The algebraic geometry of representation spaces associated to Seifert fibered homology 3-spheres (Q756782)

The paper develops an algebro-geometric approach to the study of the representation spaces \(R(\Sigma)=Hom^*(\pi_ 1(\Sigma),SU(2))/adSU(2)\) of the fundamental group of a 3-dimensional Seifert manifold \(\Sigma\) which is a \({\mathbb{Z}}\)-homology sphere and finding its Floer homology. Previous study was initiated by Fintushel-Stern who also put forward the conjecture that R(\(\Sigma\)) admits a Morse function with only even index critical points. The authors' key tool is to interpret the representation space of the homology sphere \(\Sigma\) as the representation space of the fundamental group of a certain Dolgachev surface, which in turn is interpreted as the moduli space of certain vector bundles on it (more precisely the moduli space of rank 2 algebraic bundles with vanishing both Chern classes which are stable relative to an ample bundle). The authors show that these moduli spaces are smooth rational varieties possessing stratification such that all strata are locally closed smooth subvarieties isomorphic to the open subsets in the projective space. Moreover in this picture the complements to these strata are identified with the cones on secant varieties. This has the following corollaries. Description of the connected components as moduli spaces of representations with the same rotation numbers. It proves the above conjecture on existence of Morse function in dimensions less or equal than six (uses birational invariance of the torsion in \(H^ 3)\) and shows that this conjecture is equivalent to the vanishing of \(H^ i(R(\Sigma),{\mathbb{Z}})\) in odd dimensions. The latter is proved up to torsion. In fact the authors compute the Betti numbers of R(\(\Sigma\)) using Deligne's solution of the Weil conjectures. This calculation gives calculation of the Floer homology of all Seifert manifolds which are homology spheres modulo the above conjecture. As the authors added in proof the conjecture actually is proven by \textit{P. A. Kirk} and \textit{E. P. Klassen} [Topology 30, No.1, 77-95 (1991)], and Furuta-Steer, so the present paper does give the answer to the question: what are the Floer homology of the Seiferts \({\mathbb{Z}}\)-spheres.