Pseudo-Anosov flows in toroidal manifolds (Q357731)

The authors prove two significant rigidity results for pseudo-Anosov flows in toroidal manifolds; one for Seifert fibered 3-manifolds and one for 3-manifolds with a virtually solvable fundamental group. They also present many new examples of pseudo-Anosov flows and generalized pseudo-Anosov flows with one-prong singularities, and they study the optimal position of immersed and embedded incompressible tori with respect to arbitrary pseudo-Anosov flows. \smallpagebreak The authors' first main strong rigidity theorem is the following: \smallpagebreak { Theorem 4.1.} Let \(\Phi\) be a pseudo-Anosov flow in a Seifert fibered 3-manifold. Then up to finite covers, \(\Phi\) is topologically equivalent to a geodesic flow in the unit tangent bundle of a hyperbolic surface. \smallpagebreak The second strong rigidity theorem is as follows: \smallpagebreak { Theorem 5.7.} Suppose that \(\Phi\) is a pseudo-Anosov flow in a 3-manifold with virtually solvable fundamental group. Then \(\Phi\) has no singularities and is topologically equivalent to a suspension Anosov flow. \smallpagebreak The first rigidity theorem is proved by showing that a fiber cannot be homotopic to a closed orbit of the flow and showing that when no fiber is homotopic to a closed orbit, then there are no singularities and the stable/unstable foliations are slitherings. The second theorem is proved by first assuming that the fundamental group is solvable and has a normal rank-two abelian subgroup that acts nonfreely on the orbit space. In this case they prove that the action of the fundamental group preserves a structure in the universal cover they call a chain of lozenges. This leads to a contradiction, from which they conclude that the rank-two abelian subgroups act freely on the orbit space, and Theorem 5.7 then follows from earlier results in a paper by the second author [Geom. Dedicata 99, 61--102 (2003; Zbl 1047.37017)] and some additional arguments. \smallpagebreak The authors' main theorem on embedded incompressible tori is the following: \smallpagebreak { Theorem 6.10.} Suppose that \(M\) is orientable and that the pseudo-Anosov flow is not topologically equivalent to a suspension Anosov flow. Let \(T\) be an embedded, incompressible torus in \(M\). Then either {\parindent=6mm \begin{itemize} \item[(1)] \(T\) is isotopic to an embedded Birkhoff torus, \item [(2)] \(T\) is homotopic to a weakly embedded Birkhoff torus \(T'\) and \(T\) (or \(T'\)) is contained in a periodic Seifert fibered piece, or \item [(3)] \(T\) is isotopic to the boundary of the tubular neighborhood of an embedded Birkhoff-Klein bottle contained in a free Seifert piece. \end{itemize}} \smallpagebreak As a consequence of this theorem the authors prove the following result. \smallpagebreak { Proposition 6.9.} Let \(\alpha\) be a singular orbit of a pseudo-Anosov flow. Then \(\alpha\) is homotopic into a piece of the torus decomposition of the manifold. \smallpagebreak They then prove the following theorem concerning canonical flow neighborhoods associated to periodic Seifert fibered pieces. \smallpagebreak { Theorem B.} Let \(\Phi\) be a pseudo-Anosov flow in \(M\) orientable and let \(P\) be a periodic Seifert fibered piece of the torus decomposition of \(M\). Then there is a finite union \(Z\) of Birkhoff annuli, which is embedded except perhaps at the boundaries of the Birkhoff annuli and which is a model for the core of \(P\): a sufficiently small neighborhood of \(Z\) is a representative for the Seifert piece \(P\). The finite union \(Z\) is well defined up to flow isotopy. \smallpagebreak Using the structure given by Theorem B, the authors produce their main family of examples of generalized pseudo-Anosov flows. Their main theorems concerning their new examples are as follows. \smallpagebreak { Theorem C.} There is a large family of (possibly one-prong) pseudo-Anosov flows in graph manifolds and manifolds fibering over the circle with fiber a torus, where the flows are obtained by gluing simple blocks. The building blocks are homeomorphic to solid tori and they are canonical flow neighborhoods of intrinsic (embedded) Birkhoff annuli. The building blocks have tangential boundary, transverse boundary and only two periodic orbits. A collection of blocks is first glued along annuli in their tangential boundary to obtain Seifert fibered manifolds with boundary, and which have a semiflow transverse to the boundary with finitely many periodic orbits. Under very general and specified conditions these can be glued along their boundaries (transverse to the flow) to produce (possibly one-prong) pseudo-Anosov flows in the resulting closed manifolds. In addition one can do any Dehn surgery (except for one) in the periodic orbits of the middle step to obtain new (possibly one-prong) pseudo-Anosov flows. \smallpagebreak { Theorem D.} (1) There is an infinite family of one-prong pseudo-Anosov flows with two one-prong singular orbits and no other singular orbits where the manifold is Seifert fibered. They are doubly branched covered by the Handel-Thurston examples [\textit{M. Handel} and \textit{W. P. Thurston}, Invent. Math. 59, 95--103 (1980; Zbl 0435.58019)]. \smallpagebreak (2) There are also infinitely many examples of one-prong pseudo-Anosov flows which are doubly branched covered by a geodesic flow in a hyperbolic surface and where the original manifolds are not irreducible. \smallpagebreak { Theorem E.} There are examples of pseudo-Anosov flows in graph manifolds with one periodic piece and an arbitrary number of free pieces.