Around the de Rham-Betti conjecture

From MaRDI portal
Publication:6402345

arXiv2206.08618MaRDI QIDQ6402345FDOQ6402345


Authors: Tobias Kreutz, Mingmin Shen, Charles Vial Edit this on Wikidata


Publication date: 17 June 2022

Abstract: A de Rham-Betti class on a smooth projective variety X over an algebraic extension K of the rational numbers is a rational class in the Betti cohomology of the analytification ofX that descends to a class in the algebraic de Rham cohomology of X via the period comparison isomorphism. The period conjecture of Grothendieck implies that de Rham-Betti classes should be algebraic. We prove that any de Rham-Betti class on a product of elliptic curves is algebraic. This is achieved by showing that the Tannakian torsor associated to a de Rham-Betti object is connected, and by exploiting the analytic subgroup theorem of W"ustholz. In the case of products of non-CM elliptic curves, we prove the stronger result that overlinemathdsQ-de Rham-Betti classes are overlinemathdsQ-linear combinations of algebraic classes by showing that the period comparison isomorphism generates the torsor of motivic periods. A key step consists in establishing a version of the analytic subgroup theorem with overlinemathdsQ-coefficients. Finally, building on results of Deligne and Andr'e regarding the Kuga-Satake correspondence, we further show that any de Rham-Betti isometry between the second cohomology groups of hyper-K"ahler varieties, with second Betti number not 3, is Hodge. As two applications we show that codimension-2 de Rham-Betti classes on hyper-K"ahler varieties of known deformation type are Hodge and we obtain a global de Rham-Betti Torelli theorem for K3 surfaces over overlinemathdsQ.













This page was built for publication: Around the de Rham-Betti conjecture

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6402345)