Representations of fundamental groups of 3-manifolds into \(\mathrm{PGL}(3,\mathbb C)\): exact computations in low complexity (Q2516428)

Thurston started to work on hyperbolic structures on \(3\)-manifolds in connection with the \textit{Geometrization Conjecture} in 1970's. Then \textit{geometric structures} (but not only hyperbolic ones) on a smooth \(n\,(\geq 2)\)-manifold came to our interest around that period, especially conformally flat structures, spherical \(CR\)-structures, real projective structures, etc. The present paper is concerned with the existence of geometric structures on \(3\)-manifolds, namely a hyperbolic structure \((\text{PSL}(2,{\mathbb C}),{\mathbf H}^3_{\mathbb R})\), a spherical \(CR\)-structure \((\text{PU}(2,1),S^3)\), and a flag structure \((\text{PGL}(3,{\mathbb R}),\text{PSL}(3,{\mathbb R})/{\mathbf B})\). Let \((G,X)\) be one of the above geometries. Then \(G\) is a subgroup of \(\text{PGL}(3,{\mathbb C})\). To study a geometric structure on a manifold \(M\) one may consider discrete faithful representations \(\rho: \pi_1(M)\rightarrow G\) combined with developing immersions. The authors consider the problem whether there is a representation of the fundamental group of complete volume finite hyperbolic manifolds into \(G\). R. Schwartz considered the problem first whether a hyperbolic manifold can be realized as a spherical \(CR\)-manifold, that is, a \((\text{PU}(2,1),S^3)\)-structure. He showed that \textit{There exists a discrete subgroup \(\Gamma\leq\text{PU}(2,1)\) such that the spherical \(CR\) quotient \(\Gamma\backslash S^3 - \Lambda\) admits a hyperbolic structure.} Here \(\Lambda\) is the limit set of \(\Gamma\). \(\Gamma\) is a torsionfree finite index subgroup of a group generated by three complex reflections. In this direction but close to Thurston's method, E. Falbel together with co-authors has shown that the figure-eight knot complement \(M\) admits a spherical \(CR\)-structure in a series of papers [\textit{M. Deraux} and \textit{E. Falbel}, Geom. Topol. 19, No. 1, 237--293 (2015; Zbl 1335.32028), \textit{N. Bergeron} et al., Geom. Topol. 18, No. 4, 1911--1971 (2014; Zbl 1365.57023), \textit{E. Falbel}, J. Differ. Geom. 79, No. 1, 69--110 (2008; Zbl 1148.57025) and references therein]. \(M\) is obtained from two tetrahedra by gluing corresponding faces. Each \(4\)-tuple of points in \(S^3\) is assigned to a vertex of a tetrahedron of \(S^3\), each edge is assigned to a segment of either \({\mathbb R}\)-circles or \({\mathbb C}\)-circles, and faces are also assigned (but a bit complicated). Once two tetrahedra are realized in \(S^3\), glue them by the appropriate \(CR\)-diffeomorphisms whose quotient provides a spherical \(CR\)-manifold homeomorphic to \(M\). More precisely, there is a discrete representation \(\rho:\pi_1(M)\rightarrow \text{PU}(2,1)\) such that \(\rho(\pi_1(M))\backslash S^3-\Lambda\) is homeomorphic to \(M\). The difficulty to construct is the consistency relations: edge equations, face equations, cross-ratio equations and holonomy equations (which produce a Dirichlet domain for \(\Gamma\)). The authors of the present paper do the same game to the above three geometric structures following the computation program: {\parindent=6mm \begin{itemize} \item[-] Solving the constraints on gluing decorated tetrahedra such as the above equations. \item [-] Distinguishing hyperbolic, real flags and \(CR\) solutions. \item [-] Computing the resulting volumes. \item [-] Computing representations of the fundamental group. \end{itemize}} They produce more examples of \(3\)-dimensional spherical \(CR\)-manifolds than those obtained in the previous papers. The results of real flags modeled on \((\text{SL}(3,{\mathbb R}), \text{SL}(3,{\mathbb R})/{\mathbf B})\) seem to be sporadic but a new challenge.