Finite quotients of three-dimensional complex tori (Q2188229)

Let \(X\) be a normal compact Kähler space such that the first Chern class \(c_1(X)\) vanishes. If \(X\) is smooth, the Beauville-Bogomolov decomposition theorem determines the structure of \(X\) up to finite étale cover. If \(X\) is projective and has at most klt singularities, recent results [\textit{S. Druel}, Invent. Math. 211, No. 1, 245--296 (2018; Zbl 1419.14063); \textit{D. Greb} et al., Geom. Topol. 23, No. 4, 2051--2124 (2019; Zbl 1423.14110); \textit{A. Höring} and \textit{T. Peternell}, Invent. Math. 216, No. 2, 395--419 (2019; Zbl 07061101)] establish a similar decomposition, up to a finite cover that is étale in codimension one. In the special case where the second (orbifold) Chern class also vanishes, one expects that \(X\) is a quotient of a torus \(T\) by a finite group \(G\) acting freely in codimension one. If \(X\) is projective, this was shown by \textit{S. Lu} and \textit{B. Taji} [Int. Math. Res. Not. 2018, No. 1, 292--319 (2018; Zbl 1423.14255)], generalising earlier results of Shepherd-Barron and Wilson in the threefold case [\textit{N. I. Shepherd-Barron} and \textit{P. M. H. Wilson}, J. Algebr. Geom. 3, No. 2, 265--281 (1994; Zbl 0807.14031)]. In this paper, the authors generalise the result of Shepherd-Barron and Wilson to the Kähler case. More precisely, assume that \(X\) is a compact Kähler threefold with at most canonical singularities. If the first Chern class \(c_1(X)\) vanishes and the second orbifold Chern class \(\tilde c_2(X)\) is orthogonal to some Kähler class, then \(X\) is isomorphic to a torus quotient \(T/G\). For the proof one uses abundance on Kähler threefolds to reduce to the case when the canonical sheaf is trivial. Since a non-algebraic compact Kähler threefold always carries a non-zero holomorphic two-form, a Riemann-Roch computation shows that there are also non-zero holomorphic one-forms. Thus one can study the variety via the Albanese map, this ultimately leads to the desired structure result. While this general approach is rather classical, the technical setup for an actual proof in the analytic setting is not available in the literature. Therefore the authors develop the whole theory from scratch, their careful exposition makes this paper a very useful reference and a pleasure to read.