scientific article; zbMATH DE number 4025538

From MaRDI portal
Publication:3767418





Let \(K\) be a fixed quadratic number field, and let \(\zeta_K(s)\) be its Dedekind zeta-function, so that \(\zeta_K(s)=\zeta(s)L(s,\chi)\) for an appropriate real Dirichlet character \(\chi\). The total number \(N_K(T)\) of zeros of \(\zeta_K(s)\) with \(0<t\leq T\) is asymptotically \((T/\pi)\log T\). Recently in part I of this paper, the authors [Invent. Math. 86, 563--576 (1986; Zbl 0604.12013)] showed that \(\zeta_K(s)\) has \(\gg T^{6/11}\) simple zeros up to height \(T\). They now prove that, if the Riemann Hypothesis holds for \(\zeta(s)\), then \(\zeta_k(s)\) has asymptotically at least \(1/54\) of its zeros simple. To obtain this result they show, assuming the Riemann Hypothesis, that \(L(s,\chi)\) vanishes for at most \(2/3\) of the zeros of \(\zeta(s)\), asymptotically (here \(\chi\) may be any non-principal character). One then combines this with the result [the authors (to appear)] that at least \(19/27\), asymptotically, of the zeros of \(\zeta(s)\) are simple, if the Riemann Hypothesis holds. The proof uses the quantities \(N=\sum L(\rho,\chi)A(\rho)\) and \(D=\sum | L(\rho,\chi)A(\rho)|^2,\) where \(\rho\) runs over zeros of \(\zeta(s)\) with \(0<\nu \leq T\), and \(A(s)\) is a suitable mollifier for \(L(s,\chi)\). Cauchy's inequality then shows that \(L(\rho,\chi)\) is non-vanishing at least \(| N|^2/D\) times. The bulk of the proof is taken up with the estimation of \(N\) and \(D\), which proceeds along lines laid down by the authors in their work on simple zeros of \(\zeta(s)\) (loc. cit.). For the entire collection see [Zbl 0618.00005].











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3767418)