On rigid compact complex surfaces and manifolds (Q1649346)

The article investigates the rigidity of compact complex manifolds. First, the different flavors of rigidity are introduced: {\parindent=0.6cm\begin{itemize}\item[--] A compact complex manifold \(X\) is said to be \textit{globally rigid} if any \(X'\) which is deformation equivalent to \(X\) is in fact isomorphic to \(X\). Here, deformation equivalence means that \(X\) and \(X'\) appear as fibres of a proper smooth holomorphic map \(\mathfrak X \to B\), where \(B\) is connected (but possibly non-reduced). \item[--] \(X\) is said to be \textit{(locally) rigid} if for every deformation with \(X\) as its central fibre, all fibres in some open neighborhood of the central fibre are isomorphic to \(X\). \item[--] \(X\) is said to be \textit{infinitesimally rigid} if \(H^1(X, T_X) = 0\), where \(T_X\) is the tangent sheaf of \(X\). \item[--] \(X\) is said to be \textit{strongly rigid} if the set of all compact complex manifolds \(Y\) which are homotopy equivalent to \(X\) consists of a finite set of isomorphism classes of globally rigid varieties. \item[--] \(X\) is said to be \textit{étale rigid} if every finite étale cover \(Y\) of \(X\) is rigid. Similarly for \textit{étale globally rigid} etc. \end{itemize}} The authors then proceed to discuss the relations between the different notions of rigidity, as well as some general deformation-theoretic results with an emphasis on the case of surfaces. The main results of the paper are as follows: Theorem 1.1. The BCD surfaces (certain families of surfaces first constructed by Bauer, Catanese and Dettweiler) are infinitesimally rigid and rigid. Their Albanese map gives rise to a rigid curve inside the moduli stack of stable curves. Theorem 1.1 is in fact a consequence of Theorem 1.2. The Hirzebruch-Kummer surfaces \(\mathrm{HK}_{CQ}(n)\) are infinitesimally rigid and rigid for all \(n \geq 4\). Theorem 1.3 is a classification of all rigid compact complex surfaces. In particular, surfaces of Kodaira dimension \(0\) or \(1\) are never rigid. In higher dimensions, we have Theorem 1.4. For each \(n \geq 3\) and \(k = -\infty, 0, 2, \dots, n\), there exists a rigid projective variety \(X\) of dimension \(n\) and Kodaira dimension \(k\). The authors conclude by posing several interesting questions for future research.