Abstract
We consider a compact orientable hyperbolic 3-manifold with a compressible boundary. Suppose that we are given a sequence of geometrically finite hyperbolic metrics whose conformal boundary structures at infinity diverge to a projective lamination. We prove that if this limit projective lamination is doubly incompressible, then the sequence has compact closure in the deformation space. As a consequence we generalise Thurston’s double limit theorem and solve his conjecture on convergence of function groups affirmatively.
Similar content being viewed by others
References
Agol, I.: Tameness of hyperbolic 3-manifolds. arXiv preprint arXiv:math.GT/0405568
Bers, L.: On boundaries of Teichmüller spaces and on Kleinian groups: I. Ann. Math. 91(2), 570–600 (1970)
Bestvina, M.: Degenerations of the hyperbolic space. Duke Math. J. 56, 143–161 (1988)
Bonahon, F.: Bouts des variétés hyperboliques de dimension \(3\). Ann. Math. 124(2), 71–158 (1986)
Bonahon, F., Otal, J.-P.: Laminations mesurées de plissage des variétés hyperboliques de dimension 3. Ann. Math. 160(2), 1013–1055 (2004)
Bridgeman, M., Canary, R.D.: From the boundary of the convex core to the conformal boundary. Geom. Dedicata 96, 211–240 (2003)
Bromberg, K.: Projective structures with degenerate holonomy and the Bers’ density conjecture. Ann. Math. 166, 77–93 (2007)
Brock, J., Bromberg, K.: On the density of geometrically finite hyperbolic 3-manifolds. Acta Math. 192, 33–93 (2004)
Bromberg, K., Souto, J.: The density conjecture: a prehistoric approach (in preparation)
Brock, J., Canary, R., Minsky, Y.: The classification of Kleinian surface groups, II: the ending lamination conjecture. Ann. Math. 176, 1–149 (2012)
Calegari, D., Gabai, D.: Shrinkwrapping and the taming of hyperbolic 3-manifolds. J. Am. Math. Soc. 19, 385–446 (2006)
Canary, R.: Algebraic convergence of Schottky groups. Trans. Am. math. Soc. 337, 235–258 (1993)
Canary, R.D., Epstein, D.B.A., Green, P.: Notes on notes of Thurston. In: Analytical and geometric aspects of hyperbolic space (Coventry/Durham, 1984), London Math. Soc. Lecture Note Ser, vol.111, pp. 3–92. Cambridge Univ. Press, Cambridge (1987)
Culler, M., Morgan, J.: Group actions on R-trees. Proc. London Math. Soc. 55(3), 571–604 (1987)
Fathi, A., Laudenbach, F., Ponéaru, V.: Travaux de Thurston sur les surfaces, pp. 66–67. Séminaire Orsay, Astérisque No (1979)
Floyd, W.J.: Group completions and limit sets of kleinian groups. Invent. Math. 57, 205–218 (1980)
Jaco, W., Shalen, P.B.: Seifert fibered spaces in 3-manifolds. Memoirs Am. Math. Soc. 21(220), pp. viii+192 (1979)
Johannson, K.: Homotopy equivalence of 3-manifolds with boundary. Lect. Notes Math. 761, (1979)
Hempel, J.: 3-manifolds. Ann. Math. Stud. 86, pp. xii+195. Princeton Univ. Press/University of Tokyo Press, Princeton, N. J/Tokyo (1976)
Kapovich, M.: Hyperbolic manifolds and discrete groups. Progress Math. 183, pp. xxvi+467. Birkhäuser Boston Inc., Boston, MA (2001)
Kapovich, M., Leeb, B.: On asymptotic cones and quasi-isometry classes of fundamental groups of 3-manifolds. Geom. Funct. Anal. 5, 582–603 (1995)
Kleineidam, G., Souto, J.: Algebraic convergence of function groups. Comment. Math. Helv. 77, 244–269 (2002)
Lecuire, C.: Plissage des variété hyperboliques de dimension 3. Invent. Math. 164(1), 85–141 (2006)
Lecuire, C.: An extension of Masur domain, Spaces of Kleinian groups. London Math. Soc. Lec. Notes 329, 49–73 (2006)
Masur, H.: Measured foliations and handlebodies. Ergodic Theory Dynam. Syst. 6, 99–116 (1986)
Minsky, Y.: The classification of Kleinian surface groups, I: models and bounds. Ann. Math. 171(2), 1–107 (2010)
Morgan, J.W., Otal, J.-P.: Relative growth rates of closed geodesics on surfaces under varying hyperbolic structures. Comment. Math. Helv. 68, 171–208 (1993)
Morgan, J.W., Shalen, P.B.: Degenerations of hyperbolic structures I: valuations, trees and surfaces. Ann. Math. 120, 401–476 (1984)
Morgan, J.W., Shalen, P.B.: Degenerations of Hyperbolic structures III: actions of 3-manifolds groups on trees and Thurston’s compactness theorem. Ann. Math. 127, 457–519 (1988)
Namazi, H., Souto, J.: Non-realizability and ending laminations: proof of the density conjecture. Acta Math. 209, 323–395 (2012)
Ohshika, K.: On limits of quasi-conformal deformations of Kleinian groups. Math. Z. 201, 167–176 (1989)
Ohshika, K.: Limits of geometrically tame Kleinian groups. Invent. Math. 99, 185–203 (1990)
Ohshika, K.: Ending laminations and boundaries for deformation spaces of Kleinian groups. J. London Math. Soc. 42(2), 111–121 (1990)
Ohshika, K.: A convergence theorem for Kleinian groups which are free products. Math. Ann. 309(1), 53–70 (1997)
Ohshika, K.: Realising end invariants by limits of minimally parabolic, geometrically finite groups. Geom. Topol. 15, 827–890 (2011)
Otal, J.-P.: Courants géodésiques et produits libres. Université Paris-Sud, Orsay, Thèse d’Etat (1988)
Otal, J.-P.: Sur la dégénérescence des groupes de Schottky. Duke Math. J. 74, 777–792 (1994)
Otal, J.-P.: Le théorème d’hyperbolisation pour les variétés fibrées de dimension 3. Astérisque 235, pp. x+159 (1996)
Paulin, F.: Topologie de Gromov équivariante, structures hyperboliques et arbre réels. Invent. Math. 94, 53–80 (1988)
Penner, R.C., Harer, J.L.: Combinatorics of train tracks. Ann. Math. Stud. 125, pp. xii+216. Princeton University Press, Princeton, NJ, (1992)
Skora, R.K.: Splitting of surfaces. J. Am. Math. Soc. 9, 605–616 (1996)
Thurston, W.P.: Three-dimensional manifolds, Kleinian groups and hyperbolic geometry. Bull. Am. Math. Soc. (N.S.) 6, 357–381 (1982)
Thurston, W.P.: Hyperbolic structures on \(3\)-manifolds. I: deformation of acylindrical manifolds. Ann. Math. 124, 203–246 (1986)
Thurston, W.P.: Hyperbolic structures on 3-manifolds, II: surface groups and 3-manifolds which fiber over the circle (1998). arXiv:math.GT/9801045
Thurston, W.P.: Hyperbolic structures on 3-manifolds, III: deformations of 3-manifolds with incompressible boundary (1998). arXiv:math.GT/9801058
Acknowledgments
We thank the referees for numerous comments which helped us to make the paper more readable.
Author information
Authors and Affiliations
Corresponding author
Additional information
I. Kim gratefully acknowledges the partial support of Grant (NRF-2014R1A2A2A01005574). C. Lecuire acknowledges the partial support of Grant ANR-12-BS01-0003, GDSous. K. Ohshika acknowledges the partial support of JSPS Grants-in-Aid 22244005.
Appendix
Appendix
For the sake of completeness, we shall give brief proofs of some propositions which we cited in previous sections.
Proposition 2.2
[23], Sect. 2 paragraphs after Lemme 2.7] Let \(\Sigma \) and \(\Sigma '\subset \partial _{\chi <0} M\) be two compact, connected, incompressible surfaces which are disjoint or equal and do not contain any essential closed curve which can be homotoped into \( \partial _{\chi =0} M\). Let \(\tilde{\Sigma }\subset \partial \tilde{M}\) (resp. \(\tilde{\Sigma }'\)) be a connected component of the preimage of \(\Sigma \) (resp. \(\Sigma '\)) and let \(\Gamma \subset \rho (\pi _1(M))\) (resp. \(\Gamma '\)) be the stabiliser of \(\tilde{\Sigma }\) (resp. \(\Gamma '\)).
Then \(\overline{\tilde{\Sigma }} \cap \overline{\tilde{\Sigma }}'\) is either empty or equal to the limit set of \(\Gamma \cap \Gamma '\).
In the latter case, if \(\Gamma \cap \Gamma '\) is not cyclic, then it is the fundamental group of a (possibly twisted) I-bundle which is a connected component of a characteristic submanifold of \((M,\Sigma \cup \Sigma ')\). If \(\Gamma \cap \Gamma '\) is cyclic, then it is a finite index subgroup of a solid torus which is a connected component of a characteristic submanifold of \((M,\Sigma \cup \Sigma ')\).
Proof
Suppose that \(\xi \) belongs to the limit set of \(\Gamma \,\cap \,\Gamma '\). Since we are considering only geometrically finite groups, both \(\Gamma \) and \(\Gamma '\) are convex cocompact. Let l be a geodesic ray from the origin \(O\in {\mathbb H}^3\) to \(\xi \). Let F be a fundamental domain of the convex core of \(\Gamma \) containing O, and \(F'\) a fundamental domain of the convex core of \(\Gamma '\) containing O. Since \(\Gamma \) and \(\Gamma '\) are convex cocompact, the diameters of F and \(F'\) are bounded. Choose \(g_n\in \Gamma , g_n'\in \Gamma '\) such that \(g_n F\cap l\ne \emptyset , g_n'F'\cap l\ne \emptyset \) and \(g_nF\cap g_n'F' (\ne \emptyset ) \longrightarrow \xi \). Since the diameters of F and \(F'\) are bounded, we have
for all n and a fixed K. By discreteness of \(\rho (\pi _1(M))\), after passing to a subsequence, \(g_n^{-1}g_n'=g\) for all n. Then \(g_n^{-1}g_n'=g_m^{-1}g_m'\), i.e., \(g_m g_n^{-1}=g_m'g_n'^{-1}\in \Gamma \cap \Gamma '\). Since \(\Gamma \) and \(\Gamma '\) are convex cocompact, these elements are hyperbolic and the limit set of \(\Gamma \cap \Gamma '\) contains at least two points. Let \(h\in \Gamma \cap \Gamma '\) be a hyperbolic element. Then the invariant geodesics \(\tilde{c}, \tilde{c}'\) of h in \(\tilde{\Sigma },\tilde{\Sigma }'\) descend to closed geodesics c and \(c'\) in \(\Sigma \) and \(\Sigma '\). Hence c and \(c'\) bound an annulus A (not necessarily embedded) which is not homotoped into \(\partial M\). Hence A must be contained in some characteristic submanifold of \((M,\Sigma \cup \Sigma ')\). If \(\Sigma =\Sigma '\), then \(c=c'\) and A is a Möbius band.
Let \(F=\overline{\tilde{\Sigma }} \cap \overline{\tilde{\Sigma }}'\) and \(C,C'\) be the projections of the convex hulls of F in \(\tilde{\Sigma }\) and \(\tilde{\Sigma }'\). Then \(\pi _1(C)=\Gamma \cap \Gamma '\) and \(C\cup C'\) is the boundary of an I-bundle \(C\times I\) which is essential in \((M,\Sigma \cup \Sigma ')\). If \(\Sigma =\Sigma '\), then it is a twisted I-bundle with the boundary C. \(\square \)
Next we summarise some proofs from [22] for the reader’s convenience.
Proposition 8.1
Suppose that the action of \(\pi _1(M)\) on an \({\mathbb R}\)-tree \(\mathcal T\) is minimal and small. Let \(S\subset \partial M\) be a compact compressible surface which has (possibly empty) geodesic boundary with respect to a hyperbolic metric on \(\partial M\). Suppose that \(\phi :\mathcal T_\mu {\rightarrow }\mathcal T\) is a \(\pi _1(M)\)-equivariant morphism which folds only at complementary regions. If \(\mu \subset S\) is in tight position with respect to a meridian \(m\subset S\), then \(|\mu |\) can be extended to a geodesic lamination which contains a homoclinic leaf h in S.
Proof
Fix a hyperbolic metric on \(\partial M\). The universal cover \(\tilde{S}\subset {\mathbb H}^2\) is a convex subset of \({\mathbb H}^2\). Let \(\bar{m}\) represent an element of \(\pi _1(S)\) corresponding to the meridian m, which leaves invariant a lift \(\tilde{m}\subset \tilde{S}\). By equivariance, for \(x\in \tilde{m}\),
since \(\bar{m}\) is trivial in \(\pi _1(M)\), where \(\pi :\tilde{S}{\rightarrow }\mathcal T_\mu \) is an equivariant map. Since \(\phi \) folds \([\pi (x), \pi (\bar{m} x)]\) only finitely many times, one can find segments \(\tilde{I}_1, \tilde{I}_2 \subset \tilde{m}\) such that \(\tilde{I}_1\cap \tilde{I}_2=y\in \tilde{m}\) and \(\phi \) folds \(\pi (\tilde{I}_1)\) and \(\pi (\tilde{I}_2)\) along \(\pi (y)\). For \(x\in m\cap \mu \), let \(\mu _x^+\) denote a half leaf of \(\mu \) starting from x to a chosen positive direction. Then using Skora’s idea, Kleineidam and Souto showed [22] (Proposition 3) that there are \(z_i\in I_i\cap \mu \) such that the lifts of \(\mu _{z_1}^+\) and \(\mu _{z_2}^+\) to \(\partial \tilde{M}\) have the same endpoints. Let C be the complementary region of \(\tilde{\mu }\) in \(\tilde{S}\), which contains the folding point y. Let \(\tilde{\mu }_1,\tilde{\mu }_2\) be boundary leaves of C. Up to reversing the orientation, we can assume that \(\tilde{\mu }_1^+,\tilde{\mu }_2^+\) are not asymptotic in \(\tilde{S}\). Since the lifts of \(\mu _{z_1}^+\) and \(\mu _{z_2}^+\) to \(\partial \tilde{M}\) have the same end points, by shrinking the intervals, we can see that the projection of \(\tilde{\mu }_1^+,\tilde{\mu }_2^+\) to \(\partial \tilde{M}\) have the same endpoint. Let l be the geodesic in \(\tilde{S}\) joining the end points of \(\tilde{\mu }_1^+,\tilde{\mu }_2^+\). The projection of l to S becomes a homoclinic leaf disjoint from \(\mu \). \(\square \)
Proposition 8.2
Let S be a compressible surface in \(\partial M\), which contains a homoclinic leaf h. Then there is a sequence of meridians whose Hausdorff limit does not cross h.
Proof
Since a homoclinic leaf cannot be contained in an incompressible surface by Lemma 2.1, \(S(\bar{h})\) must contain a meridian m.
If h contains infinitely many homotopy classes of m-waves. Then there are \((x_i),(y_i)\subset {\mathbb R}\) such that \(h(x_i),h(y_i)\in m\) and \(h[x_i,y_i]\) are non-homotopic m-waves. Since m is compact, after passing to a subsequence, we may assume that \(h(x_i)\) and \(h(y_i)\) converge. Hence for any \(\epsilon >0\), we can choose i, j such that the lengths of segments \([h(x_i),h(x_j)],[h(y_i),h(y_j)]\subset m\) are less than \(\epsilon \). Then \(h[x_i,y_i]\cup h[x_j,y_j]\cup [h(x_i),h(x_j)]\cup [h(y_i),h(y_j)]\) is a meridian whose geodesic representative lies nearby the homoclinic leaf.
If h contains only finitely many homotopy classes of m-waves. then there is a meridian m and two disjoint half-leaves \(h^+\) and \(h^-\) of h such that \(h^+\) and \(h^-\) are in tight position with respect to m. Considering the intersections of m and \(h^+\) and \(h^-\) respectively, one obtains a picture similar to . 8, namely there is an arc \(k\subset h^+\) and an arc \(k'\subset h^-\) which nearly bounds an annulus and a wave between k and \(k'\). These arcs can be used to construct a sequence of meridians whose Hausdorff limit does not cross h as explained in the proof of [22, Proposition 1]. \(\square \)
With more work, one could probably prove that \(\bar{h}\) is a Hausdorff limit of meridians. On the other hand, in all the situations we have used Proposition 8.2, with only little changes, we could have replaced it with the following weaker result whose proof is easier.
Lemma 8.3
Let S be a compressible surface which contains a homoclinic leaf h, and let \(\beta \) be a measured lamination which does not cross h. Then there is a sequence of meridians whose Hausdorff limit does not cross \(\beta \).
Proof
Using cut-and-paste operation as in Lemma 2.3, we construct a sequence of meridians \(m_i\) such that \(i(m_i,\beta )\longrightarrow 0\). Start with a meridian \(m\subset S(\bar{h})\). Since h is homoclinic, it contains an m-wave. Using this wave as in the proof of Lemma 2.3, we get a meridian \(m_1\) such that \(i(m_1,\beta )\le \frac{1}{2} i(m,\beta )\). Then we do the same again on \(m_1\). Repeating this, we get a sequence of meridians \(m_i\) such that \(i(m_i,\beta )\le \frac{1}{2^i} i(m,\beta )\). \(\square \)
Rights and permissions
About this article
Cite this article
Kim, I., Lecuire, C. & Ohshika, K. Convergence of freely decomposable Kleinian groups. Invent. math. 204, 83–131 (2016). https://doi.org/10.1007/s00222-015-0609-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00222-015-0609-5