Free Seifert pieces of pseudo-Anosov flows (Q2048409)

This article explores the interactions between pseudo-Anosov flows and the topology of 3-manifolds. In three dimensions, pseudo-Anosov flows were introduced by \textit{W. P. Thurston} in the 1980's [Bull. Am. Math. Soc., New Ser. 19, No. 2, 417--431 (1988; Zbl 0674.57008)]. \textit{L. Mosher} then generalized this notion in [Duke Math. J. 65, No. 3, 449--500 (1992; Zbl 0754.58030)] and defined general pseudo-Anosov flows in 3-manifolds: they are locally like the suspension flows of pseudo-Anosov homeomorphisms. This class of flows appears, for example, in the proof of the hyperbolization theorem of manifolds fibered over \(\mathbb{S}^1.\) Their existence implies, among other things, that the manifold is irreducible, but not necessarily \textit{atoroidal}; atoroidal means that it contains no embedded incompressible torus. In this paper a number of references and examples on Anosov flows restricted to Seifert-fibered spaces (i.e., admitting a 1-dim foliation by circles), and on \(\mathbb{R}\)-covered Anosov flows (i.e., the universal cover of either the stable or the unstable foliation is homeomorphic to \(\mathbb{R}\)), are mentioned in the introduction and scattered here and there at times. The core purpose of the manuscript is the treatment of the much more general class of pseudo-Anosov flows and their connection with the \textit{torus decomposition} (or JSJ decomposition, due to \textit{W. Jaco} and \textit{P. B. Shalen} [in: Geometric topology, Proc. Conf., Athens/Ga. 1977, 91--99 (1979; Zbl 0471.57001)] and \textit{K. Johannson} [Homotopy equivalences of 3-manifolds with boundaries. Berlin-Heidelberg-New York: Springer-Verlag (1979; Zbl 0412.57007)]) into Seifert-fibered pieces of a compact, irreducible and toroidal 3-manifold. The article, however, does not address the open problem about the characterization of those 3-manifolds that admit pseudo-Anosov flows. Rather, the central question is whether the pieces are \textit{periodic} (i.e., the regular fibers are homotopic to periodic orbits, up to finite order), or otherwise \textit{free} (non-periodic). In a previous publication the authors described the structure of pseudo-Anosov flows restricted to a periodic piece, up to isotopy [Geom. Topol. 17, No. 4, 1877--1954 (2013; Zbl 1317.37038)]. The existence of nontrivial free Seifert pieces that are orbitally equivalent to a geodesic flow piece (with countably many periodic orbits, and uncountably many nonperiodic orbits entirely contained in the piece) was also known. Both the Bonatti-Langevin flows, and a class of manifolds for which all the pieces of the JSJ decomposition are Seifert-fibered (graph manifolds) and periodic, are cited as being examples of the former category (periodic). The case of geodesic flows and the \(\mathbb{R}\)-covering of the Handel-Thurston flows constitute examples of the latter (free). The article is strongly influenced by the works of Calegari, Mosher, Thurston and others. \textit{T. Barbot} [Ann. Inst. Fourier 46, No. 5, 1451--1517 (1996; Zbl 0861.58028)] had already determined that if the flow is \(\mathbb{R}\)-covered, then its restriction to the Seifert piece in an appropriate neighborhood, is orbitally equivalent to a \textit{geodesic flow piece} (i.e., a finite cover of the geodesic flow of a compact hyperbolic surface, typically with boundary). In this article the question whether the flow in a free Seifert piece is orbitally equivalent to a geodesic flow piece, is closely examined. The exposition is clear and enriched by a number of examples and constructions of pseudo-Anosov flows via Dehn surgery. The examples show how different adjoining pieces of the JSJ decomposition can behave with respect to a pseudo-Anosov flow in several categories: free-to-free pieces, free-to-periodic pieces, free-to-atoroidal, atoroidal-to-atoroidal, periodic-to-atoroidal, periodic-to-periodic, and the case of adjoining two transverse tori. The paper presents quite a few remarkable novelties in the proof of the main theorem as well. The use of almost \(k\)-convergence groups, the concept of hyperbolic blow-ups of actions and the alternative model of the geodesic flow on a hyperbolic surface are important innovations in the literature that may be helpful in other contexts.