Infinitely many universally tight torsion free contact structures with vanishing Ozsváth-Szabó contact invariants (Q443938)

For the classification of contact structures on \(3\)-manifolds, \textit{Y. Eliashberg} [Invent. Math. 98, No. 3, 623--637 (1989; Zbl 0684.57012)] introduced a dichotomy between tight and overtwisted contact structures, and gave a complete classification of overtwisted contact structures on \(3\)-manifolds in terms of homotopy theory. A contact structure \(\xi\) on a \(3\)-manifold \(M\) is said to be overtwisted if there is an embedded disk \(D\) which is everywhere tangent to \(\xi\) along \(\partial D\). Non-overtwisted contact structures are called tight. The latter notion was further divided into universally tight and virtually overtwisted. Two algebraic objects play an important role in studying the classification of tight contact structures. The first one is \textit{E. Giroux}'s torsion introduced in [Invent. Math. 141, No. 3, 615--689 (2000; Zbl 1186.53097)]. The \(k\pi\)-torsion of a contact manifold \((M, \xi)\) was defined to be the supremum of all integers \(n\geq 1\) such that there exist a contact embedding of \[ (T^2 \times [0,1], {\text{ ker}}(\cos(nk\pi z z) dx -\sin(nk\pi z)dy)),\quad (x,y,z)\in T^2\times[0, 1] \] into the interior of \((M, \xi)\) or zero if no such integer \(n\) exists. It is invariant under isomorphisms, and always infinite for the overtwisted structures. Another object is \textit{P. Ozsváth} and \textit{Z. Szabó}'s contact invariant in Heegaard Floer homology of the ambient manifold introduced in [Duke Math. J. 129, No. 1, 39-61 (2005; Zbl 1083.57042)]. This invariant vanishes for the overtwisted contact structures. Hence these two invariants are a powerful tool to detect tightness of contact structures on \(3\)-manifolds. \textit{P. Ghiggini, K. Honda} and \textit{J. Van Horn-Morris} [``The vanishing of the contact invariant in the presence of torsion'', preprint', \url{arXiv:0706.1602}] proved that the Ozsváth-Szabó contact invariant over \(\mathbb{Z}\)-coefficients vanishes whenever Giroux torsion is non zero. The main result of the paper under review shows that the converse does not hold. More precisely, Massot proves that every Seifert manifold whose base has genus at least three supports infinitely many (explicit) isotopy classes of universally tight torsion free contact structures whose Ozsváth-Szabó invariant over \(\mathbb{Z}\) coefficients vanishes. In addition there are two interesting results in this article. The first one states that there exists a specific example of overtwisted contact manifold containing a null-homologous strongly non loose Legendrian knot whose sutured invariant vanishes. This shows that the converse of Corollary 1.2 of \textit{A. I. Stipsicz} and \textit{V. Vértesi} [Pac. J. Math. 239, No. 1, 157--177 (2009; Zbl 1149.57031)] does not hold. The second one claims that there is a unique up to sign \(H_1(T^3)\)-equivariant isomorphism between \(\widehat{HF}(T^3)\) and \(H^1(T^3)\oplus H^1(T^3)\), under which the Ozsváth-Szabó invariant of a torsion free contact structure on \(T^3\) is sent to the Poincaré dual of its Giroux invariant. This proves Conjecture 7.13 in [\textit{K. Honda, W. H. Kazez} and \textit{G. Matić}, ``Contact structures, sutured Floer homology and TQFT'', preprint, \url{arXiv:0807.2431}]. Shortly after the first version of the paper under review, closely related results were obtained by \textit{C. Wendl} [``Holomorphic curves in blown up open books'', preprint, \url{arXiv:1001.4109}] using contact fiber sums of open book decompositions. He defined an infinite hierarchy of filling obstructions for contact \(3\)-manifolds, called planar \(k\)-torsion for integers \(k\geq 0\), showed that planar \(0\)-torsion is equivalent to overtwistedness and that every contact manifold with Giroux torsion also has planar \(1\)-torsion, and also constructed examples of contact manifolds that have planar \(k\)-torsion for any \(k\geq 2\) but no Giroux torsion.