Modular forms and K3 surfaces

From MaRDI portal
Modular forms and \(K3\) surfaces



Abstract: For every known Hecke eigenform of weight 3 with rational eigenvalues we exhibit a K3 surface over QQ associated to the form. This answers a question asked independently by Mazur and van Straten. The proof builds on a classification of CM forms by the second author.


This paper gives an affirmative answer to the question of geometric realizations of Hecke eigenforms of weight \(>2\). This question (known as the converse problem) has been raised by Mazur and van Straten independently. More specifically, this paper solves the geometric realization problem in weight \(3\). The main result is formulated in the following theorem.NEWLINENEWLINETheorem. Assume the validity of the extended Riemann Hypothesis (ERH) for odd real Dirichlet characters. Then every Hecke eigenform \(f\) of weight \(3\) with rational eigenvalues is associated to a singular \(K3\) surface \(X\) defined over \(\mathbb{Q}\).NEWLINENEWLINE\(K3\) surfaces \(S\) with Picard number \(20\) are often referred to as \textit{singular} \(K3\) surfaces. Previously, it is known by \textit{R. Livné} [Isr. J. Math. 92, No. 1--3, 149--156 (1995; Zbl 0847.11035)] that every singular \(K3\) surface \(X\) over \(\mathbb{Q}\) is modular, that is, there is a Hecke eigenform \(f\) with system of eigenvalues \([c_p]\) such that \(\text{trace}(\text{Frob}_p)=c_p\) for almost all prime \(p\). The eigenform \(f\) has weight \(3\) and CM by the imaginary quadratic field \(K=\mathbb{Q}(\sqrt{d})\) where \(d=\text{Disc}(\mathrm{NS}(X))<0\) is the discriminant of \(X\). Thus, the above theorem presents the converse result to Livné's theorem. The ERH assumption is needed to establish the finiteness of imaginary quadratic fields with class group exponent \(2\) (\(65\) in total), which correspond bijectively to CM newforms of weight \(3\) with rational eigenvalues, up to twisting.NEWLINENEWLINEThe main technique is to construct one-parameter families of \(K3\) surfaces with Picard number \(\geq 19\), and then search for singular specializations over \(\mathbb{Q}\) (i.e., with Picard number \(20\)). For such families of \(K3\) surfaces, elliptic \(K3\) surfaces (with section) with Picard number \(\geq 19\) are used, and an algorithm is presented for the search of extra divisors to attain Picard number \(20\).



Cites work


Cited in
(30)








This page was built for publication: Modular forms and \(K3\) surfaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q390723)