🔢 mathematics

An integral analogue of Fontaine's crystalline functor

This paper constructs an integral crystalline functor $\mathbb{D}_\mathrm{crys}$ that establishes an equivalence between prismatic $F$ -gauges with Hodge--Tate weights in $[0,p-2]$ and Fontaine--Laffaille modules, thereby clarifying the relationship between prismatic Dieudonné theory and the classification of $p$ -divisible groups via Breuil--Kisin modules.

Original authors: Naoki Imai, Hiroki Kato, Alex Youcis

Published 2026-07-08

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Naoki Imai, Hiroki Kato, Alex Youcis

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 complex, hidden city (let's call it the "World of $p$ -adic Numbers"). Mathematicians have been building different maps of this city for decades, but they are drawn in different styles, using different languages, and based on different landmarks. Sometimes, a map drawn by one group doesn't seem to match the map drawn by another, even though they are describing the same place.

This paper, written by Naoki Imai, Hiroki Kato, and Alex Youcis, is about building a universal translator that connects these different maps. Specifically, it connects three major ways mathematicians study "good reduction" (a fancy way of saying "how a shape behaves when we zoom in on its smallest, most fundamental parts").

Here is the breakdown of the three maps and how the authors connect them:

The Three Maps (The Approaches)

The Crystal Map (The "Old School" Map):
This is the original map created by Jean-Marc Fontaine and others. It uses "filtered crystals." Think of this as a map drawn on a rigid, crystalline grid. It's very precise but can be hard to work with when you are dealing with whole numbers (integers) rather than fractions. It's like trying to draw a smooth curve using only square pixels.
The Prism Map (The "New School" Map):
Recently, mathematicians discovered "Prisms." This is a newer, more flexible way to look at the city. It uses "Prismatic F-crystals." Imagine this as a map drawn on a flexible, stretchy rubber sheet. It's very powerful and handles the "whole number" problems better, but it speaks a different language than the Crystal Map.
The Gauge Map (The "Syntomic" Map):
This is a variation of the Prism Map, using "Prismatic F-gauges." Think of this as the Prism Map with a specific set of rulers (filters) attached to it. It's a middle ground that tries to capture the best of both worlds.

The Problem

For a long time, mathematicians knew these maps were related, but they didn't have a perfect, step-by-step instruction manual to convert a drawing from the "Crystal Map" directly into the "Prism Map" (and vice versa) without losing information. They knew the maps matched when you zoomed out (using fractions), but they wanted to know if they matched exactly when you zoomed in (using whole numbers).

The Solution: The "Dcrys" Translator

The authors build a new tool called $D_{crys}$ . You can think of this as a universal translator or a magic lens.

How it works: The authors realized that every object in the "Prism Map" (Prismatic F-crystals) has a hidden "Crystal" side and a hidden "De Rham" side (a type of smooth, fluid description). They found a way to compare these two sides directly.
The Magic: By comparing these two sides, they can take a "Prism" object and instantly generate its corresponding "Crystal" object. This is the $D_{crys}$ functor.

The Big Discovery: The "Perfect Match" Zone

The paper proves that this translator works perfectly in a specific, important zone called the Fontaine-Laffaille range (roughly, when the numbers involved are small enough, specifically between 0 and $p-2$ ).

In this zone, the authors prove a "Theorem A":

The Crystal Map, the Prism Map, and the Gauge Map are all equivalent.
It's like proving that if you translate a sentence from English to French, and then from French to Spanish, and then back to English, you get the exact same sentence you started with. No information is lost.
This confirms that the "Old School" Crystal theory and the "New School" Prism theory are actually describing the exact same reality in this specific range.

Why This Matters for "Divisible Groups"

The paper also applies this translator to a specific type of mathematical object called $p$ -divisible groups. You can think of these as special shapes that can be divided by the number $p$ infinitely many times.

The Old Way: There was a way to describe these shapes using "Crystalline Dieudonné theory" (the Crystal Map).
The New Way: There was a newer way using "Prismatic Dieudonné theory" (the Prism Map).
The Connection: The authors use their $D_{crys}$ translator to show that the New Way is just a different view of the Old Way. They can take a shape described by the Prism Map and instantly convert it into the Crystal Map description, preserving all the intricate details (like the "filtration" or the layers of the shape).

The "Kim" Connection

Finally, the paper connects this to the work of a mathematician named Kim, who used "Breuil-Kisin modules" (another type of map) to classify these shapes. The authors show that their new translator can bridge the gap between the Prism Map and Kim's map, proving that they are all just different ways of looking at the same underlying structure.

Summary Analogy

Imagine you have three different teams of architects designing a skyscraper:

Team Crystal draws the building using rigid, stone blocks.
Team Prism draws it using flexible, transparent glass panels.
Team Gauge draws it using glass panels with specific measurement grids.

For years, the teams argued about whether their designs were the same. This paper builds a 3D scanner ( $D_{crys}$ ) that can scan the Glass building and instantly print out the exact Stone blueprint. The authors prove that for buildings of a certain height (the Fontaine-Laffaille range), the scanner is perfect: the Stone blueprint and the Glass design are identical. This unifies the work of all three teams and gives mathematicians a single, powerful tool to study these complex shapes.

Technical Summary: An Integral Analogue of Fontaine's Crystalline Functor

Problem Statement
In $p$ -adic Hodge theory, a central question concerns the "good reduction" of a $\mathbb{Q}_p$ -local system $V$ on a smooth formal scheme $X$ over the Witt vectors $W(k)$ of a perfect field $k$ . While Fontaine established that crystalline local systems admit a model as a filtered $F$ -isocrystal $D_{\text{crys}}(V)$ , arithmetic applications (e.g., to Shimura varieties) require an integral analogue. Specifically, given a $\mathbb{Z}_p$ -lattice $L \subseteq V$ , one seeks a precise object modeling $L$ that captures its reduction properties.

Currently, three distinct approaches exist to model such lattices:

Crystalline Approach: Utilizes strongly divisible filtered $F$ -crystals (Fontaine–Laffaille modules) in the range $[0, p-2]$ .
Prismatic Approach: Uses prismatic $F$ -crystals on $X$ .
Syntomic Approach: Uses prismatic $F$ -gauges.

While relationships between the prismatic and syntomic approaches are understood via a forgetful functor, a rigorous integral comparison between the crystalline approach and the prismatic/syntomic approaches has been lacking. The paper addresses the construction of an integral analogue of the crystalline realization functor, $D_{\text{crys}}$ , mapping from prismatic objects to filtered crystalline objects, and investigates its properties, particularly its ability to induce equivalences in specific weight ranges.

Methodology
The authors construct the functor $D_{\text{crys}}$ by leveraging a canonical crystalline-de Rham comparison isomorphism.

Crystalline and de Rham Realizations:
- For a prismatic $F$ -crystal $(E, \phi_E)$ , the crystalline realization $D_{\text{crys}}(E, \phi_E)$ is defined as the $F$ -crystal associated to the restriction of $E$ to the crystalline site.
- The de Rham realization $D_{\text{dR}}(E, \phi_E)$ is defined via pullback along the prismatic de Rham point $\rho^{\Delta}_{\text{dR}, X}$ .
- The core technical step (Theorem 1.19) establishes a canonical isomorphism $\iota_X: D_{\text{crys}}(E, \phi_E)|_{X_{\text{Zar}}} \xrightarrow{\sim} D_{\text{dR}}(E, \phi_E)$ of vector bundles on $X$ . This isomorphism is constructed using Breuil and Breuil–Kisin prisms, providing a concrete coordinate-free description.
Defining the Integral Functor:
- The functor $D_{\text{crys}}$ is defined by taking the underlying $F$ -crystal of the crystalline realization and endowing it with a filtration induced by the inverse of the comparison isomorphism $\iota_X$ applied to the Nygaard filtration on the de Rham realization.
- The authors analyze the properties of this functor on various subcategories, specifically focusing on locally filtered free (lff) prismatic $F$ -crystals and strongly divisible objects.
Comparison with Existing Theories:
- The paper utilizes the work of Tsuji, Faltings, and Anschütz–Le Bras to relate the constructed functor to étale realization functors and Dieudonné theories.
- It employs stack-theoretic arguments (using the Nygaard de Rham point and Rees stacks) to establish the compatibility of filtrations and the behavior of the functor under base change.

Key Contributions and Results

Construction of $D_{\text{crys}}$ :
The paper defines a $\mathbb{Z}_p$ -linear tensor functor $D_{\text{crys}}: \text{Vect}_{\phi}(X^{\Delta}) \to \text{Vect}_{W}^{F\phi}(X_{\text{crys}})$ (and similarly for prismatic $F$ -gauges). This functor serves as an integral lattice in the rational crystalline functor.
Equivalence in the Fontaine–Laffaille Range:
The primary result (Theorem A / Theorem 3.1) states that for a smooth formal $W$ -scheme $X$ and $p > 2$ , the functor $D_{\text{crys}}$ induces an equivalence of categories between:
- The category of prismatic $F$ -gauges on $X$ with Hodge–Tate weights in $[0, p-2]$ .
- The category of Fontaine–Laffaille modules (strongly divisible filtered $F$ -crystals) on $X$ with weights in $[0, p-2]$ .
This equivalence commutes with the étale realization functors, thereby unifying the crystalline, prismatic, and syntomic approaches in this specific weight range.
Properties of $D_{\text{crys}}$ :
- Lattice Property: Theorem C establishes that $D_{\text{crys}}[1/p]$ is canonically identified with the composition of the crystalline realization and the étale realization, confirming $D_{\text{crys}}$ as a functorial lattice.
- Exactness: The functor is shown to be exact (and bi-exact for the forgetful functor from gauges) when restricted to the subcategory of strongly divisible, locally filtered free prismatic $F$ -crystals (Proposition 2.26, Proposition 2.27).
- Detection of lffness: The functor $D_{\text{crys}}$ detects whether a prismatic $F$ -crystal is locally filtered free (Proposition 2.22).
Applications to $p$ -divisible Groups:
The authors apply $D_{\text{crys}}$ to clarify the relationship between different Dieudonné theories for $p$ -divisible groups ( $p > 2$ ):
- They prove that the prismatic Dieudonné functor $M_{\Delta}$ (Anschütz–Le Bras) composed with $D_{\text{crys}}$ recovers the classical filtered crystalline Dieudonné functor $D$ (Grothendieck–Messing) (Theorem 4.8).
- They establish a canonical identification between the prismatic Dieudonné functor and the Breuil–Kisin–Kim Dieudonné functor $M$ (Proposition 4.16), resolving the relationship between these modern and historical classifications.
- This leads to a strengthening of a result by Ito, establishing equivalences between categories of $p$ -divisible groups, prismatic gauges, and Kisin modules with connections (Corollary 4.18).

Significance
The paper claims to complete the comparison of the three major approaches to integral $p$ -adic Hodge theory (crystalline, prismatic, syntomic) by explicitly constructing the integral crystalline functor $D_{\text{crys}}$ . Its significance lies in:

Unification: Providing a bridge that allows one to translate between prismatic/syntomic objects and classical filtered crystalline objects, particularly in the well-behaved Fontaine–Laffaille range where an equivalence is established.
Clarification of Dieudonné Theory: Resolving the structural relationships between the prismatic Dieudonné theory of Anschütz–Le Bras and the Breuil–Kisin–Kim theory, showing how the latter can be recovered from the former via $D_{\text{crys}}$ .
Integral Refinement: Offering an integral refinement of known rational comparisons (such as those by Kisin and Gao) and providing a framework for studying $p$ -divisible groups and local systems with good reduction in a unified categorical setting.

The authors note that while the equivalence in the Fontaine–Laffaille range is the tightest connection, the functor $D_{\text{crys}}$ is defined more generally, though its behavior outside the range $[0, p-2]$ is more complex and not fully equivalent to the crystalline category.