🔢 mathematics

Logarithmic Crystalline Representations

This note explicitly constructs logarithmic Fontaine-Faltings modules and logarithmic crystalline representations, thereby providing the detailed extension of Faltings' 1989 comparison theorem to the logarithmic context that he previously mentioned without elaboration.

Original authors: Zhenmou Liu, Jinbang Yang, Kang Zuo

Published 2026-02-05

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Zhenmou Liu, Jinbang Yang, Kang Zuo

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

The Big Picture: Connecting Two Different Worlds

Imagine you are trying to understand a complex landscape. You have two different maps of this same land:

Map A (The "Etale" Map): This map is like a satellite photo. It sees the land from high up, capturing the big picture and the "holes" or gaps in the terrain. In math, this is called étale cohomology.
Map B (The "Crystalline" Map): This map is like a microscopic view of the soil itself. It looks at the tiny crystals and structures that make up the ground. In math, this is crystalline cohomology.

For a long time, mathematicians knew these two maps were related, but the connection was tricky. In 1989, a mathematician named Faltings built a bridge between them. He showed that if you have a specific type of mathematical object (a "Fontaine-Faltings module"), you can translate it perfectly into a representation of the landscape's "holes" (a "crystalline representation").

The Problem: Faltings' bridge worked great for smooth, perfect landscapes. But what if the landscape has sharp edges, cracks, or "singularities" (like a cliff or a riverbank)? In math, these are called logarithmic structures. Faltings claimed he could extend his bridge to these rougher landscapes, but he didn't write down the blueprints. He just said, "It can be done."

The Solution: This paper by Liu, Yang, and Zuo is the blueprint. They explicitly build the bridge for these "rough" landscapes. They show exactly how to construct the mathematical objects needed to cross the gap when the terrain has cracks.

The Key Concepts (Simplified)

1. The "Fontaine-Faltings Module" (The Mathematical Object)

Think of a Fontaine-Faltings module as a specialized suitcase.

Inside the suitcase, you have a bundle of data (a vector bundle).
This data is organized into layers (a filtration), like clothes folded by size.
There is a "connection" (a rule for how the data moves) that tells you how to walk through the landscape without tearing the clothes.
There is a "Frobenius" (a magical shuffling machine) that rearranges the contents of the suitcase in a specific way.

In the old theory, this suitcase had to be perfectly smooth. The authors of this paper say: "Let's make a suitcase that can handle a bumpy ride." They modify the "connection" rule so it can slide along the cracks (the logarithmic poles) without breaking.

2. The "D-Functor" (The Translator)

The D-functor is the translator or the bridge.

You take your "suitcase" (the Fontaine-Faltings module).
You run it through the D-functor.
It comes out the other side as a "crystalline representation" (a description of the holes in the landscape).

The paper proves that this translator works perfectly even when the suitcase is designed for bumpy roads.

3. The "Logarithmic" Twist (Handling the Cracks)

Imagine you are walking through a forest.

Smooth Case: You walk on a flat path.
Logarithmic Case: You are walking near a river or a cliff. The path gets tricky.

In the smooth case, the translator (D-functor) works by looking at the path directly. But near the river (the "logarithmic" part), the path is undefined.

The Authors' Trick:
Instead of trying to translate the suitcase directly while it's on the edge of the cliff, they do something clever:

They take the suitcase and pull it back onto a temporary, smooth platform (a "cover") that is built right over the cliff.
On this smooth platform, the "cracks" disappear (the logarithmic poles vanish).
They use the old, trusted translator (Faltings' original D-functor) to translate the suitcase on this smooth platform.
They then push the translation back down to the original cliff edge.

They prove that this "push-pull" method works perfectly and gives the same result as if they had built a new translator from scratch.

The Main Results (What They Actually Proved)

Construction: They wrote down the exact rules for what a "Logarithmic Fontaine-Faltings module" looks like. It's the same as the old one, but the rules for moving data (the connection) are tweaked to handle the "cracks" in the math.
The Bridge (Dlog-functor): They built a new translator, called Dlog, that takes these new "logarithmic suitcases" and turns them into "logarithmic crystalline representations."
Consistency: They proved that if you take a logarithmic suitcase, remove the "cracks" (go to the smooth part), and translate it, you get the exact same result as if you had translated the whole thing first and then removed the cracks. The two methods match perfectly.
Reliability (Full Faithfulness): They proved that this new translator is perfect.
- Faithful: If two suitcases are different, the translator will definitely produce two different results. It never confuses them.
- Full: If you have two translated results that look compatible, there is guaranteed to be a suitcase that created them. You never get a "ghost" translation that didn't come from a real suitcase.

Summary Analogy

Imagine you have a recipe book (the Fontaine-Faltings modules) that tells you how to bake a cake.

Faltings (1989) showed that if you follow the recipe, you get a specific taste (the crystalline representation).
The Problem: Faltings said, "If you add a weird ingredient like 'spicy chili' (the logarithmic part), the recipe still works, but I didn't write down the instructions."
Liu, Yang, and Zuo (This Paper): They wrote the instructions. They showed you exactly how to mix the chili into the batter without ruining the cake. They proved that if you follow their new "Chili Recipe," the resulting taste is still perfectly predictable and matches the taste of the original cake if you take the chili out.

In short: They filled in the missing details of a famous mathematical theory, allowing it to work on "rougher" mathematical landscapes without breaking the rules.

Technical Summary: Logarithmic Crystalline Representations

Problem Statement
In 1989, Faltings established a comparison theorem between étale cohomology and crystalline cohomology by introducing the category of Fontaine-Faltings modules ( $MF^{\nabla}$ ) and a fully faithful functor ( $D$ ) mapping these modules to representations of the étale fundamental group of the generic fiber. While Faltings asserted in [Fal89] that this theory extends to the logarithmic context (involving normal crossing divisors), he provided no details. Consequently, the explicit construction of logarithmic Fontaine-Faltings modules and the corresponding logarithmic comparison functor remained absent from the literature. This paper aims to fill that gap by explicitly constructing these objects and proving the full faithfulness of the resulting functor.

Methodology
The authors proceed in two main stages, first reviewing the non-logarithmic theory and then extending it to the logarithmic setting.

Review of Non-Logarithmic Theory: The paper begins by recalling the definition of Fontaine-Faltings modules over a small affine base and a global base. These are quadruples $(V, \nabla, \text{Fil}, \phi)$ consisting of a de Rham module, a Hodge filtration, and a semi-linear Frobenius isomorphism. The authors review Faltings' $D$ -functor, which utilizes period rings ( $B^+(R)$ ) and the construction of a horizontal action of the étale fundamental group to map these modules to continuous representations. A key technical tool is the "gluing functor," which establishes that the category of modules is independent of the choice of Frobenius lifting $\Phi$ (provided the Hodge-Tate weight range $b-a \leq p-2$ ).
Construction of Logarithmic Modules:
- Definition: A logarithmic Fontaine-Faltings module over a pair $(Y, Z)$ (where $Z$ is a normal crossing divisor) is defined similarly to the classical case, but the connection $\nabla$ is required to have logarithmic poles along $Z$ . The Frobenius lift $\Phi$ must respect the logarithmic structure (i.e., $\Phi(T_i) = w_i T_i^p$ ).
- Gluing Theorem: The authors prove a logarithmic version of Faltings' gluing theorem (Theorem 2.3). They construct a canonical isomorphism $\alpha_{\Phi, \Psi}$ between modules defined by different Frobenius lifts using a logarithmic Taylor formula involving logarithmic derivations $\partial_i = T_i \frac{\partial}{\partial T_i}$ . This ensures the category of logarithmic modules is well-defined independent of the choice of $\Phi$ .
Construction of the Logarithmic $D_{\log}$ -Functor:
- Local Construction: For a logarithmically small affine $(Y, Z)$ , the authors construct the local functor $D_{\log}$ . The core idea is to exploit the fact that for a $p$ -torsion module $M$ , the logarithmic poles of its pullback along a specific tower of finite flat covers (involving $p^n$ -th roots of the local coordinates defining $Z$ ) vanish.
- Descent: On these covers, the pullback becomes a standard (non-logarithmic) Fontaine-Faltings module. The authors apply the classical Faltings $D$ -functor to obtain a local system on the cover. Using the fact that the fundamental group of the complement $Y^\circ_K$ acts on these systems, they descend the local system from the cover back to $Y^\circ_K = Y_K - Z_K$ .
- Globalization: By covering the global scheme $(Y, Z)$ with logarithmically small affines, the authors glue these local $D_{\log}$ functors. They utilize faithfully flat descent for rigid spaces to ensure the resulting local systems glue to form a global $p$ -adic étale local system on $Y^\circ_K$ .
Full Faithfulness: The authors prove that the global $D_{\log}$ functor is fully faithful. The proof relies on the full faithfulness of the non-logarithmic $D$ -functor. They show that a morphism (or isomorphism) between logarithmic modules can be detected locally on the complement $U = Y - Z$ , where the logarithmic structure is trivial, and then extended back to the whole space using the abelian nature of the category of torsion modules and the vanishing of cokernels on the open dense set.

Key Contributions and Results

Explicit Construction: The paper provides the first explicit definition of logarithmic Fontaine-Faltings modules and the construction of the logarithmic $D_{\log}$ -functor.
Theorem 2.10: The main result establishes a functor $D_{\log}$ $D_{l o g}$ from the category of logarithmic Fontaine-Faltings modules over $(Y, Z)$ $(Y, Z)$ to the category of continuous representations of $\pi_1^{\text{ét}}(Y^\circ_K)$ $π_{1}^{\overset{e}{ˊ} t} (Y_{K}^{\circ})$ .
- This functor is compatible with the classical Faltings $D$ -functor: $D_{\log}(M)|_{U_K} \cong D(M|_U)$ .
- If $Y$ is proper over $W$ , the functor algebraizes to a representation of $\pi_1^{\text{ét}}(Y^\circ_K)$ .
Full Faithfulness (Proposition 2.11): The functor $D_{\log}$ is proven to be fully faithful. Consequently, representations in the essential image are defined as "logarithmic crystalline representations."
Independence of Frobenius: Theorem 2.3 confirms that the category of logarithmic Fontaine-Faltings modules is independent of the choice of Frobenius lifting, analogous to the non-logarithmic case.

Significance
The paper claims to complete the theoretical framework initiated by Faltings in 1989 regarding the logarithmic extension of crystalline representations. By explicitly detailing the construction of the $D_{\log}$ -functor and proving its full faithfulness, the authors provide the necessary tools to study $p$ -adic representations arising from logarithmic geometry (such as those associated with varieties with semistable reduction) in a manner parallel to the classical Fontaine-Laffaille theory. The work serves as a foundational reference for the logarithmic context, which was previously only alluded to without rigorous construction.