Symplectic, Poisson, and contact geometry on scattering manifolds (Q2658487)

The inverse of a symplectic form is a non-degenerate Poisson structure. Manifolds often do not admit symplectic structures, but any manifold admits at least the zero Poisson structure (the most degenerate Poisson structure possible). The goal of this paper is to introduce and investigate a class of Poisson structures, called scattering-symplectic, which are both close enough to being non-degenerate to facilitate their study, and general enough to include many non-symplectic examples. Their degeneracy locus is a hypersurface \(Z\) where the degeneracy behaviour is controlled. The first goal is achieved as the paper describes for scattering-symplectic manifolds, among other things: their local structure, their Poisson cohomology (quite hard to compute in the general Poisson case), and relations with fillings of contact manifolds. As for the second goal, some examples in the paper are spheres of any even dimension, and gluings of certain contact fillings. Let us describe in some more detail the contents of the paper. In this review ``sc-symplectic'' will stand for scattering symplectic. Section 2 describes the general framework used to study sc-symplectic structures. This includes a comparison with other structures defined by closed \(2\)-forms with mild degeneracies, namely folded symplectic and \(b\)-symplectic forms (and more generally \(b^k\)-symplectic forms). Sc-symplectic forms are viewed as closed non-degenerate forms on an appropriate Lie algebroid substitute \(A\) for the tangent bundle \(TM\), i.e., as \(A\)-symplectic forms. To describe this \(A\), the author extends a rescaling procedure for vector bundles, Proposition 8.1 of [\textit{R. B. Melrose}, The Atiyah-Patodi-Singer index theorem. Wellesley, MA: A. K. Peters, Ltd. (1993; Zbl 0796.58050)], to the context of Lie algebroids. The scattering tangent bundle \(A= {}^{sc}TM\) is constructed as a rescaling of the \(b\)-tangent bundle \({}^{b}TM\), itself a rescaling of \(TM\). Within this framework, the author explores \(A\)-symplectic and \(A\)-Poisson structures, an \(A\)-symplectic Moser lemma, relations with almost regular Poisson manifolds, and techniques for computing Poisson cohomology of \(A\)-symplectic manifolds. Section 3 is devoted to a detailed study of sc-symplectic manifolds. For such an \((M,\omega)\) the author describes the Lie algebroid cohomology of \({}^{sc}TM\), a cooriented contact structure induced by \(\omega\) on its degeneracy locus \(Z\), and the local structure of \(\omega\) -- via both a sc-Darboux Theorem and a tubular neighbourhood theorem for \(\omega\) near \(Z\). There is also the construction of sc-symplectic forms on any even dimensional sphere \(S^{2n}\), such that the degeneracy locus is the equator \(S^{2n-1}\), with the standard contact structure. This is contrasted with \(b\)-symplectic geometry, where \(S^{2n}\) has no \(b\)-symplectic structure for \(n>1\) [\textit{I. Mărcuţ} and \textit{B. O. Torres}, J. Symplectic Geom. 12, No. 4, 863--866 (2014; Zbl 1321.53100)]. The relation with contact geometry is further explored: the paper characterizes (in cohomological terms) which sc-symplectic manifolds arise as convex strong symplectic fillings of a contact manifold (in the sense that the components of the complement of the degeneracy locus are the fillings). Convex and concave strong symplectic fillings of the same contact manifold can be glued together to obtain a symplectic manifold. The author shows that gluing instead two convex fillings, the result is a sc-symplectic manifold. This provides more examples of sc-symplectic manifolds, and some are described in detail in this section. It is also shown that two concave fillings can be glued to obtain a folded symplectic manifold. Finally, there is the description of the symplectic leaves, and the computation of Poisson cohomology of a sc-symplectic manifold. This is done both in an invariant manner, without choices, and also in a more explicit description (using a tubular neighborhood), exhibiting the cohomology as a sum of smaller subspaces.