A sharp cusp count for complex hyperbolic surfaces and related results (Q403039)

Let \(H^{2}\) be the two-dimensional complex hyperbolic space. Given a torsion-free non-uniform lattice \( \Gamma\) in PU(1,2), then \(H^{2}/{\Gamma}\) is a non-compact finite volume complex hyperbolic surface with cuspidal ends. It is known that if the parabolic elements in \(\Gamma\) have no rotational parts, then \(H^{2}/{\Gamma}\) admits a compactification \((X, D)\) consisting of a smooth projective surface \(X\) and an exceptional divisor \(D\). This pair \((X, D)\) is known as the toroidal compactification of \(H^{2}/{\Gamma}\). In a previous paper [``Effective results for complex hyperbolic manifolds'', \url{arXiv:1212.0501}] the authors showed that in any dimension, the log-canonical divisor \(K_{X} + D\) is not only big and nef but it is a limit of ample divisors of the form \(K_{X} + \alpha D,\) where \(\alpha <1\). The purpose of the paper under review is to improve the results in their previous paper above and to give a sharp cusp count in complex dimension 2. The main result is the following: Let \(k\) be the number of components of \(D\). Then \(k \leq (4/3)(K_{X}+D)^{2}\). Moreover, there exists a smooth toroidal compactification which saturates this bound. The authors use this result to study in details the geometry of cusped complex hyperbolic surfaces and their compactifications. As one of the applications, they prove that \(k + 1\leq \rho(X) \leq 4 + (13/3)(K_{X}+D)^{2}\), where \(\rho(X)\) is the Picard number of \(X\). Then, they show that there exists a smooth toroidal compactification which saturates the lower bound.