The transcendental lattice of the sextic Fermat surface (Q2511479)

Let \(S\) be a smooth projective surface defined over the complex numbers. The lattice \(A_S=H^{1,1}(S)\cap H^2(S,\mathbb{Z})/(\text{torsion})\) is called the algebraic part of \(H^2(S,\mathbb{Z})\) and its orthogonal \(A_S^{ \bot} \) with respect to the usual bilinear form \(Q\) on \(H^2(S,\mathbb{Z})\) given by cup product is called the transcendental part \(T_S\) of \(H^2(S,\mathbb{Z})\). \(T_S\) has a natural polarized integral Hodge structure. \textit{V. S. Kulikov} [Math. Notes 83, No. 1, 57--64 (2008; Zbl 1147.14018); translation from Mat. Zametki 83, No. 1, 61--68 (2008)] conjectured that for every smooth projective surface \(S\) defined over the complex numbers, \(T_S\) is idecomposable. This conjecture implies that the very general cubic fourfold is irrational. In this paper it is proved that if \(S\subset \mathbb{P}^3\) is the sextic Fermat surface given by \(x_0^6+x_1^6+x_2^6+x_3^6=0\), then the polarized integral structure on the transcendental part \(T_S\) of \(H^2(S,\mathbb{Z})\) is decomposable. This result gives a counterexample to the aforementioned conjecture of V. Kulikov.