Ngô Support Theorem and Polarizability of Quasi-Projective Commutative Group Schemes

Basic Information

• Paper ID: 2304.07729
• Title: Ngô support theorem and polarizability of quasi-projective commutative group schemes
• Authors: Giuseppe Ancona (IRMA, Strasbourg), Dragoş Frăţilă (IRMA, Strasbourg)
• Classification: math.AG (Algebraic Geometry)
• Published Journal: Épijournal de Géométrie Algébrique, Volume 8 (2024), Article No. 11
• Paper Link: https://arxiv.org/abs/2304.07729

Abstract

This paper proves that any commutative group scheme of finite type over a base scheme over a field, if it has connected fibers and admits a relatively ample line bundle, is polarizable in the sense of Ngô. More precisely, the authors associate a polarization to any relatively ample line bundle. This extends the applicability of Ngô's support theorem to new situations, such as Lagrangian fibrations with integral fibers, and has implications for the construction of algebraic classes.

Research Background and Motivation

Problem Background

1. Importance of Ngô's Support Theorem: Ngô's support theorem is an important tool in algebraic geometry, with broad applications particularly in the study of Lagrangian fibrations and hyperkähler varieties. A key assumption of this theorem is the polarizability of the underlying group scheme.
2. Existing Limitations: In Ngô's original work, the condition of polarizability restricted the scope of applications of the support theorem. In particular, for Lagrangian fibrations with integral fibers, it is difficult to verify the polarizability of the corresponding group schemes.
3. Theoretical Significance: The theory of polarization of group schemes is a fundamental concept in algebraic geometry, connecting the algebraic structure of group schemes with their geometric properties.

Research Motivation

The core motivation of this paper is to extend the applicability of Ngô's support theorem by proving that a large class of group schemes (quasi-projective commutative group schemes) are polarizable, thereby enabling the application of the support theorem to a wider range of geometric situations.

Core Contributions

1. Main Theorem: Proves that any quasi-projective commutative group scheme is polarizable; more precisely, the first Chern class of any relatively ample line bundle induces a polarization.
2. Construction Method: Provides an explicit method for constructing polarizations from relatively ample line bundles, with this construction possessing functorial properties compatible with base change.
3. Extended Applications: Extends the applicability of Ngô's support theorem to Lagrangian fibrations with integral fibers.
4. Technical Innovation: Combines algebraic and analytic methods, particularly establishing connections between Chern classes and classical polarizations through the Appell-Humbert theorem.

Detailed Methodology

Problem Setup

Given a commutative group scheme $\pi: G \to B$, where $B$ is a finite type scheme over a field and $G$ has connected fibers, the task is to prove that if there exists a relatively ample line bundle $L$ on $G$, then one can construct a polarization $\eta_L: T(G) \otimes T(G) \to \mathbb{Q}_B(1)$ where $T(G) = R^{2d-1}\pi_!\mathbb{Q}_G(d)$ is the relative Tate module.

Core Construction

1. Construction of Polarization

For any cohomology class $\omega \in H^2(G,\mathbb{Q})(1)$, the authors construct the map: $\eta_\omega: \Lambda^2 T(G) \to \mathbb{Q}_B(1)$

Construction process:

• View $\omega$ as a map in the derived category: $\omega: \mathbb{Q}_G \to \mathbb{Q}_G[2](1)$
• Utilize relative Poincaré duality: $\mathbb{Q}_G = \pi^!\mathbb{Q}_B[-2d](-d)$
• Obtain via adjunction $(R\pi_!, \pi^!)$: $\omega: R\pi_!\mathbb{Q}_G \to \mathbb{Q}_B[-2d+2](-d+1)$
• Take the $2d-2$ cohomology sheaf to obtain the desired pairing

2. Functorial Properties

Lemma 2.1: This construction is compatible with base change. For base change $f: B' \to B$, we have $f^*(\eta_\omega) = \eta_{\omega'}$, where $\omega'$ is the pullback of $\omega$.

Technical Route

Step One: Reduction to the Absolute Case

Using functorial properties, the problem reduces to the case $B = \text{pt}$, i.e., studying polarizations induced by ample line bundles on commutative algebraic groups $G$.

Step Two: Chevalley Structure Theorem

For a connected commutative algebraic group $G$, there exists an exact sequence: $1 \to L \to G \xrightarrow{p} A \to 1$ where $L$ is an affine algebraic group and $A$ is an abelian variety.

Step Three: Relations in the Picard Group

Proposition 4.1: The projection $p: G \to A$ induces a surjection $p^*: \text{Pic}(A) \to \text{Pic}(G)$.

Proposition 4.5: If $p^*L_{(H,\rho)}$ is ample on $G$, then $H$ is a non-degenerate Hermitian form.

Step Four: Abelian Variety Case

Using the Appell-Humbert theorem, establish connections between Chern classes and the imaginary part of Hermitian forms, proving that the pairing induced by an ample line bundle is non-degenerate.

Technical Innovations

1. Unified Construction: Provides a unified method for constructing polarizations, independent of specific fiber structures.
2. Algebraic-Analytic Combination: Cleverly combines abstract methods from algebraic geometry with concrete calculations from complex analytic geometry.
3. Functoriality: The construction possesses good functorial properties, enabling treatment of families.

Experimental Setup

This is a pure theoretical mathematics paper with no numerical experiments. All results are obtained through rigorous mathematical proofs.

Main Results

Core Theorem

Theorem 1.2: Any quasi-projective group scheme is polarizable. More precisely, let $L$ be a relatively ample line bundle on $G \to B$, then its first Chern class induces a polarization: $\eta_L: T(G) \otimes T(G) \to \mathbb{Q}_B(1)$

Important Applications

Corollary 1.3: Let $X$ be a projective hyperkähler variety and $f: X \to B$ a Lagrangian fibration with integral fibers. Then all perverse sheaves appearing in the decomposition theorem of $f$ have dense support.

Technical Results

Theorem 3.6: For an ample line bundle $L_{(H,\rho)}$ on an abelian variety $A = V/\Gamma$, the pairing constructed from its Chern class $\eta_\omega: \Lambda^2 T(A) \to \mathbb{Q}(1)$ is non-degenerate.

Proposition 6.4: Let $L$ be an ample line bundle on $G$. If $L \cong p^*M$ (where $M$ is a line bundle on $A$), then $M$ is also ample.

Related Work

This paper builds upon the following important works:

1. Ngô's Fundamental Lemma: Provides key tools for the Langlands program, with the support theorem being an important component.
2. Chevalley Structure Theorem: Describes the structure of commutative algebraic groups, forming the basis of the technical approach in this paper.
3. Appell-Humbert Theorem: Provides explicit descriptions of line bundles on abelian varieties, serving as a bridge connecting algebraic and analytic methods.
4. Hyperkähler Geometry: Particularly the study of Lagrangian fibrations, providing important application background for this paper.

Conclusions and Discussion

Main Conclusions

1. Proves that polarizability of quasi-projective commutative group schemes is a very general property.
2. Provides an explicit method for constructing polarizations with good functorial properties.
3. Extends the applicability of Ngô's support theorem, particularly for applications to Lagrangian fibrations.

Technical Contributions

1. Methodology: Demonstrates how to combine abstract derived category methods with concrete algebraic geometry calculations.
2. Theoretical Unification: Unifies the concept of polarization in different contexts into a single framework.

Future Directions

1. Further investigation of polarization theory for non-commutative group schemes.
2. Exploration of applications in arithmetic geometry.
3. Development of more general support theorems.

In-Depth Evaluation

Strengths

1. Theoretical Depth: The paper resolves an important theoretical problem, filling a gap in the applications of Ngô's support theorem.
2. Technical Innovation: Cleverly combines multiple mathematical tools, including derived categories, Hodge theory, and complex geometry.
3. Application Value: Directly extends the applicability of an important theorem, with significant impact on hyperkähler geometry and related fields.
4. Writing Quality: The paper is well-structured with clear explanations from motivation to technical details.

Technical Advantages

1. Uniformity: Provides a unified construction method, avoiding case-by-case analysis.
2. Functoriality: The construction possesses good categorical properties.
3. Completeness: Addresses both the complex number field case and extends to more general $\ell$-adic settings.

Limitations

1. Technical Complexity: The proofs involve multiple deep mathematical theories with a high barrier to understanding.
2. Scope of Applications: While extending applicability, it remains limited to specific geometric situations.
3. Computational Complexity: Practical construction of polarizations may be complex in specific cases.

Impact Assessment

1. Theoretical Impact: Provides new solutions to important problems in algebraic geometry.
2. Application Prospects: Significantly advances research on Lagrangian fibrations and hyperkähler geometry.
3. Methodological Contribution: Exemplifies the combined use of different tools in modern algebraic geometry.

Applicable Scenarios

1. Lagrangian Fibrations: Particularly those with integral fibers.
2. Hyperkähler Geometry: Applications in the construction of algebraic classes.
3. Group Scheme Theory: Provides new tools for studying commutative group schemes.
4. Arithmetic Geometry: Potential applications in research related to the Langlands program.

References

The paper cites important literature in algebraic geometry, including:

• Works related to Ngô's fundamental lemma
• Mumford's classical textbooks on abelian varieties
• SGA series on group schemes and étale cohomology
• Recent research on hyperkähler geometry and Lagrangian fibrations

Overall Assessment: This is a high-quality theoretical mathematics paper that resolves an important and technically demanding problem in algebraic geometry. The contribution lies not only in the specific results but also in demonstrating the methodological value of synergistic use of multiple tools in modern algebraic geometry. For mathematicians working in related fields, this represents an important theoretical advance.