🔢 mathematics

Generalised Kato classes on CM elliptic curves of rank 2

This paper constructs a generalised Kato class for CM elliptic curves of analytic rank at least two and proves that its nonvanishing is equivalent to the Selmer group having dimension two, thereby establishing a rank-two analogue of Kolyvagin's result through a novel connection to anticyclotomic Iwasawa theory.

Original authors: Francesc Castella

Published 2026-02-17

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Francesc Castella

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

Imagine you are trying to solve a massive, cosmic puzzle. The puzzle is an Elliptic Curve, a specific type of mathematical shape that looks like a twisted loop. Mathematicians have been trying to understand the "rational points" on these loops (points with simple fraction coordinates) for centuries.

The Birch and Swinnerton-Dyer (BSD) Conjecture is the "Holy Grail" of this field. It predicts a deep connection between the shape of the curve and a mysterious function called the L-function.

Here is the simple breakdown of what this paper achieves, using a few analogies:

1. The Problem: The "Rank 2" Mystery

Think of the L-function as a musical note.

If the note is silent (the function equals zero) at a specific spot ( $s=1$ ), it tells us something about the curve.
If the note is silent once, we know how to find the "hidden treasures" (rational points) on the curve. This is the "Rank 1" case, which was solved decades ago.
The New Challenge: What if the note is silent twice (or more)? This is the "Rank 2" case. It's like trying to find two hidden treasures when you only have a map for one. For a long time, mathematicians didn't know how to construct these extra points.

2. The Tool: "Generalised Kato Classes"

The author, Francesc Castella, introduces a new tool called a Generalised Kato Class (let's call it $\kappa_p$ ).

The Analogy: Imagine you are trying to find a specific key in a giant, dark room. In the past, you had a flashlight that only worked if the room was empty (Rank 1). Now, you have a new, high-tech sonar device ( $\kappa_p$ ).
This device is built by combining three different "families" of mathematical objects (called Hida families) in a very specific way. It's like mixing three colors of paint to create a new, unique shade that reveals hidden structures.

3. The Big Discovery: The "If and Only If" Rule

The paper proves a powerful rule about this new sonar device ( $\kappa_p$ ) when the curve has Rank 2 (meaning there are exactly two independent hidden treasures).

The Rule: The sonar device ( $\kappa_p$ ) will "ping" (be non-zero) if and only if the treasures are actually "visible" to a local observer at the prime number $p$ .
The Metaphor: Imagine the two treasures are buried in a field.
- If the treasures are buried so deep that a local detector at the edge of the field ( $p$ ) can't sense them, your high-tech sonar ( $\kappa_p$ ) will stay silent, even though the treasures exist.
- But if the local detector can sense them, your sonar will definitely ping.
- Crucially: If the sonar pings, it proves for a fact that there are exactly two treasures (Rank 2).

4. The Secret Weapon: The "Anticyclotomic" Connection

How did the author prove this?

He connected two different worlds of mathematics that usually don't talk to each other:
1. The World of Curves: The elliptic curve and its points.
2. The World of "Anticyclotomic" Theory: A complex branch of number theory involving "imaginary quadratic fields" (a special type of number system).
The Analogy: It's like realizing that the pattern of leaves on a tree (the curve) is actually a direct reflection of the wind patterns in a distant valley (the anticyclotomic theory). By studying the wind, he could predict exactly where the leaves would fall.
He used a "Main Conjecture" (a grand theory about how these winds behave) to prove that his sonar device works exactly as predicted.

5. Why Does This Matter?

Solving the Puzzle: This is a major step toward solving the BSD conjecture for Rank 2 curves. It gives mathematicians a concrete way to construct the second hidden point, rather than just guessing it exists.
A New Map: The paper provides a "basis" (a set of coordinates) for the space of these points. It's like finally drawing a map that shows exactly where the two treasures are, rather than just saying "they are somewhere in this region."
Bridging Gaps: It fixes a gap in previous methods. Earlier attempts to solve Rank 2 problems failed for curves with "Complex Multiplication" (CM). This paper specifically targets those tricky CM curves and solves the problem for them.

Summary

Francesc Castella built a new mathematical "sonar" ( $\kappa_p$ ) to hunt for hidden points on elliptic curves. He proved that this sonar works perfectly for curves with two hidden points, but only if those points are "visible" to a specific local detector. By linking this to a grand theory about imaginary number systems, he provided a rigorous proof that helps us understand the deepest secrets of these mathematical shapes.

1. Problem Statement and Context

The paper addresses a central open problem in the arithmetic of elliptic curves: the systematic construction of rational points (or non-torsion Selmer classes) when the $L$ -function of an elliptic curve $E/\mathbb{Q}$ vanishes to order $\ge 2$ at $s=1$ .

Background: The Gross–Zagier and Kolyvagin theorems (1980s) successfully link the non-vanishing of Heegner points to the rank of $E$ when the analytic rank is 0 or 1. However, for higher ranks (specifically rank 2), constructing such points remains a major challenge.
Specific Setting: The author focuses on CM (Complex Multiplication) elliptic curves $E$ defined over $\mathbb{Q}$ , where $E$ has CM by an imaginary quadratic field $K$ . The prime $p \ge 5$ is assumed to be of good ordinary reduction and splits in $K$ .
Hypothesis: The $L$ -function $L(E, s)$ has a functional equation sign of $+1$ and vanishes at $s=1$ , implying the analytic rank is at least 2.
Goal: To construct a "generalised Kato class" $\kappa_p$ in the $p$ -adic Selmer group $\text{Sel}(\mathbb{Q}, V_p E)$ and prove that its non-vanishing implies the algebraic rank is exactly 2, providing a rank-2 analogue of Kolyvagin's theorem.

2. Methodology

The proof relies on a sophisticated interplay between Hida families, Iwasawa theory, and diagonal cycles. The methodology can be broken down into four main pillars:

A. Construction via Hida Families and Diagonal Cycles

The author modifies the construction of Darmon and Rotger (DR) for generalised Kato classes.

Triple Product Setup: Instead of working directly with $E$ $E$ , the author considers a triple of Hida families $(\phi, g, h)$ $(ϕ, g, h)$ where:
- $\phi$ is a CM Hida family associated with the Hecke character $\psi_E$ of $E$ .
- $g$ and $h$ are CM Hida families associated with finite-order Hecke characters $\lambda_1, \lambda_2$ .
Big Diagonal Class: A "big diagonal class" $\kappa(\phi, g, h)$ is constructed in the cohomology of a triple tensor product of Galois representations. This class interpolates $p$ -adic étale Abel–Jacobi images of generalised Gross–Kudla–Schoen diagonal cycles.
Specialisation: By choosing specific Hecke characters satisfying a self-duality condition ( $\psi_E = \lambda_0 \lambda_1^c \lambda_2^c$ ), the author specialises the big diagonal class to obtain a specific class $\kappa_p \in H^1(\mathbb{Q}, V_p E)$ .

B. Factorisation of $p$ -adic $L$ -functions

A critical technical step is the factorisation of the triple product $p$ -adic $L$ -function $L_p^\phi(\phi, g, h)$ .

The author proves that in the CM setting, the square of this triple product $L$ -function factors into a product of four anticyclotomic Katz $p$ -adic $L$ -functions associated with specific Hecke characters derived from $\lambda_0, \lambda_1, \lambda_2$ .
This factorisation is crucial because it allows the author to relate the complex triple product setting to classical anticyclotomic Iwasawa theory, where results are better understood.

C. Decomposition of Selmer Groups

The paper establishes a decomposition of the balanced Selmer group $\text{Sel}_{\text{bal}}(\mathbb{Q}, V^\dagger_{\phi gh})$ associated with the triple product.

Using Shapiro's lemma and the specific structure of CM representations, the global Selmer group decomposes into a direct sum of anticyclotomic Selmer groups over the field $K$ .
Crucially, one of these summands corresponds to a "reversed" local condition Selmer group ( $\text{Sel}_{0, \emptyset}$ ), which differs from the standard Bloch–Kato definition by swapping local conditions at the primes above $p$ .

D. The Main Conjecture Link

The proof hinges on the Big Diagonal Class Main Conjecture (Conjecture A).

The author proves a two-variable specialisation of this conjecture in the CM case.
By combining the factorisation of the $L$ -function with the decomposition of the Selmer group, the author deduces that the non-vanishing of the big diagonal class is equivalent to the non-vanishing of the corresponding anticyclotomic $p$ -adic $L$ -functions.
This link allows the translation of the non-vanishing of the $L$ -function (guaranteed by non-vanishing theorems of Greenberg and Rohrlich) into the non-vanishing of the geometric class $\kappa_p$ .

3. Key Contributions and Results

Theorem A: The Main Conjecture in the CM Case

The author proves Conjecture A for a specific triple of CM Hida families.

Result: The big diagonal class $\kappa(\phi, g, h)$ is not torsion, and the characteristic ideal of the dual Selmer group is generated by the square of the triple product $p$ -adic $L$ -function.
Significance: This extends previous results (which excluded the CM case due to "big image" hypotheses) to the CM setting, utilizing the specific factorisation properties of CM forms.

Theorem B: Rank 2 Analogue of Kolyvagin's Theorem

This is the paper's primary arithmetic result.

Statement: Let $E/\mathbb{Q}$ $E / Q$ be a CM elliptic curve with $L(E, 1) = 0$ $L (E, 1) = 0$ and sign $+1$ $+ 1$ . Let $\kappa_p$ $κ_{p}$ be the generalised Kato class constructed above.
1. Non-vanishing implies Rank 2: If $\kappa_p \neq 0$ , then $\dim_{\mathbb{Q}_p} \text{Sel}(\mathbb{Q}, V_p E) = 2$ .
2. Converse: If $\dim_{\mathbb{Q}_p} \text{Sel}(\mathbb{Q}, V_p E) = 2$ , then $\kappa_p \neq 0$ if and only if the restriction map $\text{res}_p: \text{Sel}(\mathbb{Q}, V_p E) \to E(\mathbb{Q}_p) \hat{\otimes} \mathbb{Q}_p$ is non-zero.
Novelty: Unlike the rank 1 case where Heegner points are non-zero automatically if the rank is 1, in the rank 2 case, the non-vanishing of $\kappa_p$ depends on the non-vanishing of the restriction map (essentially, the class must not be "trivial" at $p$ ).

Theorem C: Basis for the Selmer Group

Under the assumption that the $p$ -adic $L$ -function vanishes to order 2, the author constructs an explicit basis for the 2-dimensional Selmer group:
$\text{Sel}(\mathbb{Q}, V_p E) = \mathbb{Q}_p \kappa_p \oplus \mathbb{Q}_p x_p^{(2)}$
where $x_p^{(2)}$ is a "derived elliptic unit" (related to the derivative of the $p$ -adic $L$ -function). This provides a concrete $p$ -adic regulator for the rank 2 case.

4. Significance and Impact

Resolution of the CM Case: Previous approaches to generalised Kato classes (e.g., by Darmon–Rotger and Castella–Hsieh) relied on the non-CM assumption to ensure the residual Galois representation is "big" (absolutely irreducible with large image). This paper bypasses that obstruction by exploiting the specific structure of CM representations and the factorisation of $L$ -functions.
New Proof for Non-CM Case: The techniques developed here (specifically the decomposition of Selmer groups via anticyclotomic Iwasawa theory) provide a new, independent proof of the non-CM results previously obtained by Castella and Hsieh.
Advancement of BSD: The results offer a rigorous framework for the Birch–Swinnerton-Dyer conjecture in the rank 2 setting for CM curves. It confirms that the generalised Kato classes serve as the correct $p$ -adic regulators in this context, provided the local restriction condition is met.
Iwasawa Theory: The paper establishes a deep connection between the non-vanishing of diagonal cycles and the anticyclotomic Iwasawa main conjecture for Hecke characters, demonstrating how geometric constructions can be controlled by analytic $p$ -adic $L$ -functions even in the presence of complex multiplication.

In summary, Castella successfully constructs a rank-2 $p$ -adic regulator for CM elliptic curves, proving that its non-vanishing is equivalent to the Selmer group having dimension 2 (under a mild local condition), thereby extending the paradigm of Kolyvagin's Euler systems to higher ranks in the CM setting.