Products of primes in arithmetic progressions

From MaRDI portal



Abstract: A conjecture of ErdH{o}s states that, for any large prime q, every reduced residue class pmodq can be represented as a product p1p2 of two primes p1,p2leqq. We establish a ternary version of this conjecture, showing that, for any sufficiently large cube-free integer q, every reduced residue class pmodq can be written as p1p2p3 with p1,p2,p3leqq primes. We also show that, for any varepsilon>0 and any sufficiently large integer q, at least (2/3−varepsilon)varphi(q) reduced residue classes pmodq can be represented as a product p1p2 of two primes p1,p2leqq. The problems naturally reduce to studying character sums. The main innovation in the paper is the establishment of a multiplicative dense model theorem for character sums over primes in the spirit of the transference principle. In order to deal with possible local obstructions we use bounds for the logarithmic density of primes in certain unions of cosets of subgroups of mathbbZqimes of small index and study in detail the exceptional case that there exists a quadratic character psipmodq such that psi(p)=−1 for almost all primes pleqq.


Let \(\mathbb{Z}_q^{\times}\) be the set of reduced residue classes \(\pmod q\). For integers \(k,q\geq 1\) and a real number \(N \geq 1\), let \(E_k(N)\) be the set of integers \(a \in \mathbb{Z}_q^{\times}\) such that \(a \equiv p_1 \dotsm p_k \pmod{q}\) for some primes \(p_1, \dotsc, p_k \leq N\). The main purpose of the paper under study is to approximate the size of \(E_k(N)\) for \(k\in \{2,3\}\). The case \(k=2\) concerns with a conjecture due to Erdős stated in [\textit{P. Erdős} et al., Period. Math. Hungar. 18, 229--239 (1987; Zbl 0625.10035)] asserting that \(E_2(q)=\mathbb{Z}_q^{\times}\) for all large enough primes \(q\). This conjecture is still open, even under assuming generalized Riemann hypothesis. In this paper the authors establish a ternary variant of Erdős' conjecture by showing that if \(q \in \mathbb{N}\) is cube-free and sufficiently large, then \(E_3(q) = \mathbb{Z}_q^{\times}\). Moreover, letting \(\varepsilon>0\) and \(q \in \mathbb{N}\) be sufficiently large in terms of \(\varepsilon\), they prove that \(E_3(q^{1+\varepsilon}) = \mathbb{Z}_q^{\times}\). In another direction to consider Erdős' conjecture, the authors consider the problem of lower-bounding the density of \(E_2(q)\) inside \(\mathbb{Z}_q^{\times}\), and prove that letting \(\varepsilon>0\) and \(q \in \mathbb{N}\) be sufficiently large in terms of \(\varepsilon\), then one has \(|E_2(q)| \geq (2/3-\varepsilon) \varphi(q)\), and further, assuming \(q\) is cube-free, then \(|E_2(q)| \geq (11/16-\varepsilon) \varphi(q)\).











This page was built for publication: Products of primes in arithmetic progressions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6203236)