\(\mathrm{SO}(p,q)\)-Higgs bundles and higher Teichmüller components (Q2322817)

Since this very interesting article, in particular its introduction, is clearly written, we only add some background motivation. To this end, let us briefly review the classical relationship between representations of the fundamental group \(\pi_1(S)\) of a closed (real) surface \(S\) of genus \(g_S \geq 2\) and geometric structures on \(S\). More precisely, let \(\mathcal{R}(S, G)\) be the representation variety parametrizing conjugacy classes of representations of \(\pi_1(S)\) into a Lie group \(G\). Now consider a (marked) conformal structure on \(S\). Any such conformal structure is determined by a unique hyberbolic structure, i.e. a metric of constant curvature \(-1\). The holonomy of this hyperbolic structure gives a faithful representation \(\phi\colon \pi_1(S)\to \mathrm{PSL}(2, \mathbb{R})\) with discrete image. Conversely, any such representation \(\rho\) determines a hyperbolic structure on \(S\) and hence an identification of \(S\) with the quotient of the hyperbolic plane \(\mathbb{H}^2\) by the image of \(\rho\) in \(\mathrm{PSL}(2, \mathbb{R})\). These discrete and faithful representations determine two connected components in \(\mathcal{R}(S, \mathrm{PSL}(2, \mathbb{R}))\) distinguished by choosing an orientation. In particular, one such component identifies with the Teichmüller space \(\mathcal{T}(S)\) which is homeomorphic to \(\mathbb{R}^{6g-6}\). As it turns out, discrete and faithful representations of \(\pi_1(S)\) define intriguing and new connected components in representations varieties \(\mathcal{R}(S, G)\), called \emph{higher Teichmüller components}, for more general Lie groups \(G\). Similar to Teichmüller space, these components are closely related to geometric structures on \(S\) making them particularly attractive from both the representation-theoretic and geometric point of view. Moreover, these connected components are not detactable by topological invariants. For that reason, the authors of this article call such components, including higher Teichmüller components, \emph{exotic}. (As an aside, note that this does not occur for compact or complex simple Lie groups \(G\) where \(\pi_0(\mathcal{R}(S, G)) \cong \pi_1(G)\).) Prior to the present article, two families of Lie groups giving rise to higher Teichmüller components were known \begin{itemize} \item split real semisimple Lie groups \(G\): the corresponding components are called \emph{Hitchin components} which are homeomorphic to a Euclidean space in analogy to Teichmüller space and were first discovered by Hitchin; \item Lie groups of Hermitian type \(G\): the corresponding components are defined by representations to \(G\) with maximal so-called Toledo invariant. \end{itemize} It is desirable to give a uniform characterization of such representations, possibly enabling us to find further examples where \(\mathcal{R}(S, G)\) has more connected components than expected. This was partially achieved by Labourie by introducing so-called \emph{Anosov representations}. Guichard and Wienhard further refined this notion by adding a positivity condition. This led to the conjecture that besides the above Lie groups, \(\mathcal{R}(S, \mathrm{SO}(p,q))\) for \(1 < p < q\) might have exotic connected components. The very intriguing present article proves this deep conjecture. More precisely, it shows: \begin{itemize} \item the existence of explicitly described exotic components of \(\mathcal{R}(S, \mathrm{SO}(p,q))\) for \(1<p \leq q\), \item that each exotic component contains positive Anosov representations, \item the non-existence of any other exotic components for \(p=1\) and \(2<p \leq q\). \item \end{itemize} In particular, the authors are able to count all connected components of \(\mathcal{R}(S, \mathrm{SO}(p,q))\) for \(2 < p \leq q\). The proof of these results mainly uses the non-abelian Hodge theorem giving a homeomorphism between \(\mathcal{R}(S, \mathrm{SO}(p,q))\) and the moduli space \(\mathcal{M}(X, \mathrm{SO}(p,q))\) of \(\mathrm{SO}(p,q)\)-Higgs bundles on a compact Riemann surface \(X\) with underlying topological surface \(S\). The latter moduli space comes with additional tools to study its topology which was pioneered by Hitchin. More precisely, the authors study critical points of the so-called Hitchin function and the natural \(\mathbb{C}^*\)-action on \(\mathcal{M}(X, \mathrm{SO}(p,q))\). If the moduli space of Higgs bundles is smooth, then the Hitchin function is a perfect Morse-Bott function whose critical points (which occur at fixed points of the \(\mathbb{C}^*\)-action) can be used to determine connected components. In the present situation, the moduli spaces are mildly singular but the authors are still able to execute a similar strategy by using Kuranishi theory. The article is an important contribution to the exciting interplay between surface group representations and Higgs bundles. Since the article provides necessary background to both subjects, it is well suited for readers of both areas.