Documentation

TauCeti.Topology.Algebra.Nonarchimedean.Completion.Basic

Completions of nonarchimedean groups and rings #

Three facts about the Hausdorff completion that need only the additive, resp. ring, structure: the closure of the image of an open additive subgroup is open, the kernel of the completion map is the closure of the zero ideal, and integral closedness of an open subring survives completion.

They are stated here rather than alongside the Huber-ring theory that uses them, since none mentions a pair of definition or an adic topology, and they live in the UniformSpace.Completion namespace of the construction they describe rather than in a TauCeti one.

The last of the three is Huber's Lemma 2.4.3(iv), Wedhorn's Lemma 7.47(4): if G is open and integrally closed in A, the closure Ĝ of its image in  is integrally closed in Â. The proof below is Huber's. The integral closure H of Ĝ in  contains the open subring Ĝ, hence is open, so every neighbourhood of a point of H meets the image of A inside H. For such an image i b, an integral relation over Ĝ is perturbed one coefficient at a time into a relation whose coefficients come from G; openness of Ĝ keeps the value of the perturbed relation at i b inside Ĝ, and an open G is closed, hence exactly the preimage of Ĝ, so that value is the image of an element of G. Subtracting it from the constant term leaves an integral relation for b over G, so b ∈ G, and i b therefore lies in the image of G.

Main results #

References #

The closure in the completion of the image of an open additive subgroup is open. Only the additive structure is involved.

An open additive subgroup is recovered from the closure of its image. For G open, the preimage under A → Â of the closure of the image of G is G itself.

For a uniform additive group, this is the injectivity half of the bijection G ↦ closure (ι '' G) between the open subgroups of A and those of  (Wedhorn, Example 5.33); isOpen_closure_image_coe says the map lands in open subgroups. No separation assumption on A is needed. Separate continuity of addition suffices for this preimage equality: an open subgroup is closed, and the completion map induces the topology on A.

@[simp]

The completion map A →+* Â, read as a function, is the coercion A → Â. Mathlib's Filter.Germ.coe_coeRingHom is the same statement for germs.

The kernel of the completion map A → Â is the closure of the zero ideal.

The closure Ĝ in  of the image of a subring G of A is, as a set, the closure of the image of G under the completion coercion. This unfolds Subring.topologicalClosure and Subring.map in one step; it is how every argument below passes between Ĝ and that closure.

@[simp]

Membership in the closure Ĝ in  of the image of a subring G of A, in the form the closure API consumes. This is the mem_-half of the pair whose coe_ half is above; the proofs below pass between the two forms at three separate sites.

The closure in  of the image of an open subring of A is open: the subring form of isOpen_closure_image_coe.

Huber's Lemma 2.4.3(iv): the closure in  of the image of an open subring G of A that is integrally closed in A is integrally closed in Â.