A vanishing theorem for the Gromov volume of 3-manifolds with an application to circle foliations (Q1598179)

Let \(M\) be a compact orientable manifold of dimension \(n\). The Gromov volume of \(M\), denoted by \(||M||\), is defined as the infimum of the \(\ell^1\)-norms \(\sum_{j=1}^k |\lambda_j|\) of all cycles \(\sum_{j=1}^k \lambda_j\sigma_j\), where \(\sigma_j\) is a singular \(n\)-simplex in \(M\) and \(\lambda_j\) are real numbers, representing the fundamental class \([M, \partial M]\). The vanishing theorem of Gromov says that a closed orientable \(n\)-manifold has zero Gromov volume if it has an open covering of dimension \(<\dim M\) by amenable sets. The paper under review generalizes this vanishing theorem in the case of compact orientable weakly irreducible \(3\)-dimensional manifolds allowing some non-amenable member in the covering. The extra assumption on the non-amenable subsets is that they have disjoint \(d\)-cores (defined in the paper) with zero Gromov volume. The conclusion of the theorem says that the component of \(M\) either has zero Gromov volume or is a generalized graph manifold. There are some interesting consequences of the theorem. Recall that a circle foliation on a \(3\)-manifold \(M\) gives a Seifert fibered structure on \(M\) if there are no leaves of the foliation with infinite holonomy. And the foliation is called of Epstein length at most \(1\) if either the foliation is a Seifert fibration or the restriction of the foliation on the union of leaves of infinite holonomy does not have any leave of infinite holonomy. Also it is known that Seifert fibered spaces have Gromov volume zero. A consequence (Corollary 1) of the authors' above theorem (Covering Theorem) is that if \(M\) is a compact orientable weakly irreducible \(3\)-manifold such that either \(\partial M\) has no \(2\)-sphere component or the components of \(\partial M\) of genus \(>1\) are incompressible in \(M\) and int\((M)\) supports a circle foliation of Epstein length at most \(1\) then the Gromov volume of \(M\) is zero. Some other consequences are characterization of handlebodies and generalized graph manifolds in terms of existence of circle foliation of Epstein length at most \(1\) on the complement of certain subset of the manifolds. The proof of the covering theorem is by induction on a certain complexity of \(3\)-manifold, defined in the paper, and reducing the general case to the case when all the open sets in the covering are amenable.