Twisted Fermat curves over totally real fields (Q2494228)

Let \(F\) be a totally real number field of odd degree and \(p\) a prime such that \([F(\mu_p):F]=2\). For \(\delta\in F^*\) consider the plane curve \[ W_\delta/F\, :\, X^p+Y^p=\delta. \] It is clear that \(W_\delta\) only depends on the class of \(\delta\) in \(F^*/F^{*,p}\). The main result of the paper is that there are infinitely many classes \(\delta\in F^*/F^{*,p}\) such that \(W_\delta\) has no \(F\)-rational point. It is pointed out that the result is indeed effective. The main ingredients in the proof are firstly to establish a part of the conjecture of Birch and Swinnerton-Dyer in the rank zero case for the Jacobian of a suitable quotient of \(W_\delta\), this uses Euler systems of Heegner-points the non-triviality of which follows from S. Zhang's generalization of the Gross-Zagier formula. Secondly, the requisite non-vanishing of an \(L\)-series at \(s=1\) is established using double Dirichlet series.