Dehn filling of cusped hyperbolic \(3\)-manifolds with geodesic boundary. (Q1778012)

Let \(\Delta\) denote the standard tetrahedron, and let \(\widehat\Delta\) be \(\Delta\) with its vertices removed. An ideal triangulation of a compact connected orientable 3-manifold \(M\) with boundary is a realization of the interior of \(M\) as a gluing of a finite number of copies of \(\widehat\Delta\) induced by a simplicial face-pairing of the corresponding \(\Delta\)'s. Let \(\Sigma_g\) be the closed orientable surface of genus \(g\). For each \(g\geq 2\) and \(k\geq 0\) the authors define a class \({\mathcal M}_{g,k}\) of compact connected orientable 3-manifolds having an ideal triangulation with \(g+ k\) tetrahedra, and with \(k\) toric cusps and a connected totally geodesic boundary \(\Sigma_g\) of genus \(g\). Manifolds in \({\mathcal M}_{g,k}\) have Matveev complexity \(g+ k\) and Heegaard genus \(g+ 1\). The authors prove that such manifolds are hyperbolic, and show that their homology, volume and Turaev-Viro invariants depend only on \(g\) and \(k\). In particular, if \(M\in{\mathcal M}_{g,k}\), then \(H_1(M;\mathbb{Z})\cong \mathbb{Z}^{g+k}\). It follows from [\textit{Y. Miyamoto}, Topology 33, No. 4, 613--629 (1994; Zbl 0824.53038)] that \({\mathcal M}_{g,0}\) is precisely the class of hyperbolic 3-manifolds \(M\) having minimal volume among those with \(\partial M= \Sigma_g\). In addition, every manifold \(M\) in \({\mathcal M}_{g,k}\) does not contain closed essential surfaces: in fact, every essential surface in \(M\) has nonempty boundary which intersects \(\Sigma_g\) (see Theorem 1.2 of the paper). The set \({\mathcal M}_{g,k}\) is nonempty precisely for \(g> k\) or \(g= k\) and \(g\) even. The cardinality of \({\mathcal M}_{g,k}\) for a fixed \(k\) has growth type \(g^g\). The authors completely describe the non-hyperbolic Dehn fillings of each \(M\) in \({\mathcal M}_{g,k}\), and show that, on any cusp of every hyperbolic 3-manifold obtained by partially filling \(M\) there are precisely six non-hyperbolic Dehn fillings: three of them contain essential discs and the others contain essential annuli. To state explicitly the result we need some notations. A slope in a torus \(T\) is an isotopy class of simple closed essential curves; it is represented by a number in \(\mathbb{Q}\cup\{\infty\}\) whenever a \(\mathbb{Z}\)-basis is fixed in \(H_1(T; \mathbb{Z})\). If \(M\) is a compact connected orientable 3-manifold with \(k\) boundary tori, and \(\alpha_1,\dots, \alpha_h\) are slopes in some \(h\leq k\) of these tori, let \(M(\alpha_1,\dots,\alpha_h)\) denote the result of Dehn filling these tori along \(\alpha_1,\dots, \alpha_h\). For \(g\geq 2\) and \(0\leq k\leq g\) let \(H_{g,k}\) be the compact 3-manifold obtained from the orientable handlebody of genus \(g\) by drilling out \(k\) tunnels along \(k\) different 1-handles. Theorem 1.7 of the paper says that if \(M\in{\mathcal M}_{g,k}\) with boundary \(\partial M\) homeomorphic to a disjoint union of \(\Sigma_g\) and \(k\) tori \(T_i\), \(i= 1,\dots,k\), there exists a \(\mathbb{Z}\)-basis of \(H_1(\bigcup^k_{i=1} T_i;\mathbb{Z})\) such that \(N= M(\alpha_1,\dots,\alpha_h)\) is as follows: 1) If \(\alpha_i\in\in\{0,1,\infty\}\) for some \(i\), then \(N= H_{g,k-h}\), and it is not hyperbolic; 2) If \(\alpha_i\in\{-1,{1\over 2},2\}\) for some \(i\), then \(N\) contains a Möbius strip or non-separating annulus \(R\) with \(\partial R\subset\Sigma_g\), and cutting \(N\) along \(R\) gives \(H_{g,k-h}\). Again \(N\) is not hyperbolic; 3) If \(\alpha_i\not\in\{-1,0,{1\over 2},1,2,\infty\}\) for all \(i\), then \(N\) is hyperbolic and has Heegaard genus \(g+ 1\); 4) If \(\alpha_i\in\{-2, -{1\over 2}, {1\over 3}, {2\over 3}, {3\over 2}, 3\}\) for all \(i\), then \(N\) belongs to \(M_{g,k-h}\). Moreover, every essential surface in \(N\) has nonempty boundary which intersects \(\Sigma_g\). The theorem gives an infinite class of large hyperbolic manifolds (in the sense of \textit{Y. Q. Wu} [J. Differ. Geom. 48, No. 3, 407--437 (1998; Zbl 0917.57015)] with \(\partial\)-reducible and annular Dehn fillings having distance 2, and allows to prove that the corresponding upper bound found by Wu is sharp. This answers a question raised by Wu in the quoted paper, concerning the distance between non-hyperbolic fillings of a large 3-manifold.