8} product with the one-EE-string elliptic genus.

§ 2 and § 3 relate the heights of sections to their ambient curve classes and describe the moduli space over a fixed divisor. § 4 computes the perverse generating series, § 5 proves the rigidity and orientation statements, and § 6 gives the E8E_{8} specialization.

Heights of sections and ambient curve classes

On the smooth surface YY, we regard curves as divisors. Let NS⁡(Y)\operatorname{NS}(Y) denote the Néron–Severi group of line bundles on YY modulo algebraic equivalence, endowed with the intersection pairing. Relative duality for p:Y→Dp:Y\to D, adjunction on BB and XX, and the zero section O=S∣YO=S|_{Y} yield the surface identities

R1p∗OY=KB∣D,KY=NY/X=p∗ND/B,R^{1}p_{*}\mathcal{O}_{Y}=K_{B}|_{D},\qquad K_{Y}=N_{Y/X}=p^{*}N_{D/B},
O2=−χY,O⋅F=1,F2=0,D2−χY=2g−2.(10)O^{2}=-\chi_{Y},\qquad O\cdot F=1,\qquad F^{2}=0,\qquad D^{2}-\chi_{Y}=2g-2. \tag*{(10)}

Since deg⁡(KB∣D)=−χY<0\deg(K_{B|D})=-\chi_Y<0, we have H0(D,KB∣D)=0H^0(D,K_{B|D})=0. The Leray spectral sequence for pp therefore gives h1(Y,OY)=gh^1(Y,\mathcal{O}_Y)=g, and pullback along pp induces an isomorphism on identity components of Picard schemes,

p∗:Pic⁡0(D)→∼Pic⁡0(Y),(11)p^*: \operatorname{Pic}^0(D) \xrightarrow{\sim} \operatorname{Pic}^0(Y), \tag*{(11)}

with inverse given by restriction to OO.

Proposition 2.1. For a section C∈MW⁡(Y/D)C\in\operatorname{MW}(Y/D), define

φ(C):=[C]−[O]−r(C)F,Λ:={A∈NS⁡(Y)∣A⋅O=A⋅F=0}.\varphi(C):=[C]-[O]-r(C)F,\qquad\Lambda:=\{A\in\operatorname{NS}(Y)\mid A\cdot O=A\cdot F=0\}.

Then φ\varphi gives an isomorphism of groups MW⁡(Y/D)→∼Λ\operatorname{MW}(Y/D)\xrightarrow{\sim}\Lambda under which the height pairing takes the form ⟨C1,C2⟩=−φ(C1)⋅φ(C2)\langle C_1,C_2\rangle=-\varphi(C_1)\cdot\varphi(C_2). Moreover,

NS⁡(Y)=⟨O,F⟩⊕Λ,φ(C)2=−2r(C),i∗[C]=σ∗[D]+r(C)f.(12)\operatorname{NS}(Y)=\langle O,F\rangle\oplus\Lambda,\qquad\varphi(C)^2=-2r(C),\qquad i_*[C]=\sigma_*[D]+r(C)f. \tag*{(12)}

The group MW⁡(Y/D)\operatorname{MW}(Y/D) is torsion-free and contains only finitely many sections of bounded height, and every nonzero section C≠OC\ne O satisfies r(C)≥χYr(C)\ge\chi_Y.

Proof. Since every fiber of pp is integral, the local correction terms from reducible fibers in Shioda’s height formula vanish, giving ⟨C,C⟩=2χY+2C⋅O\langle C,C\rangle=2\chi_Y+2C\cdot O [49 Theorem 8.6]. For C≠OC\ne O, we have C⋅O≥0C\cdot O\ge0 and hence r(C)≥χY>0r(C)\ge\chi_Y>0, which shows that MW⁡(Y/D)\operatorname{MW}(Y/D) has no torsion. The Shioda–Tate theorem identifies NS⁡(Y)/⟨O,F⟩\operatorname{NS}(Y)/\langle O,F\rangle with MW⁡(Y/D)\operatorname{MW}(Y/D) [49 Theorem 1.3].

The classes O,FO,F span a free subgroup, and the quotient MW⁡(Y/D)\operatorname{MW}(Y/D) is torsion-free. Hence NS⁡(Y)\operatorname{NS}(Y) is torsion-free as well.

With respect to the basis O,FO,F, the intersection form on ⟨O,F⟩\langle O,F\rangle has matrix

(−χY110)\begin{pmatrix} -\chi_Y & 1\\ 1 & 0 \end{pmatrix}

which has determinant −1-1 and signature (1,1)(1,1). Orthogonal projection off ⟨O,F⟩\langle O,F\rangle is therefore defined over Z\mathbb{Z}, yielding the direct sum decomposition NS⁡(Y)=⟨O,F⟩⊕Λ\operatorname{NS}(Y)=\langle O,F\rangle\oplus\Lambda together with the projection φ\varphi and self-intersection φ(C)2=−2r(C)\varphi(C)^2=-2r(C). By the Hodge index theorem, the intersection form on Λ\Lambda is negative definite, so only finitely many elements of MW⁡(Y/D)\operatorname{MW}(Y/D) can have bounded height.

Finally, for any class A∈NS⁡(Y)A\in\operatorname{NS}(Y), combining the projection formula with (1.2) gives

i∗A=(A⋅F)σ∗[D]+(A⋅O+χY(A⋅F))f.(13)i_*A=(A\cdot F)\sigma_*[D]+(A\cdot O+\chi_Y(A\cdot F))f. \tag*{(13)}

Indeed, pairing i∗Ai_*A with pullbacks of divisors from BB determines the coefficient of σ∗[D]\sigma_*[D], while intersecting with SS and using S∣S=KBS|_S=K_B determines the coefficient of ff. It follows that i∗Λ=0i_*\Lambda=0, and evaluating (2.4) on [C][C] yields i∗[C]=σ∗[D]+r(C)fi_*[C]=\sigma_*[D]+r(C)f. □

Proposition 2.1 applies to a section of YY defined only over DD. To compare with the height divisor of a global section, let π~:X~→B\widetilde{\pi}:\widetilde{X}\to B be a smooth elliptic Calabi–Yau threefold with zero section S0S_0 and no reducible fibers in codimension one. Wazir defines a pairing of divisors on X~\widetilde{X} with values in Pic⁡(B)\operatorname{Pic}(B) by (A,A′)↦π~∗(A⋅A′)(A,A')\mapsto\widetilde{\pi}_*(A\cdot A') [53 Definition 3.2]. Write PP also for the divisor of a section. In the convention of Morrison and Park [39 equations (3.4)–(3.8)], the Shioda class of PP and the height pairing of P,QP,Q are

Σ(P)=P−S0−π~∗π~∗((P−S0)⋅S0),b(P,Q)=−π~∗(Σ(P)⋅Σ(Q)).\Sigma(P)=P-S_0-\widetilde{\pi}^*\widetilde{\pi}_*((P-S_0)\cdot S_0),\qquad b(P,Q)=-\widetilde{\pi}_*(\Sigma(P)\cdot\Sigma(Q)).

Thus the divisor-valued height pairing is the negative of Wazir’s pairing on the Shioda classes. Put b(P)=b(P,P)b(P)=b(P,P). If Y~=π~−1(D)\widetilde{Y}=\widetilde{\pi}^{-1}(D) is smooth with integral fibers and PD=P∣DP_D=P|_D, then Σ(P)∣Y~=φ(PD)\Sigma(P)|_{\widetilde{Y}}=\varphi(P_D) in NS⁡(Y~)\operatorname{NS}(\widetilde{Y}). The projection formula gives

12D⋅b(P)=−12(Σ(P)∣Y~)2=r(PD).\frac{1}{2}D\cdot b(P)=-\frac{1}{2}\bigl(\Sigma(P)|_{\widetilde{Y}}\bigr)^2=r(P_D).

The number D⋅b(P)/2D\cdot b(P)/2 is the Jacobi index of the charge variable in base class [D][D] [37 equation (3.45)]. Fierro Cota, Klemm, and Schimannek give the corresponding index matrix for several sections [13 equations (4.32)–(4.33)]. Schimannek uses the convention b=−b(P)b=-b(P) and writes the same index as −D⋅b/2-D\cdot b/2 [47 equations (4.3), (4.8)].

In (1.2), H2(X,Z)H_2(X,\mathbb{Z}) is torsion-free and r(C)r(C) records the ordinary fiber degree. The discrete charges in Schimannek’s genus-one examples arise from torsion curve classes, weighted by characters of a flat BB-field [47 equations (3.12)–(3.13)]. The Tate–Shafarevich group parametrizes the locally trivial genus-one torsors of the Jacobian [12]. In Schimannek’s example over P2\mathbb{P}^2, the two Z/2\mathbb{Z}/2 sectors in fiber degree one and base degree zero sum to the smooth-model invariant 284+256=540284+256=540 [47 §6.3, Tables 6–7]. Toda computes the local GVGV invariants in every pure fiber class of the smooth Weierstrass model [51 Theorem 3.5]. For the torsion grading, Katz, Klemm, Schimannek, and Sharpe propose a definition using stable sheaves on small resolutions of the nodal model [30 §3.5].

Moduli of stable sheaves over a fixed curve

We use the relative Fourier–Mukai equivalence of Bridgeland and Maciocia [7], with the normalization and moduli comparison of Diaconescu [11 §3.1 and Theorem 1.1]. Let p1,p2:X×BX→Xp_1,p_2:X\times_B X\to X be the two projections and Δ⊂X×BX\Delta\subset X\times_B X the relative diagonal with ideal sheaf IΔI_\Delta. Define the normalized kernel and integral transform by

P:=IΔ⊗p1∗OX(S)⊗p2∗OX(S)⊗(π∘p1)∗KB−1,Φ(−):=Rp2∗(Lp1∗(−)⊗LP).(14)\mathcal{P}:=I_\Delta\otimes p_1^*\mathcal{O}_X(S)\otimes p_2^*\mathcal{O}_X(S)\otimes(\pi\circ p_1)^*K_B^{-1},\qquad\Phi(-):=Rp_{2*}(Lp_1^*(-)\otimes^{L}\mathcal{P}). \tag*{(14)}

Here XX is identified with the relative compactified degree-zero Picard scheme over BB, which parametrizes rank-one torsion-free sheaves of degree zero on the fibers, via x↦Ix⊗OXπ(x)(S)x\mapsto I_x\otimes\mathcal{O}_{X_{\pi(x)}}(S), where IxI_x is the ideal of xx in its fiber. On a smooth fiber over b∈Bb\in B, this is x↦OXb(σ(b)−x)x\mapsto\mathcal{O}_{X_b}(\sigma(b)-x). See [11 §3.1] for this normalization and the resulting derived equivalence.

For k≥0k\geq0, put βk:=σ∗[D]+kf\beta_k:=\sigma_*[D]+kf. Diaconescu’s comparison theorem [11 Theorem 1.1 and equations (3.8)–(3.9)] identifies MβkM_{\beta_k} with the moduli space of adiabatically semistable pure two-dimensional sheaves EE on XX with Chern character

ch⁡1(E)=[Y],ch⁡2(E)=(1−χY2)f,ch⁡3(E)=−k,(15)\operatorname{ch}_1(E)=[Y],\qquad\operatorname{ch}_2(E)=\left(1-\frac{\chi_Y}{2}\right)f,\qquad\operatorname{ch}_3(E)=-k, \tag*{(15)}

where we identify ch⁡1\operatorname{ch}_1 and ch⁡2\operatorname{ch}_2 with homology classes via Poincaré duality and view ch⁡3\operatorname{ch}_3 as a rational number by taking degree. Concretely, if we fix polarizations ω^=S−sπ∗KB\widehat{\omega}=S-s\pi^*K_B and ωt=tS−sπ∗KB\omega_t=tS-s\pi^*K_B with s≫0s\gg0 and 0<t≪10<t\ll1 chosen relative to the numerical class, adiabatic semistability means Gieseker semistability with respect to ωt′\omega_{t'} for all 0<t′≤t0<t'\leq t. This comparison is an isomorphism of moduli stacks compatible with flat families, while MβkM_{\beta_k} itself does not depend on ω^\widehat{\omega}. The pushforward i∗Ji_*J of any rank-one torsion-free sheaf JJ on the smooth integral surface YY is Gieseker stable with respect to every ample divisor on XX, and hence belongs to the adiabatically stable locus whenever its Chern character is (3.2).

Next, we construct the projection to the linear system ∣D∣:=PH0(B,OB(D))|D|:=\mathbb{P}H^0(B,\mathcal{O}_B(D)),

sk:Mβk⟶∣D∣,Mk(D):=sk−1(D),(16)s_k:M_{\beta_k}\longrightarrow|D|,\qquad M_k(D):=s_k^{-1}(D), \tag*{(16)}

where Mk(D)M_k(D) denotes the scheme-theoretic fiber over D∈∣D∣D\in|D|. For a family FT\mathcal{F}_T of stable sheaves on X×TX\times T flat over TT, the relative transform with kernel (3.1) gives a TT-flat family of pure codimension-one sheaves on XX. The divisorial support of each fiber is the sum of its support divisors weighted by the lengths at their generic points. In families these divisors form a relative effective Cartier divisor on X×TX\times T whose formation commutes with base change [45 §11, Corollary 11.8][36, Lemmas 7.29 and 7.32]. Since H1(X,OX)=0H^1(X,\mathcal{O}_X)=0, the line bundle associated with this relative divisor agrees with π∗OB(D)\pi^*\mathcal{O}_B(D) up to twist by a line bundle from TT. Because H0(X,π∗OB(D))=H0(B,OB(D))H^0(X,\pi^*\mathcal{O}_B(D))=H^0(B,\mathcal{O}_B(D)), the divisor is pulled back from a unique relative effective divisor on B×TB\times T in ∣D∣|D|. The construction descends to sks_k because it commutes with base change and twisting a family by a line bundle from TT. To identify its value at F\mathcal{F}, localize at the generic point of each component Γ\Gamma of π∗[F]\pi_*[\mathcal{F}]. Filter the localized finite-length sheaf by powers of a uniformizer of OB,η\mathcal{O}_{B,\eta}. Over an algebraic closure of the residue field, the graded pieces have a total of mult⁡Γ(π∗[F])\operatorname{mult}_{\Gamma}(\pi_*[\mathcal{F}]) point factors. Each point factor transforms to a rank-one sheaf on the elliptic fiber [11 Lemma 3.1]. Hence the multiplicity of π−1(Γ)\pi^{-1}(\Gamma) in the divisorial support is the same. Components over closed points of BB contribute no divisor, so sk(F)=π∗[F]s_k(\mathcal{F})=\pi_*[\mathcal{F}].

On an elliptic surface with integral fibers, Yoshioka uses the shift d(C)=(C−O)⋅O=r(C)d(C)=(C-\mathcal{O})\cdot O=r(C) and a subscheme of length k−d(C)k-d(C) in his Fourier–Mukai correspondence [54 §3.2, equations (3.13)–(3.15) and Theorem 3.15]. For a rational elliptic surface, Hosono, Saito, and Takahashi give the support morphism in a fixed section class and attribute the description to Yoshioka [22 Proposition 4.8]. We identify the scheme-theoretic fiber in the compact threefold moduli space.

Theorem 3.1. The scheme-theoretic fiber Mk(D)M_k(D) in (3.3) is smooth and decomposes as

Mk(D)≅∏C∈MW⁡(Y/D)r(C)≤kPic⁡g+r(C)(D)×Y[k−r(C)].(17)M_k(D)\cong\prod_{\substack{C\in\operatorname{MW}(Y/D)\\ r(C)\leq k}}\operatorname{Pic}^{g+r(C)}(D)\times Y^{[k-r(C)]}. \tag*{(17)}

Under this isomorphism, a point (N,Z)(N,Z) in the component indexed by CC corresponds to a stable sheaf F\mathcal{F} with Fourier–Mukai transform

Φ(F)=i∗J,J=IZ⊗OY(O−C)⊗p∗N,deg⁡N=g+r(C),length⁡Z=k−r(C),(18)\Phi(\mathcal{F})=i_*J,\qquad J=I_Z\otimes\mathcal{O}_Y(O-C)\otimes p^*N,\qquad\deg N=g+r(C),\qquad\operatorname{length}Z=k-r(C), \tag*{(18)}

where IZ⊂OYI_Z \subset\mathcal{O}_Y is the ideal sheaf of ZZ. The multiplicities in the support cycle [F][F] of (1.6) are determined by ∑x∈Dmx[x]=ρk−r(C)(Z)=p∗[Z]\sum_{x \in D} m_x[x] = \rho_{k-r(C)}(Z) = p_*[Z].

Proof. Let FF be a geometric point of Mk(D)M_k(D) and set E:=Φ(F)E := \Phi(F). Because the divisorial support of EE is YY with multiplicity one, EE has length one at the generic point of YY and is annihilated there by any local defining equation of YY. Multiplication on EE by a local equation of YY therefore has image supported in dimension at most one. Purity forces this image to vanish, so E=i∗JE = i_*J for a rank-one torsion-free sheaf JJ on YY. Since YY is a smooth surface, L:=J∨∨L := J^{\vee\vee} is a line bundle and J≅IZ⊗LJ \cong I_Z \otimes L for a unique zero-dimensional subscheme Z⊂YZ \subset Y.

Put δ:=c1(L)\delta:= c_1(L), ν:=c1(NY/X)=D2F\nu:= c_1(N_{Y/X}) = D^2F, and n:=length⁡Zn := \operatorname{length} Z. Applying Grothendieck–Riemann–Roch to i:Y↪Xi : Y \hookrightarrow X yields

ch⁡2(i∗J)=i∗(δ−ν/2),ch⁡3(i∗J)=∫Y(δ2/2−n−δν/2+ν2/6).(19)\begin{aligned} \operatorname{ch}_2(i_*J) &= i_*(\delta-\nu/2),\\ \operatorname{ch}_3(i_*J) &= \int_Y(\delta^2/2-n-\delta\nu/2+\nu^2/6). \tag*{(19)} \end{aligned}

Combining (15), (10), and (13), we find

i∗δ=(1−χY2+D22)f=gf,δ⋅F=0.i_*\delta= \left(1-\frac{\chi_Y}{2}+\frac{D^2}{2}\right)f = gf,\qquad\delta\cdot F=0.

By Proposition 2.1, there exists a unique section C∈MW⁡(Y/D)C \in\operatorname{MW}(Y/D) such that

δ=−φ(C)+gF=O−C+(g+r(C))F,\delta=-\varphi(C)+gF=O-C+(g+r(C))F,

which by (11) is equivalent to L≅OY(O−C)⊗p∗NL \cong\mathcal{O}_Y(O-C)\otimes p^*N for a unique line bundle N∈Pic⁡g+r(C)(D)N \in\operatorname{Pic}^{g+r(C)}(D). Furthermore,

δ2=−2r(C),δν=ν2=0,−k=−r(C)−n,\delta^2=-2r(C),\qquad\delta\nu=\nu^2=0,\qquad-k=-r(C)-n,

which establishes the classification of closed points together with the numerical relations in (18). Conversely, every sheaf of the form (18) has Chern character (15) and is Gieseker stable for every polarization, so its inverse transform defines a point of Mk(D)M_k(D).

Now choose a Poincaré line bundle on D×Pic⁡g+r(C)(D)D \times\operatorname{Pic}^{g+r(C)}(D) and the universal ideal sheaf on Y×Y[n]Y \times Y^{[n]}. The resulting family gives a morphism to MβkM_{\beta_k}. Since the source is reduced and sks_k maps every geometric point to DD, this morphism factors through Mk(D)M_k(D). The map from the disjoint union of the products is proper and bijective on geometric points. To check the scheme structure, consider the change of rings sequence

0→Ext⁡Y1(J,J)→Ext⁡X1(i∗J,i∗J)→ηH0(Y,NY/X) ∂J↓ Ext⁡Y2(J,J)(20)\begin{CD} 0 @>>> \operatorname{Ext}^1_Y(J,J) @>>> \operatorname{Ext}^1_X(i_*J,i_*J) @>{\eta}>> H^0(Y,N_{Y/X})\\ @. @. @. @V{\partial_J}VV\\ @. @. @. \operatorname{Ext}^2_Y(J,J) \tag*{(20)} \end{CD}

where HomY(J,J)=OY\mathcal{H}om_Y(J,J)=\mathcal{O}_Y identifies the normal term. Since NY/X=p∗ND/BN_{Y/X}=p^*N_{D/B} and H1(B,OB)=0H^1(B,\mathcal{O}_B)=0,

H0(Y,NY/X)=H0(D,ND/B)≅TD∣D∣.H^0(Y,N_{Y/X})=H^0(D,N_{D/B})\cong T_D|D|.

On Y∖ZY\setminus Z, the sheaf JJ is invertible. In a first-order deformation, a local equation zz of YY acts on a lift of a local generator as multiplication by εa\varepsilon a, where ε2=0\varepsilon^2=0. The normal map η\eta sends this deformation to aa, and its divisorial support has equation z−εaz-\varepsilon a. Hence η\eta and dskds_k agree on Y∖ZY\setminus Z. They agree globally as sections of NY/XN_{Y/X}, giving TFMk(D)=ker⁡η=Ext⁡Y1(J,J)T_FM_k(D)=\ker\eta=\operatorname{Ext}^1_Y(J,J).

Since HomY(J,J)=OY\mathcal{H}om_Y(J,J)=\mathcal{O}_Y, Serre duality gives dim⁡Ext⁡Y2(J,J)=h2(OY)\dim\operatorname{Ext}^2_Y(J,J)=h^2(\mathcal{O}_Y). Riemann–Roch gives ∑j=02(−1)jdim⁡Ext⁡Yj(J,J)=χY−2n\sum_{j=0}^2(-1)^j\dim\operatorname{Ext}^j_Y(J,J)=\chi_Y-2n. With dim⁡HomY(J,J)=1\dim\mathcal{H}om_Y(J,J)=1, we obtain

dim⁡Ext⁡Y1(J,J)=g+2n.\dim\operatorname{Ext}^1_Y(J,J)=g+2n.

By Fogarty’s theorem [14 Theorem 2.4], each product Pic⁡g+r(C)(D)×Y[n]\operatorname{Pic}^{g+r(C)}(D)\times Y^{[n]} is smooth of dimension g+2ng+2n. The family map is finite by properness and quasi-finiteness. It is a homeomorphism, so the local dimension g+2ng+2n at each closed point equals the tangent dimension. Thus Mk(D)M_k(D) is regular. In characteristic zero the map is birational on each component, hence an isomorphism by normality.

It remains to determine the support cycle. Write Ψ=Φ−1\Psi=\Phi^{-1}. Restricting the kernel (14) along CC gives OY(O−C)\mathcal{O}_Y(O-C) up to a pullback from DD. Thus Ψ(i∗L)\Psi(i_*L) is a line bundle on CC, extended by zero to XX. For z∈Yz\in Y, let kzk_z be its skyscraper sheaf. Then Ψ(kz)[−1]\Psi(k_z)[-1] is rank one and torsion-free on fp(z)f_p(z) [11 Lemma 3.1 and equation (3.5)]. A composition series of L∣ZL|_Z shows that Ψ(i∗(L∣Z))[−1]\Psi(i_*(L|_Z))[-1] is a sheaf filtered by the Ψ(kz)[−1]\Psi(k_z)[-1], each with multiplicity length⁡(OZ,z)\operatorname{length}(\mathcal{O}_{Z,z}). Applying Ψ\Psi to 0→i∗J→i∗L→i∗(L∣Z)→00 \to i_*J \to i_*L \to i_*(L|_Z) \to0 therefore gives a short exact sequence

0⟶Ψ(i∗(L∣Z))[−1]⟶F⟶Ψ(i∗L)⟶0.0 \longrightarrow\Psi(i_*(L|_Z))[-1] \longrightarrow\mathcal{F} \longrightarrow\Psi(i_*L) \longrightarrow0.

Additivity of fundamental cycles gives [F]=[C]+∑z∈Supp⁡(Z)length⁡(OZ,z)[fp(z)][\mathcal{F}] = [C] + \sum_{z \in\operatorname{Supp}(Z)} \operatorname{length}(\mathcal{O}_{Z,z})[f_p(z)], as claimed. □\square

Perverse direct images and generating series

Recall that ρn:Y[n]→D(n)\rho_n : Y^{[n]} \to D^{(n)} sends a length-nn subscheme ZZ to the divisor p∗[Z]p_*[Z] under the elliptic fibration p:Y→Dp : Y \to D. Define

An:=Rρn∗QY[n][2n],Qn(y):=∑a∈Zχ(pHaAn)ya,A0:=Q,Q0(y):=1.A_n := R\rho_{n*}\mathbb{Q}_{Y^{[n]}}[2n], \qquad Q_n(y) := \sum_{a \in\mathbb{Z}} \chi({}^{p}\mathcal{H}^{a}A_n)y^a, \qquad A_0 := \mathbb{Q}, \quad Q_0(y) := 1.

Thus Qn(y)Q_n(y) records the Euler characteristics on D(n)D^{(n)} of the perverse cohomology sheaves of AnA_n. Hosono, Saito, and Takahashi discuss the extension of their rational elliptic surface calculation to other elliptic surfaces in the remark following the proof of [22, Theorem 4.7].

Lemma 4.1. With g=g(D)g = g(D) and χY=χ(OY)\chi_Y = \chi(\mathcal{O}_Y), one has

∑n≥0Qn(y)qn=∏m≥1(1+yqm)2g−2(1+y−1qm)2g−2(1−qm)12χY+4g−4.(21)\sum_{n \ge0} Q_n(y)q^n = \prod_{m \ge1} \frac{(1+yq^m)^{2g-2}(1+y^{-1}q^m)^{2g-2}}{(1-q^m)^{12\chi_Y+4g-4}}. \tag*{(21)}

Proof. Let j:U↪Dj : U \hookrightarrow D denote the open locus over which pp is smooth, and set V:=(R1p∗QY)∣UV := (R^1p_*\mathbb{Q}_Y)|_U with intersection complex IC⁡D(V):=j!∗V[1]\operatorname{IC}_D(V) := j_{!*}V[1]. The decomposition theorem of Beilinson, Bernstein, Deligne, and Gabber [4 Theorem 6.2.5], applied to p:Y→Dp : Y \to D, gives

A1=Rp∗QY[2]≅QD[2]⊕IC⁡D(V)⊕QD.(22)A_1 = Rp_*\mathbb{Q}_Y[2] \cong\mathbb{Q}_D[2] \oplus\operatorname{IC}_D(V) \oplus\mathbb{Q}_D. \tag*{(22)}

At x∈Dx \in D, the multiplicity of a skyscraper summand is one less than the number of irreducible components of fxf_x, hence zero. This also follows by comparing degree-two stalks, since H2(fx,Q)=QH^2(f_x,\mathbb{Q}) = \mathbb{Q}, and applying Verdier duality.

By the canonical bundle formula for elliptic surfaces and Noether’s formula, the topological Euler characteristic of YY is e(Y)=12χYe(Y) = 12\chi_Y. Since b1(Y)=dim⁡H1(Y,Q)=2gb_1(Y) = \dim H^1(Y,\mathbb{Q}) = 2g, we obtain b2(Y)=12χY+4g−2b_2(Y) = 12\chi_Y + 4g - 2. Taking hypercohomology in (22) shows that Hb(D,IC⁡D(V))H^b(D,\operatorname{IC}_D(V)) vanishes for b≠0b \ne0 and has dimension 12χY+4g−412\chi_Y + 4g - 4 for b=0b = 0. The nonzero bigraded hypercohomology spaces Hb(D,pHaA1)H^b(D,{}^p\mathcal{H}^a A_1) are therefore recorded by

(a,b)dim⁡Hb(D,pHaA1)a+b(mod2)(1,1),(1,−1),(−1,1),(−1,−1)1 for each pair0(1,0),(−1,0)2g for each pair1(0,0)12χY+4g−40(23)\begin{array}{c|c|c} (a,b) & \dim H^b(D,{}^p\mathcal{H}^a A_1) & a+b \pmod2 \\ \hline (1,1),(1,-1),(-1,1),(-1,-1) & 1\ \text{for each pair} & 0 \\ (1,0),(-1,0) & 2g\ \text{for each pair} & 1 \\ (0,0) & 12\chi_Y+4g-4 & 0 \tag*{(23)} \end{array}

For a partition λ=(1m12m2⋯ )\lambda= (1^{m_1}2^{m_2}\cdots) of nn, consider the finite map

μλ:∏j:mj>0D(mj)⟶D(n),(ξj)j⟼∑jξj.\mu_\lambda: \prod_{j:m_j>0} D^{(m_j)} \longrightarrow D^{(n)}, \qquad(\xi_j)_j \longmapsto\sum_j \xi_j.

Pushing forward the Hilbert–Chow decomposition [17 Theorem 3 and Proposition 1] (in the form of [10 Theorem 4.1.1]) to D(n)D^{(n)} gives an isomorphism of rational constructible complexes

An≅⨁λ⊢nRμλ∗(⊠j≥1Sym⁡mjA1).(24)A_n \cong\bigoplus_{\lambda\vdash n} R\mu_{\lambda*}\left(\boxtimes_{j \ge1}\operatorname{Sym}^{m_j} A_1\right). \tag*{(24)}

where Sym⁡mK\operatorname{Sym}^m K denotes the invariant summand of the finite pushforward of K⊠mK^{\boxtimes m} to D(m)D^{(m)} with the usual Koszul signs. We omit Tate twists for rational complexes. For the induced perverse filtration on the Hilbert scheme, see [55 Proposition 4.12] and [48 Proposition 2.2].

Since finite pushforwards are perverse exact [4 §2.2.6], external products and the Künneth formula reduce the calculation to the bigraded vector space in (23), with permutation parity a+ba+b and Euler characteristic sign (−1)b(-1)^b. For each part size mm, the three rows in (23) contribute (1+yqm)−2(1+y−1qm)−2(1+yq^m)^{-2}(1+y^{-1}q^m)^{-2}, (1+yqm)2g(1+y−1qm)2g(1+yq^m)^{2g}(1+y^{-1}q^m)^{2g}, and (1−qm)−(12χY+4g−4)(1-q^m)^{-(12\chi_Y+4g-4)}, respectively, and multiplying over m≥1m \ge1 proves (21). $\square`

Setting y=−1y=-1 in Lemma 4.1 recovers the Euler characteristic specialization of Göttsche’s formula [16],

Qn(−1)=e(Y[n]),∑n≥0e(Y[n])qn=∏m≥1(1−qm)−12χY.Q_n(-1)=e(Y^{[n]}),\qquad\sum_{n\geq0}e(Y^{[n]})q^n=\prod_{m\geq1}(1-q^m)^{-12\chi_Y}.

For a fixed section CC and nn added fibers, the total fiber degree is k=r(C)+nk=r(C)+n. The morphism

τC,n:D(n)⟶Chow⁡βk(X),∑xmx[x]⟼[C]+∑xmx[fx]\tau_{C,n}:D^{(n)}\longrightarrow\operatorname{Chow}_{\beta_k}(X),\qquad\sum_x m_x[x]\longmapsto[C]+\sum_xm_x[f_x]

is injective on geometric points since the one-cycle recovers its fiber multiplicities, and it is finite by properness and quasi-finiteness. On the component Pic⁡g+r(C)(D)×Y[n]\operatorname{Pic}^{g+r(C)}(D)\times Y^{[n]} of Theorem 3.1, the support map is hC,nD=τC,n∘ρn∘pr⁡2h^D_{C,n}=\tau_{C,n}\circ\rho_n\circ\operatorname{pr}_2, where pr⁡2\operatorname{pr}_2 is projection to Y[n]Y^{[n]}. The Künneth formula gives

R(hC,nD)∗IC⁡≅(τC,n)∗(RΓ(Pic⁡g+r(C)(D),Q)[g]⊗An).(25)R(h^D_{C,n})_*\operatorname{IC}\cong(\tau_{C,n})_*\left(R\Gamma(\operatorname{Pic}^{g+r(C)}(D),\mathbb{Q})[g]\otimes A_n\right). \tag*{(25)}

Here IC⁡\operatorname{IC} is the constant sheaf on the product shifted by g+2ng+2n. Since finite pushforward is perverse exact and

∑i=02gdim⁡Hi(Pic⁡d(D),Q)yi−g=y−g(1+y)2g=(y+2+y−1)g=wg,\sum_{i=0}^{2g}\dim H^i(\operatorname{Pic}^d(D),\mathbb{Q})y^{i-g}=y^{-g}(1+y)^{2g}=(y+2+y^{-1})^g=w^g,

the contribution of the intersection complex on Pic⁡g+r(C)(D)×Y[n]\operatorname{Pic}^{g+r(C)}(D)\times Y^{[n]} is wgQn(y)w^gQ_n(y).

Proposition 4.2. Fix βk=σ∗[D]+kf\beta_k=\sigma_*[D]+kf. On the fiber Mk(D)=sk−1(D)M_k(D)=s_k^{-1}(D) over the projected curve DD, let IC⁡Mk(D)\operatorname{IC}_{M_k(D)} be the sum of the constant sheaves on the components of {(xref)}(3.4), each shifted by the dimension of that component. Define

PD,k(y):=∑aχ(p ⁣HaR(hβk∣Mk(D))∗IC⁡Mk(D))ya,ΘY(q):=∑C∈MW⁡(Y/D)qr(C).P_{D,k}(y):=\sum_a\chi\left({}^p\!H^aR(h_{\beta_k}|_{M_k(D)})_*\operatorname{IC}_{M_k(D)}\right)y^a,\qquad\Theta_Y(q):=\sum_{C\in\operatorname{MW}(Y/D)}q^{r(C)}.

Then

∑k≥0PD,k(y)qk=wgΘY(q)∏m≥1(1+yqm)2g−2(1+y−1qm)2g−2(1−qm)12χY+4g−4.(26)\sum_{k\geq0}P_{D,k}(y)q^k=w^g\Theta_Y(q)\prod_{m\geq1}\frac{(1+yq^m)^{2g-2}(1+y^{-1}q^m)^{2g-2}}{(1-q^m)^{12\chi_Y+4g-4}}. \tag*{(26)}

Proof. By Proposition 2.1, the height bound ensures that each coefficient of ΘY(q)\Theta_Y(q) is finite. Summing (25) over the components of {(xref)}(3.4) gives

PD,k(y)=wg∑C∈MW⁡(Y/D)r(C)≤kQk−r(C)(y).P_{D,k}(y)=w^g\sum_{\substack{C\in\operatorname{MW}(Y/D)\\r(C)\leq k}}Q_{k-r(C)}(y).

Multiplying by qkq^k and summing over kk gives {(xref)}(4.6) by Lemma 4.1. □

Formula {(xref)}(4.6) concerns the intrinsic intersection complexes on the fixed scheme-theoretic fiber Mk(D)M_k(D). When DD or CC deforms in XX, this fiber need not be open in MβkM_{\beta_k}, and the next section determines when the component calculation computes the ambient GV invariants {(xref)}(1.5).

Normal deformations and vanishing cycles

Proposition 5.1. Let J=IZ⊗OY(O−C)⊗p∗NJ=I_Z\otimes\mathcal{O}_Y(\mathcal{O}-C)\otimes p^*N be the sheaf on YY from {(xref)}(3.5), with n=length⁡Zn=\operatorname{length}Z. Then

dim⁡Ext⁡X1(i∗J,i∗J)=g+2n+h0(C,NC/X).(27)\dim\operatorname{Ext}^1_X(i_*J,i_*J)=g+2n+h^0(C,N_{C/X}). \tag*{(27)}

Proof. The differential of pp splits along CC, so pp is smooth there. By adjunction, NC/Y=KC⊗KY−1∣C=KD⊗ND/B−1=KB∣DN_{C/Y}=K_C\otimes K_Y^{-1}|_C=K_D\otimes N_{D/B}^{-1}=K_B|_D. The normal bundle sequence for C⊂Y⊂XC\subset Y\subset X is therefore

0⟶KB∣D⟶NC/X⟶ND/B⟶0.(28)0\longrightarrow K_B|_D\longrightarrow N_{C/X}\longrightarrow N_{D/B}\longrightarrow0. \tag*{(28)}

Let δC:H0(D,ND/B)⟶H1(D,KB∣D)\delta_C:H^0(D,N_{D/B})\longrightarrow H^1(D,K_B|_D) be the connecting map. As H0(D,KB∣D)=0H^0(D,K_B|_D)=0, its kernel is H0(C,NC/X)H^0(C,N_{C/X}). Moreover, the trace map tr⁡:Ext⁡Y2(J,J)⟶H2(Y,OY)\operatorname{tr}:\operatorname{Ext}^2_Y(J,J)\longrightarrow H^2(Y,\mathcal{O}_Y) is an isomorphism, since its Serre dual is the isomorphism H0(Y,KY)⟶Hom⁡Y(J,J⊗KY)H^0(Y,K_Y)\longrightarrow\operatorname{Hom}_Y(J,J\otimes K_Y) given by the identity on JJ.

For v∈H0(Y,NY/X)v\in H^0(Y,N_{Y/X}), let YvY_v be the corresponding deformation over C[ε]/(ε2)\mathbb{C}[\varepsilon]/(\varepsilon^2). Choose a local equation z=0z=0 for YY and a lift aa of v(z)v(z), so that YvY_v has equation z−εa=0z-\varepsilon a=0. For a flat deformation EE of i∗Ji_*J, multiplication by ε\varepsilon identifies E/εEE/\varepsilon E with εE\varepsilon E. Under this identification, multiplication by zz induces the endomorphism of JJ given by η(E)(z)\eta(E)(z), where η\eta is the normal map in {(xref)}(3.7). Thus $\eta(E)=v precisely when (z−εa)E=0(z-\varepsilon a)E=0 on each such chart. Equivalently, EE is the pushforward of a flat lift of JJ to YvY_v. Exactness of (20) gives

oJ(v)=0  ⟺  J lifts to Yv.o_J(v)=0 \iff J\text{ lifts to }Y_v.

Write κ(v)∈Ext⁡Y1(ΩY,OY)\kappa(v)\in\operatorname{Ext}^{1}_{Y}(\Omega_Y,\mathcal{O}_Y) for the Kodaira–Spencer class of YvY_v and At⁡(J)∈Ext⁡Y1(J,J⊗ΩY)\operatorname{At}(J)\in\operatorname{Ext}^{1}_{Y}(J,J\otimes\Omega_Y) for the Atiyah class of JJ. Since YY is smooth, JJ is perfect for every ZZ. Its lifting obstruction is (1J⊗κ(v))∘At⁡(J)(1_J\otimes\kappa(v))\circ\operatorname{At}(J), whose trace is the obstruction κ(v)⌟c1(det⁡J)\kappa(v)\mathbin{\lrcorner}c_1(\det J) to lifting det⁡J\det J [27 Corollary 3.4 and §4.2]. The trace is an isomorphism, so JJ lifts if and only if det⁡J\det J lifts. Now

det⁡IZ=OY,det⁡J=OY(O−C)⊗p∗N.\det I_Z=\mathcal{O}_Y,\qquad\det J=\mathcal{O}_Y(O-C)\otimes p^*N.

The factor OY(O)\mathcal{O}_Y(O) extends from OX(S)\mathcal{O}_X(S), and NN extends to the induced deformation of DD since H2(D,OD)=0H^2(D,\mathcal{O}_D)=0. Hence oJ(v)=0o_J(v)=0 precisely when OY(C)\mathcal{O}_Y(C) lifts to YvY_v. We show that this is equivalent to lifting the divisor CC. Let sCs_C be its canonical section. On every integral genus-one fiber OY(C)\mathcal{O}_Y(C) has degree one, so sCs_C spans H0H^0 and H1=0H^1=0. Cohomology and base change give

p∗OY(C)=OD,R1p∗OY(C)=0,H1(Y,OY)→sCH1(Y,OY(C)) is an isomorphism.p_*\mathcal{O}_Y(C)=\mathcal{O}_D,\qquad R^1p_*\mathcal{O}_Y(C)=0,\qquad H^1(Y,\mathcal{O}_Y)\xrightarrow{s_C} H^1(Y,\mathcal{O}_Y(C))\text{ is an isomorphism}.

Indeed, H0(D,R1p∗OY)=0H^0(D,R^1p_*\mathcal{O}_Y)=0 by (10).

If OY(C)\mathcal{O}_Y(C) lifts to YvY_v over C[ε]/(ε2)\mathbb{C}[\varepsilon]/(\varepsilon^2), choose a lift LvL_v. The obstruction to lifting sCs_C lies in H1(Y,OY(C))H^1(Y,\mathcal{O}_Y(C)). Twisting LvL_v by a line bundle with transition functions 1+εuij1+\varepsilon u_{ij} changes that obstruction by the class of (uijsC)(u_{ij}s_C). The displayed isomorphism allows us to cancel it. The lifted section cuts out a relative Cartier divisor lifting CC. Conversely, a lifted Cartier divisor gives a lift of OY(C)\mathcal{O}_Y(C).

The normal sequence (28) says that CC lifts inside YvY_v precisely when δC(v)=0\delta_C(v)=0. Hence

ker⁡oJ=ker⁡δC≅H0(C,NC/X).(29)\ker o_J=\ker\delta_C\cong H^0(C,N_{C/X}). \tag*{(29)}

Substitution into (20) proves (27).

Proof of Theorem 1.1. By Theorem 3.1, the product MC,nM_{C,n} in (7) is a smooth projective connected component of the moduli fiber Mr(C)+n(D)M_{r(C)+n}(D). Since the Fourier–Mukai equivalence preserves Ext⁡1\operatorname{Ext}^1, Proposition 5.1 and the vanishing H0(C,NC/X)=0H^0(C,N_{C/X})=0 imply that the ambient tangent dimension at each closed point of MC,nM_{C,n} is g+2ng+2n. At such a point the local dimension lies between dim⁡MC,n=g+2n\dim M_{C,n}=g+2n and the tangent dimension. The ambient local ring is therefore regular. The ideal of MC,nM_{C,n} in each such regular local ring is zero, since a nonzero ideal in a local domain lowers the dimension of the quotient. Thus the closed immersion MC,n↪M[C]+nfM_{C,n}\hookrightarrow M_{[C]+nf} is also open and identifies MC,nM_{C,n} with a connected component of M[C]+nfM_{[C]+nf}.

Let pr⁡2:MC,n→Y[n]\operatorname{pr}_2:M_{C,n}\to Y^{[n]} be the second projection. The line bundle ND/B⊠nN_{D/B}^{\boxtimes n} on DnD^n descends to a line bundle ND/B(n)N_{D/B}^{(n)} on D(n)D^{(n)} because point stabilizers act trivially on its fibers. Since the Picard variety has trivial canonical bundle, KY=p∗ND/BK_Y=p^*N_{D/B}, and the Hilbert–Chow morphism is crepant, the local construction of the canonical bundle of Y[n]Y^{[n]} in [3 §6] shows that KY[n]K_{Y^{[n]}} is the pullback of the symmetric descent of KY⊠nK_Y^{\boxtimes n} on Y(n)Y^{(n)}, and hence

KMC,n=pr⁡2∗KY[n]=(ρn∘pr⁡2)∗ND/B(n).(30)K_{M_{C,n}}=\operatorname{pr}_2^*K_{Y^{[n]}}=(\rho_n\circ\operatorname{pr}_2)^*N_{D/B}^{(n)}. \tag*{(30)}

The map ρn∘pr⁡2\rho_n\circ\operatorname{pr}_2 sends (N,Z)(N,Z) to p∗[Z]p_*[Z]. Over x1+⋯+xnx_1+\cdots+x_n with distinct xi∈Dx_i\in D and smooth fibers p−1(xi)p^{-1}(x_i), its fiber is Pic⁡g+r(C)(D)×∏ip−1(xi)\operatorname{Pic}^{g+r(C)}(D)\times\prod_i p^{-1}(x_i). Thus the proper surjective map has geometrically connected generic fiber, also for n=0n=0. Since its source is smooth and connected and its target D(n)D^{(n)} is normal, Stein factorization gives

(ρn∘pr⁡2)∗OMC,n=OD(n).(\rho_n\circ\operatorname{pr}_2)_*\mathcal{O}_{M_{C,n}}=\mathcal{O}_{D^{(n)}}.

The map τC,n\tau_{C,n} from D(n)D^{(n)} to the Chow variety adds [C][C] to the fiber cycle. It is finite, so this is the Stein factorization of the support morphism. On a smooth dd-critical scheme, the virtual canonical line bundle is KMC,n⊗2K_{M_{C,n}}^{\otimes2} [29 Theorem 2.28], and (30) shows that its square root KMC,nK_{M_{C,n}} descends to the Stein factor and therefore defines a Calabi–Yau orientation in the sense of [38 Definition A.4]. With this orientation, the smooth critical chart with zero potential gives [6 Theorem 6.9]

ϕMC,n=QMC,n[g+2n].\phi_{M_{C,n}}=\mathbb{Q}_{M_{C,n}}[g+2n].

Consequently, the left-hand side of (5) is wgQn(y)w^gQ_n(y), and Lemma 4.1 proves (8).

Finally, each Qn(y)Q_n(y) is a symmetric Laurent polynomial with exponents between −n-n and nn, and hence is an integral polynomial in w=y+2+y−1w=y+2+y^{-1} of degree at most nn. Multiplying by wgw^g gives the genus bound, while extracting the coefficient of qq in (21) gives 12χY+(2g−2)w12\chi_Y+(2g-2)w, proving (9). ∎

Remark 5.2. For B=P2B = \mathbb{P}^{2} and a smooth plane curve DD of degree dd, we have g=(d−1)(d−2)/2g = (d-1)(d-2)/2 and χY=3d\chi_{Y} = 3d. In this case, the vanishing condition in Theorem 1.1 is equivalent to the injectivity of

H0(D,OD(d))→δCH1(D,OD(−3))H^{0}(D,\mathcal{O}_{D}(d)) \xrightarrow{\delta_{C}} H^{1}(D,\mathcal{O}_{D}(-3))

from (28), where both source and target have dimension d(d+3)/2d(d+3)/2. By Proposition 2.1, every nonzero section satisfies r(C)≥3dr(C) \geq3d, which for d=1d=1 matches the threshold for additional sections in HKKW [25 §5.1.7].

Example 5.3. Let [s:t:u][s:t:u] be homogeneous coordinates on B=P2B = \mathbb{P}^{2} and let D=(u=0)D = (u=0). Choose Weierstrass coefficients

a=t12+uA11,b=s18+us7t10+u2B16.a=t^{12}+uA_{11}, \qquad b=s^{18}+us^{7}t^{10}+u^{2}B_{16}.

where A11A_{11} and B16B_{16} are general homogeneous polynomials of the indicated degrees. Along DD, the discriminant restricts to 4t36+27s364t^{36}+27s^{36}, which has 36 simple zeros, and hence DD meets the discriminant transversely. Off DD, varying A11A_{11} and B16B_{16} prescribes the first jets of a,ba,b independently. For general coefficients the pair (a,b)(a,b) is transverse to the smooth part of 4a3+27b2=04a^{3}+27b^{2}=0 and to its cusp point. Thus XX is smooth with the required discriminant, and Y=π−1(D)Y=\pi^{-1}(D) is smooth.

The smooth Weierstrass models form an open subset of the affine space of coefficients. Its general member is the crepant resolution of a degree-18 hypersurface in P(1,1,1,6,9)\mathbb{P}(1,1,1,6,9) [24 §2]. The five rays corresponding to the vertices generate the toric lattice Z5/Z(1,1,1,6,9)\mathbb{Z}^{5}/\mathbb{Z}(1,1,1,6,9). The formulas of Batyrev and Kreuzer [2 Theorems 1.6 and 3.8] therefore give trivial fundamental and Brauer groups. In particular Pic⁡(X)\operatorname{Pic}(X) and H2(X,Z)H_{2}(X,\mathbb{Z}) are torsion-free, using the exponential and universal coefficient sequences. The same torsion vanishing holds throughout the smooth family. Diaconescu’s integral bases modulo torsion [11 Lemma 2.1] now give (2).

In the chart z=1z=1, let CϵC_{\epsilon} be the section over DD given by

x=ϵt6,y=s9,ϵ∈{0,−1,−−1}.x=\epsilon t^{6}, \qquad y=s^{9}, \qquad\epsilon\in\{0,\sqrt{-1},-\sqrt{-1}\}.

Indeed, x3+t12x=0x^{3}+t^{12}x=0 for these three values of xx. Each section is disjoint from OO, so r(Cϵ)=3r(C_{\epsilon})=3. Any two meet only at t=0t=0, where s=1s=1 and (t,x)(t,x) are local coordinates on YY since 2y≠02y\ne0. Their local intersection has length six. By Proposition 2.1 the height matrix of C0C_{0} and C−1C_{\sqrt{-1}} is

(6−3−36),\begin{pmatrix} 6 & -3 \\ -3 & 6 \end{pmatrix},

of determinant 27. Thus C0C_{0} and C−1C_{\sqrt{-1}} are independent. The three displayed points are collinear on the generic cubic, so their Mordell–Weil sum is OO.

Differentiating the Weierstrass equation along CϵC_{\epsilon} gives

0⟶NCϵ/X⟶OD(1)⊕OD(6)⊕OD(9)→dOD(18)⟶0,0 \longrightarrow N_{C_{\epsilon}/X} \longrightarrow\mathcal{O}_{D}(1)\oplus\mathcal{O}_{D}(6)\oplus\mathcal{O}_{D}(9) \xrightarrow{d} \mathcal{O}_{D}(18) \longrightarrow0,

where d=(−s7t10−ϵt6A11∣D,−(3ϵ2+1)t12,2s9)d=(-s^{7}t^{10}-\epsilon t^{6}A_{11}|_{D},-(3\epsilon^{2}+1)t^{12},2s^{9}). At A11∣D=0A_{11}|_{D}=0, the three summands on global sections span the monomials s18−jtjs^{18-j}t^{j} with j=7,8,…,18j=7,8,\ldots,18, respectively. Hence H0(d)H^{0}(d) is an isomorphism for all three sections for general A11A_{11}. Each normal bundle has degree −2-2 and no global sections, so NCϵ/X≅OD(−1)⊕2N_{C_{\epsilon}/X}\cong\mathcal{O}_{D}(-1)^{\oplus2}. Inversion gives three further rigid sections with y=−s9y=-s^{9}. Thus in the single ambient class σ∗[D]+3f\sigma_{*}[D]+3f, Theorem 1.1 gives at least six isolated points, one for each of the six sections.

More generally, write [m]Cϵ[m]C_{\epsilon} for the Mordell–Weil multiple by a nonzero integer mm. On the smooth locus of π\pi, the Weierstrass group law is regular [28 Corollary 12]. Multiplication by mm is étale there since its relative differential is multiplication by mm after translation. It maps CϵC_{\epsilon} isomorphically to [m]Cϵ[m]C_{\epsilon} over DD, identifying their normal bundles. Since r([m]Cϵ)=3m2r([m]C_{\epsilon})=3m^{2}, every such section gives a component isomorphic to Y[n]Y^{[n]} in class σ∗[D]+(3m2+n)f\sigma_{*}[D]+(3m^{2}+n)f for each n≥0n\geq0. For n=0,1,2n=0,1,2, its GV contributions are

1,36−2w,702−78w+3w2.1,\qquad36-2w,\qquad702-78w+3w^{2}.

Let C=C0C=C_{0}. At w=0w=0, (8) gives ∑n≥0n0,n6qn=∏m≥1(1−qm)−36\sum_{n\geq0}n^{6}_{0,n}q^{n}=\prod_{m\geq1}(1-q^{m})^{-36}. This is the factor for adding fibers to an isolated section over a line in Greer’s genus-zero GW calculation [18 p. 3, before Conjecture 1.6]. Greer uses Noether–Lefschetz theory to count rational curves of base degree one on very general Weierstrass fibrations [19].

Corollary 5.4. In Example 5.3, put P=C0P=C_{0} and Q=C−1Q=C_{\sqrt{-1}}. For very general A11,B16A_{11},B_{16}, every nonzero section in ZP+ZQ\mathbb{Z}P+\mathbb{Z}Q is infinitesimally rigid in XX. With qq recording total fiber degree, the combined GV contribution of these sections is

(∑(a,b)∈Z2∖{(0,0)}q3(a2−ab+b2))∑n≥0Qn(y)qn,\left(\sum_{(a,b)\in\mathbb{Z}^{2}\setminus\{(0,0)\}}q^{3(a^{2}-ab+b^{2})}\right)\sum_{n\geq0}Q_{n}(y)q^{n},

@page 10/14 starts_mid=0 ends_mid=1 Remark 5.2. For B=P2B = \mathbb{P}^{2} and a smooth plane curve DD of degree dd, we have g=(d−1)(d−2)/2g = (d - 1)(d - 2)/2 and χY=3d\chi_Y = 3d. In this case, the vanishing condition in Theorem 1.1 is equivalent to the injectivity of

H0(D,OD(d))→δCH1(D,OD(−3))H^{0}(D,\mathcal{O}_{D}(d)) \xrightarrow{\delta_C} H^{1}(D,\mathcal{O}_{D}(-3))

from (28), where both source and target have dimension d(d+3)/2d(d+3)/2. By Proposition 2.1, every nonzero section satisfies r(C)≥3dr(C) \ge3d, which for d=1d = 1 matches the threshold for additional sections in HKKW [25 §5.1.7].

Example 5.3. Let [s:t:u][s:t:u] be homogeneous coordinates on B=P2B = \mathbb{P}^{2} and let D=(u=0)D = (u = 0). Choose Weierstrass coefficients

a=t12+uA11,b=s18+us7t10+u2B16.a=t^{12}+uA_{11}, \qquad b=s^{18}+us^{7}t^{10}+u^{2}B_{16}.

where A11A_{11} and B16B_{16} are general homogeneous polynomials of the indicated degrees. Along DD, the discriminant restricts to 4t36+27s364t^{36}+27s^{36}, which has 36 simple zeros, and hence DD meets the discriminant transversely. Off DD, varying A11A_{11} and B16B_{16} prescribes the first jets of a,ba,b independently. For general coefficients the pair (a,b)(a,b) is transverse to the smooth part of 4a3+27b2=04a^{3}+27b^{2}=0 and to its cusp point. Thus XX is smooth with the required discriminant, and Y=π−1(D)Y=\pi^{-1}(D) is smooth.

The smooth Weierstrass models form an open subset of the affine space of coefficients. Its general member is the crepant resolution of a degree-18 hypersurface in P(1,1,1,6,9)\mathbb{P}(1,1,1,6,9) [24 §2]. The five rays corresponding to the vertices generate the toric lattice Z5/Z(1,1,1,6,9)\mathbb{Z}^{5}/\mathbb{Z}(1,1,1,6,9). The formulas of Batyrev and Kreuzer [2 Theorems 1.6 and 3.8] therefore give trivial fundamental and Brauer groups. In particular Pic⁡(X)\operatorname{Pic}(X) and H2(X,Z)H_{2}(X,\mathbb{Z}) are torsion-free, using the exponential and universal coefficient sequences. The same torsion vanishing holds throughout the smooth family. Diaconescu’s integral bases modulo torsion [11 Lemma 2.1] now give (2).

In the chart z=1z=1, let CϵC_{\epsilon} be the section over DD given by

x=ϵt6,y=s9,ϵ∈{0,−1,−−1}.x=\epsilon t^{6}, \qquad y=s^{9}, \qquad\epsilon\in\{0,\sqrt{-1},-\sqrt{-1}\}.

Indeed, x3+t12x=0x^{3}+t^{12}x=0 for these three values of xx. Each section is disjoint from OO, so r(Cϵ)=3r(C_{\epsilon})=3. Any two meet only at t=0t=0, where s=1s=1 and (t,x)(t,x) are local coordinates on YY since 2y≠02y\ne0. Their local intersection has length six. By Proposition 2.1 the height matrix of C0C_{0} and C−1C_{\sqrt{-1}} is

(6−3−36),\begin{pmatrix} 6 & -3 \\ -3 & 6 \end{pmatrix},

of determinant 27. Thus C0C_{0} and C−1C_{\sqrt{-1}} are independent. The three displayed points are collinear on the generic cubic, so their Mordell–Weil sum is OO.

Differentiating the Weierstrass equation along CϵC_{\epsilon} gives

0⟶NCϵ/X⟶OD(1)⊕OD(6)⊕OD(9)→dOD(18)⟶0,0 \longrightarrow N_{C_{\epsilon}/X} \longrightarrow\mathcal{O}_{D}(1)\oplus\mathcal{O}_{D}(6)\oplus\mathcal{O}_{D}(9) \xrightarrow{d} \mathcal{O}_{D}(18) \longrightarrow0,

where d=(−s7t10−ϵt6A11∣D,−(3ϵ2+1)t12,2s9)d=(-s^{7}t^{10}-\epsilon t^{6}A_{11}|_{D},-(3\epsilon^{2}+1)t^{12},2s^{9}). At A11∣D=0A_{11}|_{D}=0, the three summands on global sections span the monomials s18−jtjs^{18-j}t^{j} with j=7,8,…,18j=7,8,\ldots,18, respectively. Hence H0(d)H^{0}(d) is an isomorphism for all three sections for general A11A_{11}. Each normal bundle has degree −2-2 and no global sections, so NCϵ/X≅OD(−1)⊕2N_{C_{\epsilon}/X}\cong\mathcal{O}_{D}(-1)^{\oplus2}. Inversion gives three further rigid sections with y=−s9y=-s^{9}. Thus in the single ambient class σ∗[D]+3f\sigma_{*}[D]+3f, Theorem 1.1 gives at least six isolated points, one for each of the six sections.

More generally, write [m]Cϵ[m]C_{\epsilon} for the Mordell–Weil multiple by a nonzero integer mm. On the smooth locus of π\pi, the Weierstrass group law is regular [28 Corollary 12]. Multiplication by mm is étale there since its relative differential is multiplication by mm after translation. It maps CϵC_{\epsilon} isomorphically to [m]Cϵ[m]C_{\epsilon} over DD, identifying their normal bundles. Since r([m]Cϵ)=3m2r([m]C_{\epsilon})=3m^{2}, every such section gives a component isomorphic to Y[n]Y^{[n]} in class σ∗[D]+(3m2+n)f\sigma_{*}[D]+(3m^{2}+n)f for each n≥0n\geq0. For n=0,1,2n=0,1,2, its GV contributions are

1,36−2w,702−78w+3w2.1,\qquad36-2w,\qquad702-78w+3w^{2}.

Let C=C0C=C_{0}. At w=0w=0, (8) gives ∑n≥0n0,n6qn=∏m≥1(1−qm)−36\sum_{n\geq0}n^{6}_{0,n}q^{n}=\prod_{m\geq1}(1-q^{m})^{-36}. This is the factor for adding fibers to an isolated section over a line in Greer’s genus-zero GW calculation [18 p. 3, before Conjecture 1.6]. Greer uses Noether–Lefschetz theory to count rational curves of base degree one on very general Weierstrass fibrations [19].

Corollary 5.4. In Example 5.3, put P=C0P=C_{0} and Q=C−1Q=C_{\sqrt{-1}}. For very general A11,B16A_{11},B_{16}, every nonzero section in ZP+ZQ\mathbb{Z}P+\mathbb{Z}Q is infinitesimally rigid in XX. With qq recording total fiber degree, the combined GV contribution of these sections is

(∑(a,b)∈Z2∖{(0,0)}q3(a2−ab+b2))∑n≥0Qn(y)qn,(31)\left(\sum_{(a,b)\in\mathbb{Z}^{2}\setminus\{(0,0)\}}q^{3(a^{2}-ab+b^{2})}\right)\sum_{n\geq0}Q_{n}(y)q^{n}, \tag*{(31)}

where Qn(y)Q_n(y) is the contribution from nn added fibers in (21).

Proof. Keep YY fixed and specialize A11∣D=−1s7t4A_{11}|_D=\sqrt{-1}s^7t^4. The coordinates (−1t6,s9)(\sqrt{-1}t^6,s^9) then satisfy the Weierstrass equation modulo u2u^2, so QQ extends to a section over the doubled line 2D=(u2=0)2D=(u^2=0). The normal map for PP is independent of A11,B16A_{11},B_{16}, hence [a]P[a]P is rigid for a≠0a\ne0 by Example 5.3. Translation by the extended section [b]Q[b]Q on the smooth group locus over 2D2D identifies the first infinitesimal neighborhoods of [a]P[a]P and [a]P+[b]Q[a]P+[b]Q in XX. Indeed, the square of the ideal of DD pulls back into the square of the ideal of either section. Their normal bundles are therefore isomorphic.

For each fixed (a,b)(a,b) the curve [a]P+[b]Q⊂Y[a]P+[b]Q\subset Y is independent of the coefficients. By upper semicontinuity of h0h^0 of its normal bundle, rigidity holds on a nonempty open set when a≠0a\ne0. For a=0a=0 and b≠0b\ne0 the same conclusion follows from Example 5.3. Outside the countable union of the complementary closed sets all these sections are rigid. Their half-heights are 3(a2−ab+b2)3(a^2-ab+b^2) by the height matrix above. Theorem 1.1, with g=0g=0, gives the stated sum. □

The complete formula over a (−1)(-1)-curve

Suppose now that D=E≅P1D=E\cong\mathbb{P}^1 with E2=−1E^2=-1 and Y=π−1(E)Y=\pi^{-1}(E) smooth, so that χY=1\chi_Y=1 and YY is a rational elliptic surface. Let

E8:={v∈Z8∪(Z+12)8 | ∑i=18vi∈2Z}E_8:=\left\{v\in\mathbb{Z}^8\cup\left(\mathbb{Z}+\frac{1}{2}\right)^8\ \middle|\ \sum_{i=1}^{8}v_i\in2\mathbb{Z}\right\}

be the E8E_8 lattice with its standard Euclidean inner product, and let

ΘE8(q):=∑v∈E8q(v,v)/2=E4(q)=1+240∑m≥1(∑d∣md3)qm.(32)\Theta_{E_8}(q):=\sum_{v\in E_8}q^{(v,v)/2}=E_4(q)=1+240\sum_{m\ge1}\left(\sum_{d\mid m}d^3\right)q^m. \tag*{(32)}

be its theta series, where E4E_4 is the normalized Eisenstein series of weight four.

Corollary 6.1. For βk=σ∗[E]+kf\beta_k=\sigma_*[E]+kf, the full moduli scheme and its canonically oriented sheaf of vanishing cycles are

Mβk≅∐v∈E8(v,v)/2≤kY[k−(v,v)/2],ϕMβk∣Y[n]=QY[n][2n].(33)M_{\beta_k}\cong\coprod_{\substack{v\in E_8\\(v,v)/2\le k}}Y^{[k-(v,v)/2]}, \qquad \phi_{M_{\beta_k}}|_{Y^{[n]}}=\mathbb{Q}_{Y^{[n]}}[2n]. \tag*{(33)}

Here n=k−(v,v)/2n=k-(v,v)/2 on the component indexed by vv. With w=y+2+y−1w=y+2+y^{-1}, the GV invariants satisfy

∑k,h≥0nh,βkwhqk=ΘE8(q)∏m≥11(1−qm)8(1+yqm)2(1+y−1qm)2.(34)\sum_{\substack{k,h\ge0}}n_{h,\beta_k}w^hq^k = \Theta_{E_8}(q)\prod_{m\ge1}\frac{1}{(1-q^m)^8(1+yq^m)^2(1+y^{-1}q^m)^2}. \tag*{(34)}

In particular,

nh,βk=0(h>k),nk,βk=(−1)k(k+1),(35)n_{h,\beta_k}=0\quad(h>k),\qquad n_{k,\beta_k}=(-1)^k(k+1), \tag*{(35)}

and the genus-zero series is

∑k≥0n0,βkqk=ΘE8(q)∏m≥1(1−qm)−12=q1/2E4(q)η(q)12,η(q):=q1/24∏m≥1(1−qm).(36)\sum_{k\ge0}n_{0,\beta_k}q^k = \Theta_{E_8}(q)\prod_{m\ge1}(1-q^m)^{-12} = q^{1/2}\frac{E_4(q)}{\eta(q)^{12}}, \qquad \eta(q):=q^{1/24}\prod_{m\ge1}(1-q^m). \tag*{(36)}

Proof. Since E2=−1E^2=-1, any effective divisor on BB linearly equivalent to EE must contain EE with trivial residual divisor, so ∣E∣|E| is a point and Mk(E)=MβkM_k(E)=M_{\beta_k}. Moreover, for every section CC, the sequence (28) becomes

0⟶OP1(−1)⟶NC/X⟶OP1(−1)⟶0,0\longrightarrow\mathcal{O}_{\mathbb{P}^1}(-1)\longrightarrow N_{C/X}\longrightarrow\mathcal{O}_{\mathbb{P}^1}(-1)\longrightarrow0,

which splits and gives NC/X≅OP1(−1)⊕2N_{C/X}\cong\mathcal{O}_{\mathbb{P}^1}(-1)^{\oplus2}, so Theorem 1.1 applies to every component.

Because YY is a rational elliptic surface with integral fibers, the orthogonal complement of O,FO,F in its unimodular Néron–Severi lattice is negative definite, even (since KY=−FK_Y=-F), and unimodular of rank eight, so its Mordell–Weil lattice is isomorphic to E8E_8 [49, 22]. Proposition 2.1 therefore gives r(C)=(v,v)/2r(C)=(v,v)/2, and combining Theorem 3.1 with (8) yields (33) and (34).

At degree qkq^k in (34), the maximal exponent of yy is kk, which arises solely from the factor (1+yq)−2(1+yq)^{-2} paired with the constant terms of all other factors and has coefficient (−1)k(k+1)(-1)^k(k+1). Since wkw^k has leading term yky^k, this gives (35), while setting w=0w=0 (that is, y=−1y=-1) yields (36). □`

At k=1k=1, the 240 roots of E8\mathrm{E}_8 give the decomposition

Mβ1≅Y⊔{240 reduced points},252−2w=(12−2w)+240,M_{\beta_1}\cong Y\sqcup\{240\text{ reduced points}\},\qquad252-2w=(12-2w)+240,

which accounts for the 240 additional section contributions identified by HKKW for the exceptional curve of the Hirzebruch surface F1=PP1(O⊕O(1))\mathrm{F}_1=\mathbb{P}_{\mathbb{P}^1}(\mathcal{O}\oplus\mathcal{O}(1)) [25 §5.1.4]. The first coefficients of (34) are

k=01k=1252−2wk=25130−510w+3w2k=354760−11780w+772w2−4w3k=4419895−142330w+19467w2−1038w3+5w4.(37)\begin{aligned} k=0&\quad1\\ k=1&\quad252-2w\\ k=2&\quad5130-510w+3w^2\\ k=3&\quad54760-11780w+772w^2-4w^3\\ k=4&\quad419895-142330w+19467w^2-1038w^3+5w^4. \tag*{(37)} \end{aligned}

The factor multiplying ΘE8\Theta_{\mathrm{E}_8} in (34) is the fixed-section product of Hosono, Saito, and Takahashi [22 Theorems 4.7 and 4.10]. Summing over sections supplies the theta series, as in their earlier formula [21 equation (5.18)]. At genus zero, Bryan and Leung compute the contribution of a fixed section class on the rational elliptic surface [9 Theorem 6.2]. Rose and Yui sum over the E8\mathrm{E}_8 lattice in their compact example [44 Theorems 7.3–7.4]. Dividing (34) by ww also recovers Oberdieck and Shen’s PT series in base class EE, normalized by the pure-fiber PT series [41 Theorem 2].

Klemm, Mayr, and Vafa obtained (36) for a singly wound string [35 equation (4.1) and Table 1]. At all genera, let Z1Z_1 be the elliptic genus of one E-string with all E8\mathrm{E}_8 mass parameters zero and rotation parameters (z,−z)(z,-z), as in [20 equation (6.30)]. With q=e2πiτq=e^{2\pi i\tau} and y=−e2πizy=-e^{2\pi iz}, the Jacobi triple product gives

∑k,h≥0nh,βkwhqk=q1/2wZ1,w=4sin⁡2(πz).\sum_{k,h\geq0}n_{h,\beta_k}w^h q^k=q^{1/2}wZ_1,\qquad w=4\sin^2(\pi z).

The compact moduli spaces realizing these coefficients are given in (33).

When DD moves, Proposition 5.1 identifies the extra tangent directions at the fixed-divisor locus with H0(C,NC/X)H^0(C,N_{C/X}).

Appendix A. The direct image for a family of sections

Let U⊂∣D∣U\subset|D| parametrize smooth connected curves D′D' for which π−1(D′)\pi^{-1}(D') is smooth. Let D→U\mathcal{D}\to U be the universal curve and Y:=X×BD\mathcal{Y}:=X\times_B\mathcal{D}. Write DU(n)\mathcal{D}^{(n)}_U for the relative symmetric product and YU[n]\mathcal{Y}^{[n]}_U for the relative Hilbert scheme of nn points. The maps ρn\rho_n give the proper morphism

ρU,n:YU[n]⟶DU(n),(D′,Z)⟼(D′,π∗[Z]).\rho_{U,n}:\mathcal{Y}^{[n]}_U\longrightarrow\mathcal{D}^{(n)}_U,\qquad(D',Z)\longmapsto(D',\pi_*[Z]).

For each integer r≥0r\geq0, let SrS_r be the reduced locus in the relative Hilbert scheme of Y/U\mathcal{Y}/U parametrizing pairs (D′,C)(D',C) with CC a section satisfying ⟨C,C⟩=2r\langle C,C\rangle=2r. For n≥0n\geq0, adding nn fibers gives a cycle in class βr+n=σ∗[D]+(r+n)f\beta_{r+n}=\sigma_*[D]+(r+n)f. Define

τr,n:Sr×UDU(n)⟶Chow⁡βr+n(X),(D′,C,∑xmx[x])⟼[C]+∑xmx[fx].\tau_{r,n}:S_r\times_U\mathcal{D}^{(n)}_U\longrightarrow\operatorname{Chow}_{\beta_{r+n}}(X),\qquad(D',C,\sum_xm_x[x])\longmapsto[C]+\sum_xm_x[f_x].

At n=0n=0, τr,0(D′,C)=[C]\tau_{r,0}(D',C)=[C]. Over UU, write Kk:=Rhβk,∗ϕMβk\mathcal{K}_k:=R h_{\beta_k,*}\phi_{\mathcal{M}_{\beta_k}} for the direct image of vanishing cycles under the support map.

Conjecture A.1. For a choice of orientations that restrict to the canonical ones in Theorem 1.1, there are isomorphisms of constructible complexes on Sr×UDU(n)S_r\times_U\mathcal{D}^{(n)}_U

τr,n∗Kr+n≅pr⁡1∗(τr,0∗Kr)⊗pr⁡2∗RρU,n,∗QYU[n][2n].(38)\tau_{r,n}^*\mathcal{K}_{r+n}\cong\operatorname{pr}_1^*(\tau_{r,0}^*\mathcal{K}_r)\otimes\operatorname{pr}_2^*R\rho_{U,n,*}\mathbb{Q}_{\mathcal{Y}^{[n]}_U}[2n]. \tag*{(38)}

where pr⁡1\operatorname{pr}_1 and pr⁡2\operatorname{pr}_2 are the projections to SrS_r and DU(n)\mathcal{D}^{(n)}_U, respectively.

To compare with Theorem 1.1, fix (D,C)∈Sr(D,C)\in S_r. On {(D,C)}×D(n)\{(D,C)\}\times\mathcal{D}^{(n)}, proper base change identifies the right-hand side of (A.1) with

(Kr)[C]⊗Rρn∗QY[n][2n].(\mathcal{K}_r)_{[C]}\otimes R\rho_{n*}\mathbb{Q}_{Y^{[n]}}[2n].

If H0(C,NC/X)=0H^0(C,N_{C/X})=0, Theorem 1.1 at n=0n=0 identifies (Kr)[C](\mathcal{K}_r)_{[C]} with RΓ⁡(Pic⁡g+r(D,O)[g])\operatorname{R\Gamma}(\operatorname{Pic}^{g+r}(D,\mathcal{O})[g]). For every nn, the sheaves over these cycles form the open and closed component MC,nM_{C,n}. On this slice τr,n\tau_{r,n} is the finite injective map τC,n\tau_{C,n} of Section 4. Pulling back (25) gives (A.1). Taking the perverse Euler polynomial on D(n)D^{(n)} gives wgQn(y)w^gQ_n(y), the product of the Picard and Hilbert scheme contributions. Summing over nn gives (8). For a moving or obstructed section, (25) computes the intrinsic contribution over DD, while (A.1) retains the ambient stalk (Kr)[C](\mathcal{K}_r)_{[C]}.

For B=P2B = \mathbb{P}^{2} and DD of degree dd, a section C≠OC \ne O with nn added fibers has total fiber degree k=r(C)+nk = r(C)+n. Its cycle meets SS in degree

S⋅([C]+nf)=C⋅O+n=k−3d.S \cdot([C] + nf) = C \cdot O + n = k - 3d.

For the section over a line in Example 5.3, r(C)=3r(C) = 3. At n=0n = 0 the curve is disjoint from SS, as in HKKW’s charge-zero contribution [25 §5.1.7]. HKKW’s Wilson-loop sum ranges over residual plane degrees [25 §3.4.1, equation (3.58)]. Conjecture A.1 is the author’s proposed description of the direct image over the smooth-curve locus. Deriving the coefficients of HKKW’s sum from that direct image remains open.

Use of generative AI

The author used ChatGPT (OpenAI) to help draft and revise the manuscript and to check steps in the proofs. The author takes responsibility for the final text and its mathematical content.

Acknowledgements

I thank Sheldon Katz for introducing me to this problem. I also thank Min-xin Huang, Kaiwen Sun, Xin Wang, and Yukinobu Toda for useful discussions. This work was supported by JSPS KAKENHI Grant Number JP25K17226. Kavli IPMU is supported by the World Premier International Research Center Initiative (WPI), MEXT, Japan.

References

Kavli Institute for the Physics and Mathematics of the Universe (WPI), The University of Tokyo, Japan

Email address: [email protected]

References

  1. [1]M. Alim and E. Scheidegger, Topological strings on elliptic fibrations, Commun. Number Theory Phys. 8 (2014), 729–800. [arXiv:1205.1784].arxiv.org/abs/1205.1784
  2. [2]V. Batyrev and M. Kreuzer, Integral cohomology and mirror symmetry for Calabi–Yau 3-folds, in Mirror Symmetry V, AMS/IP Stud. Adv. Math. 38, Amer. Math. Soc., 2006, 255–270. [arXiv:math/0505432].DOI
  3. [3]A. Beauville, Variétés kählériennes dont la première classe de Chern est nulle, J. Differential Geom. 18 (1983), 755–782. [doi:10.4310/jdg/1214438181].
  4. [4]A. A. Beilinson, J. Bernstein and P. Deligne, Faisceaux pervers, Astérisque 100 (1982), 5–171.
  5. [5]O. Ben-Bassat, C. Brav, V. Bussi and D. Joyce, A ’Darboux theorem’ for shifted symplectic structures on derived Artin stacks, with applications, Geom. Topol. 19 (2015), 1287–1359. [arXiv:1213.0090].
  6. [6]C. Brav, V. Bussi, D. Dupont, D. Joyce and B. Szendrői, Symmetries and stabilization for sheaves of vanishing cycles, with an appendix by J. Schürmann, J. Singul. 11 (2015), 85–151. [arXiv:1211.3259].DOI
  7. [7]T. Bridgeland and A. Maciocia, Fourier–Mukai transforms for K3 and elliptic fibrations, J. Algebraic Geom. 11 (2002), 629–657. [arXiv:math/9908022].arxiv.org/abs/math/9908022
  8. [8]J. Bryan and M. Kool, Donaldson–Thomas invariants of local elliptic surfaces via the topological vertex, Forum Math. Sigma 7 (2019), e7. [arXiv:1608.07369].DOI
  9. [9]J. Bryan and N. C. Leung, The enumerative geometry of K3 surfaces and modular forms, J. Amer. Math. Soc. 13 (2000), 371–410. [arXiv:alg-geom/9711031].DOI
  10. [10]M. A. A. de Cataldo and L. Migliorini, The Douady space of a complex surface, Adv. Math. 151 (2000), 283–312. [arXiv:math/9811159].
  11. [11]D.-E. Diaconescu, Vertical sheaves and Fourier–Mukai transform on elliptic Calabi–Yau threefolds, Commun. Number Theory Phys. 10 (2016), no. 3, 373–431. [arXiv:1509.07749].
  12. [12]I. Dolgachev and M. Gross, Elliptic threefolds I. Ogg–Shafarevich theory, J. Algebraic Geom. 3 (1994), 39–80. [arXiv:alg-geom/9210009].
  13. [13]C. Fierro Cota, A. Klemm and T. Schimannek, Topological strings on genus one fibered Calabi–Yau 3-folds and string dualities, J. High Energy Phys. 11 (2019), 170. [arXiv:1910.01988].
  14. [14]J. Fogarty, Algebraic families on an algebraic surface, Amer. J. Math. 90 (1968), 511–521.DOI
  15. [15]R. Gopakumar and C. Vafa, M-theory and topological strings—I, II, 1998. [arXiv:hep-th/9809187], [arXiv:hep-th/9812127].
  16. [16]L. Göttsche, The Betti numbers of the Hilbert scheme of points on a smooth projective surface, Math. Ann. 286 (1990), 193–207. [doi:10.1007/BF01453572].
  17. [17]L. Göttsche and W. Soergel, Perverse sheaves and the cohomology of Hilbert schemes of smooth algebraic surfaces, Math. Ann. 296 (1993), 235–245. [doi:10.1007/BF01445104].DOI
  18. [18]F. Greer, Counting curves on a Weierstrass model, 2017. [arXiv:1701.06596].arxiv.org/abs/1701.06596
  19. [19]F. Greer, Modular forms from Noether–Lefschetz theory, Algebra Number Theory 14 (2020), no. 9, 2335–2368. [arXiv:1801.00375].DOI
  20. [20]B. Haghighat, G. Lockhart and C. Vafa, Fusing E-strings to heterotic strings: E + E → H, Phys. Rev. D 90 (2014), 126012. [arXiv:1406.0850].
  21. [21]S. Hosono, M.-H. Saito and A. Takahashi, Holomorphic anomaly equation and BPS state counting of rational elliptic surfaces, Adv. Theor. Math. Phys. 3 (1999), 177–208. [arXiv:hep-th/9901151].arxiv.org/abs/hep-th/9901151
  22. [22]S. Hosono, M.-H. Saito and A. Takahashi, Relative Jefschetz action and BPS state counting, Internat. Math. Res. Notices 2001, no. 15, 783–816. [arXiv:math/0105148].DOI
  23. [23]M.-x. Huang, S. Katz and A. Klemm, Topological string on elliptic CY 3-folds and the ring of Jacobi forms, J. High Energy Phys. 10 (2015), 125. [arXiv:1501.04891].
  24. [24]M.-x. Huang, S. Katz and A. Klemm, Towards refining the topological strings on compact Calabi–Yau 3-folds, J. High Energy Phys. 03 (2021), 266. [arXiv:2010.02910].DOI
  25. [25]M.-x. Huang, S. Katz, A. Klemm and X. Wang, Refined BPS numbers on compact Calabi–Yau threefolds from Wilson loops, J. High Energy Phys. 08 (2025), 178. [arXiv:2503.16270].DOI
  26. [26]M.-x. Huang, K. Lee and X. Wang, Topological strings and Wilson loops, J. High Energy Phys. 08 (2022), 207. [arXiv:2205.02366].DOI
  27. [27]D. Huybrechts and R. P. Thomas, Deformation-obstruction theory for complexes via Atiyah and Kodaira–Spencer classes, Math. Ann. 346 (2010), 545–569. [arXiv:0805.3527].arxiv.org/abs/0805.3527
  28. [28]J. Jin, Homogeneous division polynomials for Weierstrass elliptic curves. [arXiv:1303.4327].DOI
  29. [29]D. Joyce, A classical model for derived critical loci, J. Differential Geom. 101 (2015), no. 2, 289–367. [arXiv:1304.4508].arxiv.org/abs/1304.4508
  30. [30]S. Katz, A. Klemm, T. Schimannek and E. Sharpe, Topological strings on non-commutative resolutions, Commun. Math. Phys. 405 (2024), 62. [arXiv:2212.08655].DOI
  31. [31]S. Katz, A. Klemm and C. Vafa, M-theory, topological strings and spinning black holes, Adv. Theor. Math. Phys. 3 (1999), 1445–1537. [arXiv:hep-th/9910181].arxiv.org/abs/hep-th/9910181
  32. [32]Y.-H. Kiem and J. Li, Categorification of Donaldson–Thomas invariants via perverse sheaves, 2012. [arXiv:1212.6444].arxiv.org/abs/1212.6444
  33. [33]H.-C. Kim, M. Kim and S.-S. Kim, 5d/6d Wilson loops from blowups, J. High Energy Phys. 08 (2021), 131. [arXiv:2106.04731].DOI
  34. [34]A. Klemm, J. Manschot and T. Wotschke, Quantum geometry of elliptic Calabi–Yau manifolds, Commun. Number Theory Phys. 6 (2012), 849–917. [arXiv:1205.1795].arxiv.org/abs/1205.1795
  35. [35]A. Klemm, P. Mayr and C. Vafa, BPS states of exceptional non-critical strings, Nucl. Phys. B Proc. Suppl. 58 (1997), 177–194. [arXiv:hep-th/9607139].DOI
  36. [37]S.-J. Lee, W. Lerche and T. Weigand, Tensionless strings and the weak gravity conjecture, J. High Energy Phys. 10 (2018), 164. [arXiv:1808.05958].DOI
  37. [38]D. Maulik and Y. Toda, Gopakumar–Vafa invariants via vanishing cycles, Invent. Math. 213 (2018), 1017–1097. [arXiv:1610.07303].arxiv.org/abs/1610.07303
  38. [39]D. R. Morrison and D. S. Park, F-theory and the Mordell–Weil group of elliptically-fibered Calabi–Yau threefolds, J. High Energy Phys. 10 (2012), 128. [arXiv:1208.2695].DOI
  39. [40]G. Oberdieck and A. Pixton, Gromov–Witten theory of elliptic fibrations: Jacobi forms and holomorphic anomaly equations, Geom. Topol. 23 (2019), 1415–1489. [arXiv:1709.01481].DOI
  40. [41]G. Oberdieck and J. Shen, Curve counting on elliptic Calabi–Yau threefolds via derived categories, J. Eur. Math. Soc. 22 (2020), no. 3, 967–1002. [arXiv:1608.07073].
  41. [42]R. Pandharipande and R. P. Thomas, Curve counting via stable pairs in the derived category, Invent. Math. 178 (2009), 407–447. [arXiv:0707.2348].DOI
  42. [43]B. Pioline and T. Schimannek, Revisiting the quantum geometry of torus-fibered Calabi–Yau threefolds, SciPost Phys. 21 (2026), 037. [arXiv:2510.23722].DOI
  43. [44]S. Rose and N. Yui, Elliptic Calabi–Yau threefolds over a del Pezzo surface, in Arbeitstagung Bonn 2013, Progr. Math. 319, Birkhäuser, 2016, 337–354. [arXiv:1312.0911].arxiv.org/abs/1312.0911
  44. [45]D. Rydh, Families of cycles, Paper IV in Families of cycles and the Chow scheme, doctoral thesis, KTH, Stockholm, 2008.
  45. [46]T. Schimannek, Modularity from monodromy, J. High Energy Phys. 05 (2019), 024. [arXiv:1902.08215].DOI
  46. [47]T. Schimannek, Modular curves, the Tate–Shafarevich group and Gopakumar–Vafa invariants with discrete charges, J. High Energy Phys. 02 (2022), 007. [arXiv:2108.09311].DOI
  47. [48]J. Shen and Z. Zhang, Perverse filtrations, Hilbert schemes, and the P = W conjecture for parabolic Higgs bundles, Alg. Geom. 8 (2021), no. 4, 465–489. [doi:10.14231/AG-2021-014].DOI
  48. [49]T. Shioda, On the Mordell–Weil lattices, Comment. Math. Univ. St. Paul. 39 (1990), no. 2, 211–240.DOI
  49. [50]C. T. Simpson, Moduli of representations of the fundamental group of a smooth projective variety I, Publ. Math. Inst. Hautes Études Sci. 79 (1994), 47–129. [doi:10.1007/BF02698877].
  50. [51]Y. Toda, Gopakumar–Vafa invariants and wall-crossing, J. Differential Geom. 123 (2023), 141–193. [arXiv:1710.01843].
  51. [52]X. Wang, Wilson loops, holomorphic anomaly equations and blowup equations, 2023. [arXiv:2305.09171].arxiv.org/abs/2305.09171
  52. [53]R. Wazir, Arithmetic on elliptic threefolds, Compos. Math. 140 (2004), no. 3, 567–580. [arXiv:math/0112259].DOI
  53. [54]K. Yoshioka, Moduli spaces of stable sheaves on abelian surfaces, Math. Ann. 321 (2001), no. 4, 817–884. [arXiv:math/0009001].
  54. [55]Z. Zhang, Multiplicativity of perverse filtration for Hilbert schemes of fibered surfaces, Adv. Math. 312 (2017), 636–679. [arXiv:1512.04643].arxiv.org/abs/1512.04643

Paper details

vlnc:W_qb8pj5je
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

ChatGPT (OpenAI)

View statement
Looking for related papers…
Contents