The Gromov width of 4-dimensional tori (Q372686): Difference between revisions

A \textit{symplectic manifold} is a \(2n\)-dimensional manifold \(M\) together with a closed 2-form \(\omega\) such that \(\omega^n\) is non-zero everywhere. For example, \(\omega_0:=dx_1\wedge dy_1+dx_2\wedge dy_2\) is a symplectic structure on \(\mathbb{R}^4\). Other examples are given by the quotient of \(({\mathbb R}^4,\omega_0)\) by a suitable lattice \(\Lambda\). Such a quotient manifold is necessarily a 4-torus, and the resulting symplectic form is called a \textit{linear symplectic form} on a 4-torus. It is not known whether every symplectic form on a 4-torus is symplectomorphic to such a form. Given \(a>0\) we consider the open ball of capacity \(a\), i.e. we consider \[ B^4(a):=\{ z\in \mathbb{C}^2\,|\, \pi(|z_1|^2+|z_2|^2|)<a\} \] as a subset of the symplectic space \(({\mathbb R}^4,\omega_0)\). The \textit{ball filling number} of a finite volume symplectic manifold \((M,\omega)\) is defined as \[ p(M,\omega):=\text{sup} \frac{ \text{Vol}(B^4(a))}{\text{Vol}(M,\omega)}, \] where the supremum is taken over all balls \(B^4(a)\) that symplectically embed into \((M,\omega)\). If \(p(M,\omega)<1\), then one says that \textit{there is a filling obstruction}, while if \(p(M,\omega)=1\), one says that \((M,\omega)\) \textit{admits a full filling by one ball}. There are symplectic manifolds which have a filling obstruction. But the main theorem of the paper says that a 4-torus with a linear symplectic form admits a full filling by one ball.