The characteristic toric splitting of irreducible compact 3-orbifolds (Q1093953)

Central for 3-manifold topology are the conjectures of Thurston that every compact 3-manifold and, more generally, every compact (good) 3- orbifold possesses a canonical decomposition into pieces which admit geometric structures. The first step in this decomposition is the decomposition along 2-spheres resp. spherical 2-suborbifolds into irreducible pieces. The second step, in the case of manifolds, is the torus-decomposition due to Johannson and Jaco-Shalen: every compact irreducible 3-manifold possesses a natural (unique up to isotopy) splitting along a system of essential tori into Seifert fiber spaces and atoroidal 3-manifolds (i.e. every incompressible torus is boundary- parallel). The main result of the present paper is the corresponding splitting result for compact irreducible 3-orbifolds. Now irreducible means that there are no ``bad'' 2-suborbifolds (i.e. with no manifold covering), and that every spherical 2-suborbifold is trivial in the sense that it bounds a cone-3-suborbifold; then the splitting is along Euclidean 2- suborbifolds into Seifert fibered orbifolds and orbifolds which contain no essential Euclidean 2-suborbifolds. It is not known whether such an orbifold possesses a finite covering by a manifold; if this is the case, the result can also be proved using the equivariant version of the torus- decomposition in the manifold-case [see \textit{W. H. Meeks} III and \textit{P. Scott}, Invent. Math. 86, 287-346 (1986; Zbl 0526.57006)]. In the present paper a direct proof is given working in the context of orbifolds; an important ingredient is an adaptation of minimal surface techniques to the orbifold setting. It is also determined how the decomposition behaves under finite coverings; for this the surprisingly difficult result has to be proved that every finite group action on torus \(\times [0,1]\) is standard, i.e. a product action (see also the above cited paper of Meeks- Scott for another proof of this fact). The most interesting special case of this orbifold splitting is the case of the orbifold \(S^ 3\) with a given knot or link as singular set, of branching order two. Then the splitting, which can be considered as a splitting of the link now, is along tori (``companionship-tori'') and, as a new phenomenon, along 2-spheres meeting the link in four points (``Conway-spheres''). The theory in this special case has been developed in an independent and more elementary way in a long-awaited monograph of the two authors (to appear in the Lond. Math. Soc. Lect. Note Ser.), with especially strong results in the case of ``algebraic'' links, i.e. links for which all components of the splitting are of Seifert fibered type. The paper gives a lot of background material about orbifolds, especially Seifert fibered orbifolds, geometric structures on 3-manifolds and 3- orbifolds and the status of Thurston's conjectures, with an extensive guide to the literature; especially the first three paragraphs can be recommended to everybody with some interest in recent developments in the theory of 3-manifolds.