Equivariant hyperbolization of 3-manifolds via homology cobordisms (Q6044438): Difference between revisions

A basic motivation of the present paper is to explore how homology cobordism interacts with geometric structures on 3-manifolds. Working in the category of smooth, compact and oriented manifolds, the main result states that every 3-manifold \(M\) with an action of a finite group \(G\) is equivariantly and invertibly homology cobordant, with twisted coefficients, to a hyperbolic manifold with an action of \(G\) by isometries (the non-equivariant version that every 3-manifold is homology cobordant to a hyperbolic 3-manifold is due to \textit{R. Myers} [Trans. Am. Math. Soc. 278, 271--288 (1983; Zbl 0532.57006)]). In particular, any finite group that acts on a homology 3-sphere acts also on a hyperbolic homology 3-sphere (with equivalent fixed-point behavior); also, there exist finite groups that are not subgroups of \(\mathrm{SO}(4)\) but act freely on some hyperbolic homology 3-sphere. These groups are generalized quaternion groups from the Milnor-list of finite groups of cohomological period 4; by the geometrization of finite group-actions on 3-manifolds, they cannot act on \(S^3\) but it is known, by heavy surgery-theoretic methods, that some of them admit actions on homology 3-spheres (infinitely many if one assumes the Generalized Riemann Hypothesis). By the main result of the present paper, these actions can be transferred then to hyperbolic homology 3-spheres. No explicit examples of such homology spheres are known (for example, do some of these groups admit free actions also on Seifert fibered homology 3-spheres?) In general, it is known that the finite groups admitting a free action on a homology 3-sphere are either subgroups of \(\mathrm{SO}(4)\) or generalized quaternion groups (and this might remain true also for arbitrary, i.e. not necessarily free actions, cf. a paper by the reviewer [Sib. Èlektron. Mat. Izv. 12, 955--959 (2015; Zbl 1347.57019)]). In contrast, by a result of \textit{D. Cooper} and \textit{D. D. Long} [Topology Appl. 101, No. 2, 143--148 (2000; Zbl 0943.57014)], every finite group admits a free action on a hyperbolic \textit{rational} homology 3-sphere. The authors discuss various other consequences of their main theorem; for example (applied in a non-equivariant setting), there are infinitely many hyperbolic homology 3-spheres that bound homology balls but do not bound contractible manifolds. Another application is the existence of hyperbolic ``equivariant corks'' for a wide class of finite groups (equivariant corks are considered in a paper by the present authors [Algebr. Geom. Topol. 17, No. 3, 1771--1783 (2017; Zbl 1382.57010)] which was one of the motivations of the present paper).