scientific article; zbMATH DE number 1789976

From MaRDI portal
Publication:4549288





This paper surveys recent examples of discrete groups \(\Gamma\subset\text{PU}(2,1)\) acting on the complex hyperbolic 2-space \(\text{H}_{\mathbb C^2}\) whose quotient spaces \(W = \text{H}_{\mathbb C^2}t/\Gamma\) has infinite volume. (Highly nontrivial examples of lattices have previously been constructed by Mostow, Deligne-Mostow, Hirzebruch, Livne, and several others. The groups described here are much different.) The most spectacular examples are groups \(\Gamma\) such that \(W\) is the interior of a compact manifold-with-boundary, whose boundary is a hyperbolic 3-manifold. Here \(\partial W\) is the quotient of a domain \(\Omega\subset\partial{\text{H}}_{\mathbb C}^2\approx S^3\) by \(\Gamma\). The technique for constructing such groups begins with the embedding of a Fuchsian subgroup \(\Gamma_0\hookrightarrow\text{SO}(2,1)\), and deforms the representation inside \(\text{Hom}(\Gamma,\text{PU}(2,1))\). Thus \(\Gamma_0\) preserves the hyperbolic plane \(\text{H}_{{\mathbb R}^2}\), but in general the deformed groups will not preserve any totally geodesic plane. The first case studied in detail is when \(\Gamma\) is the group generated by reflections \(I_1,I_2,I_3\) in three mutually asymptotic geodesics in \(\text{H}_{{\mathbb R}^2}\). Conjugacy classes of deformations \(\rho_\theta\) are parametrized by a single real parameter \(\theta\) in the interval \([-\pi,\pi]\), with \(\rho_0\) corresponding to the original representation in \(\text{SO}(2,1)\) and \(\rho\pm\pi\) corresponding to representations onto a group of order two. The reviewer and J. R. Parker proved that \(\rho_\theta\) is a discrete embedding for \(|\theta|\leq \tan^{-1}(\sqrt{35})\) and that \(\rho_\theta\) is not a discrete embedding for \(|\theta|> \tan^{-1}(\sqrt{125/3})\). The latter condition expresses the ellipticity of the product \(I_1I_2I_3\). The author settled the conjecture that \(\rho_\theta\) is a discrete embedding for all\break \(|\theta|\leq \theta_\infty = \tan^{-1}(\sqrt{125/3})\). That is, as long as \(I_1I_2I_3\) is not elliptic, the representation \(\rho\) is a discrete embedding. While all the actions \(\rho_\theta\) are topologically conjugate for \(|\theta|<\theta_0\), the limiting group \(\Gamma_\infty:=\rho_\infty(\Gamma_0)\) is not conjugate. The limit set is no longer a Jordan curve, and its complement, the domain of discontinuity \(\Delta\subset\partial\text{H}_{\mathbb C^2}\) covers a 3-orbifold \(M^3\) with spherical CR-structure. \(M^3\) is commensurable to the complement of the Whitehead link in \(S^3\). Deformations of other triangle groups exhibit different phenomena. Wyss-Gallifent's thesis discusses the \((4,4,\infty)\) triangle groups, and finds isolated points in the deformation space corresponding to non-faithful representations onto discrete subgroups. The author studied \((4,4,4)\)-triangle groups and found a deformation with discrete image \(\Gamma\) such that the quotient 3-orbifold \(M = \Delta/\Gamma\) is closed. Furthermore, \(M\) admits a hyperbolic structure (which does not seem to be closely related to the natural CR-structure on \(M\) arising as the boundary of a complex hyperbolic surface). For the entire collection see [Zbl 0993.00022].




Cited in
(47)








This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4549288)