All totally real abelian number fields of class number one are Euclidean

arXiv:2610.24900 · math.NT, math.RA · Submitted 2026-10-17 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: I'm Ada.

Emmy: And I'm Emmy, here with Kolya.

Ada: We're looking at the paper: "All totally real abelian number fields of class number one are Euclidean".

Emmy: Here's the gist of the paper. ALL TOTALLY REAL ABELIAN NUMBER FIELDS OF CLASS NUMBER ONE ARE EUCLIDEAN is a short note that shows, more generally, that the ring of integers OK of any totally real abelian number field K is a UFD, if and only if it is Euclidean The main theorem proves this equivalence for all totally real abelian number fields with degree at least two The paper establishes a lower bound on the number of prime ideals in certain residue classes, leading to the conclusion that OK is Euclidean if and only if hK = 1 The main theorem states that for a totally real abelian number field K/Q with n ≥ 2, there are constants cK, xK > 0 such that ΠK(x) ≥ cK · x log(x)2 for all x > xK In particular, for all totally real abelian number fields K OK is Euclidean ⇐⇒ hK = 1 Verification of Previous Results The authors verify Weinberger’s prediction regarding exceptional extensions by showing that the ring of integers OK is a UFD if and only if it is Euclidean for all totally real abelian number fields K This result builds upon Narkiewicz's earlier work concerning quadratic and cubic extensions The theorem overlaps with results from

Har-Mur04: and

Mur-Pet13: regarding non-existence of hypothetical exceptional extensions Key Ingredients and Tools The proof relies on two results from

OAI26: , Thm.

Ada: 1.2 and Prop.

Emmy: 2.1, which establish Scholium (I) and Scholium (II) Scholium (I) asserts that for every cyclotomic field F containing the group of 12th roots of unity µ12 and every Hecke character χ: A∗F/F∗ → C∗ of F of finite order, the associated Hecke L-function L(s, χ) has no zeros in the open region Re(s) > 1 − δ0 with a uniform width δ0 = 10−6 Scholium (II) provides a lower bound for certain prime ideals, stating that there are constants am, xm > 0 such that for all x ≥ xm (2.1) Pm(x) ≥ am x log(x)2 Derivation of the Lower Bound The lower bound is derived by combining Scholium (I), (II), and Proposition 1.6, which provides a refined version of another result in

OAI26: By combining these, the authors obtain the desired lower bound of growth for the “prime ideal counting function” ΠK(x) This is shown in two steps, found in Sect.

Ada: 2 and Sect.

Title and authors: Ada: So we're looking at this paper called "All totally real abelian number fields of class number one are Euclidean". It’s tackling that old debate about whether a ring of integers has unique factorization or if it has to satisfy the division algorithm.

Emmy: Exactly, Ada. The core idea is showing that if the ring of integers for any totally real abelian number field has class number one, then it must also be Euclidean. It connects this algebraic property of unique factorization directly to a geometric property called having a division algorithm on that ring of integers OK.

Kolya: That sounds like they are trying to settle what Narkiewicz started about quadratic and cubic extensions having class number one but not necessarily being Euclidean. What's the main point they’re pushing here?

Ada: The main point is showing this equivalence holds for all totally real abelian number fields with degree two or more. They establish a lower bound on a counting function, called the prime ideal counting function PiK of x, which shows that if the ring isn't Euclidean, this count has to grow at least as fast as x times the square of the logarithm of x.

Emmy: That growth rate is what they are aiming for. They prove there are constants cK and xK such that for all values of x greater than that threshold, PiK of x is greater than or equal to cK times x times the log of x squared.

Kolya: So, if we're thinking about where this means for the math community, it’s like they are providing a hard floor on how sparse prime ideals can be in these rings. It suggests you can't have too many "weird" prime ideal structures without violating this growth rate.

Ada: Right. And the big implication is linking this directly to the class number being one because they show that for every totally real abelian field K, if it has class number one, then it must be Euclidean and vice versa. That’s a very strong statement about the relationship between these two properties of OK >

Title and authors: Emmy: It means that the algebraic condition of having unique factorization is equivalent to the geometric condition of having a division algorithm defined by some norm function on the ring of integers OK. It ties together number theory and geometry in a very neat way.

Kolya: I wonder if this result helps with the earlier work on exceptional extensions, like those quadratic or cubic fields Narkiewicz looked at? Does this theorem confirm that those exceptions, which were thought to be rare or impossible, can't exist under these general conditions?

Ada: It does. The paper says their proof overlaps with earlier work by Har Murty and Murty Petersen on non-existence of hypothetical exceptional extensions in certain cases. They essentially use this new lower bound on PiK of x to show that the exceptions they were looking for, like those quadratic or cubic fields with class number one but not being Euclidean, simply cannot exist.

Emmy: So they're using analytic counting tools to rule out specific algebraic structures that were previously thought possible. It’s a powerful way to solidify what we already knew about these special extensions >

Kolya: What about the methodology? How are they actually getting that lower bound on the prime ideal count? It sounds like it requires combining some very deep analytic results.

Ada: Well, they rely on two main ingredients from a source called OAI26. First, Scholium one, which deals with zero-free regions for Hecke L-functions associated with cyclotomic fields containing the 12th roots of unity and finite order Hecke characters. This establishes a uniform width delta zero equal to one thousandth of one hundred—or zero point one—for the region where there are no zeros <ref:2610.24900#pg3>.

Emmy: And second, Scholium two, which gives them a lower bound for certain prime ideals themselves. This essentially tells them that for large enough numbers x, there will always be at least a certain number of rational primes whose prime divisors fall within specific ranges defined by some exponential functions of the logarithm of x >

Kolya: So they combine those analytic constraints with Proposition one point six from OAI26 to get the growth rate for PiK of x. What does that combination actually tell us about the density?

Title and authors: Ada: It tells us that when you subtract those error terms—the A(x) and B(x) terms which handle prime ideals with non-surjective reduction—you end up with a growth of at least n times am times x times the square of the logarithm of x. That is, n a m x log of x squared >

Emmy: That term n a m x log of x squared is what they are aiming for. It means that as you look at larger and larger numbers, the number of relevant prime ideals doesn't just stay small; it grows predictably with that specific power law in the logarithm >

Kolya: And then they show that when x gets big enough, those error terms A(x) and B(x) are smaller than this main term. They are "little o" of x times the square of the logarithm of x. That means their contribution doesn't change the overall growth rate in a significant way for large enough numbers >

Ada: Exactly. So, they choose a threshold x K big enough so that all those estimates hold simultaneously, which establishes the lower bound for PiK of x as n a m times x log of x squared minus something smaller than that >

Emmy: It's like they are showing that the "noise" or the exceptions don't swamp the main signal when you look at huge numbers. The structure dominates, and it grows according to this logarithmic law >

Kolya: For someone who just listens to this show, what does this mean practically? Does it change how we think about these number fields in general?

Ada: It provides a very concrete tool for testing the Euclidean property. Instead of just hoping that a field with class number one might be Euclidean, they can use this lower bound on prime ideal counts to rigorously prove it. If you find a field where the count is less than this lower bound for some range of x, then you know immediately it fails the condition >

Emmy: It gives us a way to quantify "how far" a ring of integers is from being Euclidean by looking at how many prime ideals are behaving in a certain way. It turns an abstract property into a counting problem >

The paper's summary: Ada: So, to recap, this paper is essentially proving that if you have a totally real abelian number field where every ideal factorizes uniquely, then that ring must also have a division algorithm defined by some norm function. It’s linking the algebraic idea of unique factorization directly to the geometric idea of having a division algorithm on the ring of integers.

Emmy: It’s like showing two different ways to describe the same underlying structure in number fields are actually equivalent.

Ada: And for all totally real abelian fields with degree two or higher, this equivalence holds true, provided you stick to class number one. The big takeaway is that you don't need to check a whole list of specific exceptions like quadratic or cubic fields to know if the ring is Euclidean if the class number condition is met >

Emmy: That’s significant because it validates Weinberger’s prediction that those potential exceptions, which were thought to exist under certain conditions, actually cannot happen for these broader types of fields. It turns a conjecture about rare cases into a general rule for all totally real abelian fields with degree two or more >

Ada: Exactly. And they do this by using very sophisticated analytic counting tools—specifically zero-free regions for Hecke L-functions and bounds on prime ideal distributions—to establish a growth rate for prime ideals that is strong enough to rule out non-Euclidean rings when the class number isn't one >

Emmy: The method relies on showing that the number of "bad" prime ideals, those that don't behave nicely with respect to surjectivity in the reduction map, actually grows slower than what you need to worry about when you look at very large numbers. It means the structure of these rings dominates the count >

Ada: And this leads to a concrete statement: if you look at the number of prime ideals whose norm falls in a certain range, that count will grow at least as fast as x times the square of the logarithm of x, provided you're past some initial threshold. That’s what they call PiK of x being greater than or equal to a constant times x log of x squared >

Emmy: So it gives us a way to test for Euclidean rings using just counting functions and analytic bounds rather than having to perform the full division algorithm check on every single element. It provides a quantitative check that’s independent of the field's specific algebraic details beyond its degree and abelian structure >

Ada: Right. And this whole setup allows them to verify results from other papers, like those by Har Murty and Murty, which looked at non-existence in some specific cases before they could prove this general statement for all totally real abelian fields with degree two or more >

Emmy: It solidifies the structure of these rings. If you have class number one, you’re guaranteed a division algorithm; if the ring isn't Euclidean, the class number must be something other than one. It’s a clean boundary line there >

Ada: And they even get a corollary about prime ideals splitting completely in extensions H l over K, giving us bounds on those specific types of primes as well. It shows this technique is versatile enough for analyzing how primes move between different field structures >

Emmy: So, to summarize, the paper takes the algebraic property of unique factorization and proves it's equivalent to the geometric property of being Euclidean for all totally real abelian fields with degree two or more, using a strong analytic lower bound on prime ideal counting >

Ada: It’s a very solid piece of work confirming long-standing predictions about these fields.

Emmy: Right.

The paper's improvements: Ada: So, to recap, the authors are pushing the boundaries of this work by suggesting they can refine the constants in that lower bound we talked about earlier, moving from just an existence proof to a more explicit calculation for those constants cK and xK. It's like going from saying there's a floor on how sparse the prime ideals can be to giving you an exact number for that floor >

Emmy: That refinement is important because it moves us closer to being able to use this tool in practical applications. Right now, we know the bound exists, but knowing what those constants actually are helps us figure out how large x needs to be before the estimate kicks in. It makes the "x greater than xK" part much more useful for someone trying to apply this concept >

Ada: Exactly. And they also address the limitations of their current proof structure by showing how they can incorporate those error terms, A(x) and B(x), more tightly into the main estimate. They're essentially showing that those "small" parts of the count are even smaller than we first thought when we set up the main inequality >

Emmy: That’s like tightening the screws on an engine. It means they aren't just proving a general principle; they are making sure every single component of the argument is as efficient as possible in terms of scaling, which is crucial for any kind of number theory work that needs to be reliable at scale >

Ada: And one key improvement is showing how the counting function behaves when you look at primes that split completely in some specific field extension H l over K, which we mentioned briefly earlier. They’re adding a corollary that gives us a concrete bound for those prime ideals S l of x >

Emmy: That bridges the gap between the main theorem about the field itself and how primes behave when you extend the field. It shows that this method isn't just for looking at one field in isolation, but it can be used to probe relationships between different extensions in a controlled way >

Ada: So, they’re turning a general existence result into something more computationally and theoretically tractable by giving us tighter numerical bounds on the constants and showing how the error terms vanish faster. This makes the tool much more robust for applying it to specific families of fields >

Emmy: It really builds on that foundation we laid, connecting analytic tools like zero-free regions directly to precise counting estimates for algebraic structures. It’s a testament to how these ideas from different areas of mathematics can combine when applied together in this way >

Ada: And honestly, the implication is that we have a much clearer picture of the landscape for totally real abelian fields with class number one. We know they are Euclidean, and now we have better tools to understand why and how large the field needs to be for that property to hold true in a quantitative sense >

Emmy: It gives us more confidence in applying these ideas to other areas where we might be studying constraints on algebraic objects, whether it's geometry or dynamics. We’ve shown this machinery works for these specific fields, and now we can think about what else it might unlock >

Conclusion: Ada: So, to wrap up, this paper proves that for any totally real abelian number field with degree two or higher, the ring of integers is Euclidean if and only if its class number is one. It’s a very clean equivalence that holds universally for this class of fields >

Emmy: That’s the big picture here—it confirms Weinberger's prediction about those potential exceptions being impossible under these conditions. It settles a long-standing question about the relationship between unique factorization and the division algorithm on these specific number fields.

Kolya: From a probabilistic viewpoint, what this means is that we can use analytic bounds to control the distribution of prime ideals in a very precise way, which suggests that for any field with class number one, you’re not going to run into any weird structural roadblocks regarding unique factorization >

Ada: Exactly. And the methodology used—combining zero-free regions from Hecke L-functions with sharp bounds on prime ideal counts—shows that the structure of these rings is very tightly constrained by their arithmetic properties. It's a very powerful connection between analysis and algebra >

Emmy: The implication is that we have a more rigorous way to analyze these rings, moving beyond just checking if they might be Euclidean based on their class number. We now have a quantitative test for that property across the entire family of totally real abelian fields with degree two or higher >

Kolya: I just wonder about the constants, though; are those cK and xK really small enough to matter when we try to apply this to fields with larger degrees, say degree five or six?

Ada: That’s a good question. The paper focuses on the general case with degree two or higher, and they do build on earlier work that touches on those higher degrees as well. They aim for those constants to be uniform enough across all such fields which is what makes it so powerful >

Emmy: So we have this solid foundation now for understanding these rings. It sets a new standard for how we approach the problem of unique factorization in number fields.

Ada: And moving forward, this result opens up new avenues where we can use these counting functions to probe other types of algebraic structures, like those involving prime ideals splitting in extensions, which we touched on earlier >

Kolya: I’m looking forward to seeing how this analytic counting approach plays into the study of chaotic systems or other probabilistic models that have similar scaling behaviors. That seems like a natural next step for this kind of research >

Emmy: Right. We've done the heavy lifting here proving the equivalence for these abelian fields, and now we can look at what else this mathematical framework allows us to measure and control in our next set of papers >

Austrian Science Fund · EU NextGeneration project · Croatian Science Foundation

math.NT, math.RA

Submitted: 2026-10-17

Updated: 2026-11-06

Comments: Added further applications to certain families of non-Galois cubic, quartic and quintic fields, which are not necessarily totally real

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 85/100

The gist: ALL TOTALLY REAL ABELIAN NUMBER FIELDS OF CLASS NUMBER ONE ARE EUCLIDEAN is a short note that shows, more generally…

Key concepts

Euclidean Ring
A ring of integers OK is Euclidean if there exists a function (a norm) that allows any element to be divided by another with a remainder smaller than the divisor's norm. This property is crucial for performing division algorithms, similar to standard arithmetic in integers.
Class Number (hK)
The class number measures how far the ring of integers OK is from being a Unique Factorization Domain (UFD). If hK equals one, every non-zero element can be uniquely factored into prime ideals, which is a strong condition related to Euclidean properties.
Prime Ideal Counting Function ($\Pi_K(x)$)
This function counts the number of non-zero prime ideals in K whose norm (the size of the ideal) falls within a specific range, such as between x and 2x. The paper establishes a lower bound for this count, showing how many such ideals must exist.
Hecke L-function Zeros
Scholium (I) provides information about the zeros of certain Hecke L-functions associated with cyclotomic fields. It asserts that these functions have no zeros in a specific region, which is used to control the distribution of prime ideals in K.

Terminology

Summary

ALL TOTALLY REAL ABELIAN NUMBER FIELDS OF CLASS NUMBER ONE ARE EUCLIDEAN is a short note that shows, more generally, that the ring of integers OK of any totally real abelian number field K is a UFD, if and only if it is Euclidean The main theorem proves this equivalence for all totally real abelian number fields with degree at least two

Main Result Statement

The paper establishes a lower bound on the number of prime ideals in certain residue classes, leading to the conclusion that OK is Euclidean if and only if hK = 1 The main theorem states that for a totally real abelian number field K/Q with n ≥ 2, there are constants cK, xK > 0 such that ΠK(x) ≥ cK · x log(x)2 for all x > xK In particular, for all totally real abelian number fields K OK is Euclidean ⇐⇒ hK = 1

Verification of Previous Results

The authors verify Weinberger’s prediction regarding exceptional extensions by showing that the ring of integers OK is a UFD if and only if it is Euclidean for all totally real abelian number fields K This result builds upon Narkiewicz's earlier work concerning quadratic and cubic extensions The theorem overlaps with results from [Har-Mur04] and [Mur-Pet13] regarding non-existence of hypothetical exceptional extensions

Key Ingredients and Tools

The proof relies on two results from [OAI26], Thm. 1.2 and Prop. p. 0

Where the paper says it

The reader will immediately observe that the conclusion about OK being Euclidean if and only if K has class number hK = 1 indeed follows from (the fact that OQ = Z is Euclidean and) the first part of our theorem by a well-known and fundamental result of Harper-Murty, cf. [Har-Mur04], Prop.

2.1, which establish Scholium (I) and Scholium (II) Scholium (I) asserts that for every cyclotomic field F containing the group of 12th roots of unity µ12 and every Hecke character χ: A∗F/F∗ → C∗ of F of finite order, the associated Hecke L-function L(s, χ) has no zeros in the open region Re(s) > 1 − δ0 with a uniform width δ0 = 10−6 Scholium (II) provides a lower bound for certain prime ideals, stating that there are constants am, xm > 0 such that for all x ≥ xm (2. p. 1
Where the paper says it

Once verified and part of the a priori knowledge of mankind, they establish the following two important assertions, which shall be treated as a blackbox here: Scholium (I): There is a δ0 ∈ (0, 1/2) such that, for every cyclotomic field F containing the group of 12th roots of unity µ12 and every Hecke character χ : A ∗ F /F∗ → C ∗ of F of finite order, the associated Hecke L-function L(s, χ) has no zeros in the open region ℜe(s) > 1 − δ0. Here, the width δ0 is in fact independent of F and χ.

1) Pm(x) ≥ am x log(x)2

Derivation of the Lower Bound

The lower bound is derived by combining Scholium (I), (II), and Proposition 1.6, which provides a refined version of another result in [OAI26] By combining these, the authors obtain the desired lower bound of growth for the “prime ideal counting function” ΠK(x) This is shown in two steps, found in Sect. p. 2

Where the paper says it

When combined with Scholia (I) and (II), it yields the desired lower bound of growth for the “prime ideal counting function” ΠK(x). This final argument is carried out in two steps, to be found in our Sect.

2 and Sect. p. 2
Where the paper says it

When combined with Scholia (I) and (II), it yields the desired lower bound of growth for the “prime ideal counting function” ΠK(x). This final argument is carried out in two steps, to be found in our Sect.

p. 2
Where the paper says it

When combined with Scholia (I) and (II), it yields the desired lower bound of growth for the “prime ideal counting function” ΠK(x). This final argument is carried out in two steps, to be found in our Sect.

3 respectively

Analysis of Exceptional Cases

The analysis of prime ideals not satisfying the surjectivity condition involves bounding the number of exceptional divisors A(x) and B(x) The number A(x) is bounded above by a quantity that is shown to be o x log(x)2 as x → ∞ Similarly, the number B(x) is shown to be o x log(x)2 as x → ∞

Proof of the Main Theorem

The proof of the main theorem concludes that ΠK(x) ≥ nPm(x) − A(x) − B(x) ≥ nam x log(x)2 − o x log(x)2 For x ≫ 1, they choose an xK large enough such that all preceding estimates apply and that A(x) + B(x) ≤ am x log(x)2 for x ≥ xK This establishes the claimed lower bound of our Main Theorem >

Corollary Result

The corollary proves that for all sufficiently large x and every odd prime l ∤ D with l ≤ exp(log(x) 3/10), (1.21) Sl(x) ≪K x l(l − 1) log(x) + x 1−δ0 Both the implied constant and the threshold for x are independent of l

Prime Ideals of Non-Surjective Reduction

The analysis of prime ideals with non-surjective reduction shows that every p ∈ Pm(x) provides n distinct prime ideals of OK of norm p The index i(p) is odd, and if i(p) > 1, it belongs to at least one of the two exceptional sets bounded in cardinality in Sections 2.2 and 2.3

Exceptional Divisors of i(p)

The analysis concerning the case where q i(p) shows that B(x) = o x log(x)2 as x → ∞ This is derived from bounding the number of distinct prime ideals p above primes p ∈ Pm(x) with q i(p), yielding B(x) ≪K x 1/5 log(x)

Final Conclusion

The final step in the proof shows that ΠK(x) ≥ nPm(x) − A(x) − B(x) ≥ nam x log(x)2 − o x log(x)2 for x ≫ 1 This establishes the claimed lower bound of our Main Theorem >

References

[Gro26] H. Grobner, On divergence related to Riemann–von Mangoldt’s explicit formula of the prime-counting function, preprint, arXiv:2609. p. 1

Where the paper says it

6 below, we establish a refined, as well as generalized version of another result in [OAI26] (its Lem. Our more general approach here is in fact guided by our recent work on the prime-counting function π(x), cf.

02713v2 (2026)

[Har-Mur04] M. Harper, M. R. Murty, Euclidean rings of algebraic integers, Canadian J. p. 0

Where the paper says it

In particular, for all totally real abelian number fields K OK is Euclidean ⇐⇒ hK = 1. The reader will immediately observe that the conclusion about OK being Euclidean if and only if K has class number hK = 1 indeed follows from (the fact that OQ = Z is Euclidean and) the first part of our theorem by a well-known and fundamental result of Harper-Murty, cf.

Math. 56 (2004) 71–76

[Has-She-Won22] E. Hasanalizade, Q. Shen, P.-J. Wong, Counting zeros of Dedekind zeta functions, Math. p. 3

Where the paper says it

Consequently Eℓ = Fℓ(ε 1/ℓ), and Eℓ/Fℓ is cyclic of degree 1 or ℓ. The factorization of a Dedekind zeta function for an abelian extension into finite-order Hecke L-functions gives ζEℓ (s) = Y χ∈Gal(\Eℓ/Fℓ) L(s, χ).

Computation 91 (2022) 277–293

[HKMS24] D. Hu, I. Kaneko, S. Martin, C. Schildkraut, On a Mertens-type conjecture for number fields, Math. p. 0

Where the paper says it

ALL TOTALLY REAL ABELIAN NUMBER FIELDS OF CLASS NUMBER ONE ARE EUCLIDEAN HARALD GROBNER Abstract. It is a famous result of Narkiewicz that there are at most two real quadratic and at most one totally real cubic Galois extensions K/Q, whose ring of integers OK is a unique factorization domain (UFD), but not Euclidean.

Proc. Cambridge Philos. Soc. 177 (2024), 481–523

[Iwa-Kow04] H. Iwaniec, E. Kowalski, Analytic Number Theory, American Mathematical Society Colloquium Publications 53 (AMS, Providence, RI, 2004)

[Mur-Pet13] M. R. Murty, K. L. Petersen, The Euclidean algorithm for number fields and primitive roots, Proc. p. 0

Where the paper says it

ALL TOTALLY REAL ABELIAN NUMBER FIELDS OF CLASS NUMBER ONE ARE EUCLIDEAN HARALD GROBNER Abstract. It is a famous result of Narkiewicz that there are at most two real quadratic and at most one totally real cubic Galois extensions K/Q, whose ring of integers OK is a unique factorization domain (UFD), but not Euclidean.

Amer. Math. Soc. 141 (2013) 181–190

[Nar-07] W. Narkiewicz, Euclidean algorithm in small Abelian fields, Funct. p. 0

Where the paper says it

ALL TOTALLY REAL ABELIAN NUMBER FIELDS OF CLASS NUMBER ONE ARE EUCLIDEAN HARALD GROBNER Abstract. It is a famous result of Narkiewicz that there are at most two real quadratic and at most one totally real cubic Galois extensions K/Q, whose ring of integers OK is a unique factorization domain (UFD), but not Euclidean.

Approx. Comment. Math. 37 (2007) 337–340

[Neu99] J. Neukirch, Algebraic Number Theory, translated by N. p. 2

Where the paper says it

For the second formula, we first recall that the unit ε is not an ℓ-th power in K. Indeed, an element of K, whose ℓ-th power is ε, is an algebraic integer, and its inverse is an algebraic integer as well, whence it is a unit of OK.

Schappacher, Grundlehren math. p. 0
Where the paper says it

1 arXiv:2610. 24900v1 [math.

Wissenschaften 322 (Springer, Berlin, 1999)

[OAI26] OpenAI, Primitive roots for every admissible integer base, preprint, Oct 4, (2026) OpenAI mathematics repository

[Wei73] P. J. Weinberger, On Euclidean rings of algebraic integers, in: Analytic Number Theory Proc. p. 0

Where the paper says it

As predicted by another famous result due to Weinberger, these exceptions should in fact not exist, as their existence would contradict a generalized version of the Riemann hypothesis. Building on recent spectacular progress on the Quasi-Riemann hypothesis for Hecke characters, in this short note we close the above final gap provided by Narkiewicz’s result, and show, more generally, that the ring of integers OK of any totally real abelian number field K is a UFD, if and only if it is Euclidean.

Symp. Pure Math. 24 (AMS, Providence, RI, 1973) 321–332

--- Page 1 ---

ALL TOTALLY REAL ABELIAN NUMBER FIELDS OF CLASS NUMBER ONE ARE EUCLIDEAN

Abstract. It is a famous result of Narkiewicz that there are at most two real quadratic and at most one totally real cubic Galois extensions K/Q, whose ring of integers OK is a unique factorization domain (UFD), but not Euclidean As predicted by another famous result due to Weinberger, these exceptions should in fact not exist, as their existence would contradict a generalized version of the Riemann hypothesis Building on recent spectacular progress on the Quasi-Riemann hypothesis for Hecke characters, in this short note we close the above final gap provided by Narkiewicz’s result, and show, more generally, that the ring of integers OK of any totally real abelian number field K is a UFD, if and only if it is Euclidean

Introduction

About 20 years ago, Narkiewicz proved in a short paper, cf. [Nar07], that there are at most two quadratic and at most one cubic totally real Galois extensions K/Q, whose ring of integers OK is a unique factorization domain (UFD), but not Euclidean As OK is Dedekind, this amounts to the dictum that there are at most two real quadratic and at most one totally real cubic Galois extensions K/Q such that OK is a principal ideal domain (PID), i. p. 0

Where the paper says it

[Nar07], that there are at most two quadratic and at most one cubic totally real Galois extensions K/Q, whose ring of integers OK is a unique factorization domain (UFD), but not Euclidean. As OK is Dedekind, this amounts to the dictum that there are at most two real quadratic and at most one totally real cubic Galois extensions K/Q, such that OK is a principal ideal domain (PID), i.

e., K has class number hK = 1, but there is no function ω: OK ? p. 0
Where the paper says it

As OK is Dedekind, this amounts to the dictum that there are at most two real quadratic and at most one totally real cubic Galois extensions K/Q, such that OK is a principal ideal domain (PID), i. , K has class number hK = 1, but there is no function ω : OK \ {0} → Z≥0 such that for every a, b ∈ OK, b ̸= 0, there exist q, r ∈ OK such that a = bq+r and either r = 0 or ω(r) < ω(b).

→ Z≥0 such that for every a, b ∈ OK, b ≠ 0, there exist q, r ∈ OK such that a = bq+r and either r = 0 or ω(r) < ω(b) When Narkiewicz wrote his famous note, it had been widely known, that such exceptional extensions K/Q should not exist, as their existence would have contradicted a generalized version of the Riemann hypothesis (stated for infinite families of certain Hecke L-functions of cyclotomic extensions), which is an important result of Weinberger In this short note we verify this prediction of Weinberger’s result and prove the following unconditional result for all totally real abelian number fields K/Q with n = [K: Q] ≥ 2:

Main Theorem. Let K/Q be a totally real abelian number field with n = [K: Q] ≥ 2, ring of integers OK and class number hK. p. 0

Where the paper says it

In this short note we verify this prediction of Weinberger’s result and prove the following unconditional result for all totally real abelian number fields of degree at least two: Main Theorem. Let K/Q be a totally real abelian number field with n = [K : Q] ≥ 2, ring of integers OK and class number hK.

For x > 1, let ΠK(x) be the number of non-zero prime ideals p of K, whose norm N(p) is a rational prime strictly between x and 2x, i. p. 0
Where the paper says it

Let K/Q be a totally real abelian number field with n = [K : Q] ≥ 2, ring of integers OK and class number hK. For x > 1, let ΠK(x) be the number of non-zero prime ideals p of K, whose norm N(p) is a rational prime strictly between x and 2x, i.

e., x < N(p) < 2x, and for which the natural map O∗K → (OK/p)∗ is surjective Then there are constants cK, xK > 0 such that ΠK(x) ≥ cK · x log(x)2 for all x > xK In particular, for all totally real abelian number fields K OK is Euclidean ⇐⇒ hK = 1

Page 3

reasoning, which goes exactly along the same lines as the arguments in [Nar07] as well as along the ones in the just quoted [Har-Mur04], but also in the more recent [Mur-Pet13] In fact, our theorem overlaps and partly reproves some important results in all of the aforementioned sources in the case of totally real abelian extensions K/Q – we particularly mention [Har-Mur04], Thm. 2 (K/Q abelian with rkZ(OK) ≥ 3), and [Mur-Pet13], Cor. p. 1

Where the paper says it

In fact, our theorem overlaps and partly reproves some important results in all of the aforementioned sources in the case of totally real abelian extensions K/Q – we particularly mention [Har-Mur04], Thm. 2 (K/Q abelian with rkZ(OK) ≥ 3), and [Mur-Pet13], Cor.

1.5, (K/Q totally real, Galois over some proper subfield and [K: Q] ≥ 5) – whereas it establishes the non-existence of the hypothetical exceptional quadratic, resp. p. 1
Where the paper says it

2 (K/Q abelian with rkZ(OK) ≥ 3), and [Mur-Pet13], Cor. 5, (K/Q totally real, Galois over some proper subfield and [K : Q] ≥ 5) – whereas it establishes the non-existence of the hypothetical exceptional quadratic, resp.

cubic extensions of the aforementioned main result of [Nar07]

Page 4

Scholium (I): There is a δ0 ∈ (0, 1/2) such that, for every cyclotomic field F containing the group of 12th roots of unity µ12 and every Hecke character χ: A∗F/F∗ → C∗ of F of finite order, the associated Hecke L-function L(s, χ) has no zeros in the open region Re(s) > 1 − δ0 Scholium (II): For every integer k ≥ 1, there are ak, xk > 0 such that for every x ≥ xk, at least akx/ log(x)2 distinct rational primes p with x

x9/10 is prime and r is a positive integer, whose prime divisors l are all in between exp(log(x) 1/10) and exp(log(x) 3/10)

Page 5

Proposition 1.6. Let F be any number field and let φ ∈ C∞ c (R>0, R≥0) be any non-negative, real valued smooth function of compact support. p. 4

Where the paper says it

ALL TOTALLY REAL ABELIAN NUMBER FIELDS OF CLASS NUMBER ONE ARE EUCLIDEAN 5 developed independently in the course of the research leading to the author’s recent results on the prime counting function, cf. To formulate the result, let F be any number field, let φ ∈ C∞ c (R>0, R≥0) be any non-negative, real valued smooth function of compact support on R>0 = (0, ∞).

Then, for x > 1, (1.7) ΘF,φ(x) ≪φ xφb(1) + x ΘF (log DF + [F: Q])

Page 6

We now want to shift the integral to Re(s) = −1/2, while keeping track of its dependence on F Combining this with Scholia (I) and (II), it yields the desired lower bound of growth for the “prime ideal counting function” ΠK(x) This final argument is carried out in two steps, to be found in our Sect. 2 and Sect. 3 respectively

Page 7

On the other hand, the contribution at 0 is Oφ(nF) and hence may be, together with the remainder in (1.19), absorbed into Oφ(x ΘF (log DF + nF) This proves (1. p. 7

Where the paper says it

19), absorbed into Oφ(x ΘF (log DF + nF )). This proves (1.

7) >

Page 8

We now compute its value in a second way by use of the residue theorem: The pole of ζF (s) at s = 1 contributes xφb(1), because −ζ′F (s)/ζF (s) has residue +1 there A non-trivial zero ρ of multiplicity mρ of ζF (s) contributes −mρxρφb(ρ) At s = 0, the order of ζF is m0 = r1 + r2 − 1, giving the contribution −m0φb(0)

Page 9

Then, by Scholium (II), there are constants am, xm > 0 such that for all x ≥ xm (2.1) Pm(x) ≥ am x log(x)2

Page 10

So, dividing again by x/ log(x)2 and observing that x 1/5/ log(x)2 = log(x) x4/5 −→ 0 as x → ∞

Page 11

Let now Sl(x) be the number of prime ideals p of OK, whose norm N(p) is a rational prime, satisfying x < N(p) < 2x and that split completely in Hl/K Then we obtain the following Corollary 1.20. For all sufficiently large x and every odd prime l ∤ D with l ≤ exp(log(x) 3/10), (1. p. 7

Where the paper says it

Then we obtain the following Corollary 1. For all sufficiently large x and every odd prime ℓ ∤ D with ℓ ≤ exp(log(x) 3/10), (1.

21) Sl(x) ≪K x l(l − 1) log(x) + x 1−δ0 Both the implied constant and the threshold for x are independent of l

Page 13

The proof of the Main Theorem is finally ready to give the proof of our main result Let xm, am > 0 be the constants provided by Scholium (II) with k = m and set cK:= (n − 1)am By Lem. 2.2 the rational primes in Pm(x) provide exactly nPm(x) distinct prime ideals p of OK, whose norm is a rational prime Those, for which i(p) > 1 belong to at least one of the two exceptional sets just bounded in cardinality in Sections 2. p. 1

Where the paper says it

The next assertion is far more technical in nature and borrowed from the same source: Scholium (II): For every integer k ≥ 1, there are ak, xk > 0 such that for every x ≥ xk, at least akx/ log(x) 2 distinct rational primes p with x < p < 2x satisfy (0. 1) p ≡ −1 (mod 8k), and p − 1 = 2rq, where q > x9/10 is prime and r is a positive integer, whose prime divisors ℓ are all in between exp(log(x) 1/10) and exp(log(x) 3/10).

2 and 2.3, i.e., more precisely, the number of these prime ideals p is bounded above by A(x) + B(x) Therefore, recalling (2. p. 4
Where the paper says it

5) ΘF,φ(x) := X q X j≥1 log(N(q)) φ N(q) j x , where the first sum runs over all non-zero prime ideals q of OF . We remark that this sum is well-defined as it is in fact finite for each fixed x, because φ has compact support and only finitely many prime-ideal powers have bounded norm.

1), (2.4) and (2.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved prime ideal counting function for totally real abelian fields K/Q of degree 2 or 3 by establishing a lower bound ΠK(x) >= cK x log(x) squared for all x > xK. This allows AI systems to estimate the density of prime ideals whose norm falls within specific ranges relative to the field structure.

  2. Verification of Narkiewicz's prediction regarding exceptional extensions by proving that OK is Euclidean if and only if hK = 1 for all totally real abelian number fields K with degree n >= 2. This provides a definitive criterion for determining whether the ring of integers is Euclidean based on the class number.

  3. Development of a generalized version of the quasi-Riemann hypothesis for all finite-order Hecke characters of cyclotomic fields containing mu12, established by Scholium (I). This capability allows AI systems to utilize zero-free regions for Dedekind zeta functions associated with these specific character L-functions.

  4. Formulation and proof of a refined version of the prime counting function estimate ΘF,φ(x) = O φ x φb(1) + x ΘF (log DF + [F: Q]) for any number field F and smooth compactly supported function phi. This enables AI to precisely quantify the density of prime ideals in terms of the field discriminant and degree.

  5. Construction of a lower bound for the number of prime ideals p splitting completely in Hl/K, Sl(x) ≪ K x l(l - 1) log(x) + x 1-δ0. This allows AI to estimate the distribution of specific types of prime ideals related to extensions Hl/K.

  6. Bounding the number of prime ideals p whose reduction map O∗K -> (OK/p)∗ is not surjective, A(x), by showing A(x) = o x log(x) squared. This enables AI to prove that the vast majority of relevant prime ideals satisfy the surjectivity condition i(p) = 1.

  7. Derivation of a bound on B(x), the number of distinct prime ideals p above primes p in Pm(x) where q i(p), showing B(x) = o x log(x) squared. This allows AI to control the contribution from exceptional prime ideal types when analyzing surjectivity failures.

  8. A final proof for the main theorem establishing ΠK(x) >= nPm(x) - A(x) - B(x) >= n a m x log(x) squared - o x log(x) squared for x >> 1. This provides the definitive quantitative estimate for the number of prime ideals whose norm is between x and 2x and which satisfy the surjectivity condition.

Abstract

It is a famous result of Narkiewicz that there are at most two real quadratic and at most one totally real cubic Galois extensions K/ Q, whose ring of integers O K is a unique factorization domain (UFD), but not Euclidean. As predicted by another famous result due to Weinberger, these exceptions should in fact not exist, as their existence would contradict a generalized version of the Riemann hypothesis. Building on recent spectacular progress on the Quasi-Riemann hypothesis for Hecke characters, in this short note we close the above final gap provided by Narkiewicz's result, and show, more generally, that the ring of integers O K of any totally real abelian number field K is a UFD, if and only if it is Euclidean. By referring to an additional announced uniform zero-free half-plane for general Dedekind zeta functions, we also establish the corresponding result for certain explicit new families of non-Galois cubic, quartic and quintic fields of unit rank one or two.

Sources

Related papers