Morse theory and hyperkähler Kirwan surjectivity for Higgs bundles (Q542367)

The authors develop an equivariant Morse theory on the space of Higgs bundle in order to carry out the Atiyah-Bott program for computing the cohomology of this space in the case of rank 2. Let \(\mathcal{B}\) and \(\mathcal{B}_0\) be the space of Higgs bundles respectively with non-fixed or fixed determinant; the Yang-Mills-Higgs functional \( YMH\) is defined on both these spaces. In a paper by \textit{G. Wilkin} [Commun. Anal. Geom., 16, No. 2, 283--332 (2008; Zbl 1151.58010)] it is shown that the gradient flow of \(YMH\) on \(\mathcal{B}\) and \(\mathcal{B}_0\) converges to a critical point that corresponds to the graded object of the Harder-Narasimhan-Seshadri filtration of the initial conditions. This convergence allows the authors to develop a Morse theory on \(\mathcal{B}\) and \(\mathcal{B}_0\) and to compute their cohomology. More precisely, the cohomology of \(\mathcal{M}^{\mathrm{Higgs}}(2,d)\), seen as a hyperkähler quotient, can be computed. In particular the main result in this paper is a formula for the equivariant Poincaré polynomial of the space of the semistable Higgs bundles of rank 2 and degree 0. Moreover, \(\mathcal{M}^{\mathrm{Higgs}}(2,d)\), seen as a hyperkähler quotient, has an associated Kirwan map. Another important result of this paper is that in the non-fixed determinant case this map is surjective also for degree 0 (surjectivity was already known in the case of odd degree).