Branched covers of $\mathbb{P}^1$ and divisibility in class group

The paper establishes a construction that generates $n$-torsion elements in the class group of a specific number field $K$ by utilizing $n$-torsion points from the Jacobian of an $m$-gonal curve.

The Gist

The Big Picture: The "Mathematical Treasure Hunt"

Imagine you are a treasure hunter. In the world of numbers, there is a hidden structure called the Class Group. Think of this group as a vault containing "keys" (mathematical elements). Some keys are simple, but some are very complex.

The big question mathematicians have been asking for a long time is: "How can we find a specific, complex key (an element of a certain order) inside this vault?"

For example, can we find a key that, when you turn it 5 times, it unlocks nothing (returns to zero), but turning it 1, 2, 3, or 4 times does something? This is called finding an element of "order 5."

This paper presents a new, clever way to find these keys. Instead of digging directly into the vault of numbers, the authors use a bridge built from shapes and geometry.

---

The Analogy: The "Shape-Shifting Bridge"

To understand their method, let's break down the technical terms into a story:

1. The Starting Point: A Twisted Rope (The Curve)

The authors start with a specific shape called a curve (specifically, a "super-elliptic curve"). Imagine this as a long, twisted rope floating in space.

• The Rope: This rope has a special property. If you look at it closely, it has "twists" or "knots" (torsion points) that repeat in a cycle.
• The Map: They draw a map from this rope down to a simple line (called $\mathbb{P}^1$). Imagine shining a light on the rope and seeing its shadow on the floor.

2. The Bridge: Spreading the Rope (The Fibration)

Now, they take this rope and the shadow, and they build a bridge that stretches across the entire landscape of numbers (from the rational numbers $\mathbb{Q}$ all the way to the integers $\mathbb{Z}$).

• Think of this bridge as a long hallway.
• At every specific spot (every prime number) along the hallway, the bridge has a "floor" (a fiber).
• Most of the time, the floor is smooth and perfect. These are the "good primes."
• Sometimes, the floor gets a little bumpy or broken. These are the "bad primes."

3. The Magic Trick: Dropping a Stone (Restriction)

Here is the core of their discovery:

• They take one of those special "twists" (knots) from the original rope (the Jacobian of the curve).
• They "drop" this twist down the bridge to a specific floor (a specific number field).
• The Result: If they drop it on a smooth floor (a good prime), that twist doesn't disappear. Instead, it lands in the Class Group of that specific number field.

The Metaphor: Imagine you have a specific, unique pattern of ripples in a pond (the twist on the curve). You build a canal system (the bridge) leading to hundreds of different small ponds (number fields). If you pour the water from the main pond into a small, clear pond, the ripples travel with it. The authors prove that these ripples (the torsion elements) survive the journey and become real, physical objects in the new pond.

---

The Main Discovery (Theorem 1.1)

The authors prove a powerful statement:

If your starting rope has a twist, you can find infinitely many different number fields (new ponds) that also have that same twist.

In plain English:
If you start with a curve that has a specific type of mathematical "knot" (an $n$-torsion element), you can use this bridge method to generate an infinite list of number fields. In the "vault" (Class Group) of each of these fields, you are guaranteed to find that same knot.

This is huge because finding these knots is usually very hard. This method gives a recipe to generate them endlessly.

---

The "Why It Works" (The Monodromy)

You might ask: "How do they know the twist doesn't just vanish when it hits the floor?"

The authors use a concept called Étale Monodromy.

• Analogy: Imagine walking around a tree in a forest. If you hold a ribbon and walk around the tree, the ribbon might get twisted. If you walk around a different tree, the ribbon might twist differently.
• The "Monodromy" is the study of how things twist as you move around different points in the landscape.
• The authors show that because the twist exists in the "center" of the system (the generic point), and the system is connected, the twist must appear in almost all the specific spots (the number fields) along the way. It's like saying, "If the wind is blowing in the center of the room, it must be blowing in almost every corner of the room."

---

A Real-World Example (Section 3)

The paper gives a concrete example to show this isn't just theory:

• They look at a curve defined by $y^5 = x^5 - 31$.
• This curve has a "5-fold twist."
• Using their bridge method, they prove there are infinitely many number fields related to the 5th roots of unity (a specific type of number system) where the Class Group is divisible by 5.
• The Punchline: This confirms a deep connection between the shape of the curve and the hidden structure of numbers. It even helps prove that certain famous numbers (Bernoulli numbers) must be divisible by specific primes.

Summary

1. The Problem: Finding specific "keys" (torsion elements) in the vault of number fields is difficult.
2. The Tool: Build a bridge from a geometric shape (a curve) that has these keys built-in.
3. The Action: Project the keys from the shape down onto specific number fields.
4. The Result: The keys survive the journey. You now have an infinite supply of number fields that contain these specific keys.

The authors have essentially found a factory that can mass-produce number fields with specific, complex internal structures, using geometry as the blueprint.

Technical Summary

1. Problem Statement

The central problem addressed in this paper is a classical question in algebraic number theory: How to explicitly construct elements of a specific order $n$ within the class group of a number field.

While it is known that class groups are finite, finding elements of large order or proving the existence of $n$-torsion elements in infinitely many number fields is difficult. Previous approaches have utilized:

• Algebro-geometric methods to find large order elements (Agboola and Pappus).
• Pulling back torsion line bundles to class groups (Gillibert and Levin).
• Specific constructions for hyperelliptic curves to quadratic fields (Gillibert).

This paper aims to generalize these approaches by establishing a systematic method to generate $n$-torsion elements in the class groups of number fields derived from branched covers of the projective line $\mathbb{P}^1$.

2. Methodology

The authors employ a synthesis of arithmetic geometry, specifically utilizing Chow schemes, Hilbert schemes, and étale monodromy. The methodology proceeds through the following logical steps:

A. Geometric Setup: Spreading over $\text{Spec}(\mathbb{Z})$

1. Initial Curve: Start with a smooth projective curve $C$ defined over $\mathbb{Q}$ equipped with a regular $m:1$ map to $\mathbb{P}^1_{\mathbb{Q}}$. This map is branched.
2. Integral Model: Construct an integral model of this fibration over $\text{Spec}(\mathbb{Z})$, denoted as $C \to \mathbb{P}^1_{\mathbb{Z}}$.
3. Fiber Restriction: For a "good" prime $p$ (where the fiber is smooth), the fiber $C_p$ corresponds to the spectrum of the ring of integers $\mathcal{O}_L$ of a number field $L$.
   • Example: For a super-elliptic curve $y^m = f(x)$, choosing a specific $x=p$ yields an extension of $\mathbb{Z}$ whose integral closure is the ring of integers of a number field.

B. Torsion Propagation via Chow Groups

1. Source Torsion: Identify a non-trivial $n$-torsion element $\alpha$ in the Jacobian $J(C)$ (identified with the degree 0 Chow group $A^1(C)$).
2. Spreading the Cycle: Extend this torsion element $\alpha$ to a "spread" $\tilde{\alpha}$ over the arithmetic surface $\mathbb{P}^1_{\mathbb{Z}}$.
3. Restriction: Restrict $\tilde{\alpha}$ to the fiber $C_P$ at a point $P \in \mathbb{Z}$. This restriction yields an element in the Chow group of the fiber.
4. Descent to Class Group: By restricting this element further to the affine open subset corresponding to the ring of integers $\mathcal{O}_L$, the authors obtain an element in the class group of the number field.

C. Theoretical Framework: Chow Varieties and Monodromy

To prove that the restricted torsion element remains non-trivial for infinitely many fibers, the authors utilize:

• Chow Schemes: They analyze the family of cycles parametrized by the Chow variety $C^1_{d,d}(C_U/U)$.
• Rational Equivalence: They define rational equivalence for arithmetic varieties using morphisms from $\mathbb{P}^1_U$.
• Theorem 2.1: They prove that the set of points where a torsion cycle becomes trivial (rationally equivalent to zero) forms a countable union of Zariski closed subsets within the Chow variety.
• Étale Monodromy: By considering the étale fundamental group $\pi_1(U, \bar{\eta})$ acting on the Tate module $T_l(C_{\bar{\eta}})$, they show that if the Tate module is non-trivial, the torsion property is preserved under the monodromy action. This ensures that the torsion element does not vanish for "most" fibers.

3. Key Contributions and Results

Main Theorem (Theorem 1.1)

Let $C$ be a smooth projective curve over $\mathbb{Q}$ with a branched cover to $\mathbb{P}^1$. If the Tate module $T_l(C_{\bar{\eta}})$ is non-trivial (implying $H^1(C_{\bar{\eta}}, \mathbb{Q}_l) \neq 0$), then:

There exist infinitely many number fields $L$ such that the class group of $L$ contains elements of order $l^i$ for some integer $i \geq 1$ and some prime $l$.

Technical Proofs

• Theorem 2.1: Establishes that the locus of trivial torsion in the family of fibers is "small" (a countable union of closed sets). This implies that for a generic choice of the fiber (i.e., a generic number field in the family), the torsion element remains non-trivial.
• Sheaf Theory: The proof utilizes the locally constant constructible sheaf $R^1 f_{U*} \mathbb{Z}/l^n\mathbb{Z}$ and its relation to the étale cohomology $H^1(C_{\bar{\eta}}, \mathbb{Z}/l^n\mathbb{Z})$. The non-vanishing of the Tate module ensures the existence of torsion points that persist across the family.

Concrete Example (Section 3)

The authors apply their theory to the super-elliptic curve:
$$y^5 = x^5 - 31$$

• Since the Tate module is non-trivial, there exist infinitely many finite extensions $L$ of the cyclotomic field $K = \mathbb{Q}(\zeta_5)$ where the class group has $l^i$-torsion.
• Specifically, for $l=5$, if the degree $[L:K]$ is not divisible by 5, the class group of $L$ surjects onto the class group of $K$, implying 5-divisibility in the class number of $\mathbb{Q}(\zeta_5)$.
• This connects to the Herbrand-Ribet Theorem, suggesting that for any prime $p$, there exists an extension where the class number is divisible by $p$, which implies $p$ divides a specific Bernoulli number $B_k$.

4. Significance

• Unification of Fields: The paper successfully bridges the gap between the geometry of curves (Jacobian torsion) and the arithmetic of number fields (class group structure).
• Infinite Families: Unlike previous results that might find torsion in specific fields, this method guarantees the existence of infinitely many number fields with specific torsion properties.
• Constructive Approach: It provides a concrete recipe (via branched covers and spreading torsion cycles) to generate number fields with large class group divisibility, offering a new tool for investigating the distribution of class numbers.
• Generalization: The method is not limited to hyperelliptic curves but applies to general branched covers of $\mathbb{P}^1$, broadening the scope of algebro-geometric applications in number theory.

In summary, the paper demonstrates that non-trivial torsion in the Jacobian of a curve over $\mathbb{Q}$ can be "spread" to produce non-trivial torsion in the class groups of infinitely many number fields arising as fibers of the curve's integral model.