Nonexistence for complete Kähler-Einstein metrics on some noncompact manifolds (Q1674588)

Let \(M\) be a compact Kähler manifold and \(N\) be a subvariety with codimension greater than or equal to \(2\). The authors show that there are no complete Kähler-Einstein metrics on \(M-N\). As an application, let \(E\) be an exceptional divisor of \(M\). Then \(M-E\) cannot admit any complete Kähler-Einstein metric if the blow-down of \(E\) is a complex variety with only canonical or terminal singularities. A similar result is shown for pairs. A simple application of the results obtained in this paper is in the context of desingularization of Satake compactifications [\textit{J. Igusa}, Proc. Sympos. Pure Math. 9, 301--305 (1966; Zbl 0173.33301)]. Without using any detail on modular forms, the authors can conclude that the compactification of locally symmetric spaces must proceed by adding boundary components which contain singular points, i.e., cusps. This follows from the simple fact that arithmetic quotients admit complete Kähler-Einstein metrics. The paper is organized as follows: The first section is an introduction to the subject and contains the formulation of the main theorem. Section 2 is devoted to the prove of the following result: Let \(f: M \longrightarrow X\) be two complex normal varieties which are birational, and let \(E\) the sum of the exceptional divisors. Then, if \(X\) has only canonical or terminal singularities, and the codimension-two singular locus is non-empty in \(X\), then there are no complete Kähler-Einstein metrics on \(M-E\). The proof is made by contradiction and the key point is to show the boundedness of the volume element nearby \(N\). In the proof, the reader will note that all the analysis is local. In Section 3, although the proof is local and mainly differential geometric in nature, the theorem is phrased in the right algebraic setting. The purpose of Section 4 is to generalize the above result to the case of pairs \((M, D)\) with boundary divisor \(D\). As the used analysis essentially only depends on the curvature condition, it can be carried out for these open cases as well. For pairs, the authors prove the following theorem: If the pair \((X, D)\) is Kawamata log terminal (klt), then there are no complete Kähler-Einstein metrics on \(M-E\), where \(E\) is the exceptional divisor of a log resolution \(f: M\longrightarrow X\) where \(E\cap f^{-1}(\text{supp}(D))\) is a simple normal crossing.