The Seiberg-Witten equations and the Weinstein conjecture II: More closed integral curves of the Reeb vector field (Q1006156)

This paper is a sequel to a previous paper of the author [Geom. Topol. 11, 2117--2202 (2007; Zbl 1135.57015)] in which he gave a proof of the Weinstein conjecture, which states that the Reeb vector field \(v\) of any contact 1-form on a compact 3-manifold \(M\) admits a closed integral curve. The main theorem of this paper gives a sufficient condition for a homology class on \(M\) to be represented by a positive integer weighted sum of closed integral curves of \(v\). The contact form gives a canonical isomorphism between \(H^2(M;{\mathbb Z})\) and the set of \(\mathrm{Spin}^c\) structures on \(M\); thus each element of the second cohomology has an associated Seiberg-Witten Floer homology group, as in [\textit{P. Kronheimer} and \textit{T. Mrowka}, Monopoles and three-manifolds. New Mathematical Monographs 10. Cambridge: Cambridge University Press (2007; Zbl 1158.57002)]. The theorem states that the Poincaré dual of a nonzero class \(e\in H^2(M;{\mathbb Z})\) is represented by a positive integer weighted sum of closed integral curves of \(v\) if the associated Seiberg-Witten Floer homology group is nontrivial. The zero class is represented by a positive integer weighted sum of closed integral curves of \(v\) if the associated Seiberg-Witten Floer homology group is not isomorphic to \({\mathbb Z}\). Let \(K^{-1}\) denote the subbundle of \(TM\) which is the kernel of the contact form. The proof of the above theorem in the case where \(e\) differs from the Euler class of \(K\) by a torsion class was given in the previous paper of the author mentioned above. This paper gives the proof in the case that these elements differ by a non-torsion class. The proof follows the same general strategy as the previous paper. A perturbation of the Seiberg-Witten equations is given involving a real parameter \(r\) times the contact form. Very roughly speaking, the condition on the Seiberg-Witten Floer homology ensures the existence of a solution \((A_r,\alpha_r,\beta_r)\) to these perturbed Seiberg-Witten equations, where \(\alpha_r\) is a section of the line bundle \(E\) with first Chern class given by \(e\). As \(r\to\infty\), the subset \(\alpha_r^{-1}(0)\) converges in a suitable sense to the desired set of closed integral curves. The two main differences in the current situation are that the Seiberg-Witten Floer homology groups have a relative \({\mathbb Z}/p\,{\mathbb Z}\) grading instead of a relative \({\mathbb Z}\) grading, and the functional which plays the role of a Morse function in the definition of the Floer homology groups is not fully gauge-invariant.