betaFeedback

WorksMathematics

Lagrangian tori in the three-dimensional quadric and mirror symmetry

Abstract

In the three-dimensional quadric, we study the Oakley–Usher torus and its Chekanov counterpart as defined by the authors in the companion paper \cite{OU1}. Computing the versal deformation of the ΨΨ-invariant of Shelukhin–Tonkonog–Vianna, we find that the Chekanov Oakley–Usher torus is knotted in a strong sense: it is not Hamiltonian isotopic to the Biran-lift of any Vianna torus in the two-dimensional quadric. Furthermore, we investigate enumerative properties of the Oakley–Usher torus in Q3Q^3 and prove its non-displaceability, settling an open question asked by Oakley–Usher. We relate this to mirror symmetry by proving a split-generation result for the monotone Fukaya category of Q3Q^3. This should be viewed as the analogue of a result by Abouzaid–Diogo after compactifying. Interestingly, in the compactification, the monotone copy of the (S1×S2)(S^1 \times S^2)-generators of Abouzaid–Diogo becomes displaceable. On the other hand, the Oakley–Usher torus still appears as a generator. In the appendix joint with Soham Chanda, we give an alternative construction of the Chekanov Oakley–Usher torus using a double solid torus mutation. This allows us to compute its disk-potential using wall-crossing results.

Introduction

In the companion paper [10], the authors have constructed knotted versions of the Oakley–Usher Lagrangians in [24]. The goal of this paper is to have an in-depth look at the situation in the three-dimensional quadric Q3Q^3.

Setup

We equip Q3={z02+⋯+z42=0}⊂CP4Q^3 = \{z_0^2 + \cdots+ z_4^2 = 0\} \subset\mathbb{CP}^4 with the Fubini–Study form normalized so that a line has area 2π2\pi. By [10 Proposition 3.10], this space can be viewed as the symplectic cut of a the cotangent bundle T∗S3T^*S^3. This identifies D∗S3D^*S^3 with the complement of a copy Q2Q^2 of the two-dimensional quadric. We write Σ⊂Q3\Sigma\subset Q^3 for the image of the zero-section under the symplectic cut; in this model Σ={[1:iq]:q∈S3}\Sigma= \{[1 : iq] : q \in S^3\}.

The families P1,1(r)\mathcal{P}_{1,1}(r) and P1,1Ch(a)\mathcal{P}_{1,1}^{\mathrm{Ch}}(a) arise by lifting curves under the T2=SO(2)×SO(2)T^2 = \mathrm{SO}(2) \times\mathrm{SO}(2) symplectic reduction described in § 3 in [10]. The ordinary Oakley–Usher Lagrangians P1,1(r)\mathcal{P}_{1,1}(r) are lifts of the noncontractible circles, while their Chekanov counterparts P1,1Ch(a)\mathcal{P}_{1,1}^{\mathrm{Ch}}(a) are lifts of contractible circles bounding area a/2a/2 in the smooth reduced cylinder. We use the size convention of [10]: both families have their monotone member at r=a=4π/3r = a = 4\pi/3. Throughout this paper we abbreviate

P=P1,1(4π/3),PCh=P1,1Ch(4π/3).\mathcal{P} = \mathcal{P}_{1,1}(4\pi/3), \qquad\mathcal{P}^{\mathrm{Ch}} = \mathcal{P}_{1,1}^{\mathrm{Ch}}(4\pi/3).

Both are Lagrangian three-tori. They are Lagrangian isotopic by [10, Theorem B`], but not Hamiltonian isotopic, as we shall see now.

Main results

As discussed in [24 Theorem 1.3], the torus P\mathcal{P} can be viewed as the Biran circle bundle lift of the Clifford torus in Q2≅S2×S2Q^2 \cong S^2 \times S^2 [24 Theorem 1.3]. Similarly, one can lift all Vianna tori from Q2Q^2 [33]. Our first result shows that the Chekanov OU-torus is different from tori obtained in this way.

Theorem A. The torus PCh\mathcal{P}^{\mathrm{Ch}} cannot be mapped by a symplectomorphism to P\mathcal{P} or the lift of any other Vianna torus from the divisor Q2⊂Q3Q^2 \subset Q^3.

The proof computes the versal deformations of the Ψ\Psi-invariant of Shelukhin–Tonkonog–Vianna [29]. To that end, we study a Gelfand–Cetlin fibration which has PCh\mathcal{P}^{\mathrm{Ch}} as its monotone fibre. Related lifting constructions are considered in earlier work by Chanda–Hirschi–Wang [12] and by Brendel [8]. This approach will be significantly generalized in upcoming work by Diogo–Tonkonog–Vianna–Wu [14].

Our second result concerns P\mathcal{P}. It computes the signed Maslov two disk-potential and identifies a finite set of generators for the monotone Fukaya category.

Theorem B. There is a spin structure on the Oakley–Usher torus P⊂Q3\mathcal{P} \subset Q^3, and a basis of π1(P)≅Z3\pi_1(\mathcal{P}) \cong\mathbb{Z}^3 for which the superpotential WP:Hom⁡(π1(P),Λ×)≅(Λ×)3→Λ\mathcal{W}_{\mathcal{P}} : \operatorname{Hom}(\pi_1(\mathcal{P}),\Lambda^\times) \cong(\Lambda^\times)^3 \to\Lambda is

WP(x1,x2,x3)=T2π/3(x1+x2+x3x1+x3x2+1x3).\mathcal{W}_{\mathcal{P}}(x_1,x_2,x_3) = T^{2\pi/3}\left(x_1 + x_2 + \frac{x_3}{x_1} + \frac{x_3}{x_2} + \frac{1}{x_3}\right).

where Λ\Lambda is the universal Novikov field over C\mathbb{C} defined in (18). Furthermore, it carries three critical local systems, which, when taken together with a disjoint Lagrangian sphere Σ⊂Q3\Sigma\subset Q^3, split-generate the monotone Fukaya category of Q3Q^3.

In particular, the Oakley–Usher torus P\mathcal{P} is non-displaceable, answering a question raised in [24] for this torus. The sphere Σ\Sigma accounts for the zero-eigenvalue summand of quantum multiplication by c1(Q3)c_1(Q^3), while the three critical local systems on P\mathcal{P} account for the three nonzero eigensummands. The proof of Theorem B relies on a toric degeneration of Q3Q^3 and on the disk-counting results of Nishinou–Nohara–Ueda [22, 23].

Comparison to previous work

The split-generation result in Theorem B is inspired by work of Abouzaid–Diogo [1], which treats the open case of T∗S3T^*S^3. Let us briefly compare the two results. Abouzaid–Diogo prove that the monotone Fukaya category is split-generated by the zero section equipped with an uncountable family of bounding cochains, together with either of the following two families:

  1. a one-parameter family of copies of Lagrangian S1×S2S^1 \times S^2 corresponding to the Oakley–Usher Lagrangians P0,2(r)⊂T∗S3\mathcal{P}_{0,2}(r) \subset T^*S^3, equipped with an uncountable family of suitable local systems;

  2. a one-parameter family of Oakley–Usher tori P1,1(r)⊂T∗S3\mathcal{P}_{1,1}(r) \subset T^*S^3, equipped with an uncountable family of suitable local systems.

See [24] or [10] for more details on these families. After compactification, only one member of each of the families above is monotone in Q3Q^3. Interestingly, the monotone copy of S1×S2S^1 \times S^2 is displaceable in Q3Q^3 and therefore vanishes in the Fukaya category. The monotone copy of the tori is P⊂Q3\mathcal{P} \subset Q^3 and thus appears as a generator by our result.

Kim [20] and Kawamoto [19] study a different torus in Q3Q^3. In fact, their torus can be viewed as the lift of the Chekanov torus in Q2Q^2. Enumerative properties of the OU-torus in T∗S3T^*S^3 were studied by Chan–Pomerleano–Ueda [11] and Achig-Andrango [2]; the latter also discusses the effect of compactification on the potential. Sheridan [30] established homological mirror symmetry for Fano hypersurfaces (including Q3Q^3) in complex projective space. We identify explicit geometric generators: the Oakley–Usher torus equipped with three critical local systems, together with the Lagrangian sphere Σ\Sigma.

Outline and conventions

Section 2 develops the Ψ\Psi-invariant, constructs the Gelfand–Cetlin fibration, computes the two OU Ψ\Psi-germs, and compares PCh\mathcal{P}^{\mathrm{Ch}} with lifted Vianna tori, proving Theorem A. Section 3 computes the superpotential, proves non-displaceability and split-generation, and studies the change under compactification, proving Theorem B. We use S1=R/ZS^1=\mathbb{R}/\mathbb{Z} and integer coefficients for homology unless otherwise specified. Symplectic reduction is denoted by X//μ=cGX//_{\mu=c}G. Induced moment maps on reduced or cut spaces retain the original notation. The symplectic normalization and size parameters agree with those in [10]; the Novikov field and local-system conventions are specified in Section 3.

Acknowledgments

This paper and its companion [10] grew out of a Master’s thesis written by AP under the supervision of JB at ETH Zürich. We thank Paul Biran for orchestrating this. We warmly thank Leonid Polterovich for drawing our attention to Oakley–Usher Lagrangians, and Johanna Bimmermann, Marc Fares, Johannes Hauber, David Keren Yaar, Joel Schmitz, Mike Usher, and Renato Vianna for interesting discussions, comments and questions. We are particularly grateful to Felix Schlenk for asking many interesting questions. JB is supported by SNSF Ambizione Grant PZ00P2-223460.

Statement on the use of LLMs

An earlier version of this project comprising this paper and its companion [10] was written completely without the use of large language models. In subsequent versions, GPT was used to help split the project into two papers, and proofread the text. The authors take full responsibility for the final mathematical content.

Chekanov OU-torus in the three-dimensional quadric

In this section, we exclusively work with the case k=m=1k=m=1 in the three-dimensional quadric Q3Q^3. For ease of notation, we drop the 1,11,1-subscripts and sizes of the OU-Lagrangians everywhere. In other words, we write P=P1,1(4π3)\mathcal{P}=\mathcal{P}_{1,1}\left(\frac{4\pi}{3}\right) and PCh=P1,1Ch(4π3)\mathcal{P}^{\mathrm{Ch}}=\mathcal{P}^{\mathrm{Ch}}_{1,1}\left(\frac{4\pi}{3}\right). Recall that these are the monotone OU-Lagrangian and Chekanov OU-Lagrangian, respectively.

The Ψ\Psi-invariant and its versal deformation

In order to distinguish tori up to Hamiltonian isotopy, we use the so-called Ψ\Psi-invariant, introduced by Shelukhin–Tonkonog–Vianna [29]. Let (X,ω)(X,\omega) be a symplectic manifold. For the general statements in this subsection, take XX closed, and LL closed, orientable and spin. These hypotheses suffice for all applications here; extensions to open manifolds require the appropriate compactness hypotheses for the Fukaya algebra.

The Ψ\Psi-invariant associates to every orientable spin Lagrangian L⊂XL\subset X a number Ψ(L)∈(0,+∞]\Psi(L)\in(0,+\infty] which measures the smallest symplectic area of a JJ-holomorphic disk which contributes nontrivially to some algebraic disk-count in the Fukaya A∞A_{\infty}-algebra of LL.

As we will see, Ψ\Psi itself does not distinguish the tori we are interested in. For example, we will see that Ψ(P)=Ψ(PCh)=2π3\Psi(\mathcal{P})=\Psi(\mathcal{P}^{\mathrm{Ch}})=\frac{2\pi}{3}. Therefore, we consider its versal deformation. We refer to [9 Section 3] for details on versal deformations.

Definition 2.1. Let L⊂XL \subset X be an orientable spin Lagrangian. We call Ψ-germ (of LL) the germ, in the sense of [9 Definition 3.13], of the invariant Ψ\Psi at LL. We denote it by

ψL=[Ψ]L:H1(L;R)⇢R∪{+∞},\psi_L = [\Psi]_L : H^1(L;\mathbb{R}) \dashrightarrow\mathbb{R} \cup\{+\infty\},

where the dashed arrow denotes a map defined on a neighbourhood of the origin of H1(L;R)H^1(L;\mathbb{R}).

Proposition 2.2. The Ψ\Psi-germ is of the form

ψL(a)=Ψ(L)+min⁡{⟨a,ξi⟩},(1)\psi_L(a) = \Psi(L) + \min\{\langle a,\xi_i\rangle\}, \tag*{(1)}

for some ξ1,…,ξN∈H1(L)\xi_1,\ldots,\xi_N \in H_1(L), provided Ψ(L)<+∞\Psi(L)<+\infty. In particular, the displayed piecewise-affine expression extends to ψL:H1(L;R)→R\psi_L:H^1(L;\mathbb{R})\to\mathbb{R}.

Proof. It was proved in [29] that, in a small neighbourhood of LL in the space of Lagrangians, the invariant Ψ\Psi can be written as the minimum of symplectic areas of relative classes β1,…,βN∈H2(X,L)\beta_1,\ldots,\beta_N\in H_2(X,L) with boundary on LL. More precisely, the classes βi\beta_i are those of area Ψ(L)\Psi(L) which contribute nontrivially to some symmetrized A∞A_\infty-operation on odd-degree cohomology classes, in a classically minimal model of the Fukaya algebra of LL; see [29 Theorem 4.8 and the proof of Theorem 4.9]. By Gromov’s compactness theorem, this set is finite. Setting ξi=∂βi\xi_i=\partial\beta_i and computing the versal deformation yields the claim. □\square

Instead of working with ψL\psi_L directly, it is often more convenient to work with its (super-)level sets, since these are easier to manipulate and can be represented visually.

Definition 2.3. Let L⊂XL\subset X be an orientable spin Lagrangian with Ψ(L)<+∞\Psi(L)<+\infty. Using the global piecewise-affine extension of ψL\psi_L given by Proposition 2.2, we define

ΔΨ(L):={a∈H1(L;R)∣ψL(a)≥0}.\Delta^\Psi(L):=\{a\in H^1(L;\mathbb{R})\mid\psi_L(a)\ge0\}.

This is a convex polyhedron, which we call the Ψ\Psi-polytope of LL whenever it is bounded. If Ψ(L)=+∞\Psi(L)=+\infty, the germ is identically +∞+\infty by [29 Theorem 4.9]; one may separately set ΔΨ(L)=H1(L;R)\Delta^\Psi(L)=H^1(L;\mathbb{R}).

Remark 2.4. For the Oakley–Usher torus considered in the next subsection, the relevant structure is a Gelfand–Cetlin fibration, meaning that the Ψ\Psi-polytope agrees with

  1. the polytope obtained as the base of the fibration, after a translation that shifts the image of the monotone fibre LL to the origin, if necessary.

  2. the appropriately scaled polar dual of the Newton polytope of the superpotential of that fibre.

The fact that these three polytopes agree is [29 Theorem B]. We only use Gelfand–Cetlin fibrations once to compute ψP\psi_P for the Oakley–Usher Lagrangian P⊂Q3P\subset Q^3. All other invariants will be computed by versal deformation techniques.

We use the following result to distinguish tori.

References

  1. [1]M. Abouzaid and L. Diogo. Monotone Lagrangians in cotangent bundles of spheres. Adv. Math., 427:Paper No. 109114, 45, 2023.arxiv.org/abs/2011.13478
  2. [2]S. Achig-Andrango. Singular Lagrangian torus fibrations on the smoothing of algebraic cones. J. Symplectic Geom., 22(4):847–913, 2024.
  3. [3]H. Alizadeh, M. S. Atallah, D. Cant, and J. Shang. On the spectral diameter of the Grassmannians. arXiv:2512.20125, 2025.arxiv.org/abs/2512.20125
  4. [4]A. Beauville. Quantum cohomology of complete intersections. Mat. Fiz. Anal. Geom., 2(3-4):384–398, 1995.arxiv.org/abs/alg-geom/9501008
  5. [5]P. Biran and O. Cornea. Lagrangian topology and enumerative geometry. Geom. Topol., 16(2):963–1052, 2012.DOI
  6. [6]J. Brendel. An introduction to the Chekanov torus. In Proceedings of the Gö kova Geometry-Topology Conference 2021, pages 1–56. Int. Press, Somerville, MA, 2022.
  7. [7]J. Brendel. Hamiltonian classification of toric fibres and symmetric probes. Alebr. Geom. Topol., 25(3):1839–1876, 2025.DOI
  8. [8]J. Brendel. Local exotic tori. Trans. Amer. Math. Soc., 378(6):4369–4411, 2025.arxiv.org/abs/2310.11359
  9. [9]J. Brendel, J. Hauber, and J. Schmitz. Semi-local exotic Lagrangian tori in dimension four. Ann. Inst. Fourier (Grenoble), 2026. Online first.arxiv.org/abs/2403.00408
  10. [10]J. Brendel and A. Piccirilli. On Lagrangians of Oakley–Usher type. 2026.arxiv.org/abs/2607.15136
  11. [11]K. Chan, D. Pomerleano, and K. Ueda. Lagrangian torus fibrations and homological mirror symmetry for the conifold. Comm. Math. Phys., 341(1):135–178, 2016.DOI
  12. [12]S. Chanda, A. Hirschi, and L. Wang. Infinitely many exotic Lagrangian tori in higher projective spaces. J. Fixed Point Theory Appl., 26(4):Paper No. 46, 18, 2024.arxiv.org/abs/2307.06934
  13. [13]Y. Chekanov and F. Schlenk. Notes on monotone Lagrangian twist tori. Electron. Res. Announc. Math. Sci., 17:104–121, 2010.arxiv.org/abs/1003.5960
  14. [14]L. Diogo, D. Tonkonog, R. Vianna, and W. Wu. Lifting Lagrangians from Donaldson-type divisors. Work in progress, 2026.
  15. [15]J. Evans. Lectures on Lagrangian torus fibrations, volume 105 of London Mathematical Society Student Texts. Cambridge University Press, Cambridge, 2023.arxiv.org/abs/2110.08643
  16. [16]J. D. Evans. KIAS lectures on symplectic aspects of degenerations. arXiv:2403.03519, 2024.arxiv.org/abs/2403.03519
  17. [17]M. Harada and K. Kaveh. Integrable systems, toric degenerations and Okounkov bodies. Invent. Math., 202(3):927–985, 2015.DOI
  18. [18]Y. Karshon, S. B. Tukachinsky, and Y. Zimhon. Weinstein neighborhood theorems for stratified subspaces. Journal of the London Mathematical Society, 114(2):e70673, 2026.DOI
  19. [19]Y. Kawamoto. Hofer geometry via toric degeneration. Math. Ann., 390(1):721–762, 2024.arxiv.org/abs/2210.14726
  20. [20]Y. Kim. Disk potential functions for quadrics. J. Fixed Point Theory Appl., 25(2):Paper No. 46, 31, 2023.arxiv.org/abs/2107.05839
  21. [21]D. McDuff. Displacing Lagrangian toric fibers via probes. In Low-dimensional and symplectic topology, volume 82 of Proc. Sympos. Pure Math., pages 131–160. Amer. Math. Soc., Providence, RI, 2011.arxiv.org/abs/0904.1686
  22. [22]T. Nishinou, Y. Nohara, and K. Ueda. Toric degenerations of Gelfand-Cetlin systems and potential functions. Adv. Math., 224(2):648–706, 2010.arxiv.org/abs/0810.3470
  23. [23]T. Nishinou, Y. Nohara, and K. Ueda. Potential functions via toric degenerations. Proc. Japan Acad. Ser. A Math. Sci., 88(2):31–33, 2012.arxiv.org/abs/0812.0066
  24. [24]J. Oakley and M. Usher. On certain Lagrangian submanifolds of $S^2 \times S^2$ and $\mathrm{CP}^n$. Algebraic Geom. Topol., 16(1):149–209, 2016.DOI
  25. [25]Y.-G. Oh. Floer cohomology, spectral sequences, and the Maslov class of Lagrangian embeddings. Internat. Math. Res. Notices, 1996(7):305–346, 1996.
  26. [26]J. Pascalleff and D. Tonkonog. The wall-crossing formula and Lagrangian mutations. Adv. Math., 361:106850, 67, 2020.
  27. [27]C. Pech, K. Rietsch, and L. Williams. On Landau–Ginzburg models for quadrics and flat sections of Dubrovin connections. Advances in Mathematics, 300:275–319, 2016.arxiv.org/abs/1404.4844
  28. [28]W.-D. Ruan. Lagrangian torus fibration of quintic hypersurfaces. I. Fermat quintic case. In Winter School on Mirror Symmetry, Vector Bundles and Lagrangian Submanifolds (Cambridge, MA, 1999), volume 23 of AMS/IP Stud. Adv. Math., pages 297–332. Amer. Math. Soc., Providence, RI, 2001.DOI
  29. [29]E. Elshukhin, D. Tonkonog, and R. Vianna. Geometry of symplectic flux and Lagrangian torus fibrations. J. Topol., 17(4):Paper No. e70002, 56, 2024.
  30. [30]N. Sheridan. On the Fukaya category of a Fano hypersurface in projective space. Publ. Math. Inst. Hautes Études Sci., 124:165–317, 2016.DOI
  31. [31]I. Smith. Floer cohomology and pencils of quadrics. Inventiones Mathematicae, 189:149–250, 2012.arxiv.org/abs/1006.1099
  32. [32]J. Smith. Quantum cohomology and Fukaya summands from monotone Lagrangian tori. J. Éc. polytech. Math., 12:287–318, 2025.arxiv.org/abs/2409.07922
  33. [33]R. Vianna. Infinitely many monotone Lagrangian tori in del Pezzo surfaces. Selecta Math. (N.S.), 23(3):1955–1996, 2017.arxiv.org/abs/1602.03356

Paper details

vlnc:W_9vmggak3
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 6

downloads · 0

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

  • Mathematics

GPT (multiple versions)

View statement
Looking for related papers…
Contents