Explicit families of \(K3\) surfaces having real multiplication (Q6158007): Difference between revisions

Let \(X\) be a complex \(K3\) surface and let \(T\) denote its transcendental lattice, which has a pure Hodge structure of weight \(2\). Let \(E\) denote the algebra of endomorphisms of \(T\) which respect the Hodge structure. \textit{Yu. G. Zarkhin} shows that \(E\) is a totally real or a CM field [J. Reine Angew. Math. 341, 193--220 (1983; Zbl 0506.14034)]. Indeed, in most cases \(E\) is simply \(\mathbb{Q}\). If \(E\) is a non-trivial totally real (resp. CM) field, \(X\) is said to have real (resp. complex) multiplication. While in the literature one can find several example and constructions of \(K3\) surfaces with complex multiplication, \(K3\) surfaces with real multiplication remains more elusive. \textit{B. Van Geemen} shows that there exists a one-dimensional family of \(K3\) surfaces with real multiplication by \(\mathbb{Q}(\sqrt{d})\), for every odd integer \(d\) that is the sum of two squares [Mich. Math. J. 56, No. 2, 375--399 (2008; Zbl 1161.14027)]. Van Geemen's approach is ``Hodge theoretical'' and leaves the problem of providing explicit examples open. \textit{A.-S. Elsenhans} and \textit{J. Jahnel}, building on and generalizing van Geemen's work, provide the first explicit examples of \(K3\) surfaces with real multiplication [LMS J. Comput. Math. 17A, 14--35 (2014; Zbl 1307.14063)]. In particular, they provide a one dimensional family of \(K3\) surfaces for which they are able to prove that infinitely many members have real multiplication by \(\mathbb{Q}(\sqrt{2})\). In the paper under review, the same authors give an arithmetic criterion to establish whether the generic member of a family of \(K3\) surfaces have complex or real multiplication. They also show that the endomorphism algebra \(E\) does not shrink under specialization, that is, if \(E_\eta, E_\tau\) are the endomorphism fields of the generic member and a special member of a family, respectively, then \(E_\eta\hookrightarrow E_\tau\). Using these result, they show that every member of the family provided in their previous work [loc. cit.] admits real multiplication by \(\mathbb{Q}(\sqrt{2})\). Moreover, the authors also provide a new family whose members admit real multiplication by \(\mathbb{Q}(\sqrt{5})\).