🔢 mathematics

Brauer group of moduli stacks of parabolic principal bundles over a curve

This paper establishes that the Brauer group of the moduli stack of parabolic stable $\text{PGL}(r,\mathbb{C})$ -bundles on a curve coincides with that of the smooth locus of its coarse moduli space, while demonstrating that the analytic and algebraic Brauer groups of the moduli stack of quasi-parabolic principal $G$ -bundles vanish for any simple and simply connected complex linear algebraic group $G$ .

Original authors: Indranil Biswas, Sujoy Chakraborty

Published 2026-02-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Indranil Biswas, Sujoy Chakraborty

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 an architect trying to understand the "shape" of a city. But this isn't a city of buildings; it's a city of mathematical objects called bundles. Specifically, the authors of this paper are studying a very specific type of city: the Moduli Stack of Parabolic Principal Bundles.

That sounds terrifyingly complex, so let's break it down using some everyday analogies.

1. The Setting: The Curve and the Bundles

Think of a Curve ( $X$ ) as a long, winding road (like a piece of string).

Principal Bundles: Imagine you are walking along this road, and at every step, you are carrying a backpack. But these aren't normal backpacks; they are "twisted" in specific ways. In math, these are called bundles.
Parabolic Points: Now, imagine there are specific "checkpoints" (like rest stops) along the road. At these checkpoints, the backpacks have special rules. You have to arrange the straps in a specific order (a "flag" or "filtration"). This is the Parabolic part.
Weights: At each checkpoint, some straps are more important than others. We assign a "weight" (a number) to each strap to say how important it is. This is the Weight system.

2. The Two Cities: The Stack vs. The Coarse Space

The paper studies two different ways of looking at the "city" of all possible valid backpacks (bundles) on this road.

The Moduli Stack ( $\mathcal{N}$ ): This is the High-Definition, 3D Virtual Reality version of the city. It remembers every tiny detail, including how the backpacks are twisted and how they relate to each other. It's messy, complex, and full of "symmetries" (ways you can rotate a backpack and it still looks the same).
The Coarse Moduli Space ( $N$ ): This is the Flat, 2D Map version. It ignores the tiny symmetries and just shows you the general shape of the city. It's smoother and easier to walk around, but you lose some of the fine-grained details.

The Big Question: Do these two cities have the same "hidden holes" or "twists" in their structure? In math, we measure these hidden twists using something called the Brauer Group.

3. The Main Discovery: The "Brauer Group"

The Brauer Group is like a "magnetic field" or a "topological fingerprint" of the city. It tells you if there are any invisible, non-trivial structures that prevent you from flattening the city out completely.

The Authors' First Big Result (Theorem 1.1):
They proved that for a specific type of backpack (related to the group $PGL(r, \mathbb{C})$ ), the "magnetic fingerprint" of the High-Definition Virtual City (the Stack) is exactly the same as the fingerprint of the Smooth part of the Flat Map (the Coarse Space).

The Analogy: Imagine you have a crumpled piece of paper (the Stack) and a smoothed-out version of it (the Coarse Space). Usually, crumpling a paper creates new wrinkles that change its properties. But the authors proved that for this specific type of paper, if you smooth out the "rough edges" (the singularities), the fundamental "twistiness" (the Brauer group) remains identical.
Why it matters: It means you don't need to do the super-hard math of the Virtual City to understand the twists; you can just study the easier Flat Map, as long as you avoid the broken spots.

4. The Second Discovery: The "Vanishing Act"

The authors also looked at a different kind of backpack, governed by a group called $G$ (which is "simple and simply connected"). Think of this as a very rigid, perfectly symmetrical type of backpack.

The Second Big Result (Theorem 1.2):
They proved that for these rigid backpacks, the Brauer Group vanishes. It becomes zero.

The Analogy: Imagine the city is made of pure, clear glass. No matter how you look at it, there are no hidden magnetic fields, no invisible knots, and no twists. The "Brauer Group" is empty.
Why it matters: This is a huge simplification. It tells mathematicians that for these specific, highly symmetric structures, the "twistiness" problem is solved: there is nothing to worry about. The structure is perfectly "flat" in a topological sense.

Summary of the "Story"

The Problem: Mathematicians wanted to know if the complex, detailed world of "Parabolic Bundles" (the Stack) has different "twists" (Brauer groups) than the simpler, smoothed-out world (the Coarse Space).
The Solution: They proved that for a generic set of rules (weights), the twists are identical. You can use the simpler map to understand the complex city.
The Bonus: They also proved that for a specific, highly symmetric type of bundle, there are no twists at all. The city is perfectly smooth and simple.

In a Nutshell:
The paper is like a guidebook for a complex mathematical landscape. It tells us: "Don't worry about the scary, detailed 3D version; the 2D map works just as well for understanding the hidden twists. And for some parts of the landscape, there are no hidden twists at all!"

1. Problem Statement and Context

The paper investigates the cohomological Brauer group ( $Br(Y) = H^2_{\text{\'et}}(Y, \mathbb{G}_m)_{\text{torsion}}$ ) of moduli spaces and moduli stacks associated with parabolic principal bundles on a smooth projective curve $X$ of genus $g \geq 2$ .

The authors address two specific problems:

PGL(r, C)-bundles: To generalize previous results (specifically from [BCD1]) regarding the relationship between the Brauer group of the moduli stack of parabolic stable $PGL(r, \mathbb{C})$ -bundles and the Brauer group of the smooth locus of the corresponding coarse moduli space. Previous results were limited to "full-flag" systems of weights; this paper aims to prove the result for generic systems of weights.
Simply Connected Groups (G): To determine the Brauer groups of moduli stacks of quasi-parabolic principal $G$ -bundles, where $G$ is a simple and simply connected complex linear algebraic group. The authors aim to prove that these Brauer groups vanish.

2. Methodology and Technical Framework

The paper employs a combination of algebraic geometry, representation theory, and stack theory. The methodology is divided into two main parts corresponding to the two main theorems.

Part A: Parabolic $PGL(r, \mathbb{C})$ -Bundles

Setup: The authors consider the moduli space $M^{m, \alpha}_X(r, \xi)$ of stable parabolic vector bundles of rank $r$ with fixed determinant $\xi$ and generic weights $\alpha$ . The group $\Gamma$ of $r$ -torsion line bundles on $X$ acts on this space via tensor product. The moduli stack of $PGL(r, \mathbb{C})$ -bundles is identified with the quotient stack $[M^{m, \alpha}_X(r, \xi) / \Gamma]$ .
Fixed Point Analysis: A crucial step is analyzing the fixed-point locus of the action of non-trivial elements $L \in \Gamma$ $L \in Γ$ on the moduli space.
- The authors utilize spectral curves. For a torsion line bundle $L$ of order $d$ , they construct an étale Galois cover $\gamma: Y \to X$ with Galois group $\mathbb{Z}/d\mathbb{Z}$ .
- They establish a surjective morphism from a disjoint union of moduli spaces of parabolic bundles on the spectral curve $Y$ (of rank $r/d$ ) to the fixed-point locus in $M^{m, \alpha}_X(r, \xi)$ .
- Dimensional Estimation: By carefully calculating the dimensions of these moduli spaces on $Y$ and comparing them to the dimension of the original moduli space, they prove that the codimension of the fixed-point locus (the union of all non-trivial fixed points) is at least 3.
Stack vs. Coarse Space: Using the fact that the complement of the smooth locus in the coarse moduli space has high codimension (at least 3), and that the quotient map by a finite group preserves codimension, they apply standard results on Brauer groups of Deligne-Mumford stacks (specifically that removing a closed subscheme of codimension $\geq 2$ does not change the Brauer group).

Part B: Quasi-Parabolic $G$ -Bundles (Simply Connected $G$ )

Uniformization Theorem: The authors utilize the uniformization theorem (from Laszlo and Sorger [LS]), which expresses the moduli stack $\mathcal{M}^{\text{par}}_G$ as a quotient stack:
$\mathcal{M}^{\text{par}}_G \simeq [L_X(G) \backslash Q^{\text{par}}_G]$
where $L_X(G)$ is an ind-group scheme (loop group) and $Q^{\text{par}}_G$ is an ind-scheme (related to the affine Grassmannian and flag varieties).
Cohomological Computation:
- They compute the singular cohomology $H^*_B(\mathcal{M}^{\text{par}}_G, \mathbb{Z})$ using the Cartesian diagram relating $Q^{\text{par}}_G$ to the product of the affine Grassmannian and flag varieties.
- They establish that $H^3_B(\mathcal{M}^{\text{par}}_G, \mathbb{Z}) = 0$ .
Analytic vs. Algebraic Brauer Groups:
- They invoke a proposition (from [BHol]) stating that if a stack satisfies specific conditions (classes in $H^2_B$ are represented by line bundles, and line bundles admit algebraic structures), then the algebraic Brauer group is isomorphic to the torsion part of the analytic Brauer group, which in turn is isomorphic to the torsion part of $H^3_B$ .
- They verify these conditions for $\mathcal{M}^{\text{par}}_G$ using the structure of the Picard group and the vanishing of specific cohomology groups of the underlying ind-schemes.
- Since $H^3_B = 0$ , the analytic Brauer group vanishes, implying the algebraic Brauer group vanishes.

3. Key Contributions and Results

Theorem 1.1 (Theorem 4.2)

Statement: Let $\alpha$ be a generic system of weights. Let $\mathcal{N}^{m, \alpha}_X(r, \delta)$ be the moduli stack of parabolic stable principal $PGL(r, \mathbb{C})$ -bundles, and let $N^{m, \alpha}_X(r, \delta)^{\text{sm}}$ be the smooth locus of the corresponding coarse moduli space. Then:
$Br(\mathcal{N}^{m, \alpha}_X(r, \delta)) \simeq Br(N^{m, \alpha}_X(r, \delta)^{\text{sm}})$
Significance: This generalizes the result of [BCD1] from full-flag weights to any generic system of weights. It confirms that the Brauer group of the stack is determined entirely by the smooth part of the coarse moduli space, regardless of the specific generic weight configuration.

Theorem 1.2 (Proposition 5.7 and Theorem 5.8)

Statement: Let $G$ be a simple and simply connected complex linear algebraic group. Both the analytic and algebraic Brauer groups of the moduli stack $\mathcal{M}^{\text{par}}_G(p, P)$ of quasi-parabolic principal $G$ -bundles vanish:
$Br(\mathcal{M}^{\text{par}}_G(p, P)) = 0$
Significance: This is a strong vanishing result. It implies that for simply connected groups, the moduli stack of quasi-parabolic bundles has no non-trivial Azumaya algebras (up to Morita equivalence) and no non-trivial gerbes. This contrasts with cases involving non-simply connected groups (like $PGL(r)$) where the Brauer group is often non-trivial.

4. Significance and Implications

Rationality and Birational Invariants: The Brauer group is a stable birational invariant. By computing these groups, the authors provide tools to study the rationality of these moduli spaces. The vanishing of the Brauer group for simply connected groups suggests these spaces behave "nicely" from a birational perspective.
Generalization of Weight Systems: The extension from full-flag weights to generic weights in the $PGL(r)$ case is a significant technical advancement. It removes a restrictive assumption on the parabolic structure, making the result applicable to a broader class of moduli problems.
Distinction between Simply Connected and Non-Simply Connected Groups: The paper highlights a fundamental dichotomy. For $PGL(r)$ (which is not simply connected), the Brauer group is non-trivial and related to the geometry of the coarse moduli space. For simply connected $G$ , the Brauer group vanishes, reflecting the topological triviality of the underlying loop group actions and the structure of the affine Grassmannian.
Methodological Rigor: The paper provides a robust framework combining spectral curve techniques (for fixed point analysis) and uniformization theorems (for loop group actions) to tackle Brauer group computations in the context of parabolic bundles.

In summary, the paper resolves the computation of Brauer groups for two major classes of parabolic principal bundles, establishing a precise link between stacks and coarse spaces for $PGL(r)$ and proving a complete vanishing theorem for simply connected groups.