Gaps between zeros of Dedekind zeta-functions of quadratic number fields (Q489047): Difference between revisions

Let \(K\) be a quadratic number field of discriminant \(d\), and let \(\zeta_K(s)\) be its Dedekind zeta-function. Let \(0<\gamma_1\leq\gamma_2\leq \ldots\) be the sequence of consecutive ordinates of nontrivial zeros of \(\zeta_K(s)\) and put \[ \mu_K=\liminf_{n\to\infty}{(\gamma_{n+1}-\gamma_n)\log(\sqrt{|d|}\gamma_n)\over\pi}, \] \[ \lambda_K=\limsup_{n\to\infty}{(\gamma_{n+1}-\gamma_n)\log(\sqrt{|d|}\gamma_n)\over\pi}. \] It is conjectured that \(\mu_K=0\) and \(\lambda_K=\infty\). The author shows (Corollary 2) that under the Generalized Riemann Hypothesis for \(\zeta_K(s)\) one has \(\lambda_K\geq6\). This is obtained as a corollary to the following result (Theorem 1): If \(\varepsilon>0\), \(T\geq2\) and \(|d|\leq T^{7/9-\varepsilon}\), then the interval \((T,2T)\) contains a sub-interval \(I\) of length at least equal to \[ {\sqrt 6\pi\over\log\sqrt{|d|}T}\left(1+O\left(|d|^\varepsilon/\log T\right)\right), \] such that \(\zeta_K(1/2+it)\neq0\) for \(t\in I\). The proof of this theorem is based on an asymptotical formula for the moments \[ \int_T^{2T}\zeta_K^{(\mu)}(1/2+it)\zeta_K^{(\nu)}(1/2-it)dt \] (\(\mu, \nu\geq0\), integers), which in turn is deduced from the asymptotic formula for \[ \int_T^{2T} \zeta_K(1/2+it+\alpha)\zeta_K(1/2-it+\beta) \] (with complex \(\alpha,\beta\) with \(|\alpha|,|\beta|=O(1/\log(|d|T)))\), proved by \textit{W. Heap} [Int. J. Number Theory 10, No. 1, 235--281 (2014; Zbl 1314.11054)].