Norm-Euclidean Galois fields and the generalized Riemann hypothesis (Q713181)

Let \(K\) be an abelian extension with prime degree \(\ell\) and conductor \(f\). Assuming the Generalized Riemann Hypothesis, the author shows that if \(f\) is sufficiently large (more precisely: if \(f > 38 (\ell-1)^2 (\log f)^6 \log \log f\)), then \(K\) is not norm-Euclidean. As a corollary he obtains the result that the only norm-Euclidean cyclic cubic number fields are all have conductor \(\leq 157\) (here GRH is needed only for excluding conductors between \(10^{11}\) and \(10^{70}\) (by results obtained by the author in [Int. J. Number Theory 8, No. 1, 227--254 (2012; Zbl 1290.11147)], which suggesst that the complete proof of the conjecture that \(f \leq 157\) is now just a matter of time). The Riemann hypothesis is needed for proving, given a real Dirichlet character \(\chi\) defined modulo \(m \geq 10^9\), the existence of primes \(q_1 < q_2 < 2.5 (\log m)^2\) with \(\chi(q_1)\), \(\chi(q_2) \neq 1\), and of an integer \(r < 2.5 (\ell-1)^2 (\log f)^2\) coprime to \(q_1q_2\) with prescribed Dirichlet character \(\chi(r)\).