Higher symplectic capacities

From MaRDI portal





For \(a_1,\dots ,a_n\in (0,\infty]\), let \(E(a_1,\dots ,a_n) = \{(z_1,\dots ,z_n)\in \mathbb{C}^n \mid \sum_{k=1}^n \frac{\pi |z_k|^2}{a_k}\leq 1 \}\). It has a standard symplectic structure as a domain in \(\mathbb{C}^n\). Gromov's cerebrated non-squeezing theorem shows that there is no symplectic embedding \(E(a,\dots ,a) \hookrightarrow E(a',\infty,\dots ,\infty)\) unless \(a\leq a'\). After this theorem, a natural question is the following (Question 1.1): For which \(a_1,\dots ,a_n\) and \(a'_1,\dots ,a'_n\) is there a symplectic embedding \(E(a_1,\dots ,a_n)\hookrightarrow E(a'_1,\dots ,a'_n)\)? To deal with this problem (and more general symplectic embedding problems), the notion of symplectic capacities have been used. For example, the Ekeland-Hofer capacity and the ECH (embedded contact homology) capacity assign positive real numbers to symplectic manifolds, and their monotonicity property can be used to obstruct symplectic embeddings.\N\NMore recently, the ``stabilized ellipsoid embedding problem is focused (Question 1.4): For which \(a,b,c,d\) is there a symplectic embedding \(E(a,b,\infty,\dots ,\infty)\hookrightarrow E(c,d,\infty,\dots ,\infty)\)? By several works, including [\textit{D. Cristofaro-Gardiner} and \textit{R. Hind}, Comment. Math. Helv. 93, No. 1, 1--32 (2018; Zbl 1398.53083)], there have been progress on this problem by using rigid pseudo-holomorphic curves in \(E(c,c)\setminus E(1,b)\), but this method is not always applicable due to the fact that a rigid curve might have a negative index after a stabilization taking the product of the ellipsoids and \(\mathbb{C}^{n-2}\).\N\NThis paper under review introduces a new family of capacities \(\mathfrak{g}_b\) for \(b\in \overline{S}\mathbb{K}\langle \mathfrak{t}_1, \mathfrak{t}_2,\dots \rangle\) (the reduced symmetric tensor algebra with formal variables \(\mathfrak{t}_1, \mathfrak{t}_2,\dots\)). For any Liouville domain \(X\), a positive real number \(\mathfrak{g}_b(X)\) is assigned. Each \(\mathfrak{g}_b\) satisfies the scaling and monotonicity properties, as well as other capacities, and moreover it is invariant under stabilization, that is, \(\mathfrak{g}_b(X\times \mathbb{C})=\mathfrak{g}_{b}(X)\) if \(X\) satisfies the index positivity assumption. The precise statement and a quantitative refinement of the stability property is provided in Theorem 1.8. As specializations of \(\{\mathfrak{g}_b\}\), symplectic capacities \(\mathfrak{g}_1 ,\mathfrak{g}_2,\dots\) are defined, and Theorem 1.10 gives constraints on its values \(\mathfrak{g}_k(X)\) under certain conditions on \(X\). More precisely, one condition is about the Gromov-Witten invariant of a closed symplectic manifold \(M\) admitting a symplectic embedding \(X\hookrightarrow M\) and the other is about the Reeb dynamics on \(\partial X\).\N\NThe construction of \(\{\mathfrak{g}_b\}\) is based on the rational symplectic field theory using the contact homology algebra \(\mathrm{CHA(\partial X)}\) of \(\partial X\) and its linearization \(\mathrm{CH}_{\mathrm{lin}}(X)\) together with a filtered \(\mathcal{L}_{\infty}\) structure. It is crucial to include higher operations in the \(\mathcal{L}_{\infty}\) sturcture. It should be remarked that, as mentioned in Disclaimer 1.9, the rational symplectic field theory requires virtual perturbation techniques and its full construction is not yet established, so this paper takes its existence and expected properties outlined in e.g. [\textit{Y. Eliashberg} et al., GAFA 2000, 560--673 (2000; Zbl 0989.81114)] as a black box.\N\NIn addition, Theorem 1.14 gives computations of \(\mathfrak{g}_k\) for a ball, a ellipsoid and a polydisk in \(\mathbb{C}^2\). As applications, one can get constraints on the existence of stabilized symplectic embeddings, where the ECH capacities are not applicable, and some of which (Example 1.15 and 1.17) are sharper than the constraints from the Ekeland-Hofer capacities.\N\NThe paper is organized as follows. Section 2 gives a review of filtered \(\mathcal{L}_{\infty}\) algebras. In Section 3, the rational symplectic field theory is constructed by counting (in a suitable virtual sense) pseudo-holomorphic curves in exact symplectic cobordisms. In general, the complement of a symplectic embeddings is a nonexact symplectic cobordism. In order to treat such cobordisms, we consider in Section 4 a deformation by the Cieliebak-Latschev element, which is a Maurer-Cartan element. Then, we can obtain a transfer map between the linearized complexes with \(\mathcal{L}_{\infty}\) structures for the boundary contact manifolds. In Section 5, \(\mathcal{L}_{\infty}\) augmentations on the contact homology algebras are defined by counting pseudo-holomorphic curves satisfying local tangency constraints at finitely many points. In particular, Section 5.4 gives relations between the linearized complexes for \(X\) and \(X\times B^2(S)\), where \(B^2(S)\subset \mathbb{C}\) is a ball of radius \(S>0\). Section 6 gives the construction of \(\mathfrak{g}_b(X)\) using the augmentations of Section 5, and then proves the fundamental properties in Theorem 1.8 and Theorem 1.10. The computations as in Theorem 1.14 is provided in the last Section 6.3.



Cites work









This page was built for publication: Higher symplectic capacities

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6870527)