Embedded contact homology and Seiberg-Witten Floer cohomology. I. (Q619657)

Gauge theory and symplectic topology have given rise to a wealth of new invariants of 4- and 3-dimensional manifolds in the last 3 decades. The author of this series of papers has been at the forefront of this activity and this work represents another significant contribution to our understanding. Various other works of Taubes are in the background of this work. In the 1990's he studied the Seiberg-Witten equations on closed symplectic 4-manifolds \((X,\omega)\), showing in particular that the Seiberg-Witten invariants of \(X\) are equal to a Gromov invariant given by counting pseudo-holomorphic curves for a compatible almost-complex structure \(J\) [\textit{C. H. Taubes}, Seiberg Witten and Gromov invariants for symplectic 4-manifolds. First International Press Lecture Series. 2. Somerville, MA: International Press. (2000; Zbl 0967.57001)]. The proof involved perturbing the Seiberg-Witten equations by a term containing \(r\omega\); as \(r\to\infty\) the solutions to the perturbed equations concentrate along a pseudo-holomorphic curve. In the other direction, ``almost'' solutions to the perturbed Seiberg-Witten equations, supported near a pseudo-holomorphic curve in \(X\), are constructed using solutions to the vortex equations on the complex plane, and then under certain conditions it is shown that there is an exact solution near each such almost-solution. This series of papers establishes a 3-dimensional analogue of the \(\mathrm{Gr=SW}\) result mentioned above. For a 3-manifold \(M\), solutions to the Seiberg-Witten equations on \(M\times\mathbb R\) can be interpreted as gradient flowlines of the Chern-Simons-Dirac functional. This viewpoint (and a great deal of work) gives rise to Seiberg-Witten Floer homology groups associated to each \(\text{Spin}^c\) structure on \(M\). This has been developed by various authors including [\textit{M. Marcolli} and \textit{B.-L. Wang}, Commun. Anal. Geom. 9, No. 3, 451--639 (2001; Zbl 1026.57025)] and \textit{K. A. Frøyshov} [Duke Math. J. 155, No. 3, 519--576 (2010; Zbl 1237.57033)]. The most general and authoritative account --- applicable to any closed oriented \(M\) --- is given by \textit{P. Kronheimer} and \textit{T. Mrowka} [Monopoles and three-manifolds. New Mathematical Monographs 10. Cambridge: Cambridge University Press (2007; Zbl 1158.57002)], and their invariant \(\widehat{HM}_*\) is the version of Seiberg-Witten Floer homology considered here. There is an associated Seiberg-Witten Floer cohomology group, here denoted \(\mathcal{H}^{\text{SW}}\). Given a contact form on \(M\), there is a natural way to endow \(M\times\mathbb R\) with a symplectic form and compatible almost complex structure. In this context, Michael Hutchings and various collaborators, including Taubes, have developed an invariant called \textit{embedded contact homology}, here denoted \(\mathcal{H}_{\text{ech}}\) [\textit{M. Hutchings}, J. Eur. Math. Soc. (JEMS) 4, No. 4, 313--361 (2002; Zbl 1017.58005), \textit{M. Hutchings} and \textit{M. Sullivan}, Geom. Topol. 10, 169--266 (2006; Zbl 1101.53053), \textit{M. Hutchings} and \textit{C. H. Taubes}, J. Symplectic Geom. 5, No. 1, 43--137 (2007; Zbl 1157.53047), J. Symplectic Geom. 7, No. 1, 29--133 (2009; Zbl 1193.53183)]. This is the homology of a chain complex whose differential counts pseudo-holomorphic curves \(C\) in \(M\times\mathbb R\), and whose generators are finite sets of pairs \((\gamma,m)\), each consisting of a closed integral curve of the Reeb vector field \(v\) on \(M\) and a positive integer. At each end of \(M\times\mathbb R\), these curves \(C\) asymptotically approach disjoint unions of cylinders on \(m\)-fold covers of Reeb orbits \(\gamma\). There is an \(\mathcal{H}_{\text{ech}}\) invariant associated to each \(\text{Spin}^c\) structure on \(M\). In other previous work, \textit{C. H. Taubes} gave a proof of the Weinstein conjecture by proving that solutions to perturbed Seiberg-Witten equations on \(M\) give rise to closed integral curves of \(v\) [Geom. Topol. 11, 2117--2202 (2007; Zbl 1135.57015), ibid. 13, No. 3, 1337--1417 (2009; Zbl 1200.57018)]. In the papers under review, he describes a map \(\Phi^r\) from generators of \(\mathcal{H}_{\text{ech}}\) to generators of \(\mathcal{H}^{\text{SW}}\) and also a map \(\Psi^r\) from the set of psudo-holomorphic curves on \(M\times\mathbb R\) which are counted in the differential for \(\mathcal{H}_{\text{ech}}\), to the set of Seiberg-Witten solutions on \(M\times\mathbb R\) which are counted in the differential for \(\mathcal{H}^{\text{SW}}\), and (very roughly speaking) shows that for large \(r\) these give rise to an isomorphism between the two invariants. This beautiful result proves in particular that embedded contact homology does not depend on the choice of contact form, and that it is finitely generated in each degree; both of these facts were conjectured by Hutchings. The construction of the maps \(\Phi^r\) and \(\Psi^r\) is quite involved, and is along the lines of the earlier work on closed symplectic 4-manifolds mentioned above. In the first paper of the series a review of each of \(\mathcal{H}^{\text{SW}}\) and \(\mathcal{H}_{\text{ech}}\) is given and the main theorem establishing their equality is stated, and its proof outlined modulo the description and properties of the maps \(\Phi^r\) and \(\Psi^r\). These are given in the second, third and fourth papers. In the last paper it is shown that the isomorphism intertwines extra structure including the degree two endomorphisms on each of \(\mathcal{H}^{\text{SW}}\) and \(\mathcal{H}_{\text{ech}}\), the actions on each by the first homology of \(M\), and the distinguished elements associated to a contact structure in each theory. The papers are clear, well-organised and highly readable considering their level of technical detail. A different extension of link Floer homology to singular links is given by \textit{P. Ozsváth, A. Stipsicz} and \textit{Z. Szabó} [J. Topol. 2, No. 2, 380--404 (2009; Zbl 1190.57020)].