On the motive of Kummer varieties associated to \(\Gamma_1(7)\). Supplement to the paper ``The modularity of certain non-rigid Calabi-Yau threefolds'' by R. Livné and N. Yui (Q2494632)

This article serves as a supplement to the paper ``The modularity of certain non-rigid Calabi-Yau threefolds'' by \textit{R. Livné} and \textit{N. Yui} [J. Math. Kyoto Univ. 45, No. 4, 645--665 (2005; Zbl 1106.14025)]. In that paper they consider Calabi-Yau threefolds which possess a non-isotrivial semistable \(K3\)-fibration with 6 singular fibers, the minimal number of singular fibers of such a fibration [\textit{X. Sun, S.-L. Tan} and \textit{K. Zuo}, Math. Res. Lett. 10, No. 2--3, 323--342 (2003; Zbl 1100.14004)] by the Arakelov-Yau bound. In this case the base curve must be a modular curve. The authors start with the list of 9 torsion-free genus 0 congruence subgroups of \(\text{PSL}(2,\mathbb{Z})\) of index 24 and lift them to \(\text{SL}(2,\mathbb{Z})\). In four of these examples, the lifted group \(\tilde{\Gamma}\) contains \(\Gamma(2)\). In these four cases, the authors construct Calabi-Yau threefolds using a Kummer construction and show that the resulting threefolds are associated to modular forms. In the present paper, the authors construct Calabi-Yau threefolds in the remaining five cases. The authors concentrate on the \(\Gamma_1(7)\) case, although their methods generalize to the other cases. The starting point is the universal elliptic curve \(Y(\Gamma_1(7))\) with a group of sections of order 7. The authors show that the curve \(B\) of non-zero 2-torsion points is an elliptic curve defined over \(\mathbb{Q}\) with conductor 14; its \(L\)-series is the series whose inverse Mellin transform is the weight 2 modular form \(\eta(\tau) \eta(2\tau) \eta(7\tau) \eta(14\tau)\), where \(\eta\) is the Dedekind function. The Calabi-Yau threefold \(X\) the authors construct is obtained by taking the product of \(Y(\Gamma_1(7))\) with an elliptic curve, quotienting out by the fibrewise involution \((-1)\) and blowing up the double point locus. They then show that the \(L\)-series of \(H^3(X)\) is given by \(L(g_3 \otimes g_2^E,s) L(g_2^E, s-1)^9 L(g_2^B, s-1)^4\), where \(g_2^E\) and \(g_2^B\) are the weight 2 modular forms associated to the elliptic curves \(E\) and \(B\) respectively and \(g_3\) is the weight 3 modular form associated to the lattice of transcendental cycles in \(Y(\Gamma_1(7))\). The verification of the \(L\)-series of \(X\) can also be done independently by counting points on \(X\) defined over finite fields and using the Lefschetz fixed point theorem applied to the Frobenius automorphism.