Counting immersed surfaces in hyperbolic 3-manifolds (Q2571368)

Among the closed hyperbolic \(3\)-manifolds are many examples containing no \(2\)-sided imbedded incompressible surface. On the other hand, it is conjectured that every such manifold contains immersed incompressible surfaces, each of which lifts to an imbedding in a finite-sheeted cover. Various constructions are known which produce such immersions in classes of hyperbolic \(3\)-manifolds. In this paper, the author gives a general upper bound for the number of such immersions, and a new construction which gives a lower bound for many cases. For a closed hyperbolic \(3\)-manifold \(M\), let \(s_2(M,g)\) denote the number of conjugacy classes of surface subgroups of genus at most \(g\) in \(\pi_1(M)\). W. Thurston proved that \(s_2(M,g)\) is finite, and \textit{T. Soma} [Topology Appl. 41, No. 3, 179--192 (1991; Zbl 0753.57011)] gave an upper bound for it which is doubly exponential in the genus. The first main result of the paper shows that there is a positive constant \(c_2\), depending on \(M\), such that \(s_2(M,g)<e^{c_2g\log g}\). The idea of the proof is first to move the surface by homotopy into pleated form, then find a triangulation for the surface whose edges are sufficiently short, then count the possible graphs which arise as \(1\)-skeleta. The second main result of this paper gives a general construction that produces many distinct immersed incompressible surfaces in any closed hyperbolic \(3\)-manifold \(M\) that already contains a transverse, totally geodesic immersion of some hyperbolic surface. More precisely, let \(s_1(M,g)\) be the number of surface subgroups up to conjugacy and commensurability (i.~e.~having a common subgroup of finite index), and \(s(M,g)\) the number of maximal surface subgroups up to conjugacy. Provided that \(M\) contains one totally geodesic immersion, there exists a positive constant \(c_1\) such that \(e^{c_1 g \log g} < s_1(M,g)\). Since \( s_1(M,g) \leq s(M,g) \leq s_2(M,g)\), this combines with the first main result to trap all three of these counts between bounds of the form \(e^{cglog g}\). The proof of the latter result is by explicit construction of immersed incompressible surfaces. The main difficulty is to show that the resulting surface groups are inequivalent, which is accomplished by proving that they determine hyperbolic manifold coverings of \(M\) with non-isometric convex cores. In fact, the proof shows that the surface subgroups are not even conjugate or commensurable as subgroups of \(\text{PSL}(2,\mathbb{C})\). The author also gives more explicit, combinatorial proofs of both bounds for the case when \(M\) is an orbifold produced by reflection in the sides of a right-angled hyperbolic polyhedron \(P\). For this case, he obtains the explicit values \(c_1=1\) and \(c_2=8\,c(P) + 1\), where \(c(P)\) is an integer depending on the combinatorial structure of \(P\). The author notes that his construction produces only quasifuchsian subgroups, and that the other known constructions seem to produce geometrically infinite subgroups only ``sporadically,'' so very little is known about the number of conjugacy classes of maximal, geometrically infinite subgroups.