An \(\mathrm{HP}^2\)-bundle over \(\mathrm{S}^4\) with nontrivial \(\hat{A}\)-genus (Q2658210): Difference between revisions

The seminal paper [\textit{B. Hanke} et al., Publ. Math., Inst. Hautes Étud. Sci. 120, 335--367 (2014; Zbl 1321.58008)] on spaces of Riemannian metrics of positive scalar curvature relies on the existence of smooth fiber bundles over spheres whose total spaces are spin with non-vanishing \(\hat{A}\)-genus, and whose fiber supports a metric of positive scalar curvature. Remarkably, Hanke-Schick-Steimle prove rather abstractly that such bundles exist and give no concrete description of the fiber. They also note that it seems to be a difficult problem to determine if it is possible that the fiber even admits a metric of positive sectional or Ricci curvature. Krannich-Kupers-Randal-Williams solve this problem by explaining the existence of a bundle \(\mathrm{H} P^2 \to E\to S^4\) with \(\hat{A}(E) \neq 0\). The proof is based on deep results from surgery and pseudoisotopy theory. As a corollary, they show that the space of Riemannian metrics of positive sectional or Ricci curvature on \(\mathrm{H}P^2\) has a rationally non-trivial third homotopy group. In the final remarks, the authors insinuate that more general results can be proven using similar methods. For instance, the fiber can also be a quaternionic projective space \(\mathrm{H}P^n\) with \(n>2\) and the base a sphere of bigger dimension. It seems it was paramount for the authors to write a brief paper, potentially at the cost of generality and also accessibility to unexperienced readers, which is a bit unfortunate as the ideas contained in this work are truly interesting.