Let $R$ be a Noetherian ring. Show that there is a prime ideal $\mathfrak{p}$ and a monomorphism of $R$-modules $f:R/\mathfrak{p}\to R$.


I considered the set of ideals $\{I< R: \exists \text{ monomorphism } f:R/I\to R\}$ which is not empty hence has a maximal element, say $\mathfrak{p}$.

How can I show that $\mathfrak{p}$ is prime?

  • $\begingroup$ Suppose you have two elements not in $\mathfrak p$ whose product is. CAn you construct a larger ideal adimitting a monomorphism $R/I\to I$? $\endgroup$ – Brandon Apr 3 at 19:24
  • $\begingroup$ @Brandon That is exactly what I was trying to do, but I can't $\endgroup$ – giannispapav Apr 3 at 19:36
  • 1
    $\begingroup$ Any associated prime of $R$ does the job for $p$. At least one exists as $R$ is Noetherian. $\endgroup$ – Youngsu Apr 3 at 20:35
  • 2
    $\begingroup$ Suppose $ab\in\mathfrak p$, $a\notin\mathfrak p$ and $b\notin\mathfrak p$. Then $a\in(\mathfrak p:b)$, so $\mathfrak p\subsetneq(\mathfrak p:b)\subsetneq R$. On the other side, $R/(\mathfrak p:b)$ is isomorphic to the cyclic submodule of $R/\mathfrak p$ generated by the residue class of $b$. Now restrict the injective morphism $R/\mathfrak p\to R$ to this submodule and find an injective morphism $R/(\mathfrak p:b)\to R$, a contradiction. $\endgroup$ – user26857 Apr 3 at 21:13
  • $\begingroup$ @user26857 Thank you very much! $\endgroup$ – giannispapav Apr 4 at 4:19

I would go about this in a different way. For a commutative ring $R$, prime ideal $\mathfrak p \subset R$, and $R$-module $M$, the following are equivalent:

(1) there is an injective $R$-module homomorphism $R/\mathfrak p \to M$;

(2) there is an element $u\in M$ with $\mathrm{Ann}_R(u):= \{r \in R : ru = 0\} = \mathfrak p$.

To see why (2) implies (1), notice that the $R$-module homomorphism $R\to M$ given by $r\mapsto ru$ has kernel $\mathfrak p$. So we'll go about this problem by trying to satisfy (2). Since $\mathrm{Ann}_R(0) = R$, we need to assume $M\ne 0$ if we want there to be such a prime ideal, and we obviously need $R \ne 0$ too (if we want any prime ideals at all). For us, $M=R$, but let's keep it general. $R \ne 0$ is Noetherian and $M \ne 0$ is an $R$-module.

It is in fact true that any nonzero element of $M$ has a nonzero multiple with prime annihilator. Let $0\ne u \in M$. Let $\mathcal S = \{I \subset R : I = \mathrm{Ann}_R(ru) \text{ for some } r\in R \text{ with } ru \ne 0\}$. Notice that if $I\in \mathcal S$, then $I \ne R$, since $1\notin \mathrm{Ann}_R(ru)$ when $ru \ne 0$. Zorn's lemma tells us there's a maximal element $\mathfrak p \in \mathcal S$ under containment, and $\mathfrak p = \mathrm{Ann}_R(ru)$ for some $r\in R$ with $ru\ne 0$. Suppose $a,b \in R-\mathfrak p$ are such that $ab\in \mathfrak p$. Then $bru \ne 0$ and $\mathrm{Ann}_R(bru) \in \mathcal S$ contains $a$ as well as $\mathfrak p$, contradicting maximality of $\mathfrak p$ in $\mathcal S$. So $\mathfrak p$ is prime.


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.