Gromov hyperbolicity of Denjoy domains (Q861684)

As explained in the introduction of this paper by the authors, this article is part of the study on how to decide whether a given geometric space is hyperbolic in Gromov's sense, i.e., whether there is an upper bound of the distance of every point in a side of any geodesic triangle to the union of the other two sides (Rips condition). The special case of a Riemann surface endowed with the Poincaré metric gives a negatively curved space, but not all are Gromov hyperbolic. In this article the authors study conditions implying that that such Riemann surfaces are in fact Gromov hyperbolic. In order to do it they study the hyperbolicity of Denjoy domains, i.e., a domain \(\Omega\) of the Riemann surface such that \(\partial\Omega\subset\mathbb{R}\cup\{\infty\}\). The authors obtain the result that a Denjoy domain is hyperbolic if and only if the distance to \(\mathbb{R}\) of any point in any simple closed geodesic is uniformly bounded. They actually obtain characterizations for a Schottky double of a Riemann surface, and the double of a metric space. Some of these characterizations in the authors own words are particularly surprising: It is sufficient to check Rips condition only on geodesic ``bigons'' (triangles with two vertices). This is not the case in general. Since Denjoy domains are a wide class of Riemann surfaces and the above condition is not easily verified the authors concentrate on trains and generalized trains. In both cases they obtain characterizations and sufficient conditions that enable one to guarantee or discard hyperbolicity. In trains the hyperbolicity is studied in terms of lengths of two types of their simple closed geodesics, named fundamental, whose lengths are denoted by \({l_n}\) and \({r_n}\). Most of the hyperbolicity criteria will depend on the fundamental geodesics through their lengths \({l_n}\) and \({r_n}\). Example: If \({l_n}\) is bounded, the train is always hyperbolic; if \({l_n}\) is not bounded one needs to impose \({r_n}\) to be bounded to guarantee hyperbolicity. The article is very complete in the information provided to understand the new results. In fact there is a short but complete, for the main goals of the article, introduction to Gromov and metric spaces, and Riemann surfaces before going into the details of the new theorems.