Miranda-Persson's problem on extremal elliptic \(K3\) surfaces. (Q700634): Difference between revisions

Let \(f:X \to C\) be an elliptic surface over a smooth projective curve \(C\) with a section \(O\). The authors assume in the paper under review that \(f\) has always at least one singular fiber. Let \(\text{MW}(f)\) be the Mordell-Weil group of \(f:X \to C\) and \(R\) be the subgroup of the Néron-Severi group \(\text{NS}(X)\) of \(X\) generated by the section \(O\) and all irreducible components in the fibers of \(f\). Then we have \(\text{MW}(f) \cong \text{NS}(X)/R\). In particular the rank of \(\text{MW}(f)\) is equal to \(\rho(X)- \text{rk}(R)\), where \(\rho(X)=\operatorname{rk} \text{NS}(X)\). The authors call \(f:X \to C\) ``extremal'' if \(\operatorname{rk} \text{MW}(X)=h^{1,1}(X)\) and \(\operatorname{rk} \text{MW}(f)=0\). It follows that \(f:X \to C\) is extremal if and only if \(\operatorname{rk} \text{MW}(X)=\operatorname{rk} R=h^{1,1}(X)\). The main interest of the authors is in the elliptic extremal \(K3\) cases. Mainly this is due to a previous result of \textit{R. Miranda} and \textit{V. Persson} [Math. Z. 201, No. 3, 339--361 (1989; Zbl 0694.14019) and in: Problems in the theory of surfaces and their classification (Cortona, 1988), 167--192, Sympos. Math., XXXII, Academic Press, London (1991; Zbl 0827.14013)] in which a complete classification of all possible configurations of singular fibers and a description of the Mordell-Weil group in terms of the singular fibers was given, except for 17 cases which were not settled. The main interest of the authors is indeed in these 17 cases. They settle this problem completely, i.e. they determine Mordell-Weil groups for the remaining cases. They also consider the uniqueness problem, which is formulated as follows: Let \(f_i:X_i \to \mathbb P^1\) be semi-stable extremal elliptic \(K3\) surfaces such that both \(X_1\) and \(X_2\) have the same configuration of singular fibers and \(\text{MW}(f_1) \cong \text{MW}(f_2)\). Is it true there exists an isomorphism \(\varphi:X_1 \cong X_2\) such that \(\varphi\) preserves the fibrations and the zero section of \(f_1\) maps to the zero section of \(f_2\) under \(\varphi\)? They show for instance that, if for an extremal semi-stable elliptic \(K3\) surface \(\# \text{MW}(f) \geq 4\) holds, then the answer to the previous question is affirmative with only one exceptional case.