Monotone Lagrangians in cotangent bundles of spheres (Q6098433)

Understanding the Lagrangian submanifolds of the cotangent bundle is of great interest in symplectic topology. Indeed, the cotangent bundle with its zero-section is the local symplectic model for a small neighborhood of any (embedded) closed Lagrangian submanifold. In fact, these Lagrangian submanifolds are expected to be quite nice. More precisely, the nearby Lagrangian conjecture states that every \emph{exact} closed Lagrangian submanifold of \(T^*Q\), where \(Q\) is closed and connected, is Hamiltonian isotopic to the zero-section. Although this statement has only been proven in a few specific cases, i.e., \(S^2\) [\textit{R. Hind}, Geom. Funct. Anal. 14, No. 2, 303--318 (2004; Zbl 1066.53129)], \(\mathbb{R}P^2\) [\textit{R. Hind} et al., J. Symplectic Geom. 14, No. 1, 203--226 (2016; Zbl 1355.57026)], and \({T}^2\) [\textit{G. Dimitroglou Rizell} et al., Geom. Funct. Anal. 26, No. 5, 1297--1358 (2016; Zbl 1358.53089)]), great steps have been made towards its proof, in great part by the first author of the paper. In particular, under some light topological hypotheses on \(Q\), it is known that every exact closed Lagrangian submanifold of \(T^*Q\) is quasi-isomorphic to the zero-section in the appropriate Fukaya category associated to \(T^*Q\), i.e., they are all the same from the point of view of Floer theory [\textit{K. Fukaya} et al., Invent. Math. 172, No. 1, 1--27 (2008; Zbl 1140.53036); \textit{K. Fukaya} et al., Lect. Notes Phys. 757, 1--26 (2009; Zbl 1163.53344); \textit{D. Nadler}, Sel. Math., New Ser. 15, No. 4, 563--619 (2009; Zbl 1197.53116)]. However, precious little is known about the behavior of \emph{nonexact} closed Lagrangian submanifolds of \(T^*Q\). In the present paper, the authors thus study \emph{monotone} closed Lagrangian submanifolds of \(T^*S^n\), \(n\geq 2\), i.e., closed Lagrangian submanifolds \(L\) with a number \(\tau\geq 0\) such that \begin{align*} \int_{D^2}u^*\omega_0 =\tau\mu(u) \end{align*} for all smooth maps \(u:(D^2,\partial D^2)\to (T^*Q,L)\). Here, \(\omega_0\) is the canonical symplectic form of \(T^*Q\), and \(\mu(u)\) is the Maslov index of \(u\). In the particular case that \(Q=S^n\), \(n\geq 2\), a monotone closed Lagrangian submanifold \(L\) is exact precisely when \(\tau=0\). The main result of the paper (Theorem~1.2) can be seen as an extension of the above cited results on exact Lagrangian submanifolds of \(T^*Q\) to the monotone case: if \(L\subset T^*S^n\) if a sufficiently nice monotone Lagrangian submanifolds, then either \(HF^* (L,S^n)\neq 0\) , or there is a \(\tau>0\) for which \(HF^* (L,(S^1\times S^{n-1})_\tau)\neq 0\) (where \((S^1\times S^{n-1})_\tau\) has monotonicity constant \(\tau\)). In particular, \(L\) is not Hamiltonian-displaceable from either \(S^n\) or \((S^1\times S^{n-1})_\tau\) , for some \(\tau>0\). Furthermore, in \(T^*S^3\), one can replace \((S^1\times S^2)_\tau\) with a torus \(T_\tau^3\). From this statement it follows that Floer theory can differentiate between various nonexact Lagrangian submanifolds in certain cotangent bundles. Moreover, the authors have some quite explicit geometrical descriptions of the Lagrangian submanifolds \((S^1\times S^{n-1})_\tau\) and \(T_\tau^3\), which are given in Section~2, along with some of their properties. The rest of the paper is devoted to the proof of the main theorem. More precisely, the result follows directly from showing that the zero-section and the Lagrangian submanifolds \((S^1\times S^{n-1})_\tau\) (or \(T^3_\tau\) if \(n=3\)) split-generate the monotone Fukaya category of \(T^*S^n\) (Theorem~1.4). This is the proper extension of the results in [K. Fukaya et al., loc. cit.; D. Nadler, loc. cit.] to the monotone setting. The proof of this statement, which is outlined in the introduction, can be broken down in three large steps as follows. \begin{enumerate} \item The authors first show that the appropriate version \(\mathcal{W}\) of the wrapped Fukaya category of the cotangent bundle, a variant of the usual Fukaya category which includes certain nice noncompact Lagrangian submanifolds, is generated by the fiber. This is done in Subsection~4.1 by adapting some of the first author's results. The precise versions of the wrapped Fukaya category of interest are explained in Subsection~3.4. \item The authors then show that the cohomological Yoneda functor \(Y:\mathcal{W}\to\mathrm{mod}\), which sends a Lagrangian submanifold to its wrapped Floer cohomology with the fiber, is a cohomologically full and faithful embedding. They also show that the restriction \(Y_c:\mathcal{F}\to \mathrm{mod}_{pr}\) to the full subcategory \(\mathcal{F}\subseteq\mathcal{W}\) generated by closed Lagrangian submanifolds has those properties. Here, \(\mathrm{mod}_{pr}\subseteq\mathrm{mod}\) are categories of modules over a polynomial ring. This is done in Subsection~3.5, using the previous step and the results of Section~5. \item Using the restriction of the cohomological Yoneda functor, the statement on generators is reduced to a purely algebraic statement on modules and consist of computations showing that one has indeed the right (split-)generators. The required computations on the Fukaya side are done in Subsections 4.2 and 4.3, whilst those on the module side are done in Section 6. \end{enumerate}