1
$\begingroup$

It is quite easy to show that if $A$ is a Dedekind domain and $\mathfrak{p}\in \operatorname{Spec} A$, then if $A_{\mathfrak{p}}$ is the completion of $A$ at $\mathfrak{p}$ and $A_{(\mathfrak{p})}=(A\setminus \mathfrak{p})^{-1}A$ is the localization of $A$ at $\mathfrak{p}$, then $$A_{(\mathfrak{p})}=A_{\mathfrak{p}}\cap K$$ where $K$ is the field of fractions of $A$ (essentially, by using the notion of order at $\mathfrak{p}$ of the elements of $K$). I would like to know if the same equality remains true in more general cases, assuming $A$ to be for example an order in some number field. Without prime factorization available I find the equality not easy to prove, and I wonder if it remains true or if there are contrexamples?

$\endgroup$
2
  • $\begingroup$ TeX note: use $\operatorname{Spec} A$ $\operatorname{Spec} A$, not $Spec$ $A$ $Spec$ $A$. I have edited accordingly. $\endgroup$
    – LSpice
    May 6, 2022 at 10:06
  • $\begingroup$ There is an ambiguity: what is the completion $A_{\mathfrak p}$? There are two possibilities: either take the $\mathfrak p$-adic completion, or first localize at $\mathfrak p$ then complete the local ring. It seems that A.G's answer is taking the former, but the later seems more reasonable. If you take the later, you could start with a local domain, and ask whether it is some sort of intersection of the completion and its fraction field? $\endgroup$
    – Z. M
    May 6, 2022 at 19:22

1 Answer 1

1
$\begingroup$

No. If $A$ is a complete noetherian local ring, then $A$ is complete for the adic topology defined by any other prime ideal.

$\endgroup$
8
  • $\begingroup$ Thank you for the answer, actually I was interested in something more specific, like an order in a number field. I should probably have specified better in my question, which I have now modified accordingly $\endgroup$
    – Hair80
    May 6, 2022 at 14:25
  • $\begingroup$ What I said was that, more generally, the completion at a prime does not determine the localization at that prime. But I do not understand the meaning of the equality that you suggest. $\endgroup$
    – A.G
    May 6, 2022 at 14:48
  • $\begingroup$ @Hair80 It does not make sense to modify the question so that an answer is "invalidated". $\endgroup$
    – Z. M
    May 6, 2022 at 16:00
  • $\begingroup$ That intersection could be interpreted if $A_{\mathfrak p}$ is a domain, or maybe weakened, so that the injective ring map $A\to A_{\mathfrak p}$ (injectivity by Krull's intersection theorem) induces an injective map $K\to Q(A_{\mathfrak p})$ where $Q(\cdots)$ is the total ring of fractions, where the intersection is taken. $\endgroup$
    – Z. M
    May 6, 2022 at 16:05
  • $\begingroup$ @Z.M It is not ideal, but it sometimes necessary. $\endgroup$
    – Will Sawin
    May 6, 2022 at 16:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.