On the density version of quadratic Waring’s problem and the quadratic Waring–Goldbach problem

Zi Li Lim · 2025

Read On the density version of quadratic Waring’s problem and the quadratic Waring–Goldbach problem by Zi Li Lim (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-14939.txt.

Introduction

Zi Li Lim proves new results in number theory in a 2025 arXiv paper. The paper treats density versions of two classic problems: the quadratic Waring problem and the quadratic Waring-Goldbach problem. For the first, Lim proves a sharp density theorem over cyclic groups when the number of variables is at least five. For the second, which concerns the integers, the paper reports fresh improvements on known bounds. The work is a short, technical proof paper aimed at specialists in this branch of number theory.

Catalog record

CREATORZi Li Lim
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE86 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-14939

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On the density version of quadratic Waring’s problem and the quadratic Waring–Goldbach problem Zi Li Lim Abstract . We prove a sharp density theorem for quadratic Waring’s problem over cyclic groups, when the number of variables is at least 5 5 . Also, we obtain some new improvements on the density version of the quadratic Waring–Goldbach problem over integers. 1. Introduction 1.1. Introduction Denote the set of all prime numbers by \mathcal{P} . Let A be a subset of primes, recall that the relative lower density \delta_{\mathcal{P}}(A) is defined to be \delta_{\mathcal{P}}(A)=\liminf_{N\rightarrow\infty}\frac{|A\cap[N]|}{|\mathcal{P}\cap[N]|}. Li and Pan [ LP10 ] proved a density version of Vinogradov three primes theorem, which states that if \delta_{\mathcal{P}}(A)>2/3 , then every sufficiently large odd integer can be represented as a sum of three elements in A . In this paper, we refer to quantities such as the 2/3 in Li and Pan’s result as a “density threshold”. In a breakthrough paper, Shao [ Sha14 ] proved the sharp density version of Vinogradov three primes theorem, lowering the density threshold from 2/3 to 5/8 , which led to many interesting subsequent developments, such as [ Sha14a ] , [ MS17 ] and [ MMS17 ] . Later, density type theorems were considered in other arithmetic settings, for example, Waring’s problem [ Sal21 ] , or the Waring–Goldbach problem [ Gao25 ] , [ Zha25 ] , but the optimal densities are unknown. In this paper, we present some refinements of density type theorems for quadratic Waring’s problem and the quadratic Waring–Goldbach problem. Our first theorem is a sharp density theorem for quadratic Waring’s problem over cyclic groups. Informally, it gives a density threshold of 1/s of the Waring problem for sums of s squares in a cyclic group, provided that the order of the group is not divisible by small primes, and s is at least 5 . For any positive integer W , denote the squares in \mathbb{Z}/W by (\mathbb{Z}/W)^{(2)} . Theorem 1.1 . Let s\geq 5 be an integer and \theta\in(1/s,1] . There exists a constant M(s,\theta) depending only on s and \theta such that the following holds. Suppose W=\prod_{i}p_{i}^{n_{i}} is an integer with p_{i}\geq M(s,\theta) for all i , and A is a subset of the squares in \mathbb{Z}/W with relative density greater than or equal to \theta , i.e. |A|\geq\theta|(\mathbb{Z}/W)^{(2)}| . Then, for all y\in\mathbb{Z}/W , there exist x_{1},x_{2},...,x_{s}\in A such that y=x_{1}+x_{2}+...+x_{s} . Moreover, one has the following quantitative estimate |\{(x_{1},x_{2},...,x_{s})\in A^{s}:y=x_{1}+x_{2}+...+x_{s}\}|\geq c(s,\theta)W^{-1}|(\mathbb{Z}/W)^{(2)}|^{s} for some constant c(s,\theta)>0 depending only on s and \theta . Remark 1.2 . For Theorem 1.1 , it is necessary to exclude small prime divisors. Let W be an integer coprime to 3 , and A=\{x\in(\mathbb{Z}/3W)^{(2)}:x\equiv 0\;\mathrm{mod}\;3\} , then the relative density of A is 1/2 , but the s -fold sumset of A can only contains the elements congruent to 0\;\mathrm{mod}\;3 . It might be worth to point out that the squares in \mathbb{Z}/W are fairly sparse. Suppose W=\prod_{1\leq i\leq r}p_{i} is a product of r distinct odd primes, then the absolute density of (\mathbb{Z}/W)^{(2)} in the ambient space \mathbb{Z}/W is at most \prod_{1\leq i\leq r}(p_{i}+1)/(2p_{i})\leq(2/3)^{r} , which tends to zero when r\rightarrow\infty . Hence, a subset of (\mathbb{Z}/W)^{(2)} with relative density greater than 1/s can have absolute density arbitrarily close to zero. The next theorem confirms that the density threshold 1/s is indeed sharp, except possibly at the endpoint. Theorem 1.3 . Let s\geq 5 be an integer and \tau\in[0,1/s) . For every sufficiently large (depending only on s and \tau ) prime p , there exists a subset A_{p} of the squares in \mathbb{Z}/p with relative density greater than or equal to \tau , i.e. |A_{p}|\geq\tau|(\mathbb{Z}/p)^{(2)}| , but the s -fold sumset sA_{p} is not equal to \mathbb{Z}/p . Now, we turn our attention from cyclic groups to integers. Salmensuu [ Sal21 ] proved a density theorem for Waring’s problem. In quadratic case of Salmensuu’s theorem, the number of squares needed is at least 44 and the density threshold is approximately 0.9206 . Gao [ Gao25 ] and Zhao [ Zha25 ] generalized Salmensuu’s theorem to the Waring–Goldbach problem. We choose to work with the Waring–Goldbach problem to demonstrate our method for technical reasons (it is easier to work with reduced congruence classes for corresponding local problem, see Theorem 1.6 ). We obtain some new improvements, but not the optimal density. For any integer s\geq 5 , define D_{s}\coloneqq\begin{cases}\frac{59}{60}&\text{if $s=5$}\\ \frac{7}{8}&\text{if $s=6$}\\ \frac{3}{4}&\text{if $s=7$}\\ \frac{s+13}{4s}&\text{if $8\leq s\leq 12$}\\ \frac{1}{2}&\text{if $s\geq 13$}\end{cases} Theorem 1.4 . Let s\geq 5 be an integer, and A be a subset of primes with relative lower density \delta_{\mathcal{P}}(A)>\sqrt{D_{s}} . Then, for all sufficiently large integers y that are congruent to s\;\mathrm{mod}\;24 , there exist primes p_{1},p_{2},...,p_{s}\in A such that y=p_{1}^{2}+p_{2}^{2}+...+p_{s}^{2} . Remark 1.5 . The congruence condition in Theorem 1.4 is necessary, since the square of any prime greater than 3 is congruent to 1\;\mathrm{mod}\;24 . For comparison, Gao [ Gao25 ] proved a density theorem for the quadratic Waring–Goldbach problem for s\geq 44 and density threshold 3/4 , and Zhao [ Zha25 ] improved the result to s\geq 8 and density threshold \sqrt{1-\min\{s,16\}/32} . (Gao also considered the Waring–Goldbach problem for higher powers, see relevant discussions later.) Qualitatively, Theorem 1.4 allows larger range of s . Quantitatively, the density thresholds given by \sqrt{D_{s}} improve upon previous record \sqrt{1-\min\{s,16\}/32} by Zhao. The table below is a detailed numerical comparison between the density thresholds in Theorem 1.4 and previous record by Zhao. s \sqrt{D_{s}} Previous record s \sqrt{D_{s}} Previous record 5 \sqrt{\frac{59}{60}}\approx 0.9916 unknown previously 11 \sqrt{\frac{6}{11}}\approx 0.7385 \sqrt{\frac{21}{32}}\approx 0.8101 6 \sqrt{\frac{7}{8}}\approx 0.9354 unknown previously 12 \sqrt{\frac{25}{48}}\approx 0.7217 \sqrt{\frac{5}{8}}\approx 0.7906 7 \sqrt{\frac{3}{4}}\approx 0.8660 unknown previously 13 \sqrt{\frac{1}{2}}\approx 0.7071 \sqrt{\frac{19}{32}}\approx 0.7706 8 \sqrt{\frac{21}{32}}\approx 0.8101 \sqrt{\frac{3}{4}}\approx 0.8660 14 \sqrt{\frac{1}{2}}\approx 0.7071 \sqrt{\frac{9}{16}}\approx 0.7500 9 \sqrt{\frac{11}{18}}\approx 0.7817 \sqrt{\frac{23}{32}}\approx 0.8478 15 \sqrt{\frac{1}{2}}\approx 0.7071 \sqrt{\frac{17}{32}}\approx 0.7289 10 \sqrt{\frac{23}{40}}\approx 0.7583 \sqrt{\frac{11}{16}}\approx 0.8292 \geq 16 \sqrt{\frac{1}{2}}\approx 0.7071 \sqrt{\frac{1}{2}}\approx 0.7071 The key to Theorem 1.4 is the following local result. For any integer s\geq 5 , define d_{s}\coloneqq\begin{cases}\frac{9}{10}&\text{if $s=5$}\\ \frac{5}{6}&\text{if $s=6$}\\ \frac{s+13}{4s}&\text{if $7\leq s\leq 12$}\\ \frac{1}{2}&\text{if $s\geq 13$}\end{cases} Note that D_{s}\geq d_{s}\geq 1/2 and d_{s}\geq(s+13)/(4s) for all s . Similar as before, for any positive integer W , denote the squares in (\mathbb{Z}/W)^{\times} by (\mathbb{Z}/W)^{\times(2)} . Theorem 1.6 . Let s\geq 5 be an integer, and W be a square free integer. Assume (2\cdot 3\cdot 5,W)=1 when s\geq 6 , and (2\cdot 3\cdot 5\cdot 7,W)=1 when s=5 . Let f_{j}:(\mathbb{Z}/W)^{\times(2)}\longrightarrow[0,1] be functions for 1\leq j\leq s with \frac{1}{s}(\mathop{\mathbb{E}}[f_{1}]+\mathop{\mathbb{E}}[f_{2}]+...+\mathop{\mathbb{E}}[f_{s}])>d_{s} and \mathop{\mathbb{E}}[f_{j}]\neq 0 for all 1\leq j\leq s . Then, for all y\in\mathbb{Z}/W , we have y=x_{1}+x_{2}+...+x_{s} for some x_{1},x_{2},...,x_{s}\in(\mathbb{Z}/W)^{\times(2)} such that \frac{1}{s}(f_{1}(x_{1})+f_{2}(x_{2})+...+f_{s}(x_{s}))>d_{s} and f_{j}(x_{j})\neq 0 for all 1\leq j\leq s . The essential distinction between Theorem 1.1 and Theorem 1.6 is Theorem 1.1 excludes considerably more amount of small primes. Due to this distinction, their proofs are fundamentally different. 1.2. Sketch of the proofs All density type theorems, including ours, crucially rely on the transference principle pioneered by Green and Tao in their celebrated proof that primes contain arbitrarily long arithmetic progressions [ Gre05 ] , [ GT08 ] . One of the key ingredients needed for the transference principle is restriction estimates. For Theorem 1.1 , since we are working over

Read the complete text