Where can one find a Vinogradov-Korobov zero-free region for Dirichlet L-functions? It has to be in a standard reference, but I'm having a non-trivial time finding it.
3 Answers
My paper http://arxiv.org/abs/2210.06457 establishes several explicit Vinogradov--Korobov type zero-free regions for Dirichlet $L$-functions. In particular, Theorem 1.1 states the following:
Let $q \geq 3$, and let $\chi\pmod{q}$ be a Dirichlet character. The Dirichlet $L$-function $L(\sigma+it,\chi)$ does not vanish in the region \begin{equation}\label{smallt} \sigma \geq 1-\frac{1}{10.5 \log q+61.5(\log |t|)^{2 / 3}(\log \log |t|)^{1 / 3}}, \quad|t| \geq 10. \end{equation} Also, there exists an absolute and effectively computable constant $Y > 0$ such that $L(\sigma+it,\chi)$ does not vanish in the region \begin{equation}\label{larget} \sigma \geq 1- \frac{1}{ 10.1\log q + 49.13(\log |t|)^{2/3}(\log\log|t|)^{1/3}},\qquad |t| \geq Y. \end{equation}
Additionally, a short, self-contained proof of the Vinogradov--Korobov zero-free region for Dirichlet $L$-functions (assuming an explicit upper bound for the Hurwitz zeta function $\zeta(s,u)$ proven by Ford, and a convenient form of Jensen's formula from complex analysis) is provided in the appendix of the paper (albeit with worse constants than Theorem 1.1 above).
Languasco and Zaccagnini cite Prachar, and mention Vasil'kovskaja. https://core.ac.uk/reader/81188410 https://doi.org/10.1016/j.jnt.2006.12.015
They give a zero-free region in Lemma 1 (suppressing details).
See also this question from a decade ago. Mertens-like sum in arithmetic progressions
And this one from last year (answered by Languasco): Error term in Mertens' third theorem
-
$\begingroup$ Thanks. Prachar does give a zero-free region (with proof), in Satz 6.2 of chapter VIII. That would seem sufficient. I haven't checked Vasil'kovskaja. $\endgroup$ Commented Feb 28, 2021 at 8:21
In Theorem 2 of Mark Coleman's paper below, a full proof of the Vinogradov-Korobov zero-free region for the $L$-function of a Grossencharacter twisted by a Hecke character over an arbitrary number field is provided. Restricting to $K=\mathbb{Q}$, one recovers the result for Dirichlet $L$-functions. This is the only place where I have seen a full proof of such a zero-free region published.
Coleman, M. D., A zero-free region for the Hecke L-function, Mathematika 37, No. 2, 287-304 (1990). ZBL0721.11050.
-
$\begingroup$ Thanks! As mentioned in the comments to the other answer, Prachar also gives a full proof for Dirichlet $L$-functions (or at least it looks full to me). $\endgroup$ Commented Mar 1, 2021 at 21:40
-
$\begingroup$ @HAHelfgott I just looked in Prachar. He doesn't prove the full Vinogradov-Korobov zero-free region (a weaker result is proved that still beats the standard zero-free region). If you look at the paper listed by user174996, the pages in Prachar (p. 80-81) that are referenced do not lie in the chapter with Vinogradov's method. $\endgroup$ Commented Mar 1, 2021 at 23:13
-
2$\begingroup$ @HAHelfgott To be clear, equation 6.4 on p. 295 is not considered to be THE Vinogradov-Korobov zero-free region. Prachar proves $M(k,t) = \max\{\log k, (\log(|t|+3)^{3/4} (\log\log(|t|+3))^{3/4}\}$ but one can prove $M(k,t) = \max\{\log k, (\log(|t|+3)^{2/3} (\log\log(|t|+3))^{1/3}\}$ $\endgroup$ Commented Mar 1, 2021 at 23:18
-
$\begingroup$ Odd - I should have noticed that. It would also be very odd if there is no source older than Coleman that proves the full Vinogradov-Korobov zero-free region for Dirichlet L-functions. $\endgroup$ Commented Mar 1, 2021 at 23:26
-
$\begingroup$ Well, Coleman cites Mitsui and Sokolovsky as having proved (separately) a Vinogradov-Korobov-type zero-free region for Dedekind zeta functions, and that should cover the case of a Dirichlet L-function. It would still be odd if it weren't all in some textbook, though. $\endgroup$ Commented Mar 1, 2021 at 23:31