Compact hyperbolic manifolds without spin structures (Q2224536)

An orientable Riemannian manifold \(M\) admits a spin structure if the second Stiefel-Whitney class \(w_2(M)\) vanishes (we say \(M\) is spinnable). Orientable manifolds in dimension \(2\) and \(3\) are always spinnable, but the situation becomes more complicated in higher dimensions. In the context of hyperbolic manifolds of finite volume, an old result of \textit{D. Sullivan} [Proc. Conf., Athens/Ga. 1977, 543--555 (1979; Zbl 0478.57007)] (building on his work with \textit{P. Deligne} [C. R. Acad. Sci., Paris, Sér. A 281, 1081--1083 (1975; Zbl 0317.55016)]) shows that in any dimension, every finite volume hyperbolic manifold has a finite cover that is spinnable. Examples of finite volume but non-compact hyperbolic \(n\)-manifolds for all \(n\geq 5\) which are not spinnable were constructed in [\textit{D. D. Long} and \textit{A. W. Reid}, Proc. Edinb. Math. Soc., II. Ser. 63, No. 2, 305--313 (2020; Zbl 1436.57033)], but the difficult and most interesting case is the closed case. This is where the paper under review makes progress. It constructs closed orientable hyperbolic manifolds in all dimensions \(\geq 4\) that are not spinnable. The key case is in dimension \(4\). This case can then be used to build higher dimensional examples by an embedding result of [\textit{A. Kolpakov} et al., Math. Res. Lett. 25, No. 4, 1305--1328 (2018; Zbl 1446.57022)]. Because of a lack of explicit examples in dimensions \(\geq 4\), it was unknown whether every finite volume hyperbolic manifold was spinnable or not. Now standard \(4\)-manifold topology shows that if a closed orientable 4-manifold is spinnable, it has an even intersection form. Thus an obstruction to a spin structure in dimension \(4\) is odd intersection form, and so in particular an approach to building a closed orientable hyperbolic \(4\)-manifold which is not spinnable is to construct a closed orientable hyperbolic \(4\)-manifold with an odd intersection form. This is the basis of the key result in the paper under review, allowing the authors to produce (infinitely many) closed orientable hyperbolic 4-manifolds with an embedded \(\pi_1\)-injective surface \(S\) with \(S.S=1\); i.e. the intersection form is odd. The construction of \(S\) is very beautiful, using what the authors call a \(Y\)-shaped piece which is inspired by the notion of a trisection in 4-manifold topology. The construction of the \(Y\)-shaped piece uses the combinatorics of the all right \(120\)-cell in \(\mathbb{H}^4\), and the surface \(S\) is a pleated surface built out of all right pentagons (which arise as codimension 2 strata). The construction of these closed orientable hyperbolic \(4\)-manifolds that are not spinnable, also provides the first examples of closed orientable, hyperbolic 4-manifolds \(M\) satisfying either of the conditions: \begin{itemize} \item{} \(H_2(M; \mathbb{Z})\) is not generated by totally geodesically immersed surfaces. \item{} There is a covering of \(M\) that is a non-trivial bundle over a compact surface. \end{itemize} This is a beautiful and important paper.