A cohomological interpretation of derivations on graded algebras (Q1702147)

The paper under review gives a cohomological interpretation of homogeneous derivations on an arbitrary finitely generated positively graded normal algebra over an arbitrary field \(k\). The paper's approach is based on a construction presented in [\textit{M. Demazure}, in: Introduction à la théorie des singularités. II: Méthodes algébriques et géométriques. Paris: Hermann. 35--68 (1988; Zbl 0686.14005)]; the authors trace derivations through Demazure's correspondence between a finitely generated positively graded normal \(k\)-algebras \(A\) and normal projective \(k\)-varieties \(X\) equipped with an ample \(\mathbb{Q}\)-Cartier \(\mathbb{Q}\)-divisor \(D\). The main result of the paper is obtaining a generalized Euler sequence involving a sheaf on \(X\) whose space of global sections consists of all homogeneous \(k\)-linear derivations of \(A\) and a sheaf of logarithmic derivations on \(X\). In the last part of this work, the authors apply their results to a graded isolated complete intersection singularity with negative derivation (see [\textit{M. Granger} and \textit{M. Schulze}, ``Derivations of negative degree on quasi-homogeneous isolated complete intersection singularities'', Preprint, \url{arXiv:1403.3844}]). They describe the associated projective variety \(X\) and its singularities and interpret a degree \(-1\) derivation as a twisted vector field on \(X\). As another application of their results, the authors consider the degree \(-1\) derivations on \(k[x, y, z]/(xy-z^{2})\), where \(x, y, z\) have degree \(1\), and give a cohomological explanation of the fact that there is no such derivations if \(\mathrm{Char }k = 2\), while if \(\mathrm{Char }k\neq 2\), then \(\partial_{z}\) is such a derivation.