An absolute grading on Heegaard Floer homology by homotopy classes of oriented 2-plane fields (Q2358615)

This paper refines existing gradings on the Heegaard Floer homology of a closed 3-manifold to a grading by homotopy classes of oriented 2-plane fields. Heegaard Floer homology defines, for a closed oriented 3-manifold \(Y\), various abelian groups, denoted \(\widehat{HF} (Y), HF^+ (Y), HF^- (Y)\), and \(HF^\infty (Y)\), which are powerful invariants. \textit{P. Ozsváth} and \textit{Z. Szabó} [Ann. Math. (2) 159, No. 3, 1027--1158 (2004; Zbl 1073.57009)] showed that these abelian groups are graded by \(\text{Spin}^c (Y)\), the set of spin-c structures on \(Y\), so \[ \widehat{HF} (Y) = \bigoplus_{\mathfrak{s} \in \text{Spin}^c (Y) } \widehat{HF} (Y, \mathfrak{s}). \] Further, for a fixed spin-c structure \(\mathfrak{s}\), Ozsváth and Szabó showed that the \(\mathfrak{s}\)-summand has a relative grading by \(\mathbb{Z}/d(c_1(\mathfrak{s}))\), where \(c_1(\mathfrak{s})\) is the first Chern class of \(\mathfrak{s}\) and \(d(c_1(\mathfrak{s}))\) is its divisibility. The present paper shows that these gradings can be combined into a single absolute grading by homotopy classes of oriented 2-plane fields. The homotopy classes of oriented 2-plane fields on an oriented 3-manifold are distinguished by 2-dimensional and 3-dimensional obstruction classes, corresponding to spin-c structures and a \(\mathbb{Z}/d(c_1(\mathfrak{s}))\)-valued obstruction, so this grading is indeed a refinement. More precisely, denote by \(\mathcal{P}(Y)\) the set of homotopy classes of oriented 2-plane fields on \(Y\), and \(\mathcal{P}(Y, \mathfrak{s})\) the subset with spin-c structure \(\mathfrak{s}\). For a genus-\(g\) Heegaard diagram \((\Sigma, \mathbf{\alpha}, \mathbf{\beta})\) of \(Y\), defining tori \(\mathbb{T}_\alpha\) and \(\mathbb{T}_\beta\) in \(\text{Sym}^g (\Sigma)\), the authors define a grading function \(\widetilde{\text{gr}} : \mathbb{T}_\alpha \cap \mathbb{T}_\beta \rightarrow \mathcal{P}(Y)\). This function \(\widetilde{\text{gr}}\) is shown to refine the existing gradings: if \(\mathbf{x} \in \mathbb{T}_\alpha \cap \mathbb{T}_\beta\) has spin-c grading \(\mathfrak{s}\), then \(\widetilde{\text{gr}}(\mathbf{x}) \in \mathcal{P}(Y, \mathfrak{s})\); and if \(\mathbf{x}, \mathbf{y}\) have the same spin-c grading \(\mathfrak{s}\), then \(\widetilde{\text{gr}}(\mathbf{x}) - \widetilde{\text{gr}}(\mathbf{y})\) gives the relative grading in \(\mathbb{Z}/d(c_1(\mathfrak{s}))\). The grading is shown to be compatible with 4-dimensional cobordisms, in an appropriate sense. It is also shown to be invariant under Heegaard moves, so that absolute gradings are obtained on all of \(\widehat{HF}(Y), HF^+ (Y), HF^- (Y)\) and \(HF^\infty (Y)\). The authors also show that the grading is compatible with contact invariants. \textit{P. Ozsváth} and \textit{Z. Szabó} [Duke Math. J. 129, No. 1, 39--61 (2005; Zbl 1083.57042)] showed that a contact structure \(\xi\) on \(Y\) yields an invariant \(c(\xi) \in \widehat{HF}(-Y)\). The authors show that \(c(\xi)\) has grading given by the homotopy class of \(\xi\). As corollaries, the authors are able to re-derive some known results on the number of homotopy classes of weakly fillable contact structures on \(Y\): this number is shown to be finite, and if \(Y\) is an \(L\)-space or admits a metric of constant positive curvature, then it is \(\leq |H_1 (Y; \mathbb{Z})|\). The proofs are very explicit. To define \(\widetilde{\text{gr}} : \mathbb{T}_\alpha \cap \mathbb{T}_\beta \rightarrow \mathcal{P}(Y)\), the authors consider a Morse function \(f : Y \rightarrow \mathbb{R}\) compatible with \((\Sigma, \mathbf{\alpha}, \mathbf{\beta})\), and make explicit modifications to a gradient-like vector field for \(f\) near flowlines between critical points. They use the fact that an \(\mathbf{x} \in \mathbb{T}_\alpha \cap \mathbb{T}_\beta\) determines flowlines between index 1 and 2 critical points. The orthogonal complement of the resulting nonvanishing vector field is then defined to be \(\widetilde{\text{gr}}(\mathbf{x})\). Proofs that the grading has the desired properties are obtained by careful consideration of this explicit vector field, and several further constructions associated to it, including use of the Pontryagin-Thom construction.