3
$\begingroup$

Let $R = \mathbb{Z}/2\mathbb{Z}[X,X^{1/2},X^{1/4},\cdots,X^{1/{2^n}},\cdots]$

Is $\operatorname{Spec}R$ a Noetherian topological space?

Here is what I know.

  1. $R$ is integral over $\mathbb{Z}/2\mathbb{Z}[X]$ and hence $1$-dimensional.

  2. If we replace $\mathbb{Z}/2\mathbb{Z}$ with $\mathbb{C}$, the corresponding Spectrum is non-Noetherian since $(X-1)\subset (X^{1/2}-1)\subset (X^{1/4}-1)\subset \cdots (X^{1/{2^n}}-1)\subset \cdots$ is a strictly increasing chain of radical ideals.

  3. $R$ is not a Noetherian ring.

$\endgroup$
2
  • $\begingroup$ Yes, the map $\mathbf F_2[x] \to \mathbf F_2[x^{1/2^\infty}]$ induces a universal homeomorphism on spectra, so $\operatorname{Spec} \mathbf F_2[x^{1/2^\infty}]$ is Noetherian as it is homeomorphic to $\operatorname{Spec} \mathbf F_2[x]$. See [Tags 04DF and 01S4(2)$\Leftrightarrow$(3)]. $\endgroup$ Commented May 21, 2023 at 15:20
  • 2
    $\begingroup$ Crosspost from MSE. When cross-posting, please link all versions of your qeustion together. $\endgroup$
    – KReiser
    Commented May 22, 2023 at 5:10

1 Answer 1

10
$\begingroup$

In general, if $R$ is a ring of characteristic $p > 0$, write $F_*^n R$ for the ring $R$ but viewed as $R$-algebra via the $p^n$-power Frobenius $R \to F_*^nR$, $x \mapsto x^{p^n}$. Define the perfection $R^{\text{perf}}$ as the colimit of the tower $$R \to F_*R \to F_*^2R \to \ldots.$$ It has a natural map from $R$ inserting into the $0^{\text{th}}$ term. For instance, if $k$ is a field of characteristic $p$ and $R = k[x_1,\ldots,x_n]$, then $R^{\text{perf}} = k^{\text{perf}}\bigl[x_1^{1/p^\infty},\ldots,x_n^{1/p^\infty}\bigr]$.

Lemma. The map $\phi \colon \operatorname{Spec} R^{\text{perf}} \to \operatorname{Spec} R$ is a homeomorphism.

In particular, $\operatorname{Spec} R^{\text{perf}}$ is topologically Noetherian if and only if $\operatorname{Spec} R$ is.

Proof. For any perfect field $k$ (in particular if $k$ is algebraically closed), the map $\operatorname{Hom}(R^{\text{perf}},k) \to \operatorname{Hom}(R,k)$ is a bijection, by the universal property of perfection (or: check that you can extend $f \colon R \to k$ uniquely to $R^{\text{perf}}$ by sending $r \in F_*^nR$ to $\phi(r)^{1/p^n}$). This implies that $\phi$ is a bijection. Since $R \to R^{\text{perf}}$ is integral, the map $\phi$ is closed [Tags 00HZ and 00GU], so it is a homeomorphism. $\square$

(In fact, with a little bit more work one shows that such maps are universal homeomorphisms: it remains a homeomorphism after arbitrary base change. See [Tags 04DF and 01S2] for details.)

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .