Explicit uniform bounds for Brauer groups of singular \(K3\) surfaces (Q6043801)

\(K3\) surfaces provide a remarkable example of algebraic varieties for which the Hasse principle does not hold. Although a complete explanation for the failure of the Hasse principle for \(K3\) surfaces is still missing, we know one reason that might occur: the \textit{Brauer-Manin obstruction}. This obstruction to the Hasse principle can be explained in terms of the Brauer group. More precisely, let \(k\) be a number field and \(X\) a \(K3\) surface over \(k\). The \textit{Brauer group} of \(X\), denoted by \(\mathrm{Br}\,X\) is defined as \(H^2(X,\mathbf{G}_m)\). Let us fix an algebraic closure \(\overline{k}\) of \(k\); let \(X(\mathbf{A}_k)\) denote the set of adelic points of \(X\). Manin showed that there is a pairing on \(\mathrm{Br}\,X \times X(\mathbf{A}_k)\) whose left kernel, denoted by \(X(\mathbf{A}_k)^{\mathrm{Br}}\), fits in the following chain of inclusions: \[ X(k)\subseteq X(\mathbf{A}_k)^{\mathrm{Br}} \subseteq X(\mathbf{A}_k) \subseteq X( \overline{k} )\; . \] If \(X(\mathbf{A}_k)\neq \emptyset\) and \(X(\mathbf{A}_k)^{\mathrm{Br}}=\emptyset\) then we say that \(X\) has the Brauer-Manin obstruction to the Hasse principle. This approach to the study of rational points of a \(K3\) surface led to an increasing interest in their Brauer group. The Brauer group of a \(K3\) surface \(X\) over \(k\) fits in the following sequence, where \(\overline{X}=X\times_k \overline{k}\). \[ \begin{tikzcd} \mathrm{Br}\, k \arrow[r]{}{\alpha} & \mathrm{Br}\,X \arrow[r]{}{\beta} & \mathrm{Br}\,\overline{X} \end{tikzcd} \] The above sequence gives a filtration of the Brauer group: \(\mathrm{Br}_1\, X\) and \(\mathrm{Br}_0\, X\) denote \(\ker \beta\) and \(\mathrm{im} \alpha\), respectively. Clearly \(\mathrm{Br}_0\, X \subseteq \mathrm{Br}_1\, X \subseteq \mathrm{Br}\, X\) and hence we can consider the following quotients: \(\mathrm{Br}_1\, X/ \mathrm{Br}_0\, X\) and \(\mathrm{Br}\, X/ \mathrm{Br}_1\, X\), called the algebraic and transcendental parts of \(\mathrm{Br}\, X\), respectively. Let \(X\) be a \(K3\) surface defined over a number field \(k\) and assume that \begin{center} the Néron-Severi lattice of \(X\) is isometric to the Néron-Severi lattice of the Kummer surface associated to a product of isogenous (not necessarily full) CM elliptic curves over~\(\overline{k}\). \end{center} Then, in the paper under review, the authors give an explicit upper bound for the size of the transcendental part of the Brauer group of \(X\). The bound is explicit in terms of the discriminant of the Néron-Severi lattice and of \([k:\mathbb{Q}]\). The bound is not expected to be sharp and it can be greatly improved if one allows for some extra assumptions, like: \begin{itemize} \item that \(X\) is isomorphic to the Kummer surface associated to a product of isogenous (not necessarily full) CM elliptic curves, or \item the Generalised Riemann Hypothesis. \end{itemize} Under the GRH, the authors obtain analogous results for \begin{itemize} \item singular \(K3\) surfaces whose Néron-Severi lattice is isometric to the Néron-Severi lattice of the Kummer surface associated to the product of two elliptic curves over \(\overline{k}\), and \item for Kummer surfaces associated to a product of \textit{non}-isogenous (not necessarily full) CM elliptic curves over \(\overline{k}\). \end{itemize} The proof of the main result is based on the study of the transcendental part of the Brauer group of abelian surfaces obtained as product of elliptic curves. Historically, the transcendental part of the Brauer group is the hardest part to compute, making the above result particularly relevant. The explicit nature of the result implies that if \(X\) is a singular \(K3\) surface over \(k\) that is also geometrically Kummer, then the set \(X(\mathbf{A}_k)^{\mathrm{Br}}\) is effectively computable.