Refining a theorem of Zarhin, we prove that given a $g$-dimensional abelian variety $X$ and an endomorphism $u$ of $X$, there exists a matrix $A \in \operatorname{M}_{2g}(\mathbb{Z})$ such that each Tate module $T_\ell X$ has a $\mathbb{Z}_\ell$-basis on which the action of $u$ is given by $A$, and similarly for the covariant Dieudonné module tensored with $\mathbb{Q}$ if over a perfect field of characteristic $p$.
- Paper ID: 2107.06363
- Title: Lattices in Tate modules
- Authors: Bjorn Poonen, Sergey Rybakov
- Classification: math.AG (Algebraic Geometry), math.NT (Number Theory)
- Publication Information: Published in Proc. Nat. Acad. Sciences 118 (49) e2113201118 (2021)
- Paper Link: https://arxiv.org/abs/2107.06363
This paper strengthens a theorem of Zarhin by proving that for a given g-dimensional abelian variety X and an endomorphism u of X, there exists a matrix A∈M2g(Z) such that every Tate module TℓX admits a Zℓ-basis in which the action of u is given by the matrix A. Similarly, corresponding results hold for covariant Dieudonné modules tensored with Q over complete fields of characteristic p.
- Core Problem: Investigating whether the action of endomorphisms of abelian varieties on Tate modules can be represented by a unified integer matrix
- Theoretical Importance: This problem concerns the arithmetic properties of abelian varieties, particularly the relationship between the structure of the endomorphism ring and the geometry of Tate modules
- Existing Limitations: Zarhin proved in 2020 that there exists a rational matrix A∈M2g(Q) such that for each ℓ=p, there exists a Qℓ-basis of Vℓ in which the action of u is given by A
- Research Motivation: This paper aims to strengthen Zarhin's result from the rational case to the integral case and extend it to Dieudonné modules
- Abelian varieties in characteristic p>0 cannot always be lifted to characteristic 0, preventing simple reduction from the characteristic 0 case
- Compatibility issues between Tate modules at different primes ℓ must be addressed
- The Frobenius action on Dieudonné modules introduces additional complexity
- Strengthening the Main Theorem: Strengthening Zarhin's theorem from rational matrices to integer matrices, proving the existence of A∈M2g(Z)
- Characterization of Lattice Structures: Establishing the existence of endomorphism-stable lattice structures in various settings
- Extension to Dieudonné Modules: Extending results to covariant Dieudonné modules over complete fields of characteristic p
- Unified Framework: Providing a unified framework for handling cases of different characteristics
Definition 1.1 (Lattice): Given rings R⊆R′ and corresponding modules L⊆L′, we call L an R-lattice in L′ if L has an R-basis that is simultaneously an R′-basis of L′.
Theorem 1.2: Let u∈EndX. Then:
- (a) There exists a u-stable Q-lattice V⊂V
- (b) There exists a u-stable Z-lattice T⊂T
- (c) If p>0 and k is complete, there exists a u-stable Q-lattice V⊂VW
- (d) If k=Fp, there exists a u-stable Z-lattice T⊂TW
Lemma 2.1: Using properties of the Frobenius endomorphism to prove freeness of certain modules in characteristic p. The key insight is the transitivity of the permutation group induced by the Frobenius action on ring extensions.
Lemma 2.2: Systematically treating freeness of various modules when a number field E⊂End∘X is present:
- Tate modules Vℓ are free as (E⊗QQℓ)-modules
- Dieudonné modules MQ are free as (E⊗QK)-modules
- For ℓ∤p⋅discO, Tℓ is free as an (O⊗ZZℓ)-module
The proof employs a strategy of reducing the general case to the number field case:
- Using Zarhin's structure theorem to reduce endomorphisms to product forms of matrix rings
- Reducing the problem to individual factors via isogeny decomposition
- Final reduction to the case of number field actions
- Part (a): Using Zarhin's decomposition theorem and Lemma 2.2, constructing stable lattices of the form V=P⊗QE
- Part (b): Starting from the result in (a), using Q∩Z^(p)=Z[1/p] to construct Z-lattices
- Part (c): Similar to (a), but handling cases involving Dieudonné modules
- Part (d): Crucially using the property Q+W=K when k=Fp
As a pure mathematics theory paper, this work contains no numerical experiments but rather verifies theoretical results through rigorous mathematical proofs.
- Ribet (1976): Establishing foundational results in Galois representation theory
- Zarhin (2020): Proving the rational version of the main theorem
- Chai-Conrad-Oort (2014): Providing modern treatment of complex multiplication theory
- Relative to Zarhin: Strengthening from rationals to integers, extending from single primes to compatibility across all primes
- Relative to Classical Theory: Providing finer characterization of lattice structures
- Technical Innovation: Unified treatment of cases with different characteristics
When considering Q-subalgebras R⊂End∘X, if R is contained in a subalgebra of the form ∏iMri(Ei) (where Ei are number fields), then R-stable lattices exist.
- Quaternion Algebra Case: If X is an elliptic curve and End∘X is a quaternion algebra, no End∘X-stable Q-lattice exists, since quaternion algebras cannot act on 2-dimensional Q-vector spaces.
- Commutative but Non-Number Field Case: A sophisticated counterexample is constructed where R is commutative but no R-stable lattice exists. The specific construction involves products of elliptic curves and upper triangular matrix rings.
- Integer Matrix Representation: Proving that endomorphisms of abelian varieties can be uniformly represented by integer matrices on all Tate modules
- Characteristic p Treatment: Successfully handling Dieudonné modules in characteristic p
- Existence of Lattice Structures: Establishing existence of stable lattices in various settings
- Completeness Requirement: Theorem 1.2(d) currently holds only for k=Fp; the case of general complete fields remains open
- Constructivity: The proof is existential and does not provide algorithms for explicitly constructing stable lattices
- Generalization Restrictions: Counterexamples show that results cannot be naively extended to arbitrary endomorphism subalgebras
- Algorithmic Problems: Seeking effective algorithms for computing stable lattices
- General Complete Fields: Investigating whether Theorem 1.2(d) holds for all complete fields of characteristic p
- Higher-Dimensional Generalizations: Studying whether similar results hold for more general algebraic geometric objects
- Technical Depth: Skillfully combining techniques from algebraic geometry, number theory, and ring theory
- Precision of Results: The strengthening from rationals to integers is a substantial improvement
- Uniformity: Providing a unified framework for handling cases of different characteristics
- Completeness: Not only providing positive results but also offering counterexamples that clarify the limits of generalization
- Proof of Lemma 2.1: Using transitivity of Frobenius action is the key insight
- Reduction Strategy: Systematically reducing complex general cases to manageable special cases
- Counterexample Construction: Demonstrating deep structural understanding
- Foundational Theory: Providing important tools for the arithmetic theory of abelian varieties
- Application Prospects: Potential applications in cryptography and coding theory
- Methodology: Demonstrating effective methods for handling mixed characteristic problems
- Academic Value: Solving an important problem in the field, expected to receive significant citations
- Technical Impact: The methods provided may be applicable to related problems
- Openness: The open problems left behind provide directions for future research
Core technical estimates in the paper include:
- Using the discriminant condition ℓ∤p⋅discO to ensure good properties of local rings
- Exploiting transitivity of permutation groups induced by Frobenius action
- Computing dimensions in tensor product decompositions
- Criteria for determining freeness of modules
- Intersection and generation properties of lattices
- Compatibility of module structures over different rings
This paper exemplifies the typical characteristics of fine structure theory in modern algebraic geometry: obtaining precise structural results through deep technical analysis while using counterexamples to clarify the scope of applicability of the results.