Localising an open subring at a topologically nilpotent element #
Let A be a commutative ring with a topology and separately continuous multiplication, let B
be an open subring of A, and let s : B be topologically nilpotent in A. Inverting s on
both sides does not distinguish the two rings: the induced map B_s → A_s is a ring isomorphism.
The point is that B is open, so a topologically nilpotent s absorbs every element of A into
B after enough multiplications. Passing to B_s makes that absorption invertible, which is
exactly what surjectivity needs. Injectivity is not topological at all and is Mathlib's already.
Implementation notes #
Surjectivity goes through the Mathlib criterion rather than through elements of the localisations:
IsLocalization.Away.map_surjective_iff reduces it to "for every a : A, some sᵐ * a is the
image of an element of B", which is exists_mul_pow_mem_of_isTopologicallyNilpotent verbatim,
up to commuting the product.
Only SeparatelyContinuousMul is assumed, matching the absorption lemma: continuity of addition, a
nonarchimedean neighbourhood basis and any Huber structure are all irrelevant here.
Injectivity is not proved here at all: it is Mathlib's IsLocalization.map_injective_of_injective
applied to Subring.subtype_injective, which already says an injective ring map induces an
injective map on the corresponding localisations. Nothing about the subring being open, or about
s, enters that half.
Source #
Wedhorn, Adic Spaces (arXiv:1910.05934v1), the localisation step inside the proof of Lemma 7.44(1). Only that step is formalised here — the spectral half of 7.44(1), the homeomorphism between the complements of the two closed subsets, is not proved in this file.
Main results #
TauCeti.Localization.awayMap_subtype_surjective_of_isTopologicallyNilpotent: the induced map on localisations is surjective when the subring is open andsis topologically nilpotent.TauCeti.Localization.awayRingEquivOfIsTopologicallyNilpotent: the resulting isomorphismB_s ≃+* A_s.
The localisation of an open subring surjects onto the localisation of the ring. Given
a : A, the topologically nilpotent s absorbs it into the open subring B after some number of
multiplications, and inverting s undoes them.
Inverting a topologically nilpotent element does not see an open subring. For B an open
subring of A and s : B topologically nilpotent in A, the induced map B_s → A_s is a ring
isomorphism: injective for any subring by IsLocalization.map_injective_of_injective, surjective
because s absorbs into the open B.