The virtual Thurston seminorm of 3-manifolds (Q2631630)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | The virtual Thurston seminorm of 3-manifolds |
scientific article |
Statements
The virtual Thurston seminorm of 3-manifolds (English)
0 references
15 May 2019
0 references
In this paper, the authors prove that the Thurston seminorms of all finite covers of an aspherical $3$-manifold $N$ with empty or toroidal boundary determine whether $N$ is a graph manifold, a mixed manifold or a hyperbolic manifold. \par For a $3$-manifold $N$ as above, let its Thurston seminorm be denoted by $x_N:H_1(N;\mathbb{R})\to \mathbb{R}_{\geq 0}$. The authors consider the following quantities: \[ \begin{aligned} b_1(N)&=\dim(H_1(N;\mathbb{R})), \\ k(N)&=\dim(\ker (x_N)),\\ r(N)&=\frac{k(N)}{b_1(N)}\ (\text{when}\ b_1(N)>0),\\ \widehat{r}(N)&=\sup_{\widetilde{N}\in \mathcal{C}(N)}\ r(\widetilde{N}),\\ \rho(N)&=\inf_{\widehat{N}\in \mathcal{C}^{\text{sub}}(N)}\ r(\widehat{N}), \\ \widehat{\rho}(N)&=\sup_{\widetilde{N}\in \mathcal{C}(N)}\ \rho(\widetilde{N}). \end{aligned} \] Here $\mathcal{C}(N)$ denotes all finite regular covers $\widetilde{N}$ of $N$, and $\mathcal{C}^{\text{sub}}(N)$ denotes all finite subregular covers $\widehat{N}$ of $N$. \par Then the authors prove the following trichotomy of an aspherical $3$-manifold $N$ with empty or toroidal boundary: \begin{itemize} \item[1.] $N$ is a graph manifold if and only if $\widehat{\rho}(N)=1$ (Seifert manifolds and Sol $3$-manifolds are also counted as graph manifolds). \item[2.] $N$ is a mixed manifold if and only if $\widehat{r}(N)>\widehat{\rho}(N)=0$. \item[3.] $N$ is a hyperbolic manifold if and only if $\widehat{r}(N)=0$. \end{itemize} Note that by definition, $1\geq \widehat{r}(N)\geq \widehat{\rho}(N)\geq 0$ always holds. \par For a hyperbolic $3$-manifold $N$, $\widehat{r}(N)=0$ directly follows from that all $\mathbb{Z}^2$ subgroups of $\pi_1(N)$ are peripheral. The graph and mixed manifold cases require more work and they rely on the following two statements. \begin{itemize} \item[1.] For a graph manifold $N$ and any $\epsilon>0$, there exists a finite regular cover $\widehat{N}$ of $N$ such that any finite cover $\overline{N}$ of $\widehat{N}$ satisfies $r(\overline{N})>1-\epsilon$. \item[2.] For a mixed manifold $N$ and any $\epsilon>0$, there exists a finite subregular cover $\overline{N}$ of $N$ such that $r(\overline{N})<\epsilon$. \end{itemize} In both the graph manifold and the mixed manifold case, it is important to analyse some quantity for their Seifert (for graph manifold) and hyperbolic (for mixed manifold) pieces. For a geometric piece $N_v$ of $N$, the important quantity is \[ c(N_v)=\dim(\operatorname{coker}(H_1(\partial N_v;\mathbb{R})\to H_1(N_v;\mathbb{R}))). \] Making $c(N_v)$ big is always a desired condition to prove the above two results. The difference is that, in the Seifert case, $c(N_v)$ is big implies that many cohomology classes supported in the interior of $N_v$ have vanishing Thurston seminorms; while in the hyperbolic case, $c(N_v)$ is big implies that many cohomology classes supported in the interior of $N_v$ have positive Thurston seminorms. The Seifert case follows from a classical topological construction of finite cover; while the hyperbolic case follows from the deep cube complex machinery, in particular the largeness of hyperbolic $3$-manifold groups.
0 references
3-manifolds
0 references
Thurston seminorm
0 references
covering spaces
0 references
geometric structures
0 references