0
$\begingroup$

I'd like to understand why if $X$ is a locally Noetherian scheme, then $X$ is quasiseparated. Recall that a scheme is quasiseparated if the intersection of two quasicompact open subsets is quasicompact. This is equivalent to the intersection of any two affine open subsets being a finite union of affine open subsets. Further, locally Noetherian means that $X$ can be covered by affine open sets $\operatorname{Spec}A$ where $A$ is a Noetherian ring.

So suppose that $U=\operatorname{Spec}A$ and $V=\operatorname{Spec}B$ are two open affine subsets of $X$. Further suppose that $X$ is covered by $\{\operatorname{Spec}A_i\}$, as $i$ runs over some index set $I$, and where the $A_i$ are Noetherian rings. We'd like to show that $U\cap V$ can be covered by a finite number of affine open subsets of $X$.

Recall that the intersection $U\cap V=\operatorname{Spec}A\cap\operatorname{Spec}B$ is a union of open sets that are simulaneously distinguished in both $\operatorname{Spec}A$ and $\operatorname{Spec}B$. Further, since $X$ is covered by $\{\operatorname{Spec}A_i\}_{i\in I}$, $U\cap V$ is also the union of open sets that are distinguished in some subset of $\{\operatorname{Spec}A_i\}_{i\in I}$. But how to show that this cover is finite? Or is there a better way to approach this?

$\endgroup$
4
  • 1
    $\begingroup$ If $X$ is locally Noetherian, then for any open $\mathrm{Spec}(A) \subset X$ is such that $A$ is a Noetherian ring, so $\mathrm{Spec}(A)$ is a Noetherian topological space, so any open of $\mathrm{Spec}(A)$ is quasi-compact (because of the increasing chain condition for open subsets) $\endgroup$
    – user598294
    Sep 26, 2020 at 18:07
  • $\begingroup$ If you can, see Corollary 3.22 of the book by Görtz & Wedhorn. I am a fan of this book $\endgroup$
    – user598294
    Sep 26, 2020 at 18:10
  • $\begingroup$ @AlexL So then we get that $U\cap V$ is the union of distinguished open sets of each of the $\operatorname{Spec}A_i$'s that intersect $U\cap V$, and that each of these are quasicompact. But I still dont see why there should be a finite cover, since there still might be an infinite number of $\operatorname{Spec}A_i$'s that intersect $U\cap V$ $\endgroup$
    – ponchan
    Sep 26, 2020 at 18:33
  • $\begingroup$ It is better if you know that every affine open of a locally Noetherian scheme is the spectrum of a Noetherian ring. That way you are not stuck with just the given covering $\mathrm{Spec}(A_i)$ $\endgroup$
    – user598294
    Sep 26, 2020 at 18:52

1 Answer 1

4
$\begingroup$

Let $U=\mathrm{Spec}(A) \subset X$ be an open of a locally Noetherian scheme $X$. Suppose $U_i=\mathrm{Spec}(A_i)$ is an open covering of $X$ by spectra of Noetherian rings.

As you said, $U \cap U_i$ can be cover by open that are distinguished (or principal) in both $U$ and $U_i$. In $U_i$, a distinguished open is Noetherian, being $\mathrm{Spec}((A_i)_g)$ for some $g \in A_i$. Thus $\mathrm{Spec}(A)$ is cover by (finitely many) distinguished opens $\mathrm{Spec}(A_f)$ with $A_f$ Noetherian. This implies that $A$ is Noetherian (maybe take this as an exercice). But the spectrum of a Noetherian ring is a Noetherian topological space; hence every open of $\mathrm{Spec}(A)$ is quasi-compact. In particular for another affine open $V$ of $X$, $U \cap V$ is quasicompact since it is an open of $U$

$\endgroup$
6
  • $\begingroup$ I hope this is clear enough $\endgroup$
    – user598294
    Sep 26, 2020 at 19:39
  • $\begingroup$ Thanks. But I still don't understand why $\operatorname{Spec}(A)$ is covered by finitely many distinguished opens $\operatorname{Spec}(A_f)$. I understand that each $U\cap U_i$ can be covered by these Noetherian open sets, but I don't understand why there are a finite number of them, let alone a finite number of them covering $\operatorname{Spec}(A)$ . $\endgroup$
    – ponchan
    Sep 27, 2020 at 12:59
  • $\begingroup$ So I know that every open subset of a Noetherian space is qc (and so all the $U\cap U_i$ have a finite subcover of distinguished open sets), but there still might be an infinite number of the $U\cap U_i$. $\endgroup$
    – ponchan
    Sep 27, 2020 at 13:16
  • 1
    $\begingroup$ An affine scheme is always quasi-compact $\endgroup$
    – user598294
    Sep 27, 2020 at 15:14
  • $\begingroup$ of course thanks. $\endgroup$
    – ponchan
    Sep 27, 2020 at 15:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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