Documentation

TauCeti.NumberTheory.Padics.MultiplicativeCompletion.LayerTateModule.Basic

The Tate module of a finite layer #

Let L/K be a finite Galois extension of p-adic fields with group G = Gal(L/K), and let A(L) = lim_m Lˣ/(Lˣ)^(p^m) be the p-adic completion of Lˣ, a ℤ_p[G]-module. The module Y = I_{G_K} / I_{G_L} I_{G_K} of NSW (5.6.5) is a finitely generated ℤ_p[G]-module of projective dimension at most one which is an extension

0 → A(L) → Y → I_G → 0

of the augmentation ideal I_G of ℤ_p[G] by A(L). The integral decomposition Y ≃ M₀ ⊕ ℤ_p[G]^N of the proof of NSW (7.4.1) (TauCeti.nonempty_linearEquiv_tameFrameModule_prod) uses only these properties, so a Tate module of the layer, TauCeti.LayerTateModule, is any module with exactly them; it is not identified here with the quotient I_{G_K} / I_{G_L} I_{G_K}. (The name TauCeti.TateModule is the p-adic Tate module lim_n A[p^n] of an abelian group.)

Projective dimension at most one is the theorem of Nakayama and Rim (Rep.projective_ker_of_isZero_res): a cohomologically trivial representation has a projective resolution of length one. So any cohomologically trivial extension of I_G by A(L) is a Tate module of the layer (TauCeti.LayerTateModule.nonempty_of_isZero). The splitting module of a class u ∈ H²(G, A(L)) is such an extension once u satisfies Tate's hypotheses on the p-subgroups of G (TauCeti.LayerTateModule.nonempty_of_tateHypotheses); the class of the layer, carried to A(L) along local reciprocity, is one.

Main definitions #

Main results #

References #

structure TauCeti.LayerTateModule (p : ℕ) [Fact (Nat.Prime p)] (L : Type u) [Field L] (K : Type u) [Field K] [Algebra K L] :
Type (u + 1)

A Tate module of a layer L/K: the properties of the module Y = I_{G_K}/I_{G_L} I_{G_K} of NSW (5.6.5) that the integral decomposition in the proof of (7.4.1) consumes, without an identification with that quotient. It is a finitely generated ℤ_p[Gal(L/K)]-module of projective dimension at most one, together with an extension 0 → A(L) → Y → I_G → 0 of the augmentation ideal I_G by the p-adic completion A(L) of Lˣ.

Instances For
    theorem TauCeti.LayerTateModule.exact_ι_π {p : ℕ} [Fact (Nat.Prime p)] {L : Type u} [Field L] {K : Type u} [Field K] [Algebra K L] (Y : LayerTateModule p L K) :
    Function.Exact ⇑Y.ι ⇑Y.π

    The inclusion of A(L) and the projection onto I_G form an exact sequence.

    theorem TauCeti.LayerTateModule.nonempty_of_isZero {p : ℕ} [Fact (Nat.Prime p)] {L K : Type} [Field L] [Field K] [Algebra K L] [Algebra ℚ_[p] L] [Module.Finite ℚ_[p] L] [Finite Gal(L/K)] (Y : Rep.{0, 0, 0} ℤ_[p] Gal(L/K)) (hY : ∀ (S : Subgroup Gal(L/K)) [inst : Fintype ↥S] (n : ℤ), CategoryTheory.Limits.IsZero (tateCohomology (Rep.res S.subtype Y) n)) {ι : Additive ↥(padicCompletionUnits p L) →ₗ[MonoidAlgebra ℤ_[p] Gal(L/K)] Y.ρ.asModule} {π : Y.ρ.asModule →ₗ[MonoidAlgebra ℤ_[p] Gal(L/K)] ↥(RingHom.ker (MonoidAlgebra.augmentation ℤ_[p] Gal(L/K)))} (hι : Function.Injective ⇑ι) (hπ : Function.Surjective ⇑π) (hιπ : Function.Exact ⇑ι ⇑π) :

    A cohomologically trivial extension of I_G by A(L) is a Tate module. Let L be a finite extension of ℚ_p and K a subfield with Gal(L/K) finite. Let Y be a representation of Gal(L/K) over ℤ_p whose Tate cohomology vanishes on every subgroup in every degree, and 0 → A(L) → Y → I_G → 0 an exact sequence of ℤ_p[Gal(L/K)]-modules. Then Y is finitely generated, and by the theorem of Nakayama and Rim (Rep.projective_ker_of_isZero_res) it has projective dimension at most one, so it is a Tate module of the layer.

    A degree-two class of A(L) satisfying Tate's hypotheses gives a Tate module. Let L be a finite extension of ℚ_p and K a subfield with Gal(L/K) finite, and let u ∈ H²(Gal(L/K), A(L)). Suppose that for every p-subgroup S of Gal(L/K), H¹(S, A(L)) = 0, the restriction of u generates H²(S, A(L)) as a ℤ_p-module, and H²(S, A(L)) has order #S. Then the splitting module of u, an extension 0 → A(L) → Y → I_G → 0, is cohomologically trivial (TauCeti.TateCohomology.isZero_res_splittingModule; on ℓ-subgroups for ℓ ≠ p all positive-degree group cohomology of a ℤ_p-module vanishes), so it is a Tate module of the layer.

    theorem TauCeti.LayerTateModule.nonempty_of_equiv {p : ℕ} [Fact (Nat.Prime p)] {L K : Type u} [Field L] [Field K] [Algebra K L] {L' K' : Type} [Field L'] [Field K'] [Algebra K' L'] (φ : Gal(L/K) ≃* Gal(L'/K')) (c : ↥(padicCompletionUnits p L) ≃* ↥(padicCompletionUnits p L')) (hsmul : ∀ (a : ℤ_[p]) (x : Additive ↥(padicCompletionUnits p L)), Additive.ofMul (c (Additive.toMul (a • x))) = a • Additive.ofMul (c (Additive.toMul x))) (haut : ∀ (σ : Gal(L/K)) (x : ↥(padicCompletionUnits p L)), c (((padicCompletionUnitsAut p L K) σ) x) = ((padicCompletionUnitsAut p L' K') (φ σ)) (c x)) (Y : LayerTateModule p L' K') :

    Transport of the Tate module along an isomorphism of layers. Let L/K and L'/K' be layers, L', K' in Type, with an isomorphism φ of their automorphism groups and a ℤ_p-linear isomorphism A(L) ≃ A(L') intertwining the actions of σ and φ σ. Then a Tate module of L'/K' gives one of L/K: its carrier, lifted to the universe of L, is a module over ℤ_p[Gal(L/K)] through the induced isomorphism of group algebras.