🔢 mathematics

On the level of distribution of Goldbach primes and its applications

This paper establishes that the set of Goldbach primes possesses a level of distribution of $1/6$ for almost all even integers, a result that is subsequently applied to prove that almost all even integers can be expressed as the sum of two primes satisfying specific primality conditions on linear and quadratic combinations of those primes.

Original authors: Mizuki Akeno

Published 2026-06-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Mizuki Akeno

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: The Goldbach Party

Imagine the Goldbach Conjecture as a massive, never-ending party. The host (mathematics) claims that for every even number greater than 2 (like 4, 6, 100, 1,000,000), you can always find two "prime number" guests who, when added together, equal that even number.

For example:

$4 = 2 + 2$
$10 = 3 + 7$
$100 = 47 + 53$

Mathematicians have been trying to prove this is true for every even number. While they haven't proven it for 100% of numbers yet, they have proven it for "almost all" of them.

The Problem: How Evenly Are the Guests Distributed?

The author of this paper isn't just asking if the party happens; they are asking how the guests are distributed across different neighborhoods.

Imagine you are looking for prime number pairs that add up to a specific number $N$ . You want to know: If I look at all the prime pairs in a specific "neighborhood" (defined by a mathematical rule called a "residue class"), do I find the expected number of pairs?

The Old Way: Previous mathematicians could only look at neighborhoods up to a certain size (let's say, neighborhoods up to the square root of the total crowd).
The New Discovery: Mizuki Akeno has proven that we can look much deeper. We can check neighborhoods that are 1/6th the size of the total crowd, and the distribution of prime pairs still looks perfectly even and predictable.

The Analogy:
Imagine a giant stadium filled with people.

Previous Math: "We can count how many people are in the first 10 rows, and it looks fair."
This Paper: "We can count how many people are in the first 16 rows (which is much deeper), and it still looks perfectly fair and predictable."

This "Level of Distribution" (1/6) is a new record for this specific type of problem. It's like having a super-powerful telescope that lets you see further into the crowd without the view getting blurry.

The Application: Finding Special Prime Pairs

Why does this deeper view matter? Because now the author can find prime pairs that do more than just add up to $N$ . They can find pairs that satisfy extra, tricky conditions.

The paper proves two main things about "almost all" even numbers (meaning, if you pick a random even number, it's almost guaranteed to work):

1. The "Difference" Trick

The author shows that for almost all even numbers $N$ , you can find two primes, $p_1$ and $p_2$ , such that:

$p_1 + p_2 = N$ (They add up to the target).
AND the result of $p_1 - p_2 + 1$ is a number that has at most 4 prime factors.

The Metaphor:
Imagine you are looking for two specific keys ( $p_1$ and $p_2$ ) that open a door (sum to $N$ ).

Old Result: "We can find keys that open the door."
This Paper: "We can find keys that open the door, AND if you rub them together ( $p_1 - p_2 + 1$ ), the resulting dust cloud is made of only 4 or fewer distinct types of sand grains."

2. The "Product" Trick

For numbers $N$ that are divisible by 6, the author proves you can find two primes, $p_1$ and $p_2$ , such that:

$p_1 + p_2 = N$ .
AND the number $2p_1p_2 + 1$ has at most 13 prime factors.

The Metaphor:

Old Result: "We can find keys that open the door."
This Paper: "We can find keys that open the door, AND if you multiply them together and add a little extra ( $2p_1p_2 + 1$ ), the resulting machine is made of only 13 or fewer distinct gears."

How Did They Do It? (The "Secret Sauce")

To get this result, the author had to invent a new way to measure the "noise" in the system.

The Exponential Sum Estimate: In math, "noise" often looks like waves crashing. The author developed a new formula (Theorem 2.1) to predict exactly how these waves behave when they interact with specific patterns. This formula is the engine that allows them to see the "1/6" level of distribution.
The Weighted Sieve: Think of a sieve as a colander used to separate pasta from water. In math, a "sieve" separates prime numbers from non-primes. The author used a "weighted" sieve, which is like a colander with special sensors. It doesn't just catch the primes; it weighs them to ensure the ones it catches fit the extra conditions (like the "4 factors" or "13 factors" mentioned above).

Summary

The Goal: Understand how prime numbers pair up to form even numbers.
The Breakthrough: Proved that these pairs are evenly distributed much further out than previously thought (up to 1/6 of the total range).
The Result: Because we can see further, we can now prove that for almost all even numbers, there exist prime pairs that not only add up to the number but also create specific, complex patterns when you do math with them (like having a limited number of prime factors).

The paper is a technical tour de force that pushes the boundaries of what we know about the "neighborhoods" where prime numbers live, allowing us to find more complex and interesting combinations of them.

Technical Summary: On the Level of Distribution of Goldbach Primes and Its Applications

Problem Statement
The paper addresses the distribution of Goldbach primes, specifically the set of primes $P \cap (N - P)$ for even integers $N$ . The central problem is to establish a "level of distribution" for these primes in arithmetic progressions where the residue class depends on the modulus and potentially on $N$ itself. While the Bombieri–Vinogradov theorem provides strong distribution results for primes in fixed arithmetic progressions, extending this to the Goldbach problem with residue classes $l_i$ depending on the moduli $d_i$ and the integer $N$ is significantly more challenging. The author aims to prove that for almost all even integers $N$ , the set of Goldbach primes has a level of distribution $\theta = 1/6$ . This threshold is critical for applications involving the weighted sieve, particularly in determining the number of prime factors of linear or quadratic forms associated with the Goldbach partition.

Methodology
The proof strategy combines analytic number theory techniques, specifically the circle method, exponential sum estimates, and sieve methods.

Exponential Sum Estimates (The Core Innovation):
The paper establishes a new estimate for exponential sums involving the von Mangoldt function $\Lambda(n)$ over arithmetic progressions with moduli $d$ and residue classes $l$ that may depend on $N$ . Theorem 2.1 provides a bound for sums of the form:
$\sum_{d \sim D} \sum_{b \in \mathbb{Z}/d\mathbb{Z}} \left| \sum_n \Lambda(n) e\left(\left(\alpha + \frac{b}{d}\right)n\right) w(n/X) \right|^2$
This estimate is sharper than standard large sieve bounds when the modulus $q$ of the rational approximation to $\alpha$ falls within a specific range. The proof utilizes Vaughan's identity to decompose the sum into Type I and Type II sums, followed by applications of the Poisson summation formula and bounds on the number of solutions to Diophantine inequalities involving rational approximations (Lemmas 2.2–2.4).
Level of Distribution (Theorem 1.1 and 3.1):
Using the exponential sum estimate, the author proves Theorem 1.1, which asserts that for almost all $N \in (X/2, X]$ , the error term in the distribution of Goldbach primes in arithmetic progressions with moduli up to $X^{1/6}$ is small. The proof involves:
- Major Arcs: Utilizing a mean value theorem for Dirichlet polynomials (Choi and Kumchev) and arguments by Maynard to handle the main term.
- Minor Arcs: Applying the new exponential sum estimate (Theorem 2.1) and Bessel's inequality to bound the contribution of the minor arcs.
- Generalization: Theorem 3.1 extends this result to cases where one of the $\Lambda$ factors is replaced by a structured weight $\beta(n)$ (e.g., indicator functions of rough numbers or products of primes), which is necessary for the sieve applications.
Sieve Applications:
The paper applies the derived level of distribution $\theta = 1/6$ to weighted sieve problems.
- Theorem 1.2 (Linear Form): The author uses Kuhn's weighted sieve with switching (following Matomäki and Zúñiga-Alterman) to study the equation $p_1 + p_2 = N$ with the condition $p_1 - p_2 + 1 \in P_k$ . The paper demonstrates that if the level of distribution is at least $0.1635$, one can find solutions where the linear form has at most 4 prime factors ( $k=4$ ). Since $1/6 \approx 0.1667 > 0.1635$ , the result follows.
- Theorem 1.3 (Quadratic Form): Using Richert's weighted sieve, the paper addresses $p_1 + p_2 = N$ with $2p_1p_2 + 1 \in P_k$ for $N$ divisible by 6. The analysis shows that the level of distribution $1/6$ suffices to prove the existence of solutions where the quadratic form has at most 13 prime factors.

Key Contributions and Results

Theorem 1.1: Proves that for almost all even integers $N$ , the set of Goldbach primes has a level of distribution $1/6$ . This is a significant improvement over previous results (such as Tolev's $X^{1/3}$ or Matomäki's $X^{1/2}$ under specific factorability assumptions) in the context where the residue class depends on $N$ .
Theorem 1.2: Establishes that almost all even integers $N$ can be written as $p_1 + p_2$ such that $p_1 - p_2 + 1$ has at most 4 prime factors. This improves upon previous bounds where the number of prime factors was higher (e.g., $P_5$ or $P_7$ ).
Theorem 1.3: Establishes that almost all integers $N \equiv 0 \pmod 6$ can be written as $p_1 + p_2$ such that $2p_1p_2 + 1$ has at most 13 prime factors.
Technical Lemma: Theorem 2.1 provides a novel exponential sum estimate that is crucial for achieving the $1/6$ exponent, distinguishing the minor arc analysis from previous works like Maier and Pomerance.

Significance
The paper claims that the primary significance lies in pushing the level of distribution for Goldbach primes to $1/6$ in a setting where the residue class depends on $N$ . This specific threshold is the "tipping point" that allows the weighted sieve to reduce the number of prime factors in the associated linear form $p_1 - p_2 + 1$ from 5 or more down to 4. The author notes that previous applications of the weighted sieve failed to reach $P_4$ because the required level of distribution was slightly higher than what was previously available (specifically, the criteria remained above $1/6$ ). By achieving this level, the paper resolves the solvability of these specific Goldbach-type equations for almost all $N$ with a tighter bound on the number of prime factors. The results are presented as "almost all" statements, consistent with the limitations of current analytic methods regarding the exceptional set in the Goldbach conjecture.