On the number of integral ideals in a number field (Q2674322): Difference between revisions

Let \(K\) be an algebraic number field of degree \(n \geqslant 2\), signature \((r_1,r_2)\), discriminant \((-1)^{r_2}d_K\), and let \(\zeta_K(s)\) be the Dedekind zeta function attached to \(K\) with residue at \(s=1\) denoted by \(\kappa_K\). If \(r_K(m)\) is the \(m\)th coefficient of \(\zeta_K(s)\), then the Ideal Theorem is the study of the error term \(\Delta_K (x)\) defined by \[ \Delta_K (x) := \sum_{m \leqslant x} r_K(m) - \kappa_K x \] for \(x > 0\). Estimates for this error term go back to the late of the \(19\)th century, when \textit{H. Weber} [Lehrbuch der Algebra. In zwei Bänden. 2. Band. Braunschweig: F. Vieweg und Sohn (1896; JFM 27.0056.01)] has shown that \(\Delta_K (x) \ll_K x^{1-1/n}\). This result was soon improved by \textit{E. Landau} [Einführung in die elementare und analytische Theorie der algebraischen Zahlen und Ideale. New York: Chelsea Publishing Company (1949; Zbl 0045.32202)] who used techniques from complex analysis to derive the bound \(\Delta_K (x) \ll_K x^{1-\frac{2}{n+1}}\). Landau's result was subsequently somewhat improved by \textit{W. G. Nowak} [Math. Nachr. 161, 59--74 (1993; Zbl 0803.11061)], \textit{H. Lao} [Chin. Ann. Math., Ser. B 31, No. 3, 385--392 (2010; Zbl 1221.11223)] and \textit{B. Paul} and \textit{A. Sankaranarayanan} [J. Number Theory 215, 98--119 (2020; Zbl 1448.11203)], but there are still quite few explicit results in the literature. \textit{J. S. Sunley} [Trans. Am. Math. Soc. 175, 209--232 (1973; Zbl 0289.12010)] announced a result from her thesis which is a completely explicit version of the Ideal Theorem by giving the implied constant depending on the usual invariants of \(K\). In the paper under review, the author improves Sunley's result by proving that, for all \(x > 0\), we have \[ \left| \Delta_K (x) \right| \leqslant C_K \, d_K^{\frac{1}{n+1}} (\log d_K)^{n-1} x^{1-\frac{2}{n+1}}, \] where \[ C_K := \frac{0.54 \, (3n-1) \rho_K(n)}{(n-1)^2 (\log m_K)^{n-1}} \times n^{3/2} \times n!, \] with \(m_K = (\pi/4)^{n} (n^n/n!)^{2}\) and, if \(n > 13\) \[ \rho_K(n) := (n+1)^{n-\frac{1}{2} - \frac{1}{2n}} \left( \tfrac{5}{8} + \tfrac{\pi}{2} + \tfrac{1}{n} + \tfrac{3}{8n^2} \right)^{1/2} e^{4.13 n + \frac{0.02}{n}}. \] Note that the author also gives the value of \(\rho_K(n)\) in the range \(2 \leqslant n \leqslant 13\). This interesting result may have several applications, such as getting upper bound for the class number of \(K\), or improving Mertens' theorems for number fields as established in [\textit{S. R. Garcia} and \textit{E. S. Lee}, Ramanujan J. 57, No. 3, 1169--1191 (2022; Zbl 1497.11278)]. The proofs essentially follows the ideas of Sunley's paper, but the author updates the key ingredients in the method and implements modern knowledge about the invariants of the Dedekind zeta function.