Search arXivSearch

arXiv · 2505.07803

Log-free bounds on exponential sums over primes

Abstract

We establish completely log-free bounds for exponential sums over the primes and the Möbius function. Let $0<η\leq 1/10$, and suppose $α= a/q + δ/x$, with $(a,q)=1$ and $|δ| \leq x^{1/5 + η}/q$, and set $δ_0 = \max(1, |δ|/4)$. For $x \geq x_0(η)$ sufficiently large, we show that: \begin{equation*} \Biggl| \sum_{n \leq x} Λ(n) e(nα) \Biggr| \leq \frac{q}{φ(q)} \frac{\mathscr{F}_η\bigl( \frac{\log δ_0 q}{\log x}, \frac{\log^+ δ_0/q}{\log x} \bigr) \cdot x }{\sqrt{δ_0 q}} \ \text{ and } \ \Biggl| \sum_{n \leq x} μ(n) e(nα) \Biggr| \leq \frac{\mathscr{G}_η\bigl( \frac{\log δ_0 q}{\log x}, \frac{\log^+ δ_0/q}{\log x} \bigr) \cdot x}{\sqrt{δ_0 φ(q)}}, \end{equation*} for all $1 \leq q \leq x^{2/5 - η}$, where $\log^+ z = \max(\log z, 0)$, and the functions $\mathscr{F}_η$ and $\mathscr{G}_η$ are explicitly determined, taking small to moderate values. These bounds improve substantially upon the existing results - particularly with respect to the permissible ranges of $q$, $δ$ in which log-free bounds are known to hold and potentially with respect to asymptotic functions $\mathscr{F}_η$ and $\mathscr{G}_η$ as well. Moreover, the range $1 \leq q \leq x^{2/5 - η}$ is essentially the best possible we can expect. The main innovation is a sieve-weighted version of Vaughan's identity (Lemma 2.1), which is effectively log-free. We employ several ideas and results from the pioneering work of Helfgott, and particularly, they play a central role in ensuring the log-freeness of the type-I contribution. Also, like in his work, these bounds improve as $δ$ increases.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Priyamvad Srivastav. 2026-01-27. Log-free bounds on exponential sums over primes. https://arxiv.org/abs/2505.07803

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

On the factorisation of the $p$-adic Rankin-Selberg $L$-function in the supersingular case

Given a cusp form $f$ which is supersingular at a fixed prime $p$ away from the level, and a Coleman family $F$ through one of its $p$-stabilisations, we construct a $2$-variable meromorphic $p$-adic $L$-function for the symmetric square of $F$. We prove that this new $p$-adic $L$-function interpolates values of complex imprimitive symmetric square $L$-functions, for the various specialisations of the family $F$. We use this $p$-adic $L$-function to prove a $p$-adic factorisation formula, expressing the geometric $p$-adic $L$-function attached to the Rankin--Selberg convolution of $f$ with itself as a the product of the $p$-adic symmetric square $L$-function of $f$ and a Kubota-Leopoldt $L$-function. This extends a result of Dasgupta in the ordinary case.

math.NT

Exceptional poles of archimedean Rankin-Selberg L-functions for irreducible generic representations of GL(n,R)

For irreducible generic representations $π_1$ and $π_2$ of $\operatorname{GL}_n(\mathbb R)$, we prove that the notions of exceptional pole of type $1$ and type $2$ coincide at every level. When both representations are in general position, we use this identification to express the Rankin--Selberg $L$-function $L(s,π_1\timesπ_2)$ in terms of the exceptional $L$-factors attached to the irreducible constituents of their derivatives.

math.NT