0

I recently had a bunch of questions on a problem set that I could not solve and the instructor is not providing solutions.

Let $f: A \to B$ be a surjective ring homomorphism. Prove that the induced map $g: \operatorname{Spec} B \to \operatorname{Spec} A$ is injective.

I tried using the definition of injectivity but I don't know where to go from there and how to use the surjectivity of $f$. Can someone give me a detailed explanation?

Arctic Char
  • 16,007
  • 4
    Assume $f^{-1}(\mathfrak{p})=f^{-1}(\mathfrak{q})$, then we have $\mathfrak{p}=f(f^{-1}(\mathfrak{p}))=f(f^{-1}(\mathfrak{q}))=\mathfrak{q}$ as the map is surjective. –  Nov 14 '21 at 06:42
  • 1
    @Jiacheng that looks like an answer to me - would you care to record it as such below? – KReiser Nov 14 '21 at 07:07
  • 1
    Welcome new contributor. I've added MathJax formatting to your post - please take a look at the changes with the [edit] button. For more details, check out the MathJax tutorial so you can use this in your posts in the future. – KReiser Nov 14 '21 at 07:09
  • @KReiser I think I will just leave it as it is. Thank you. –  Nov 14 '21 at 07:41

2 Answers2

4

This is a community wiki post recording the answer from the comments so that this post may be marked as answered.

Assume $f^{-1}(\mathfrak{p})=f^{-1}(\mathfrak{q})$, then we have $\mathfrak{p}=f(f^{-1}(\mathfrak{p}))=f(f^{-1}(\mathfrak{q}))=\mathfrak{q}$ as the map is surjective. – Jiacheng

KReiser
  • 65,137
3

As a bonus observation: if $f: A \to B$ is injective, then $g$ has a dense image:

Let $D(a) = \{I \in \text{Spec } A\mid a \notin I\}$ be a non-empty basic open subset in the image, for some $a \in A$. Note that $a \neq 0$ as $D(0)=\emptyset$ (all ideals contain $0$).

Then $D(a) \cap g[\text{Spec } B]$ is non-empty iff

$$\exists I' \in \text{Spec } B: g(I')=f^{-1}[I'] \in D(a)$$ or

$$\exists I' \subseteq B \text{ prime ideal }: f(a) \notin I'$$

which is true, as we have $f(a) \neq 0$ by $f$ being an injective ring homomorphism and every non-zero element of a unitary ring has some prime ideal that it is not in. (Well,not quite, I realised it's probably a bit more complicated and I have to consider the radicals of the rings instead but that's ring theory stuff and I'm but a simple topologist, but I did find this answer).

Henno Brandsma
  • 242,131