[Paper Review] Major arcs for Goldbach's problem

H. A. Helfgott|arXiv (Cornell University)|May 13, 2013

Analytic Number Theory Research21 citations

TL;DR

This paper develops explicit, unconditional bounds for exponential sums over primes weighted by Gaussian-based smoothing functions, using rigorous verification of the Generalized Riemann Hypothesis up to conductor 300,000. It establishes sharp estimates for sums on major arcs in the circle method, enabling a complete proof of the ternary Goldbach conjecture by controlling error terms via new explicit bounds on parabolic cylinder functions and Mellin transforms.

ABSTRACT

The ternary Goldbach conjecture states that every odd number $n\\geq 7$ is the sum of three primes. The estimation of the Fourier series $\\sum_{p\\leq x} e(\\alpha p)$ and related sums has been central to the study of the problem since Hardy and Littlewood (1923). Here we show how to estimate such Fourier series for $\\alpha$ in the so-called major arcs, i.e., for $\\alpha$ close to a rational of small denominator. This is part of the author's proof of the ternary Goldbach conjecture. In contrast to most previous work on the subject, we will rely on a finite verification of the Generalized Riemann Hypothesis up to a bounded conductor and bounded height, rather than on zero-free regions. We apply a rigorous verification due to D. Platt; the results we obtain are both rigorous and unconditional. The main point of the paper will be the development of estimates on parabolic cylinder functions that make it possible to use smoothing functions based on the Gaussian. The generality of our explicit formulas will allow us to work with a wide variety of such functions.

Motivation & Objective

• To provide rigorous, unconditional estimates for exponential sums over primes in the major arcs of the circle method.
• To overcome limitations of traditional zero-free regions by relying on finite, verified GRH up to conductor 300,000.
• To develop a general framework for estimating sums with smooth weights based on the Gaussian function.
• To enable the full proof of the ternary Goldbach conjecture by controlling error terms in the major arc analysis.

Proposed method

• Uses the circle method with smooth weights derived from the Gaussian $ \eta_{\heartsuit}(t) = e^{-t^2/2} $ and related functions like $ \eta(t) = t^2 e^{-t^2/2} $.
• Applies explicit formulas for sums involving the von Mangoldt function and Dirichlet characters, leveraging Mellin transforms of twisted Gaussians.
• Employs rigorous numerical verification of non-trivial zeros of Dirichlet L-functions up to conductor 300,000 and height bounded by 4r/q.
• Derives new, fully explicit bounds for parabolic cylinder functions in critical ranges to control the Mellin transform of the smoothing function.
• Uses bisection and interval arithmetic to compute tight bounds on the $ L^\infty $, $ \ell^2 $, and other norms of the smoothing functions and their derivatives.
• Establishes error bounds that decay with $ 1/\sqrt{x} $, enabling control of the main term and error in the major arc contribution.

Experimental results

Research questions

• RQ1 How can one obtain explicit, unconditional bounds for exponential sums over primes in the major arcs without relying on zero-free regions?
• RQ2 What is the optimal smoothing function for major arc estimates in the ternary Goldbach problem, and how can its Mellin transform be bounded explicitly?
• RQ3 Can a finite, rigorous verification of the Generalized Riemann Hypothesis up to conductor 300,000 replace traditional zero-free region arguments in major arc analysis?
• RQ4 How do the decay properties of the Mellin transform of a Gaussian-based smoothing function affect the error term in the circle method?
• RQ5 What is the quantitative impact of using higher-order vanishing smoothing functions (e.g., $ t^2 e^{-t^2/2} $) on the error term in major arc estimates?

Key findings

• For $ x \geq 10^8 $, the sum $ \sum_{n} \Lambda(n)\chi(n) e(\delta n / x) e^{-(n/x)^2/2} $ is bounded by $ I_{q=1} \cdot \sqrt{2\pi} e^{-2\pi^2 \delta^2} x + E x $, where $ |E| \leq 5.281 \times 10^{-22} + \frac{1}{\sqrt{x}} \left( \frac{650400}{\sqrt{q}} + 112 \right) $.
• For the weight $ \eta(t) = t^2 e^{-t^2/2} $, the error term satisfies $ |E| \leq \frac{4.269 \times 10^{-14}}{q} + \frac{1}{\sqrt{x}} \left( \frac{276600}{\sqrt{q}} + 56 \right) $, showing improved decay for $ q > 1 $.
• The method achieves unconditional results by relying on a verified GRH up to conductor 300,000, avoiding reliance on ineffective zero-free regions.
• Explicit bounds on the $ L^\infty $, $ \ell^2 $, and other norms of the smoothing functions and their derivatives are derived via bisection and interval arithmetic.
• The framework allows for the use of a wide class of smoothing functions through explicit control of the Mellin transform of the Gaussian.
• The error terms decay as $ O(1/\sqrt{x}) $, which is sufficient to close the major arc contribution in the proof of the ternary Goldbach conjecture.