$F$-divided bundles on normal $F$-finite schemes

Basic Information

• Paper ID: 2510.10582
• Title: $F$-divided bundles on normal $F$-finite schemes
• Authors: Adrian Langer, Lei Zhang
• Classification: math.AG math.AC math.AT math.NT
• Publication Date: October 14, 2025 (arXiv preprint)
• Paper Link: https://arxiv.org/abs/2510.10582

Abstract

This paper investigates $F$-divided bundles on irreducible Noetherian normal $F$-finite $\mathbb{F}_p$-schemes, proving that their Tannakian category is determined by the behavior at the generic point. In particular, if $U\subset X$ is an open subset of a normal variety defined over an algebraically closed field, then the corresponding $F$-divided fundamental group homomorphism is faithfully flat. This parallels known facts about the topological fundamental group of open subsets of normal complex analytic varieties. Using this result, the authors prove that simply connected, proper, normal varieties in positive characteristic admit no nontrivial $F$-divided bundles, generalizing earlier results of H. Esnault and V. Mehta on smooth projective varieties and resolving the Gieseker conjecture in a more general setting.

Research Background and Motivation

Problem Background

1. Topological Analogy: For a simply branched complex analytic variety $X$ and the complement $U\subset X$ of a proper closed analytic subset, there exists a surjection $\pi_1^{\text{top}}(U)\to\pi_1^{\text{top}}(X)$. This fact has a corresponding analog in algebraic geometry.
2. Stratified Fundamental Group: Grothendieck introduced coherent stratified sheaves, which form a Tannakian category on finite type schemes, leading to the definition of the stratified fundamental group $\pi_1^{\text{strat}}(X)$.
3. Positive Characteristic Analogy: In positive characteristic, coherent stratified sheaves on $F$-finite Noetherian schemes are equivalent to the category of $F$-divided bundles.

Research Motivation

1. Theoretical Completeness: Prove fundamental group properties in positive characteristic analogous to the characteristic zero case
2. Gieseker Conjecture: Resolve the conjecture concerning the existence of $F$-divided bundles on simply connected varieties in positive characteristic
3. Technical Generalization: Extend known results for smooth projective varieties to normal proper varieties

Core Contributions

1. Main Theorem 0.1: Proves that for an integral Noetherian geometrically simply branched $F$-finite $\mathbb{F}_p$-scheme $X$, the restriction $\Pi_U^{F\text{-div}}\to\Pi_X^{F\text{-div}}$ for any open subset $U\subseteq X$ is a relative gerbe over $O_X(X)^{\text{perf}}$.
2. Main Theorem 0.2: Proves the Gieseker conjecture for normal schemes: If $X$ is a proper geometrically connected normal integral scheme over a perfect field in positive characteristic, and the maximal étale quotient of the Nori fundamental group scheme vanishes, then $X$ admits no nontrivial $F$-divided bundles.
3. Technical Innovation: Establishes h-descent theory for $F$-divided bundles, generalizing Bhatt's results.
4. Structure Theory: Develops structure theory for simple $F$-divided bundles on normal projective varieties.

Methodology Details

Core Concept Definitions

$F$-divided Sheaves

Let $X$ be a locally Noetherian $\mathbb{F}_p$-algebraic stack. An $F$-divided coherent sheaf is a sequence $\{E_i,\sigma_i\}_{i\in\mathbb{Z}_{\geq 0}}$, where:

• $E_i$ is a coherent $\mathcal{O}_X$-module on $X$
• $\sigma_i: F_X^*E_{i+1}\to E_i$ is an $\mathcal{O}_X$-module isomorphism

Tannakian Gerbe

For a connected Noetherian $F$-finite $\mathbb{F}_p$-scheme $X$, the category of $F$-divided vector bundles $\text{Vect}^{\text{perf}}(X)$ is a Tannakian category over $\text{End}(1_X)=\mathcal{O}_X(X)^{\text{perf}}$, corresponding to the $F$-divided fundamental gerbe $\Pi_X^{F\text{-div}}$.

Technical Methods

1. h-Descent Theory

Theorem 2.2 (B. Bhatt): The functor $\text{Vect}^{\text{perf}}(-)$ satisfies descent with respect to h-covers of locally Noetherian $\mathbb{F}_p$-schemes.

This result is based on Halpern-Leistner and Preygel's theory of h-descent for vector bundles on locally Noetherian derived schemes.

2. Normality and Reflexivity

For a normal locally Noetherian algebraic stack $X$, define the category of $F$-divided reflexive sheaves: $\text{Ref}^{\text{perf}}(X) = \lim\left(\cdots \text{Ref}(X) \xrightarrow{F_X^{[*]}} \text{Ref}(X) \xrightarrow{F_X^{[*]}} \text{Ref}(X)\right)$

Lemma 2.6: The restriction $j^*: \text{Ref}^{\text{perf}}(X)\to\text{Vect}^{\text{perf}}(X_{\text{reg}})$ yields a categorical equivalence.

3. Generic Point Behavior

Lemma 3.1: For a Noetherian geometrically simply branched integral $\mathbb{F}_p$-algebra $R$ and its fraction field $K$, the canonical map $R^{\text{perf}}\to K^{\text{perf}}$ is an isomorphism.

Lemma 3.6: For an integral Noetherian regular $F$-finite $\mathbb{F}_p$-scheme $X$ and its generic point $\eta$, the restriction functor $\text{Vect}^{\text{perf}}(X)\to\text{Vect}^{\text{perf}}(\eta)$ is fully faithful.

Proof Strategy

Main Theorem Proof Outline

1. Normalization Reduction: Use properties of finite universal homeomorphisms to reduce the general case to the normal case
2. Regular Locus: Consider the regular locus $U=X_{\text{reg}}$, utilizing properties of codimension $\geq 2$
3. Full Faithfulness: Prove that $\text{Vect}^{\text{perf}}(X)\to\text{Vect}^{\text{perf}}(U)$ is fully faithful
4. Relative Gerbe: Apply discriminant criteria from Tannakian theory

Experimental Setup

Theoretical Verification

This paper is pure mathematical theory research with no numerical experiments. Verification methods include:

1. Constructive Proofs: Demonstrate theoretical results through explicit constructions
2. Counterexample Analysis: Analyze boundary cases and counterexamples
3. Comparison with Known Results: Consistency checks against known results by Esnault-Mehta and others

Technical Tools

1. Gabber's Alteration Theorem: Guarantees the existence of smooth alterations
2. Chow's Lemma: Reduces proper scheme problems to the projective case
3. Moduli Space Theory: Utilizes moduli spaces of Gieseker-stable vector bundles

Main Results

Core Theorems

Theorem 5.3: Let $X$ be an irreducible Noetherian geometrically simply branched $F$-finite $\mathbb{F}_p$-scheme, and let $\eta$ be the generic point of $X$. Then the induced 1-morphism $\Pi_\eta^{F\text{-div}}\to\Pi_X^{F\text{-div}}$ is a relative gerbe over the field $\mathcal{O}_X(X)^{\text{perf}}$.

Resolution of the Gieseker Conjecture

Theorem 7.3: Let $X$ be a proper geometrically connected normal integral scheme over a perfect field $k$ in positive characteristic. If for some rational point $x\in X(k)$, the maximal étale quotient $\pi_1^{N,\text{ét}}(X,x)$ of the Nori fundamental group scheme vanishes, then $X$ admits no nontrivial $F$-divided bundles.

Technical Results

1. Proposition 3.4: The restriction of the Picard group of $F$-divided line bundles on a normal integral scheme to the generic point is injective
2. Proposition 6.3: Structure theorem for $F$-divided bundles on normal projective varieties
3. Corollary 5.5: Birational morphisms induce relative gerbes

Related Work

Historical Development

1. Grothendieck (1968): Introduced the concept of coherent stratified sheaves
2. Gieseker (1975): Proposed the conjecture on flat vector bundles
3. Esnault-Mehta (2010): Proved the case of smooth projective varieties
4. Kindler (2015): Related results using D-module theory

Technical Foundations

1. Tannakian Theory: Foundational theory of Deligne-Milne
2. $F$-divided Sheaf Theory: Modern developments by Bhatt and others
3. Alteration Theory: Gabber's generalization of de Jong's results

Conclusions and Discussion

Main Conclusions

1. Establishes a complete theory of $F$-divided fundamental groups in positive characteristic
2. Resolves the Gieseker conjecture for normal proper varieties
3. Proves fundamental group properties in positive characteristic analogous to complex geometry

Theoretical Significance

1. Unification: Unifies fundamental group theory across different characteristics
2. Generalization: Extends from smooth projective to normal proper cases
3. Depth: Reveals deep connections between $F$-divided bundles and geometric properties

Technical Innovation

1. h-Descent: First establishment of h-descent theory for $F$-divided bundles
2. Non-flat Descent: Resolves technical difficulties in moduli space theory
3. Gerbe Theory: Systematic application of gerbe theory to positive characteristic geometry

In-Depth Evaluation

Strengths

1. Theoretical Depth: Resolves an important long-standing conjecture
2. Technical Innovation: Develops multiple new technical tools
3. Systematicity: Establishes a complete theoretical framework
4. Generalization: Significantly extends the scope of known results

Challenges and Limitations

1. Technical Complexity: Proofs require multiple deep modern tools
2. Applicability Conditions: Still restricted to specific types of schemes
3. Computational Complexity: Practical verification of related properties may be difficult

Impact Assessment

1. Theoretical Contribution: Important contribution to foundational algebraic geometry theory
2. Methodology: Provides new research methods for related problems
3. Future Research: May inspire broader research directions

Applicable Scenarios

1. Fundamental Research: Fundamental group theory in algebraic geometry
2. Classification Problems: Classification studies of special varieties
3. Moduli Theory: Research on moduli spaces of vector bundles

References

The paper cites key literature in the field, including:

• Grothendieck's SGA series
• Esnault-Mehta's work on the smooth case
• Bhatt's modern theory of $F$-divided sheaves
• Gabber's generalization of alteration results

This paper represents significant progress in $F$-divided bundle theory in algebraic geometry. It not only resolves a long-standing conjecture but also establishes a solid theoretical foundation for further development in the field. Its technical innovations and theoretical depth make it an important contribution to the area.