Fundamental group of moduli spaces of representations (Q744908)

Let \(\Gamma\) be the fundamental group of an algebraic curve \(X\) of genus \(g\) with \(n\) points removed, and let \(G\) be a connected reductive algebraic group over the complex numbers or a connected compact Lie group. The article states some facts about the fundamental group of the moduli space of representations \(\mathfrak{X}_\Gamma(G)=\Hom(\Gamma,G)/G\), where we identify representations by conjugation. They present two results on its fundamental group depending on the number of the points removed. In case that \(n\geq 1\), i. e. when some points are removed, authors prove that the group \(\pi_1(\mathfrak{X}_\Gamma(G))\) is isomorphic to \(\pi_1(G/[G,G])\); in case \(n=0\), which is the compact case, they states only for cases \(G=\mathrm{GL}(n,\mathbb{C})\) and \(U(n)\), where \(\pi_1(\mathfrak{X}_\Gamma(G))\cong \mathbb{Z}^{2g}\), and for cases \(G=\mathrm{SL}(n,\mathbb{C})\) and \(G=\mathrm{SU}(n)\), where \(\pi_1(\mathfrak{X}_\Gamma(G))=0\). While the first result is a generalization of a previous theorem published by the same authors (see Theorem B in [\textit{S. Lawton} and \textit{D. Ramras}, New York J. Math. 21, 383--416 (2015; Zbl 1339.57002)]), the second one is a partial result of one published earlier (see Theorem 1.1 in [\textit{I. Biswas} et al., Math. Z. 281, No. 1--2, 415--425 (2015; Zbl 1349.14042)]). In any case, the techniques presented in this paper are very different than the previous used. For the first statement (the non-compact case), the proof takes a bit more than three pages (recall that in the mentioned paper they uses more than twenty pages to prove a partial result). The key of this proof lies on the long exact sequence in homotopy \(\pi_1(\mathfrak{X}_\Gamma([G,G])\to \pi_1(\mathfrak X_\Gamma(G))\to \pi_1(\mathfrak{X}_\Gamma(G/[G,G]))\to \pi_0(\mathfrak{X}_\Gamma([G,G]))\to 0\). Once they prove \(\mathfrak{X}_\Gamma([G,G])\) is a connected simply-connected space, the result follows. They do several steps to establish this assertion: first, they consider the subset \(\Hom^g(\Gamma,G)\subset \Hom(\Gamma,G)\) of \textit{good} representations (those satisfy \(Z(\rho(\Gamma))= Z(G)\)), and they prove that the natural projection map \(\Hom^g(\Gamma,G)\to \mathfrak{X}_\Gamma(G)\) is a \(PG\)-bundle, where \(PG\) is a parabolic subgroup (see Lemma~2.2). Based in this property, they show that the map \(\Hom(\Gamma,G)\to \mathfrak{X}_\Gamma(G)\) is \(\pi_1\)-surjective. Then, since \(\pi_1([G,G]^r)\to \pi_1(\mathfrak{X}_\Gamma([G,G])\) sends to zero all map, \(\pi_1\)-surjectivity implies the simply-connectivity. A different technique is applied to prove the second result. Here, authors take into account the moduli space of Higgs bundles. They notice that the moduli space of Higgs bundles of rank \(n\) on \(X\), denoted by \(\mathcal{M}_H(X,n)\), (resp. with trivial determinant, denoted by \(\mathcal{N}_H(X,n)\)) is isomorphic to \(\mathfrak{X}_\Gamma(\mathrm{GL}(n,\mathbb{C}))\) (resp. \(\mathfrak{X}_\Gamma(\mathrm{SL}(n,\mathbb{C}))\)). They states that the map \(\mathcal{M}_H(X,n)\to \mathrm{Pic}^0(X)\) induces an isomorphism between their fundamental groups, so that \(\pi_1(\mathcal{M}_H(X,n))\cong \mathbb{Z}^{2g}\).