Moduli spaces of Higgs bundles on degenerating Riemann surfaces (Q1678152)

This paper is part of a novel and important approach to the self-dual Yang-Mills equations (SDYM), more specifically the moduli space of solutions introduced by \textit{N. J. Hitchin} [Proc. Lond. Math. Soc. (3) 55, 59--126 (1987; Zbl 0634.53045)] This approach was initiated by the author et al. [Duke Math. J. 165, No. 12, 2227--2271 (2016; Zbl 1352.53018)]. The goal is to study boundary points of the moduli space, and to achieve this with primarily analytic techniques (for an analytic approach, cf. also: [\textit{G. Daskalopoulos} and \textit{R. Wentworth}, Math. Ann. 297, No. 3, 417--466 (1993; Zbl 0788.32012)]). In this paper, the author initiates a yet novel kind of boundary analysis, in contrast to his previous work [loc. cit.] which focused on the asymptotics of ``large'' (in a suitable norm) Higgs fields. Here the supporting Riemann surface varies, and in the limit becomes a nodal curve. By sophisticated gluing arguments the author is able to show, roughly speaking, that any solution to the SDYM equations on a nodal curve is the uniform limit of a sequence of smooth solutions over smooth curves. The technicalities are lengthy to describe, but they involve the following property: The limiting curve \((\Sigma_0, J_0)\) (given as a topological surface \(\Sigma\) that resolves the nodes fo \(\Sigma_0\), and its complex structure \(J_0)\) has a (finite) set \(\mathfrak{p}\) of nodal sigularities; the solution \((A_0,\Phi_0 )\) to the SDYM equations , given by a unitary connection \(A_0\) on a Hermitian vector bundle \(E\) over \(\Sigma_0\) and an End\((E)\)-valued (1,0) form, has at most logarithmic singularities over \(\mathfrak{p}\). The technical assumptions concern the local expansion of \(A_0\) at the points in \(\mathfrak{p}\). As the author explains, he made them as weak as possible but this local behavior is what allows him to prove ``absence of so-called small eigenvalues'' for operators on solution spaces, and reach the main gluing theorem, which ensures that the given solution can be reached as a limit of solutions over smooth curves, Main Theorem: Given a sequence of Riemann surfaces \((\Sigma, J_i)\) that converges uniformly to \((\Sigma_0, J_0)\), for every sufficiently large \(i\in{\mathbb N}\) there exists a smooth solutions \((A_i,\Phi_i)\) on \((\Sigma, J_i)\) that converges to \((A_0,\Phi_0)\) uniformly on compact subsets of \(\Sigma_0\backslash\mathfrak{p}\). The proofs are a tour de force of inequalities over a deformation complex; the author introduces appropriate Hilbert spaces that contain solutions to the approximate equations, decomposes the solution into Fourier modes, controls each one of those, and ultimately reaches estimates that allows him to apply the Banach fixed-point theorem. The paper, despite being very technical, is eminently legible; in fact the Introduction provides an overview of the other contexts in which these spaces are prominent; a ``guiding example'' (namely, that of \(\mathrm{SL}(2,\mathbb R)\) Higgs bundles) provides the relationship of the moduli space with Teichmüller theory, with application to the Teichmüller theory of harmonic maps. The author painstakingly explains each special assumption he makes, and his reason for thus restricting the limiting Higgs bundles he can treat, announces forthcoming work that will complete the analytic study, and beyond that, envisions a project to describe complete hyperkähler metrics on the compactified moduli spaces.