Topological rigidity and \(H_1\)-negative involutions on tori (Q2509858)

The authors classify contractible \(n\)-manifolds with a proper action of \(\Gamma_n:=\mathbb{Z}^n\rtimes_{-1} C_2\), up to equivariant homeomorphism. Roughly speaking, the result is that (at least in high dimensions) the only ``exotic'' manifolds of this type arise from UNil-groups due to \textit{S. E. Cappell} [Bull. Am. Math. Soc. 80, 1117--1122 (1974; Zbl 0322.57020)], so there are none in half of the dimensions and countably infinitely many ones on the other half of the dimensions. The results of this paper support a conjectural picture for certain cocompact actions on Hadamard manifolds, which was sketched by \textit{F. Quinn} in his 1986 ICM address [Proc. Int. Congr. Math., Berkeley/Calif. 1986, Vol. 1, 598--606 (1987; Zbl 0673.57017)] and is made precise in the paper under review. Another interest of the present results is that they allow to compute the number of closed manifolds homotopy equivalent to the \(n\)-torus, and equipped with an involution acting by \(-1\) on the first homology group, modulo equivariant homeomorphism. The proof uses methods of surgery theory. Let \(C_2\) act on the \(n\)-torus \(\mathbb{R}^n/\mathbb{Z}^n\) by \(-1\) and denote by \(X\) the orbit space of the subset where \(C_2\) acts freely. Then \(X\) has an obvious compactification \(\bar X\) by adding an \(\mathbb{R} P^{n-1}\) at each end. The authors show that the above classification is equivalent to the computation of the classical surgery structure set of the pair \((\bar X, \partial \bar X)\). The latter can be computed by the surgery exact sequence, using the Farrell-Jones conjecture and an explicit model for the equivariant classifying space \(E_{vc}\Gamma_n\). In this context, one virtue of the present paper that the authors carefully explain the relation between the Quinn-Ranicki assembly map and the equivariant assembly map due to Davis-Lück.