🔢 mathematics

On optimal $p$ -adic uniformization of unitary Shimura curves

This paper extends the Kudla-Rapoport-Zink theory of $p$ -adic uniformization for Shimura curves by establishing results for two variants: the RSZ variant with arbitrary maximal levels at anisotropic places, and a unitary group variant derived from an explicit determination of the associated integral local Shimura variety.

Original authors: Michael Rapoport, Haining Wang

Published 2026-05-20

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Michael Rapoport, Haining Wang

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 understand a very complex, abstract shape that exists in the world of pure mathematics. This shape is called a Shimura curve. Think of it like a mysterious, multi-dimensional map that connects different worlds of numbers.

For a long time, mathematicians knew how to describe this map using "complex numbers" (the kind involving $i$ , the square root of -1). This is like looking at the map through a telescope that sees the whole picture clearly, but only from a specific, distant angle.

However, there is another way to look at this map: through the lens of $p$ -adic numbers. This is a different kind of number system, often described as a "microscope" that zooms in on the fine, grainy details of the map, specifically at a particular "prime number" location (let's call it the $p$ -spot).

The big question this paper answers is: Can we translate the clear, distant view (complex) into a sharp, close-up view ( $p$ -adic) without losing any details?

The Problem: The "Perfect" Translation Was Missing

In the past, mathematicians (specifically Cherednik) figured out how to do this translation for a specific type of Shimura curve. But there was a catch: their method only worked if the "level structure" (think of this as the grid lines or ruler markings on your map) was perfectly clean at the $p$ -spot. If the grid lines were messy or crossed the $p$ -spot, the translation broke down.

The authors of this paper, Michael Rapoport and Haining Wang, wanted to fix this. They wanted a translation that works no matter how messy the grid lines are at the $p$ -spot, as long as they are clean everywhere else. They call this an "optimal" translation.

The Solution: Two New Tools

To achieve this "optimal" translation, the authors didn't just tweak the old map; they built two new, slightly different versions of the map to work with.

1. The "RSZ" Map (The Flexible Version)

The first tool is a new version of the Shimura curve, which they call the RSZ Shimura curve.

The Analogy: Imagine you have a rigid plastic map that only works if you hold it perfectly straight. The RSZ version is like a flexible, stretchy map.
The Trick: To make it stretchy enough to handle messy grid lines at the $p$ -spot, they had to attach a slightly larger "handle" to the map. In math terms, they had to extend the "reflex field" (a specific number field that acts like the map's coordinate system).
The Result: With this flexible map, they successfully proved that you can translate the complex view to the $p$ -adic view perfectly, even when the grid lines are messy at the $p$ -spot. This is a major improvement over previous work, which required the grid to be clean everywhere.

2. The "Unitary Group" Map (The Standard Version)

The second tool deals with the original, standard Shimura curves attached to unitary groups.

The Analogy: This is like trying to translate a standard, rigid map. The authors managed to build a "canonical" (standardized) version of this map that works over the integers (the whole numbers), which is necessary for the $p$ -adic view.
The Catch: Just like with the RSZ map, to get the translation to work perfectly, they had to use a slightly larger coordinate system (a bigger extension of the reflex field).
The Result: They proved that even for these standard curves, you can get a perfect $p$ -adic translation, provided you are willing to use this slightly larger coordinate system. They also identified exactly what this "local" version of the map looks like when you zoom in all the way to the $p$ -spot.

The "Drinfeld Upper Half Plane" Connection

A key part of their proof involves a famous mathematical object called the Drinfeld upper half plane.

The Analogy: Think of the complex view of the map as a smooth, continuous landscape. The $p$ -adic view is like a landscape made of discrete, stepping stones. The Drinfeld upper half plane is the perfect stepping-stone path that connects the two worlds.
The authors show that their new, optimal maps can be built entirely out of these stepping stones, arranged in a specific pattern. This proves that the complex map and the $p$ -adic map are actually two sides of the same coin.

Summary of Achievements

Optimality: They removed the restriction that the "grid lines" (level structure) had to be clean at the $p$ -spot. Now, the translation works even if the grid is messy there, as long as it's clean elsewhere.
Two Classes: They solved this for two different families of these mathematical curves (the RSZ variants and the standard Unitary Group curves).
The Cost: The price for this perfection is that they had to use a slightly larger "coordinate system" (a bigger field extension) to define the map. Without this larger system, the translation wouldn't be as clean.

What They Did Not Do

The paper is purely theoretical mathematics. They did not:

Apply these results to physics or engineering.
Predict future technological breakthroughs.
Discuss clinical uses or medical applications.
Claim that this solves the Riemann Hypothesis or other famous open problems (though it contributes to the broader field of number theory).

In short, they built better, more flexible blueprints for translating between two different mathematical worlds, allowing mathematicians to see the fine details of these abstract shapes in a way that was previously impossible.

Technical Summary: Optimal p-Adic Uniformization of Unitary Shimura Curves

Problem Statement
The paper addresses the problem of establishing optimal $p$ -adic uniformization theorems for Shimura curves attached to unitary groups. Specifically, it seeks to generalize Cherednik's theorem, which describes the $p$ -adic uniformization of Shimura curves associated with quaternion algebras over totally real fields, to the setting of unitary groups.

A primary challenge identified in the authors' previous work [20] (concerning KRZ Shimura curves) was that the optimal form of the uniformization theorem—providing integral models and explicit descent data—was only achievable when the level structure was prime to all $p$ -adic places. The paper aims to extend this optimality to cases where there are multiple $p$ -adic places, including "banal" places (where the level structure is not necessarily prime to $p$ ), and to treat Shimura curves associated directly with unitary groups (which are of abelian type but not PEL type).

Methodology
The authors employ a combination of moduli-theoretic constructions, local-global principles, and the theory of integral local Shimura varieties.

RSZ Shimura Curves (PEL-type Variants):
- The authors utilize the RSZ Shimura curves introduced in [23] and [24]. These are variants of the KRZ curves defined via a fiber product group $\tilde{G} = G \times_{\mathbb{G}_m} Z_{\mathbb{Q}}$ , where $G$ is the unitary similitude group and $Z_{\mathbb{Q}}$ is a specific torus.
- They construct $p$ -integral models by formulating a moduli problem over the ring of integers of a reflex field extension. This moduli problem involves tuples of abelian schemes with specific CM-types, polarizations, and level structures (including "strict" and "relaxed" CL-level structures).
- The proof relies on analyzing the formal moduli spaces of CM-triples. The authors distinguish between "special" CM-types (where the local CM-type is non-trivial at the distinguished place $v_0$ ) and "banal" CM-types (where it is trivial).
- A key technical step involves the "polarized contraction functor" to relate moduli spaces of $p$ -divisible groups to formal schemes of the form $\hat{\Omega} \times \mathbb{Z}$ or quotients of $p$ -adic groups.
Unitary Group Shimura Curves (Non-PEL):
- For Shimura curves attached directly to unitary groups (not similitude groups), the authors assume the existence of canonical integral models in the sense of [21].
- They reduce the global problem to a local one involving integral local Shimura varieties (LSV) in the sense of Scholze-Weinstein [27].
- The local result is proven by relating the integral LSV for the unitary group to the product of an integral LSV for a torus and the RSZ formal moduli space studied in Part 2. This allows them to transfer the explicit uniformization results from the PEL case to the non-PEL case.
Descent Data and Reflex Fields:
- A significant portion of the methodology involves making the Weil descent data explicit. In previous works, the descent data for banal cases were not fully explicit. Here, the authors introduce "special elements" in the center of the relevant $p$ -adic groups (derived from reflex norms) to explicitly describe the descent data.
- To achieve this explicitness, the authors often require extending the reflex field. For the RSZ curves, the reflex field is enlarged to $E = E_r E_{\Phi^+}$ . For the unitary group curves, a finite extension $E(\phi_0)$ of the reflex field $K$ is required.

Key Contributions and Results

Optimal Uniformization for RSZ Curves:
- Theorem 2.3.3: The authors prove an optimal $p$ -adic uniformization theorem for the RSZ Shimura curves. They construct a projective flat integral model $\hat{\mathcal{A}}_{\tilde{K}\tilde{G}}$ over the ring of integers of the reflex field.
- The formal completion of this model along its special fiber is isomorphic to a quotient of the product of the formal Drinfeld upper half-plane $\hat{\Omega}_{F_{v_0}}$ and the adelic double coset space $\tilde{G}(\mathbb{A}_f)/\tilde{K}_{\tilde{G}}$ .
- Crucially, this result holds for level structures that are prime to $v_0$ but arbitrary at other $p$ -adic places (the "banal" places). This resolves the limitation of [20] where optimality required levels prime to all $p$ -adic places.
- The theorem provides an explicit description of the Weil descent datum on the right-hand side using a "special element" $w$ in the center of $\tilde{G}(\mathbb{Q}_p)$ .
Uniformization for Unitary Group Curves:
- Theorem 2.4.3: The authors establish a $p$ -adic uniformization theorem for Shimura curves attached to unitary groups (of abelian type).
- They construct an integral model and prove that its formal completion is isomorphic to a quotient involving the Drinfeld upper half-plane and the unitary group adelic space, provided one passes to a finite extension $E(\phi_0)$ of the reflex field $K$ .
- Theorem 2.4.5: They identify the integral local Shimura variety for an anisotropic unitary group over a finite extension of $\mathbb{Q}_p$ with the formal Drinfeld upper half-plane (after base change to $E(\phi_0)$ ). This provides an explicit model for a class of integral local Shimura varieties that were previously not fully understood in this generality.
Comparison of KRZ and RSZ Curves:
- In Part 3, the authors revisit the KRZ Shimura curves from [20]. They show that by extending the reflex field by a degree $\le 2$ extension, one can make the descent datum for KRZ curves explicit, similar to the RSZ case. This clarifies the relationship between the two types of curves and explains the "deficiencies" of the KRZ formulation in the presence of banal primes.

Significance and Claims
The paper claims to provide a more general and "optimal" version of Cherednik's $p$ -adic uniformization theorem for unitary Shimura curves.

Generality: It extends the scope of optimal uniformization to include level structures at banal $p$ -adic places, a setting where previous results were incomplete.
Explicitness: It succeeds in making the descent data explicit in cases where it was previously unknown or non-explicit, particularly for the banal case.
Integral Models: It constructs and characterizes $p$ -integral models for these curves, allowing for an explicit interpretation of points in the reduction modulo $p$ .
Local-Global Connection: The work bridges the gap between global Shimura curves and local integral Shimura varieties, providing explicit identifications for the latter in the unitary group case.

The authors note that the necessity of extending the reflex field (from $K$ to $E(\phi_0)$ ) is a consequence of their proof method. They conjecture that this extension is not necessary for the optimal result (i.e., the isomorphism should descend to $K$ ), but they do not prove this conjecture. Similarly, the existence of the canonical integral models for the unitary group curves in the ramified case or when $p=2$ is assumed rather than proven, relying on the conjectural framework of [21].

The paper does not claim immediate arithmetic applications (such as level raising or Selmer group bounds) but notes that such applications are expected to follow from these uniformization results, referencing [19] for the context of such potential applications.