Some remarks on Brauer groups of \(K3\) surfaces (Q2570072): Difference between revisions

In this inspired paper the author explores the interplay between Brauer groups, Hodge theory, and genus-one fibrations for complex projective \(K3\) surfaces. The results mainly apply to general elliptic \(K3\) surfaces, which have Picard number two. There are also remarks on the case of Picard number one. The point of departure is the following: Let \(T_X\) be the transcendental lattice of an arbitrary \(K3\) surface \(X\). The classes \(\alpha\) of order \(b\) in the Brauer group of \(X\) might be viewed as surjective homomorphisms \(\alpha:T_X\rightarrow\mathbb{Z}/b\mathbb{Z}\). The kernel of such a homomorphism is a polarized sub-Hodge structure \(T_{\langle\alpha\rangle}\) of finite index inside the transcendental lattice \(T_X\). Any primitive embedding of the lattice \(T_{\langle\alpha\rangle}\) into the standard \(K3\)-lattice yields a new polarized Hodge structure on the standard \(K3\) lattice, whence by the Torelli theorem a new \(K3\) surface \(Y\). Using lattice theoretic arguments, the author observes that if \(X\) is a general elliptic \(K3\) surface, then such embeddings of \(T_{\langle\alpha\rangle}\) always exists. They are, however, not unique in general. The paper contains a detailed analysis of the resulting \(K3\) surfaces \(Y\): The Kähler cone is computed, and from this the automorphism groups, the nature of extremal rays, and projective models are deduced. Furthermore, the author gives some general results on elliptic surfaces \(X\) viewed as a ramified double covering of an elliptic rational surface. In particular, he relates the 2-torsion in the Jacobian of the branch curve to the 2-torsion in the Brauer group of \(X\).