Degenerating the complex hyperbolic ideal triangle groups (Q5936375)

This paper describes a concrete example of a complex hyperbolic Kleinian group, generated by inversions in three mutually asymptotic complex geodesics in complex hyperbolic 2-space \(\mathbb C \mathbb {H}^2\). A triple of mutually asymptotic complex geodesics is parametrized by three points on the boundary \(S^3\asymp \mathbb C \mathbb H^2\), and up to congruence by a real parameter, Cartan's angular invariant \(s\in[-\pi,\pi]\). Such a triple defines a representation \(\varphi_s\) of the free product \(\mathbb Z/2\star\mathbb Z/2\star\mathbb Z/2\) into the group \(\mathbb{PU} (2,1)\) of biholomorphic isometries of \(\mathbb C \mathbb H^2\). When \(s=0\), the three complex geodesics are orthogonal to a totally real totally geodesic \(2\)-plane, and \(\varphi_0\) is the embedding of an ideal triangle group in \(\mathbb{SO}(2,1) \hookrightarrow\mathbb{PU} (2,1)\) stabilizing the real hyperbolic plane \(\mathbb R \mathbb H^2\subset\mathbb C \mathbb H^2\). In [J. Reine Angew. Math. 425, 71-86 (1992; Zbl 0739.53055)] the reviewer and \textit{J. R. Parker} showed that \(\varphi_s\) is not a discrete embedding for \[ |s|> s_0 := \tan^{-1}\sqrt{125/3} \] and conjectured that this necessary condition is also sufficient. In [Ann. Math. (2) 153, 533-598 (2001)] the author proved, using an involved numerical argument, that all such \(\varphi_s\) are discrete embeddings. The present paper gives a direct construction of a fundamental domain for the limiting case \(\varphi_{s_0}\), and hence an alternate proof of discreteness in this case. Furthermore the Kleinian 3-manifold (the quotient of the domain of discontinuity in \(S^3\) of the image \(\Gamma\) of \(\varphi_{s_0}\)) is identified as a 3-manifold commensurable with the complement of the Whitehead link. The author gives an explicit geometric construction of this manifold, relating the combinatorics of the fundamental domain for the real hyperbolic structure on the Whitehead link complement to a polyhedron in \(S^3\). The author also shows that the limit sets of the groups \(\Gamma_s\) for \(s<s_0\) are topological circles, and that there exists \(\delta>0\) such that the actions on \(S^3\) determined by \(\varphi_s\) are all topologically conjugate for \(s_0-\delta<s< s_0\). Furthermore the Kleinian 3-manifold is doubly covered by a product \(S^1\times S_3^2\), where \(S_3^2\) is the thrice-punctured sphere. More recently, the author has constructed (by similar methods) a complex hyperbolic Kleinian group whose Kleinian 3-manifold is homeomorphic to a closed real hyperbolic 3-manifold [Real hyperbolic on the outside, complex hyperbolic on the inside, Invent. Math. (to appear)]. In a recent preprint [A circle of modular groups] \textit{F. Falbel} and \textit{J. R. Parker} show that \(\Gamma\) is also commensurable to a limit of discrete groups generated by reflections in totally real subspaces.