Explicit quadratic large sieve inequality
Zihao Liu
TL;DR
This work provides an explicit version of Heath-Brown's quadratic large sieve inequality for quadratic Dirichlet characters, delivering precise $\varepsilon$-dependent bounds via an iterated-exponential framework. The authors develop a Heath-Brown–style recursive method, combining Poisson summation, dual sums, and careful control of main terms and errors to obtain a bound of the form $\mathcal{B}(M,N) \le \exp_4(C\varepsilon^{-1}) (MN)^{\varepsilon} (M+N)$ for $MN \ge \exp_4(1)$. They then derive concrete corollaries: an explicit quadratic large sieve inequality, fourth-moment bounds for Dirichlet $L$-functions twisted by quadratic characters, and asymptotics for rational primes splitting in quadratic fields, with quantified exceptional sets. The results hold without any structure on the coefficients $\{a_n\}$, making them broadly applicable to sieve-constructed sequences and enabling unconditional applications in moments and arithmetic statistics. Overall, the paper provides a robust, explicitly quantified baseline for quadratic large-sieve applications, potentially guiding further refinements to reduce the iterated-exponential growth.
Abstract
In this article, we obtain an explicit version of Heath-Brown's large sieve inequality for quadratic characters and discuss its applications to $L$-functions and quadratic fields.
