A note on zero-cycles on bielliptic surfaces

This paper investigates the Chow group of zero-cycles on bielliptic surfaces over arbitrary fields of characteristic not 2 or 3, proving that the kernel of the Albanese map is a torsion group with a specific exponent and providing explicit examples over $p$-adic fields where this kernel contains nontrivial elements arising from abelian surfaces.

The Gist

Imagine you are a detective trying to solve a mystery about the "shape" of a very strange, multi-layered world. In the world of mathematics, this world is called a bielliptic surface.

Here is a simple breakdown of what Evangelia Gazaki's paper is about, using everyday analogies.

1. The Setting: A Twisted World

Think of a bielliptic surface as a fancy, twisted piece of fabric.

• The Fabric: It's made by taking two loops of string (mathematicians call these elliptic curves, which look like donuts) and weaving them together to make a grid.
• The Twist: Now, imagine you have a magical robot (a finite group $G$) that grabs this grid and folds it up, gluing certain points together. The result is a new, smaller shape called a bielliptic surface.
• The Rules: The robot is picky. It only glues points in specific ways (translations and rotations) so that the final shape is smooth and doesn't have any sharp corners.

2. The Mystery: The "Zero-Cycle" Kernel

Mathematicians are interested in counting "points" on this shape. They group these points into categories called Chow groups.

• The Degree Map: Imagine you have a pile of stones (points). You can count them. If you have 5 stones, the "degree" is 5.
• The Kernel (The Mystery): What happens if you have a pile of stones that looks like it has zero weight, but isn't actually empty?
  • Think of it like a bank account. You might have a transaction that looks like it balances to zero, but there's actually a hidden fee or a ghost transaction inside.
  • In math, this hidden "ghost" part is called the Albanese kernel ($T(S)$). It represents zero-cycles that are "algebraically trivial" (they cancel out in a big picture) but might still have some hidden structure.

The Big Question: If you find a "ghost" point in this twisted world, how "heavy" is it? Is it a tiny speck, or a massive boulder? Can you keep dividing it forever, or does it eventually hit a limit?

3. The Main Discovery: The "Explosion Limit"

Gazaki's first major finding (Theorem 1) answers the question: "How big can these ghost points get before they disappear?"

• The Analogy: Imagine these ghost points are balloons. You can blow them up, but there is a maximum size they can reach before they pop (become zero).
• The Result: The paper proves that no matter how you twist the fabric (depending on the type of robot $G$ you used), these balloons cannot get bigger than a specific size.
  • If the robot is a "Type 2" robot (involving numbers divisible by 2), the balloon pops if you try to inflate it more than $4 \times |G|$ times.
  • If the robot is a "Type 3" robot (involving numbers divisible by 3), the limit is $9 \times |G|$.
• Why it matters: Before this, mathematicians knew these ghosts were finite (they weren't infinite), but they didn't know exactly how big the "pop limit" was. Gazaki gave us the exact number.

4. The Second Discovery: The "Bad Neighborhood"

The second part of the paper (Theorem 2) asks: "Do these ghost points actually exist, or are they just theoretical?"

• The Setting: The author looks at these surfaces over p-adic fields. Think of these as a specific type of mathematical "neighborhood" (like the 11-adic numbers).
• The Strategy: To prove the ghosts exist, the author builds a specific example where the "fabric" (the elliptic curves) is broken (has "bad reduction").
  • Analogy: Imagine trying to find a hidden treasure in a perfectly smooth, pristine park. It's hard to find anything unusual there. But if you go to a park that is under construction, with holes and broken fences (bad reduction), it's much easier to find hidden things.
• The Tool: The author uses a "detective tool" called the Brauer-Manin pairing.
  • Think of this as a special magnet. The "ghost points" are made of iron, and the "Brauer group" is a magnet. If you bring them close, they stick together.
  • If the magnet pulls on the ghost point, it proves the ghost point is real and not just zero.
• The Result: The author successfully built a "broken" surface where the magnet does stick. This proves that these hidden ghost points actually exist and are not just mathematical fantasies.

5. The Twist: Good vs. Bad Roads

The paper ends with a fascinating observation:

• If the surface is "smooth" (good reduction), the ghost points might be so small they are invisible (divisible by 2 forever).
• But if the surface is "broken" (bad reduction), the ghost points become visible and have a definite size.

Summary

In simple terms, this paper is about measuring the invisible.

1. We proved that the "invisible" parts of these twisted surfaces have a strict size limit (they are finite and bounded).
2. We built a model showing that these invisible parts are real, but you only see them clearly when the surface is "broken" or imperfect.

It's like discovering that while a perfect mirror reflects nothing new, a cracked mirror reveals hidden patterns that were always there, just waiting to be measured.

Technical Summary

Here is a detailed technical summary of the paper "A note on zero-cycles on bielliptic surfaces" by Evangelia Gazaki.

1. Problem Statement

The paper investigates the structure of the Chow group of zero-cycles, denoted $CH_0(S)$, for bielliptic surfaces $S$ defined over an arbitrary field $k$ (with characteristic $\neq 2, 3$).

Specifically, the author focuses on the Albanese kernel, denoted $T(S)$. For a smooth projective variety $X$, there is an Albanese map:
$$\text{alb}_X : A_0(X) \to \text{Alb}_X(k)$$
where $A_0(X)$ is the subgroup of zero-cycles of degree 0 (algebraically trivial cycles). The kernel of this map, $T(X) = \ker(\text{alb}_X)$, is the object of study.

• Context: When $k$ is algebraically closed, Bloch, Kas, and Lieberman proved $T(S) = 0$. However, for non-closed fields (finite, number, or $p$-adic fields), $T(S)$ is known to be a torsion group, but its specific exponent and structure were not fully explicit for bielliptic surfaces.
• Goal: Determine the precise exponent of the torsion group $T(S)$ for bielliptic surfaces over arbitrary fields and construct explicit examples where $T(S)$ is non-trivial over $p$-adic fields.

2. Methodology

The paper employs a combination of geometric reduction, intersection theory, and arithmetic geometry (specifically the Brauer-Manin obstruction).

A. Geometric Reduction via Étale Covers

Bielliptic surfaces are defined as étale quotients $S = (E_1 \times E_2)/G$, where $E_1, E_2$ are elliptic curves and $G$ is a finite abelian group acting on $E_1$ by translations and on $E_2$ by automorphisms.

• Classification: The author utilizes the Bagnera-De Franchis classification, which identifies 7 types of bielliptic surfaces based on the group $G$ (orders ranging from 2 to 9).
• Reduction Strategy: The proof of the main theorem relies on reducing the general case to two specific "primitive" types (Type 1 and Type 5) using intermediate étale covers.
  • If $|G|$ is composite, there exists a bielliptic surface $\tilde{S}$ such that $X \to \tilde{S} \to S$.
  • The map $\pi': \tilde{S} \to S$ induces a push-forward on Chow groups. The author shows that the exponent of $T(S)$ is controlled by the degree of this map and the exponent of $T(\tilde{S})$.
  • This reduces the problem to analyzing surfaces where $G \cong \mathbb{Z}/2\mathbb{Z}$ (Type 1) or $G \cong \mathbb{Z}/3\mathbb{Z}$ (Type 5).

B. Analysis of Zero-Cycles on Products of Elliptic Curves

The author analyzes the kernel $T(X)$ for $X = E_1 \times E_2$.

• Generators: $T(X)$ is generated by zero-cycles of the form $z_{P,Q} = [P, Q] - [P, O_2] - [O_1, Q] + [O_1, O_2]$.
• Bilinearity: These generators are bilinear in $P$ and $Q$.
• Push-forward Calculation: The core technical work involves computing the push-forward $\pi_*(z_{P,Q})$ under the quotient map $\pi: X \to S$. By exploiting the group action (e.g., $\sigma(P, Q) = (P+P_0, -Q)$ for Type 1), the author derives relations showing that specific linear combinations of these cycles are torsion.

C. Brauer-Manin Pairing for $p$-adic Examples

To prove the existence of non-trivial elements in $T(S)$ over $p$-adic fields, the author utilizes the Brauer-Manin pairing:
$$\langle \cdot, \cdot \rangle : CH_0(S) \times \text{Br}(S) \to \mathbb{Q}/\mathbb{Z}$$

• Strategy: Construct a pair of elliptic curves $E_1, E_2$ with split multiplicative reduction over a $p$-adic field $k$.
• Transcendental Brauer Group: The author analyzes the transcendental part of the Brauer group $\text{Br}(S)/\text{Br}_1(S)$, showing it is isomorphic to $\text{Hom}_{\text{Gal}}(E_1[n], E_2[n])$.
• Non-triviality: By choosing curves with specific reduction properties, the author constructs a zero-cycle $z \in T(X)$ and a Brauer class $\alpha$ such that their pairing is non-zero. This implies the push-forward $\pi_*(z)$ is non-trivial in $T(S)$.

3. Key Contributions and Results

Theorem 1: Exponent of the Albanese Kernel

The paper establishes a precise bound on the exponent of the torsion group $T(S)$ for any bielliptic surface $S$ over a field $k$ (char $\neq 2, 3$).

• Result: $T(S)$ is a torsion group with exponent:
  • $2^2 \cdot |G| = 4|G|$if$2$divides$|G|$.
  • $3^2 \cdot |G| = 9|G|$if$2$does not divide$|G|$(implying$3$divides$|G|$).
• Significance: This provides a uniform, explicit bound for all 7 types of bielliptic surfaces, refining previous knowledge that only established finiteness or torsion properties.

Theorem 2: Non-triviality over $p$-adic Fields

The paper proves the existence of bielliptic surfaces over $p$-adic fields where the Albanese kernel is non-trivial.

• Result: There exist bielliptic surfaces $S$ of Type 1 over $p$-adic fields such that $T(S)$ contains non-trivial 2-torsion elements.
• Construction: The examples are constructed using products of elliptic curves with bad reduction (specifically split multiplicative reduction). The non-triviality is detected via the Brauer-Manin pairing.

Corollary 16: The Good Reduction Case

A significant contrast is drawn between bad and good reduction.

• Result: If $E_1$ and $E_2$ have good reduction over a $p$-adic field ($p \geq 5$), then $T(S)$ is purely 2-torsion (for Type 1) or 3-torsion (for Type 5).
• Implication: The non-trivial elements in the general case (Theorem 2) arise specifically from the interaction between the zero-cycles and the bad reduction of the underlying elliptic curves.

4. Significance

1. Refinement of Bloch's Conjecture Context: While Bloch's conjecture (vanishing of $T(S)$) holds for algebraically closed fields, this paper quantifies the "failure" of this vanishing over non-closed fields by providing exact exponents.
2. Explicit Arithmetic Geometry: The paper moves beyond abstract existence proofs (like those for finite fields using class field theory) to provide explicit computational bounds and constructive examples over $p$-adic fields.
3. Role of Reduction Type: The work highlights a crucial distinction in arithmetic geometry: the structure of zero-cycles on bielliptic surfaces is heavily dependent on the reduction type of the underlying elliptic curves. The existence of non-trivial cycles is linked to the presence of bad (multiplicative) reduction, which allows for non-trivial pairings in the Brauer group.
4. Methodological Bridge: The paper effectively bridges the gap between the geometry of bielliptic surfaces (via étale covers and group actions) and arithmetic obstructions (via the Brauer-Manin pairing), offering a template for studying zero-cycles on other quotient surfaces.

In summary, Gazaki provides a complete description of the torsion exponent of the Albanese kernel for bielliptic surfaces and demonstrates that this kernel can be non-trivial over $p$-adic fields, specifically when the surface arises from elliptic curves with bad reduction.