Finite algebras over a Huber ring #
Let A be a Huber ring and B a commutative A-algebra that is finite as an A-module, with
its natural topology (IsModuleTopology A B). Then B is a topological ring
(IsModuleTopology.isTopologicalRing), and this file shows that it is again a Huber ring, and a
Tate ring when A is. No completeness, noetherian, flatness or separatedness hypothesis is
needed. This is how a finite étale algebra over a complete Tate ring becomes a Tate ring, the first
step in making it a Huber pair over (A, A⁺).
Fix a pair of definition (A₀, I) of A and a finitely generated A₀-submodule M₀ ⊆ B
containing 1 and spanning B over A. The submodules Iⁿ • M₀ are a neighbourhood basis of
0 (TauCeti.Huber.PairOfDefinition.hasBasis_nhds_zero_idealOfDefinition_pow_smul). The
multiplier ring B₀ = {b ∈ B | b M₀ ⊆ M₀} is a subring of B (AddSubgroup.multiplierRing).
Since M₀ * M₀ is finitely generated, a power Iʲ carries it into M₀
(TauCeti.Huber.PairOfDefinition.exists_pow_smul_le), so Iʲ • M₀ ⊆ B₀
(Submodule.smul_subset_multiplierRing) and B₀ is open. With the extension J of I to
B₀, one gets I^(n + j) • M₀ ⊆ Jⁿ ⊆ Iⁿ • M₀ (Submodule.smul_subset_coe_map_pow,
Submodule.coe_map_pow_subset), the second inclusion because 1 ∈ M₀ and B₀ M₀ ⊆ M₀. So the
topology of B₀ is the J-adic one and (B₀, J) is a pair of definition of B. When A is
Tate, the image of a pseudouniformiser of A is one of B.
Main results #
TauCeti.Huber.PairOfDefinition.exists_pairOfDefinition_isIntegral: every pair of definition(A₀, I)ofAgives a pair of definition ofBwhose ring of definition contains the image ofA₀and is integral overA₀.TauCeti.Huber.IsHuberRing.of_isModuleTopology: a finite algebra over a Huber ring, with its natural topology, is a Huber ring.TauCeti.Huber.IsTateRing.of_isModuleTopology: a finite algebra over a Tate ring, with its natural topology, is a Tate ring.TauCeti.Huber.IsPowerBounded.algebraMap_of_isModuleTopology: the structure mapA → Bpreserves power-boundedness.TauCeti.Huber.isOpen_integralClosure_map_algebraMap: the integral closure inBof the image of an open subring ofAis open.
References #
- R. Huber, A generalization of formal schemes and rigid analytic varieties, Math. Z. 217 (1994), §2, before Lemma 2.3.
- K. S. Kedlaya and R. Liu, Relative p-adic Hodge theory: foundations, Astérisque 371 (2015), §2.2, Lemma 2.2.12, and §2.4, Remark 2.4.2.
The pair of definition #
Pairs of definition of a finite algebra. For every pair of definition (A₀, I) of A, a
finite A-algebra B with its natural topology has a pair of definition (B₀, J) whose ring of
definition contains the image of A₀ and is integral over A₀. It is the multiplier ring of a
finitely generated spanning lattice containing 1, with the extension of I to it.
A finite algebra over a Huber ring is a Huber ring. If B is a commutative A-algebra,
finite as an A-module and carrying its natural topology, then B is a Huber ring.
A finite algebra over a Tate ring is a Tate ring. If B is a commutative A-algebra,
finite as an A-module and carrying its natural topology, then B is a Tate ring; the image of a
pseudouniformiser of A is a pseudouniformiser of B.
Power-bounded and integral elements #
The structure map of a finite algebra preserves power-boundedness. A power-bounded a
lies in a ring of definition A₀ of A, and algebraMap A B carries A₀ into a ring of
definition of B.
The integral closure of the image of an open subring is open. If C is an open subring
of a Huber ring A, the integral closure of its image in a finite A-algebra B with its natural
topology is open: it contains a ring of definition of B that is integral over a ring of
definition A₀ ⊆ C of A.