Complex singularities and the framed cobordism class of compact quotients of 3-dimensional Lie groups by discrete subgroups (Q809418)

If G is a connected Lie group and \(\Gamma\) a discrete subgroup, the tangent bundle of G/\(\Gamma\) admits a natural trivialization. Hence, if G/\(\Gamma\) is compact and if one fixes an orientation of G/\(\Gamma\), the Pontryagin construction yields an element [G/\(\Gamma\) ] in the stable homotopy group of a sphere. In the present paper, the authors complete their project of computing explicitly all the classes [G/\(\Gamma\)) for 3- dimensional groups G. The relevant homotopy group is \(\pi^ S_ 3\cong {\mathbb{Z}}/24\) which is mapped injectively into \({\mathbb{Q}}/{\mathbb{Z}}\) by the e- invariant. The Lie groups which have to be considered are the universal coverings of \(SL_ 2({\mathbb{R}})\), SU(2), \({\mathbb{R}}^ 3\), the Heisenberg group H, the group \(E^+(2)\) of orientation-preserving affine motions of \({\mathbb{R}}^ 2\), and the inhomogeneous Lorentz group E(1,1). The computation of e[G/\(\Gamma\) ] is achieved by interpreting G/\(\Gamma\) in terms of singularity theory: For G locally isomorphic to \(SL_ 2({\mathbb{R}})\), SU(2), H, or E(1,1), the quotient G/\(\Gamma\) is known to be diffeomorphic to the link of a normal Gorenstein singularity, and for \(E^+(2)\) there is a similar description. The case of \(SL_ 2({\mathbb{R}})\) is the most complex one; the treatment of E(1,1) is based on Hirzebruch's work on Hilbert modular surfaces. Also included are the values of the Rochlin invariant and the signature defect.