Hostname: page-component-586b7cd67f-l7hp2 Total loading time: 0 Render date: 2024-11-24T01:11:45.989Z Has data issue: false hasContentIssue false

EULER–KRONECKER CONSTANTS FOR CYCLOTOMIC FIELDS

Published online by Cambridge University Press:  30 May 2022

LETONG HONG
Affiliation:
Department of Mathematics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA e-mail: [email protected]
KEN ONO*
Affiliation:
Department of Mathematics, University of Virginia, Charlottesville, VA 22904, USA
SHENGTONG ZHANG
Affiliation:
Department of Mathematics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA e-mail: [email protected]
*
Rights & Permissions [Opens in a new window]

Abstract

The Euler–Mascheroni constant $\gamma =0.5772\ldots \!$ is the $K={\mathbb Q}$ example of an Euler–Kronecker constant $\gamma _K$ of a number field $K.$ In this note, we consider the size of the $\gamma _q=\gamma _{K_q}$ for cyclotomic fields $K_q:={\mathbb Q}(\zeta _q).$ Assuming the Elliott–Halberstam Conjecture (EH), we prove uniformly in Q that

$$ \begin{align*} \frac{1}{Q}\sum_{Q<q\le 2Q} |{\gamma_q - \log q}| = o(\log Q).\end{align*} $$

In other words, under EH, the $\gamma _q /\!\log q$ in these ranges converge to the one point distribution at $1$ . This theorem refines and extends a previous result of Ford, Luca and Moree for prime $q.$ The proof of this result is a straightforward modification of earlier work of Fouvry under the assumption of EH.

Type
Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
© The Author(s), 2022. Published by Cambridge University Press on behalf of Australian Mathematical Publishing Association Inc.

1 Introduction

For a number field K, the Euler–Kronecker constant $\gamma _K$ is given by

$$ \begin{align*} \gamma_K:=\lim_{s \to 1^+}\bigg(\frac{\zeta_K'(s)}{\zeta_K(s)} + \frac{1}{s-1}\bigg), \end{align*} $$

where $\zeta _K(s)$ is the Dedekind zeta-function for $K.$ The Euler–Mascheroni constant $\gamma =0.5772\ldots \!$ is the $K={\mathbb Q}$ case, where $\zeta _{{\mathbb Q}}(s)=\zeta (s)$ is the Riemann zeta-function. We consider the constants $\gamma _q=\gamma _{K_q}$ for cyclotomic fields $K_q:={\mathbb Q}(\zeta _q)$ , where $q\in {\mathbb Z}^{+}$ and $\zeta _q$ is a primitive qth root of unity.

The recent interest in the distribution of the $\gamma _q$ is inspired by work of Ihara [Reference Ihara and Ginzburg4, Reference Ihara and Rodier5]. He proposed, for every $\varepsilon> 0$ , that there is a $Q (\varepsilon )$ for which

$$ \begin{align*} (c_1-\varepsilon)\log q\le \gamma_q\le (c_2+\varepsilon)\log q \end{align*} $$

for every integer $q \ge Q (\epsilon )$ , where $0 < c_1 \le c_2 < 2$ are absolute constants. This conjecture was disproved by Ford et al. in [Reference Ford, Luca and Moree2] assuming a strong form of the Hardy–Littlewood k-tuple conjecture. However, assuming the Elliott–Halberstam conjecture (see [Reference Elliott and Halberstam1]), these same authors also proved that the conjecture holds for almost all primes $q,$ with $c_1 = c_2 = 1.$ We recall the Elliott–Halberstam Conjecture as formulated in terms of the Von Mangoldt function $\Lambda (n),$ the Chebyshev function $\psi (x)$ and Euler’s totient function $\varphi (n).$

Elliott–Halberstam Conjecture (EH).

If we let

$$ \begin{align*}E(x;m,a):=\sum_{\substack{p\equiv a \pmod m\\ p\leq x\ {\mathrm{prime}}}} \Lambda(p) -\frac{\psi(x)}{\varphi(m)},\end{align*} $$

then for every $\varepsilon>0$ and $A>0$ , we have

$$ \begin{align*}\sum_{m\le x^{1-\varepsilon}}\max_{(a,m) = 1}\lvert{E(x;m,a)}\rvert\ll_{A,\varepsilon} \frac{x}{(\log x)^A}.\end{align*} $$

Assuming EH, Ford et al. proved (see [Reference Ford, Luca and Moree2, Theorem 6(i)]), for every $\varepsilon> 0$ , that

$$ \begin{align*}1-\varepsilon < \frac{\gamma_q}{\log q} <1+\varepsilon \end{align*} $$

for almost all primes q (that is, the number of exceptional $q \le x$ is $o(\pi (x))$ as $x\to \infty $ ). Here we extend and refine this result to all integers $q.$

Theorem 1.1. Under EH, for $Q\rightarrow +\infty $ , we have

$$ \begin{align*} \frac{1}{Q}\sum_{Q<q\le 2Q} \lvert{\gamma_q - \log q}\rvert = o(\log Q), \end{align*} $$

where the sum is over integers q.

Remark 1.2. Theorem 1.1 shows that EH implies that the distribution of $\gamma _q /\!\log q$ in $[Q, 2Q]$ converges to the one point distribution supported on $1$ .

To prove Theorem 1.1, we use the work of Fouvry [Reference Fouvry3] that allowed him to unconditionally prove that

$$ \begin{align*}\frac{1}{Q}\sum_{Q<q\le 2Q}\gamma_q=\log Q+O(\log\log Q).\end{align*} $$

Our conditional result is a point-wise refinement of Fouvry’s asymptotic formula under EH.

2 Proof of Theorem 1.1

For brevity, we shall assume that the reader is familiar with Fouvry’s paper [Reference Fouvry3]. The key formula is (see (3) of [Reference Fouvry3]) the following expression for $\gamma _q$ in terms of logarithmic derivatives of Dirichlet L-functions:

(2.1) $$ \begin{align} \gamma_q = \gamma + \sum_{1<q^*|q\ }\sum_{\chi^*\!\bmod q^*}\frac{L'(1,\chi^*)}{L(1,\chi^*)}. \end{align} $$

Here the inner sum runs over the primitive Dirichlet characters $\chi ^*$ modulo $q^*$ .

We follow the strategy and notation in [Reference Fouvry3], which makes use of the modified Chebyshev function

$$ \begin{align*} \psi(x;q,a):=\sum_{\substack{n\leq x\\ n\equiv a \pmod q}} \Lambda(n), \end{align*} $$

and the integral

$$ \begin{align*} \Phi_{\chi^*}(x):=\frac{1}{x-1}\int_{1}^x \bigg(\sum_{n\leq t}\frac{\Lambda(n)}{n}\chi^*(n)\bigg) \,dt. \end{align*} $$

However, we replace the sums $\Gamma _i(Q)$ and $\Gamma _{1,j}(Q)$ defined in [Reference Fouvry3] with the pointwise terms $\gamma _i(q)$ and $\gamma _{1,j}(q)$ . Following the approach in [Reference Fouvry3], which is based on (2.1), we have

$$ \begin{align*}\gamma_q = \gamma + A(q) + B(q) - \gamma_2(q) - \gamma_3(q) - (\gamma_{1,1}(q) + \gamma_{1,2}(q) + \gamma_{1,3}(q)),\end{align*} $$

where

$$ \begin{align*}A(q)&=\sum_{q^*|q} \,{\sum_{\chi^*\mod q^*}} \frac{L'}{L}(1,\chi^*)+\Phi_{\chi^*}(x), \\B(q)&=\sum_{\substack{\chi \bmod q\\ \chi\neq \chi_0}} \Phi_{\chi}(x)-\sum_{q^*|q}\,{\sum_{\chi^*\mod q^*}} \Phi_{\chi^*}(x), \\\gamma_2(q)&=\frac{1}{x-1}\int_1^x \frac{\varphi(q)\psi(t;q,1)-\psi(t)}{t} \,dt, \\\gamma_3(q)&=\frac{1}{x-1}\int_1^x \sum_{\substack{n\leq t\\(n,q) \neq 1}} \frac{\Lambda(n)}{n}\,dt, \\\gamma_{1,1}(q)&=\frac{1}{x-1}\int_1^x\int_1^{\min(q,t)}\bigg(\frac{\varphi(q)\psi(u;q,1)-\psi(u)}{u^2}\,du\bigg)\,dt, \\\gamma_{1,2}(q)&=\frac{1}{x-1}\int_1^x\int_{\min(q,t)}^{\min(x_1,t)}\bigg(\frac{\varphi(q)\psi(u;q,1)-\psi(u)}{u^2}\,du\bigg)\,dt, \\\gamma_{1,3}(q)&=\frac{1}{x-1}\int_1^x\int_{\min(x_1,t)}^t \bigg(\frac{\varphi(q)\psi(u;q,1)-\psi(u)}{u^2}\,du\bigg)\,dt. \end{align*} $$

To complete the proof, for $\varepsilon>0$ , we let $x := q^{100}$ and $x_1 := q^{1 + \varepsilon }$ . Apart from $\gamma _{1,1}(q),$ which gives the $-\log q$ terms in Theorem 1.1, we shall show that these summands are all small.

Estimation of $A(q)$ : By Proposition 1 and Remark (i) of [Reference Fouvry3],

$$ \begin{align*}\sum_{q=Q}^{2Q} \lvert{A(q)}\rvert =O(Q).\end{align*} $$

Estimation of $B(q)$ : For $B(q)$ , by (26) and Lemma 3 of [Reference Fouvry3], we simplify

$$ \begin{align*}B(q)&=-\frac{1}{x-1}\int_{1}^x \sum_{q^*|q}\ \sum_{\chi^*\,\mod q^*} \sum_{\substack{n \leq t\\(n,q)>1}} \frac{\Lambda(n)\chi^*(n)}{n} \,dt \\&= -\frac{1}{x-1}\int_{1}^x \sum_{q^*|q}\ \sum_{\chi^*\,\mod q^*}\ \sum_{\substack{p^v\leq t\\ p|q}} \frac{\log p\cdot\chi^*(p^v)}{p^v} \,dt \\&= -\frac{1}{x-1}\int_1^x \sum_{q^*|q}\ \sum_{\substack{p^v\leq t \\ p|q\\ p\nmid q^*}}\ \sum_{\substack{d|(p^v-1,q^*)}}\frac{\log p}{p^v} \cdot \varphi(d) \mu\bigg(\frac{q^*}{d}\bigg) \,dt \\&= -\frac{1}{x-1}\int_1^x \sum_{\substack{p^v\leq t\\ p|q}}\ \sum_{\substack{d|p^v-1}}\frac{\log p}{p^v}\cdot\varphi(d) \sum_{\substack{q^*|q\\d|q^*\\ p\nmid q^*}}\mu\bigg(\frac{q^*}{d}\bigg) \,dt. \end{align*} $$

We note that the innermost sum

$$ \begin{align*}\sum_{\substack{q^*|q\\ d|q^* \\ p\nmid q^*}}\mu\bigg(\frac{q^*}{d}\bigg)\end{align*} $$

is always $0$ or $1$ , so we conclude that $B(q) \leq 0$ for any q. Proposition 2 of [Reference Fouvry3] gives

$$ \begin{align*}\sum_{q=Q}^{2Q}B(q)=O(Q),\end{align*} $$

and so we have

$$ \begin{align*}\sum_{q=Q}^{2Q}\lvert{B(q)}\rvert =O(Q).\end{align*} $$

Estimation of $\gamma _2(q)$ : By Lemma 8 of [Reference Fouvry3], uniformly in Q with $u\geq 1,$ we have

$$ \begin{align*}\sum_{q=Q}^{2Q} \psi(u;q,1) \ll u.\end{align*} $$

Therefore,

$$ \begin{align*}\sum_{q=Q}^{2Q} \lvert{\varphi(q)\psi(t;q,1)-\psi(t)}\rvert =O(Qt),\end{align*} $$

and so we conclude that

$$ \begin{align*}\sum_{q=Q}^{2Q} \lvert{\gamma_{2}(q)}\rvert = O(Q).\end{align*} $$

Estimation of $\gamma _3(q)$ : By definition, $\gamma _3$ is positive, so by (36) of [Reference Fouvry3],

$$ \begin{align*}\sum_{q=Q}^{2Q} \lvert{\gamma_{3}(q)}\rvert =O(Q).\end{align*} $$

Estimation of $\gamma _{1,1}(q)$ : Since $\psi (u;q,1)=0$ for $u<q$ , we have

$$ \begin{align*} \gamma_{1,1}(q)=-\frac{1}{x-1}\int_1^x \bigg(\int_1^{\min(q,t)}\frac{\psi(u)}{u^2} \,du \bigg) \,dt. \end{align*} $$

Dividing both sides of (41) of [Reference Fouvry3] by Q,

$$ \begin{align*}\gamma_{1,1}(q)=-\log q + O(1).\end{align*} $$

Estimation of $\gamma _{1,2}(q)$ : By the same proof as (42) of [Reference Fouvry3], we have

$$ \begin{align*} \sum_{q=Q}^{2Q} \lvert{\gamma_{1,2}(q)}\rvert \ll \varepsilon Q\log Q. \end{align*} $$

Summing the above estimates, we conclude unconditionally that

$$ \begin{align*} \frac{1}{Q}\sum_{q=Q}^{2Q} \lvert{\gamma_q-\log q}\rvert =\frac{1}{Q}\sum_{q=Q}^{2Q} \lvert{\gamma_{1,3}(q)}\rvert + O(\varepsilon \log Q). \end{align*} $$

Estimation of $\gamma _{1,3}(q)$ : If we assume Conjecture EH holds, then we have (as in Lemma 7 of [Reference Fouvry3]) that

$$ \begin{align*} \sum_{\substack{q\le 2Q\\ (q,a)=1}}\varphi(q)\bigg\vert{\psi(x;q,a)-\frac{\psi(x)}{\varphi(q)}}\bigg\vert=O_A\big(Qx(\log x)^{-A+2}\big). \end{align*} $$

Therefore,

$$ \begin{align*} \frac{1}{Q} \sum_{q=Q}^{2Q}\lvert{\gamma_{1,3}(q)}\rvert=O_{\epsilon,A}(\log^{-A} Q). \end{align*} $$

By combining these estimates, we obtain the main result

$$ \begin{align*} \frac{1}{Q}\sum_{q=Q}^{2Q}\lvert{\gamma_q-\log q}\rvert=o(\log Q). \end{align*} $$

Acknowledgements

The authors thank Pieter Moree for helpful discussions regarding his work with Ford and Luca. We thank the referee for suggestions that improved the exposition in this paper.

Footnotes

The authors acknowledge the Thomas Jefferson Fund and the NSF (DMS-2002265 and DMS-2055118) for their generous support.

References

Elliott, P. D. T. A. and Halberstam, H., ‘A conjecture in prime number theory’, Symp. Math. 4 (1969), 5971.Google Scholar
Ford, K., Luca, F. and Moree, P., ‘Values of the Euler $\varphi$ -function not divisible by a given odd prime, and the distribution of Euler–Kronecker constants for cyclotomic fields’, Math. Comput. 83(287) (2014), 14471476.CrossRefGoogle Scholar
Fouvry, É., ‘Sum of Euler–Kronecker constants over consecutive cyclotomic fields’, J. Number Theory 133(4) (2013), 13461361.CrossRefGoogle Scholar
Ihara, Y., ‘On the Euler–Kronecker constants of global fields and primes with small norms’, in: Algebraic Geometry and Number Theory: In Honor of Vladimir Drinfeld’s 50th Birthday, Progress in Mathematics, 253 (ed. Ginzburg, V.) (Birkhäuser Boston, MA, 2006), 407451.CrossRefGoogle Scholar
Ihara, Y., ‘The Euler–Kronecker invariants in various families of global fields’, in: Proceedings of Arithmetic Geometry and Coding Theory (AGCT 2005), Séminaires et Congrès 21 (ed. Rodier, F.) (Soc. Math. France, Paris, 2009), 79102.Google Scholar