Four-dimensional Haken cobordism theory (Q2358261)

In this paper, the authors prove that any two closed Haken \(3\)-manifolds are Haken cobordant to each other, and in particular any closed Haken 3-manifold is the boundary of a Haken \(4\)-manifold. The concept of Haken \(3\)-manifold is fundamental and important in \(3\)-manifold theory, since it admits the Haken hierarchy, and this structure makes inductive proofs available for Haken \(3\)-manifolds. In the authors' previous work [Contemp. Math. 560, 71--84 (2011; Zbl 1335.57038)], they defined Haken \(n\)-manifolds and Haken cobordism. Roughly speaking, a Haken \(n\)-manifold is an \(n\)-manifold that can be inductively splitted along codimension-\(1\) essential submanifolds to eventually get a collection of \(n\)-cells. A Haken cobordism between two Haken \(n\)-manifolds \(M\) and \(M'\) is a Haken \((n+1)\)-manifold \(W\) such that \(\partial W=M\cup M'\), and \(\pi_1(M)\to \pi_1(W),\pi_1(M')\to \pi_1(W)\) are injective. In the first author's thesis, he made the following fundamental observation on Haken cobordism: if \(M\) is a closed Haken \(3\)-manifold, \(F\) is an incompressible surface in \(M\), and \(N\) is obtained from \(M\) by splitting along \(F\) and regluing, then \(M\) is Haken cobordant to the union of \(N\) and a mapping torus with fibered surface \(F\) (and monodromy being the regluing map). By applying this observation with various structural results of \(3\)-manifolds, the authors could pass to a simpler Haken \(3\)-manifold via Haken cobordism, and eventually solve the problem. Since Gabai's internal hierarchy implies that any Haken \(3\)-manifold can be simplified via the cut-regluing process to a surface cross circle in finitely many steps, and a surface cross circle obviously bounds a Haken \(4\)-manifold, the Haken cobordism problem is reduced to the surface bundle over circle case. Since mapping class groups are generated by Dehn twists, the observation is applied again and the Haken cobordism problem is further reduced to the case that the monodromy of a fibered \(3\)-manifold is a Dehn twist. The higher genus case can be reduced to the genus one case, and the authors solve the genus one case by a very careful and hands-on construction.