Zeros of partial sums of the Dedekind zeta function of a cyclotomic field (Q2637182)

Let \(K\) be an algebraic number field of a degree \(n_0=r_1 + 2r_2\) over \(\mathbb{Q}\), let \(a(n)\) be the number of integral ideals of \(K\) with the norm equal to \(n\) and let \[ \zeta_K(s)= \sum_{n=1}^{\infty} \frac{a(n)}{n^s}, \] be the corresponding Dedekind zeta function. The approximate functional equation for the function \(\zeta_K(s)\), proved by \textit{K. Chandrasekharan} and \textit{R. Narasimhan} [Math. Ann. 152, 30--64 (1963; Zbl 0116.27001)] states that for \(s=\sigma + it\) with large \(t\) \[ \zeta_K(s)= \sum_{n \leq X} \frac{a(n)}{n^s} + B^{2s-1}\frac{A(1-s)}{A(s)}\sum_{n \leq Y} \frac{a(n)}{n^{1-s}} + O(X^{1-\sigma-1/n_0} \log X), \] where \(A(s)=(\Gamma(s/2))^{r_1} (\Gamma(s))^{r_2}\), \(B=2^{r_2} \pi^{n_0/2}/ \sqrt{|d_K|},\) \(X>H>0\), \(Y>H>0\), \(XY= |d_K|(|t|/2\pi)^{n_0}\), \(C_1<X/Y <C_2\) for some constants \(C_1, C_2 >0\) and \(d_K\) is the discriminant of \(K\). Therefore, it is of interest to investigate properties of the partial sums of the function \(\zeta_K(s)\), i.e. properties of the function \[ \zeta_{K,X}(s):= \sum_{n\leq X} \frac{a(n)}{n^s} \] that appears in the approximate functional equation above. The authors prove that for an arbitrary number field \(K\) of a degree \(n_0\) over \(\mathbb{Q}\) and \(X>2\) there exist real numbers \(\alpha\) and \(\beta\) with \(\alpha\) depending only on \(n_0\) and \(X\), and \(\beta\) depending only on \(n_0\), such that the zeros of \(\zeta_{K,X}(s)=\zeta_{K,X}(\sigma+it)\) all lie within the rectilinear strip in the complex plane given by \(\alpha <\sigma< \beta\). An admissible choice for \(\alpha\) and \(\beta\) is given in the paper. Furthermore, in the special case of a cyclotomic field \(K=\mathbb{Q}(\zeta_q)\), where \(\zeta_q\) is primitive \(q\)th root of unity, for \(T,X\geq 3\) the asymptotic formula \[ N_{K,X}(T)=\frac{T}{2\pi} \log N + O_q\left( X\left( \frac{\log\log X}{\log X}\right) ^{1-1/\phi(q)}\right) \] for the number \(N_{K,X}(T)\) of non-real zeros of \(\zeta_{K,X}\) with imaginary parts in the interval \([0,T]\) is proved. Here, \(N\) is the largest integer less than or equal to \(X\) such that \(a(N)\neq 0\) and \(\phi\) denotes the Euler totient function. In the special case of the Riemann zeta function, the above asymptotic formula is sharper than the one proved in [\textit{S. M. Gonek} and \textit{A. H. Ledoan}, Int. Math. Res. Not. 2010, No. 10, 1775--1791 (2010; Zbl 1198.11075)].