A surface with representable $\text{CH}_{0}$-group but no universal zero-cycle

Here is an explanation of the paper "A Surface with Representable CH0-Group but No Universal Zero-Cycle" by Theodosios Alexandrou, translated into everyday language with creative analogies.

The Big Picture: Finding a "Universal Key"

Imagine you have a complex, multi-story building (a mathematical object called a surface). Inside this building, there are many different "rooms" (mathematical points).

Mathematicians have a special tool called a Universal Zero-Cycle. Think of this as a Master Key or a Universal Remote Control.

If a building has a Master Key, you can use it to control every single room perfectly. You can map the entire building to a central control room (called the Albanese variety) and know exactly where every point is relative to the center.
For a long time, mathematicians thought: "If we can organize the rooms in a certain neat way (called a representable group), then a Master Key must exist."

The Paper's Discovery:
The author, Theodosios Alexandrou, built a specific, tricky building (a Surface) where the rooms are organized neatly, but no Master Key exists. You can organize the data, but you can't find a single, perfect tool to unlock the whole system at once.

This is a big deal because it breaks a rule that mathematicians thought was true for 2D shapes (surfaces), proving that things are more chaotic than we thought.

The Characters and the Plot

1. The Building (The Surface)

The author didn't just pick any building; he built a very specific type called a Bielliptic Surface of Type 2.

Analogy: Imagine a building made by taking two loops of string (elliptic curves) and twisting them together with a specific pattern of cuts and glues.
The Twist: This specific type of building is famous for being "well-behaved" in most ways. Its internal structure is so tidy that mathematicians assumed it must have a Master Key.

2. The "Master Key" (Universal Zero-Cycle)

The Goal: To find a cycle (a collection of points) that acts as a perfect translator between the building and its central control room.
The Problem: In 3D buildings (threefolds), we already knew some existed without a Master Key. But in 2D buildings (surfaces), everyone thought, "If the building is tidy, the key must exist."
The Result: Alexandrou proved this is wrong. He found a tidy 2D building that refuses to give up its Master Key.

3. The Detective Work (The Degeneration Strategy)

How did he prove the key doesn't exist? He didn't just look at the finished building. He used a technique called Degeneration.

The Analogy: Imagine you want to prove a bridge is unstable. Instead of just standing on it, you slowly melt the ground underneath it until it collapses into a pile of rubble.
The Process:
1. Alexandrou took his "Type 2" building and slowly "melted" it (mathematically speaking) until it fell apart into a chain of simpler pieces (like a row of connected tents).
2. He looked at the "ruins" (the special fiber).
3. He checked the "Albanese Maps" (the blueprints) of these ruins.
4. The Clue: He discovered that the blueprints of the ruins were "broken." The pieces of the ruins couldn't be stitched back together to form a path to the central control room.
5. The Logic: If the ruins are broken in a way that prevents a Master Key from existing, then the original, perfect building couldn't have had one either.

4. The "Magic Numbers" (Heegner Fields)

The author had to be very picky about which building he built. He needed the "loops of string" (elliptic curves) to have very specific mathematical properties.

He used curves related to special numbers called Heegner numbers (like 1, 2, 3, 11, 19, etc.).
Analogy: It's like trying to build a house out of wood. Most wood works, but to get this specific "unlocked" effect, you need a very rare, magical type of wood that only grows in specific forests. If you use the wrong wood, the trick doesn't work.

Why Does This Matter? (The "So What?")

1. Breaking a Rule

For a long time, there was a question: "If a surface is tidy (representable), does it have a Master Key?"

Answer: No.
This is the first time this has been proven for a 2D surface. It shows that even in simple, flat worlds, there are hidden complexities that prevent perfect organization.

2. The Hodge Conjecture (The "Ghost" Problem)

The paper ends with a bonus discovery. It uses this broken building to create a "Ghost."

In math, there's a famous guess called the Integral Hodge Conjecture. It asks: "If we see a shadow (a Hodge class) that looks like it belongs to a physical object (an algebraic cycle), is it actually a physical object?"
Usually, the answer is "Yes."
Alexandrou used his broken building to create a Ghost Shadow that looks like a physical object but isn't.
The Twist: Previous examples of these "Ghosts" were "torsion" (they were like ghosts that only appeared for a split second or were very small). Alexandrou found a non-torsion Ghost—a massive, permanent shadow that is definitely not a physical object. This is a major breakthrough in understanding the limits of geometry.

Summary in One Sentence

The author built a mathematically "perfect" 2D surface that, despite its tidy structure, lacks a universal tool to map its points, proving that even in simple shapes, perfect order doesn't guarantee a universal solution, and using this to find a massive "ghost" in geometry that defies our expectations.

Here is a detailed technical summary of the paper "A Surface with Representable $CH_0$ -Group but No Universal Zero-Cycle" by Theodosis Alexandrou.

1. Problem Statement and Context

The paper addresses a fundamental question in algebraic geometry regarding the relationship between the representability of the Chow group of 0-cycles and the existence of a universal 0-cycle.

Background: For a smooth projective complex variety $X$ , the Abel-Jacobi map $\alpha_X: CH_0(X)_{\text{hom}} \to \text{Alb}(X)$ is a regular surjective homomorphism. $X$ is said to have a representable $CH_0$ -group if $\alpha_X$ is an isomorphism.
Universal 0-cycle: A variety $X$ admits a universal 0-cycle if there exists a codimension- $d$ cycle $[\Gamma] \in CH_d(\text{Alb}(X) \times X)$ such that the induced morphism $\psi(\text{Alb}(X), [\Gamma]): \text{Alb}(X) \to \text{Alb}(X)$ is the identity map.
The Question: Voisin recently constructed a 3-fold with a representable $CH_0$ -group but no universal 0-cycle. Colliot-Thélène asked whether this phenomenon persists in lower dimensions, specifically for surfaces (Question 1.2).
The Gap: While ruled surfaces (and curves) always admit universal 0-cycles, it was unknown if a surface with $p_g=0, q \neq 0$ (where $CH_0$ is representable) could fail to admit a universal 0-cycle.

2. Methodology

The author employs a degeneration strategy combined with obstruction theory for universal cycles.

A. The Obstruction (Section 3)

The core theoretical tool is Theorem 3.1, which establishes a necessary condition for the existence of a universal 0-cycle on a variety $X$ with a representable $CH_0$ -group.

Setup: Consider a flat, projective degeneration $p: \mathcal{X} \to \Delta$ (where $\Delta = \text{Spec } k[[t]]$ ) with a regular total space. Let the special fiber $\mathcal{X}_0$ be a reduced simple normal crossings divisor with components $R_i$ forming a chain.
The Condition: If the generic fiber $\mathcal{X}_K$ admits a universal 0-cycle (implying a splitting of the pushforward map to the Jacobian/Albanese), then the sum morphism of the Albanese varieties of the special fiber components must admit a section:
$\sum \text{alb}_i: \bigoplus \text{Alb}(R_i) \to \text{Jac}(C)$
(where $C$ is the base curve of the fibration).
Corollary 3.2: If the endomorphism ring of the Jacobian is a Principal Ideal Domain (PID), the non-existence of a section for this sum morphism implies the non-existence of a universal 0-cycle on the generic fiber.

B. The Construction (Section 4)

The author constructs a specific degeneration of a bielliptic surface of Type 2.

Base Objects: An elliptic curve $E$ and a minimal regular model $F$ of an elliptic curve with a special fiber of Kodaira type $I_4$ (a cycle of 4 rational curves).
Group Action: A finite group $G$ (generated by involutions) acts on $E \times F$ . The quotient yields a bielliptic surface.
Degeneration: By carefully choosing the group action involving 2-torsion points and the specific structure of the $I_4$ $I_{4}$ fiber, the author constructs a family $\mathcal{S} \to \Delta$ $S \to Δ$ where:
1. The generic fiber $S_K$ is a bielliptic surface of Type 2.
2. The special fiber $S_0$ consists of two minimal ruled surfaces $R_1, R_2$ intersecting along a smooth elliptic curve $C$ .
3. The Albanese map of the special fiber components induces maps $\text{Alb}(R_i) \to E$ corresponding to étale double covers $E/a \to E$ and $E/b \to E$ .

C. The Algebraic Obstruction

The proof hinges on showing that the sum morphism $\text{Alb}(R_1) \oplus \text{Alb}(R_2) \to E$ admits no section.

This reduces to an arithmetic problem on the elliptic curve $E$ : Can the identity map on $E$ be written as a sum of two isogenies factoring through $E/2$ -torsion quotients?
Case (i): If $\text{End}(E) \cong \mathbb{Z}$ , the homomorphisms are restricted to integer multiples of the quotient maps. The composition with the sum map vanishes on 2-torsion, preventing a section.
Case (ii): If $E$ has Complex Multiplication (CM) by the maximal order of a Heegner field $\mathbb{Q}(\sqrt{-c})$ (for specific $c$ ), the ring of integers is a PID. The author uses the Stark-Heegner theorem and ideal theory to show that any such decomposition of the identity would imply $1$ lies in the prime ideal above 2, a contradiction.
Note: The case $c=-7$ is explicitly excluded because the ring $\mathbb{Z}[\frac{1+\sqrt{-7}}{2}]$ allows a decomposition of the identity, and indeed, Type 2 surfaces with $c=-7$ do admit universal 0-cycles (Example 5.2).

3. Key Results

Theorem 1.3 (Main Result)

There exists a smooth projective complex surface $S$ such that:

The Chow group of 0-cycles $CH_0(S)$ is representable (i.e., the Abel-Jacobi map is an isomorphism).
$S$ does not admit a universal 0-cycle.

Conditions: This holds if the Albanese variety $\text{Alb}(S) \cong E$ is an elliptic curve where either $\text{End}(E) \cong \mathbb{Z}$ or $E$ has CM by the maximal order of $\mathbb{Q}(\sqrt{-c})$ for $c \in \{1, 2, 3, 11, 19, 43, 67, 163\}$ .

Corollary 1.5 (Integral Hodge Conjecture)

The construction yields a smooth projective threefold $X = E \times S$ (where $S$ is as above) with Kodaira dimension zero.

The Integral Hodge Conjecture fails for 1-cycles on $X$ .
Specifically, there exists a non-torsion integral Hodge class in $H^4(E \times S, \mathbb{Z})$ that is not algebraic.
Significance: Previous counterexamples (Benoist-Ottem) relied on torsion classes in degree 4. This is the first example of a non-torsion counterexample in degree 4 for a variety of Kodaira dimension zero.

4. Significance and Contributions

Resolution of Question 1.2: The paper definitively answers "Yes" to the question of whether a surface with a representable $CH_0$ -group can lack a universal 0-cycle. This provides a 2-dimensional analogue to Voisin's 3-fold counterexample.
New Obstruction Mechanism: The paper introduces a novel obstruction based on the degeneration of the surface (rather than the base curve). It links the existence of universal cycles to the splitting properties of Albanese maps in the special fiber of a degeneration.
Refinement of the Integral Hodge Conjecture: By producing a non-torsion counterexample in degree 4 for a variety with $\kappa=0$ , the paper expands the known landscape of Hodge conjecture failures. It demonstrates that the failure is not limited to torsion phenomena or high Kodaira dimensions.
Arithmetic Geometry Connection: The proof relies heavily on the arithmetic properties of elliptic curves (specifically Heegner fields and the structure of their endomorphism rings), bridging complex algebraic geometry with number theory.

5. Conclusion

Theodosis Alexandrou successfully constructs a bielliptic surface of Type 2 that serves as a counterexample to the persistence of universal 0-cycles under representability assumptions. By leveraging a degeneration argument and the arithmetic of elliptic curves, the author not only solves the surface case of Colliot-Thélène's question but also generates a new class of counterexamples to the Integral Hodge Conjecture involving non-torsion classes. This work highlights the subtle interplay between the geometry of algebraic cycles, the structure of Albanese varieties, and the arithmetic of the underlying fields.

A surface with representable CH0\text{CH}_{0}CH0​-group but no universal zero-cycle