The symplectic topology of some rational homology balls (Q466216)

From MaRDI portal
scientific article
Language Label Description Also known as
English
The symplectic topology of some rational homology balls
scientific article

    Statements

    The symplectic topology of some rational homology balls (English)
    0 references
    0 references
    0 references
    0 references
    24 October 2014
    0 references
    The main results of this paper concern a family of symplectic 4-manifolds \(B_{p,q}\), which exist for relatively prime integers \(p>q>0\). These manifolds are rational homology balls, affine varieties, and naturally admit Stein structures. The boundary of \(B_{p,q}\) is the Lens space \(L(p^2, pq-1)\), and \(B_{p,q}\) is \(p\)-fold covered by the Milnor fibre \(S_{p-1}\) of the \(A_{p-1}\) singularity. These manifolds were first introduced by \textit{A. J. Casson} and \textit{J. L. Harer} [Pac. J. Math. 96, 23--36 (1981; Zbl 0483.57017)] and were used in the rational blowdown construction of \textit{R. Fintushel} and \textit{R. J. Stern} [J. Differ. Geom. 46, No. 2, 181--235 (1997; Zbl 0896.57022)]. The main theorem 1.3 states that for \(p \neq 2\), \(B_{p,q}\) has no closed exact Lagrangian submanifolds, but does contain a Floer-theoretically essential Lagrangian torus, and that its symplectic cohomology \(SH^*(B_{p,q})\) is \(\neq 0\). A theorem of \textit{C. Viterbo} [Geom. Funct. Anal. 9, No. 5, 985--1033 (1999; Zbl 0954.57015)] states that a Stein manifold with a closed exact Lagrangian submanifold has non-vanishing symplectic cohomology. The result in this paper thus demonstrates non-vanishing symplectic cohomology without invoking Viterbo's theorem. In fact, the manifolds \(B_{p,q}\) present the first known examples of Stein surfaces with no closed exact Lagrangian submanifolds, but with non-vanishing symplectic cohomology. The authors first study the Milnor fibre \(S_n\), which is given by the equation \(z^{n+1} + 2xy = 1\) in \(\mathbb{C}^3\) and, via projection to the \(z\)-coordinate, admits the structure of an exact Lefschetz fibration \(\Pi_n: S_n \rightarrow \mathbb{C}\). They consider certain Lagrangian tori \(\mathbb{T}_n\) in \(S_n\) which project to loops under \(\Pi_n\), which they call \textit{matching tori}. These tori are shown to be monotone and have minimal Maslov index 2; in the case \(n=1\) they are related to Polterovich's construction of a monotone Lagrangian torus in \(T^* S^2\) [\textit{P. Albers} and \textit{U. Frauenfelder}, Commun. Pure Appl. Math. 61, No. 8, 1046--1051 (2008; Zbl 1142.53067)]. Studying holomorphic discs bounded by \(\mathbb{T}_n\) and applying a result of \textit{P. Biran} and \textit{O. Cornea} [Geom. Topol. 13, No. 5, 2881--2989 (2009; Zbl 1180.53078)], the authors calculate \(HF^*(\mathbb{T}_n; \mathbb{Z}_2) \cong H^*(\mathbb{T}_n; \mathbb{Z}_2)\) for \(n>0\). The authors then turn their attention to the quotient \(B_{p,q}\) of \(S_{p-1}\), in the process giving Legendrian surgery diagrams for both \(S_{p-1}\) and \(B_{p,q}\). To show that \(B_{p,q}\) does not contain a closed exact Lagrangian submanifold \(L\), they consider the preimage of \(L\) in \(S_{p-1}\) under the covering map, and invoke a theorem of \textit{A. F. Ritter} [Geom. Funct. Anal. 20, No. 3, 779--816 (2010; Zbl 1228.53092)] to deduce that this preimage must consist of spheres, so that \(L\) is either a sphere or \(\mathbb{R}P^2\). Lagrangian spheres are ruled out in rational homology 4-balls, so \(L\) must be an \(\mathbb{R}P^2\). When \(p\) is odd, this case is also immediately ruled out, leaving only the case \(p\) even. In this case the authors invoke results of [\textit{A. Ishii} et al., J. Differ. Geom. 84, No. 1, 87--126 (2010; Zbl 1198.14020); \textit{A. Ishii} and \textit{H. Uehara}, J. Differ. Geom. 71, No. 3, 385--435 (2005; Zbl 1097.14013)] that any spherical object in the exact Fukaya category of \(S_{p-1}\) is isomorphic to a matching sphere. For \(p>2\) then the authors consider two disjoint matching spheres in \(S_{n-1}\) related by a deck transformation; a result of \textit{M. Khovanov} and \textit{P. Seidel} [J. Am. Math. Soc. 15, No. 1, 203--271 (2002; Zbl 1035.53122)] reduces the rank of their Floer cohomology (which is zero as they are disjoint) to the geometric intersection number of curves in the complex plane, which is shown to be nonzero, giving a contradiction. To show that \(B_{p,q}\) contains a Floer-theoretically essential Lagrangian torus, the authors consider the quotient \(\mathbb{T}_{p,q}\) of \(\mathbb{T}_{p-1}\) in \(S_{p-1}\) under the covering map. It is shown that \(HF^*(\mathbb{T}_{p,q};\mathbb{Z}_2) \cong H^*(\mathbb{T}_{p,q};\mathbb{Z}_2)\), using the previous computation of \(HF^*(\mathbb{T}_{p-1};\mathbb{Z}_2)\). Finally, they adapt a result of \textit{P. Seidel} [in: Current developments in mathematics, 2006. Somerville, MA: International Press. 211--253 (2008; Zbl 1165.57020)] to show that this Floer-theoretically essential monotone Lagrangian 2-torus implies that the symplectic cohomology \(SH^*(B_{p,q})\) is nonzero.
    0 references
    0 references
    0 references
    0 references
    0 references
    Stein manifolds
    0 references
    exact Lagrangian submanifolds
    0 references
    monotone Lagrangian tori
    0 references
    symplectic cohomology
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references