Classifying spaces for the family of virtually abelian subgroups of orientable 3-manifold groups (Q2171916)

In this paper, the authors compute the \(\mathcal{F}_n\)-dimension of \(3\)-manifold groups for all \(n\geq 2\). For a group \(\Gamma\), a collection \(\mathcal{F}\) of its subgroups is called a family if it is non-empty, closed under conjugation and taking subgroups. A \(\Gamma\)-CW-complex \(X\) is a model for the classifying space \(E_{\mathcal{F}}\Gamma\) if every isotropy group of \(X\) belongs to \(\mathcal{F}\) and the fixed point set \(X^H\) is contractible if \(H\) lies in \(\mathcal{F}\). The model for \(E_{\mathcal{F}}\Gamma\) always exists and is unique up to \(\Gamma\)-homotopy equivalence. The \(\mathcal{F}\)-geometric dimension of \(\Gamma\) is defined to be the minimal dimension of models for \(E_{\mathcal{F}}\Gamma\). This paper considers the case that \(\Gamma\) is the fundamental group of a connected, closed, oriented \(3\)-manifold, and \(\mathcal{F}=\mathcal{F}_n\) consists of all subgroups of \(\Gamma\) that contain \(\mathbb{Z}^k\) as a finite index subgroup for some \(k\leq n\). The \(\mathcal{F}_1\)-geometric dimensions of \(3\)-manifold groups were computed by \textit{K. Joecken} et al. [Groups Geom. Dyn. 15, No. 2, 577--606 (2021; Zbl 1478.55009)]. Since any abelian subgroup of a \(3\)-manifold group has rank at most \(3\), this paper only needs to consider the cases \(n=2,3\). There are three main results in this paper. Theorem 1.1 considers the behavior of \(\mathcal{F}_2\)- and \(\mathcal{F}_3\)-geometric dimensions under prime decomposition, and the \(\mathcal{F}_2\)- (\(\mathcal{F}_3\)-) geometric dimension of a connected sum is usually the maximum of the \(\mathcal{F}_2\)- (\(\mathcal{F}_3\)-) geometric dimensions of its prime summands. Theorem 1.2 considers the behavior of \(\mathcal{F}_2\)- and \(\mathcal{F}_3\)-geometric dimensions under JSJ decomposition, and the \(\mathcal{F}_2\)- (\(\mathcal{F}_3\)-) geometric dimension of a prime \(3\)-manifold is usually the maximum of the \(\mathcal{F}_2\)- (\(\mathcal{F}_3\)-) geometric dimensions of its JSJ pieces. The third result is given in Table 1, which computes the \(\mathcal{F}_2\)- and \(\mathcal{F}_3\)-geometric dimensions of geometric \(3\)-manifolds with empty or tori boundary. The proofs of these three results are arranged in reverse order. The authors first compute \(\mathcal{F}_2\)- and \(\mathcal{F}_3\)-geometric dimensions of geometric \(3\)-manifolds. Besides cases that were previously known, the lower bounds are given by push-out constructions (Theorems 2.8 and 2.9), and the upper bounds are given by computing cohomology dimensions. For the prime decomposition and JSJ decomposition, the authors first study the corresponding \(\Gamma\)-action on the Bass-Serre tree, and use it to compute the \(\mathcal{F}\)-geometric dimension for another family of subgroups \(\mathcal{F}\) containing \(\mathcal{F}_2\) and \(\mathcal{F}_3\). Then they use Proposition 2.1 that estimates the change of geometric dimensions when we pass from a bigger family of subgroups to a smaller family.