\(K3\) surfaces with maximal finite automorphism groups containing \(M_{20}\) (Q2073612)

Mukai proved that if \(G\) is a finite group acting faithfully and preserving the symplectic form on a \(K3\) surface, then there is an upper bound for the order of \(G\), namely \(|G| \leq 960\). Moreover he proved that if \(|G|=960\) then \(G\) is isomorphic to the Mathieu group denoted by \(M_{20}\). The authors define a \textit{maximal finite group} \(\Gamma\) acting faithfully on a \(K3\) surface surface as a group such that if \(\Gamma'\) is another finite group acting faithfully on a \(K3\) surface, then there are no proper subgroups of \(\Gamma'\) isomorphic to \(\Gamma\). It is an interesting purpose classifying maximal finite groups acting faithfully on a \(K3\) surface. In this direction \textit{S. Kondō} [Am. J. Math. 121, No. 6, 1245--1252 (1999; Zbl 0978.14043)] showed that the maximum order of a finite group acting faithfully on a \(K3\) surface is 3840 and this group contains \(M_{20}\) as a subgroup of index four. The main result of this paper is Theorem 6.3 where the authors prove that there are only three finite groups \(\Gamma\) that contain strictly \(\Gamma_0 = M_{20}\) as the normal subgroup of \(\Gamma\) acting faithfully and symplectically, and only three \(K3\) surfaces acted on by such a \(\Gamma\). This result is achieved due to the description of the possible transcendental lattices for such \(K3\) surfaces, and this is the content of Section 2 and in particular of Theorem 2.7. This result was obtained independently in a paper of \textit{S. Brandhorst} and \textit{K. Hashimoto} [Ann. Henri Lebesgue 4, 785--809 (2021; Zbl 1485.14061)] where they compute all the finite groups acting faithfully on \(K3\) surfaces with maximal symplectic part. Moreover among the three \(K3\) surfaces described in the paper, one was constructed by Kondō and it is denoted by \(X_{Ko}\), the second one was constructed by Mukai and it is denoted by \(X_{Mu}\), and the existence of the last one was shown by Brandhorst-Hashimoto and it is denoted by \(X_{BH}\). In the first case \(M_{20}\) is a subgroup of index four and in the other cases \(M_{20}\) is a subgroup of index two. The transcendental lattice of \(X_{Ko}\) is already computed by \textit{S. Kondō} [Am. J. Math. 121, No. 6, 1245--1252 (1999; Zbl 0978.14043)], whereas the other two transcendental lattices are not computed before, although that of \(X_{Mu}\) was already known by Mukai. Moreover the authors compute the equations for the three surfaces, namely \(X_{Mu}\) was already computed by Mukai as a smooth quartic surface in \(\mathbb{P}^{3}(\mathbb{C})\), for \(X_{Ko}\) they give a (singular) equation as a complete intersection of two quartics in weighted projective space \(\mathbb{P}(1,1,2,2,2)\), and \(X_{BH}\) is a smooth complete intersection of three quadrics in \(\mathbb{P}^{5}(\mathbb{C})\). Moreover the authors prove in Corollary 2.5 that all these three \(K3\) surfaces are Kummer surfaces of abelian surfaces that are the product of two elliptic curves. Section 3, 4 and 5 are devoted to treat the Kondō's example, the Mukai's example and the Brandhorst-Hashimoto's example respectively, Section 6 contains the main theorem i.e. Theorem 6.3 and other final remarks.