Monopole Floer homology and Legendrian knots (Q422839)

The author uses Monopole Floer Homology for sutured manifolds to construct invariants of unoriented Legendrian knots in a contact 3-manifold. These invariants assign to a Legendrian knot \(\mathcal{K}\subset Y\) elements of the Monopole Knot Homology \(KHM(-Y,K)\). Recall that a knot \(\mathcal{K}\) in a contact manifold \((Y,\xi)\) is said to be \textit{Legendrian} if the tangent vectors to \(\mathcal{K}\) lie in the contact planes \(\xi\). In [J. Eur. Math. Soc. (JEMS) 11, No. 6, 1307--1363 (2009; Zbl 1232.57017)], \textit{P. Lisca}, et al., constructed an invariant using open book decompositions. In [\textit{T. S. Mrowka} and \textit{Y. Rollin}, ``Contact Invariants and Monopole Floer Homology'', in preparation], Kronheimer-Mrowka defined a knot invariant from Monopole Floer Homology, they used a version of \textit{A. Juhász}'s sutured Floer homology [Algebr. Geom. Topol. 6, 1429--1457 (2006; Zbl 1129.57039)] in order to define the Monopole Knot Homology \(KHM(Y,K)\) as the sutured invariant of the complement of \(K\). The purpose of the author is to define an analogue to the invariant defined in [\textit{P. Lisca}, et al., op. cit.] without making use of Heegard diagrams or open books but instead he proceeds by embedding the knot complement in a closed 3-manifold. In this way he defines to an unoriented Legendrian knot \(\mathcal{K}\subset (Y<\xi)\) \(l_{g}\in KHM(-Y,K)\), where (i) the isotopy class of \(\mathcal{K}\) in \(Y\) is denoted \(K\), (ii) \(KHM(-Y,K)\) is defined using local coefficients [\textit{P. Kronheimer} and \textit{T. S. Mrowka}, Monopoles and three-manifolds. New Mathematical Monographs 10. Cambridge: Cambridge University Press. (2007; Zbl 1158.57002)] and (iii) \(g\geq 2\) is an integer. These elements are obtained by choosing a particular contact structure \(\bar{\xi}\) on the closed manifold \(\bar{Y}\), so that \((Y-\mathcal{K},\xi\mid_{Y-\mathcal{K}})\) is a contact submanifold of \((\bar{Y},\bar{\xi})\), and letting \(l_{g}(\mathcal{K})\) be the monopole contact invariant of \(\bar{\xi}\). The author stresses that \(l_{g}\) presents some advantages and disavantages over the invariant defined in [\textit{P. Lisca}, et al., op. cit.]. It is hard to compute, does not distinguish between different orientations of \(\mathcal{K}\) and it does not come with a natural bigrading the way elements of Knot Floer Homology do. However, some vanishing and nonvanishing results have very simple proofs. The following theorems are proved in the context of Monopole Floer Homology, they are known to be true for the invariant in [\textit{P. Lisca}, et al., op. cit.]. {Theorem 1:} If the complement of \(\mathcal{K}\) is overtwisted, then \(l_{g}(\mathcal{K})=0\), for all \(g\). {Theorem 2:} Let \(S_{+}(\mathcal{K})\) and \(S_{-}(\mathcal{K})\) denote the positive and negative stabilizations of a Legendrian knot \(\mathcal{K}\). Then \(l_{g}(S_{+}S_{-}(\mathcal{K}))=0\) for all \(\mathcal{K}\), for all \(g\). {Theorem 3:} Let \(\mathcal{K},\mathcal{S}\subset (Y,\xi)\) be disjoint Legendrian knots, and le \(\mathcal{K}_{\mathcal{S}}\subset Y_{\mathcal{S}}\) denote the image of \(\mathcal{K}\) in the contact manifold \(Y_{\mathcal{S}}\) obtained by performing a contact \((+1)\)-surgery on \(\mathcal{S}\). Then there is a map \(KHM(-Y,K)\to KHM(-Y_{\mathcal{S}},K_{\mathcal{S}})\) sending \(l_{g}(\mathcal{K})\) to \(l_{g}(\mathcal{K}_{\mathcal{S}})\). By using results of \textit{T. S. Mrowka} and \textit{Y. Rollin} [Algebr. Geom. Topol. 6, 1--69 (2006; Zbl 1207.57015), ``Contact Invariants and Monopole Floer Homology'', in preparation] on the monopole contact invariant, not known to be true in Heegard Floer Homology, the author proves an entirely new property of \(l_{g}(\mathcal{K})\); {Theorem 4:} Let \(\mathcal{K}_{0},\mathcal{K}_{1}\subset (Y,\xi)\) be Legendrian knots, with \(Y\) a homology \(3\)-sphere, and suppose that \(\mathcal{K}_{0}\) is a Lagrangian concordant (defined in [\textit{B. Chantraine}, Dokl. Math. 81, No. 1, 63--65 (2010); translation from Dokl. Akad. Nauk, Ross. Akad. Nauk 430, No. 4, 444--447 (2010; Zbl 1208.15013)]) to \(\mathcal{K}_{1}\). Then, there is a map \[ KHM(-Y,K_{1})\to KHM(-Y,K_{0}) \] \noindent such that \(l_{g}(\mathcal{K}_{1})\to l_{g}(\mathcal{K}_{0})\).