Theorem 11.2 (Matsumura's Commutative Ring theory) gives us equivalent conditions for a ring $R$ to be considered a DVR. I was stuck while reading the proof of $(4) \implies (3)$,

(3) $R$ is a Noetherian local ring, dim $R > 0$ and the maximal ideal $m_R$ is principal;

(4) $R$ is a one-dimensional normal Noetherian local ring

Let $K$ be the field of fractions of $R$ and let $m$ be the maximal ideal of $R$. Let $x \in m - m^2$ and let $y \in R$ such that $xR : y = m$. We have $a = yx^{-1}$.

Question 1

Suppose that $am \subset m$. How does this imply that $a$ is integral over $R$ by Theorem 2.1?

Theorem 2.1 Suppose that $M$ is an $A$-module generated by $n$ elements, and that $\phi \in$ Hom$_A (M,M)$; let $I$ be an ideal of $A$ such that $\phi(M) \subset IM$. Then there is a relation of the form $$\phi^n + b_1 \phi^{n-1} + ... + b_{n-1} \phi + b_n = 0$$

Since $m$ is an ideal of $R$, it is definitely an $R$-module. It is finitely generated, since $R$ is Noetherian. We can define $\phi : m \rightarrow m$ by $r \mapsto ar$ for all $r \in m$. But then, this only gives $$a^nr + b_1 a^{n-1}r + ... + b_{n-1} ar + b_n = 0$$

We cannot get $$a^n + b_1 a^{n-1} + ... + b_{n-1} a + b_n = 0$$

Since $1 \not\in m$. Then how does this theorem imply that $a$ is integral over $R$?

Question 2

The definition of "normal" for Matsumura is that for any prime ideal "p" of a ring $R$, we have that $R_p$ is integrally closed. How does this imply that $R$ itself is integrally closed?

Thanks in advance

  • $\begingroup$ For the question 2, you should know that being integrally closed is a local property on $\textrm{Spec}(A)$. I tried to find a reference for it (as it is well-known) and came across this : proofwiki.org/wiki/Integrally_Closed_is_Local_Property which made me laugh SO MUCH. ;-) $\endgroup$ – ujsgeyrr1f0d0d0r0h1h0j0j_juj Mar 4 '15 at 12:45
  • $\begingroup$ @RobertGreen Looks like the margin wasn't big enough to fit the full proof? ;) $\endgroup$ – rschwieb Mar 4 '15 at 12:52
  • $\begingroup$ @rschwieb Fermat was a mathematician, despite of the margin situation. ;-) Here we had someone who IMO realized that (s)he didn't know how to prove (3)=> (1) or (2) only when (s)he arrived to the point where (s)he had to show it... $\endgroup$ – ujsgeyrr1f0d0d0r0h1h0j0j_juj Mar 4 '15 at 12:57
  • $\begingroup$ Ok, I will answer to prove question at least. $\endgroup$ – ujsgeyrr1f0d0d0r0h1h0j0j_juj Mar 4 '15 at 13:02

For your question 1, you made a mistake, for theorem 2.1 will imply that $a^nr + b_1 a^{n-1} + ... + b_{n-1} ar + b_n r = 0$, there's an $r$ in front of $b_n$, because $\Phi^0 = \textrm{Id}_M$. From there, do you see how to conlude ?

For the question 2, here you go :

Lemma 1. Let $\rho : A \to B$ be a ring morphism making $B$ an $A$ algebra and $S$ be a multiplicatively closed subset of $A$. Let $C$ be the integral closure of $A$ in $B$. Then (the image of) $S^{-1} C$ is the integral closure of (the image of) $S^{-1}A$ in $S^{-1}B$. One says that localization commutes with integral closure.

Proof. First of all that (the image of) $S^{-1} C$ is integral over $S^{-1}A$ (I drop the image of's from here) is easy. Now let $x\in S^{-1}B$ a root of a monic polynomial $P$ in $S^{-1}A[T]$. Let $s\in S$ be the product of "denominators" of $x$ and coefficients of $P$. Then $sx$ verifies a monic relation over $A$, so that it is in $C$ by definition of $C$, so that $x$ is in $S^{-1}C$. $\square$

Proposition. Let $A$ be a domain. The following are equivalent :

  1. $A$ is integrally closed ;
  2. For all $\mathfrak{p}\in\textrm{Spec}(A)$ the localization $A_{\mathfrak{p}}$ is integrally closed ;
  3. For all $\mathfrak{p}\in\textrm{Spec}(A)$ that is maximal, the localization $A_{\mathfrak{p}}$ is integrally closed.

Proof. As $A$ is a domain, everything (localizations $A_{\mathfrak{p}}$'s included) will take place in $K = \textrm{Frac}(A)$. Then $\textrm{Frac}(A_{\mathfrak{p}}) = K$ for all $\mathfrak{p}\in\textrm{Spec}(A)$, so that the lemma shows that (1) implies (2). No comment one (2) implies (3). Now suppose (3). Note $C$ the integral closure of $A$ in $K$. By the lemma, the morphism (the inclusion) $C\to A$ induces an isomorphism $C_{\mathfrak{p}}\to A_{\mathfrak{p}}$ for $\mathfrak{p}\in\textrm{Spec}(A)$ that is maximal. The conclusion results from the following lemma. $\square$

Lemma 2. Let $\varphi : M\to N$ be an morphisms of $A$-modules. The following are equivalent :

  1. $\varphi$ is an isomorphism ;
  2. $\varphi_{\mathfrak{p}} : M_{\mathfrak{p}}\to N_{\mathfrak{p}}$ for all $\mathfrak{p}\in\textrm{Spec}(A)$ ;
  3. $\varphi_{\mathfrak{p}} : M_{\mathfrak{p}}\to N_{\mathfrak{p}}$ for all $\mathfrak{p}\in\textrm{Spec}(A)$ that is maximal.

Proof. Results immediately from the more general follwoing lemma. $\square.$

Lemma 3. Let $A$ be a ring and $S$ be a multiplicatively closed subset of $A$. Then $M\mapsto S^{-1}M$ is an exact functor.

Proof. $S^{-1} M \simeq M \otimes_A S^{-1} A$ functorialy so that $M\mapsto S^{-1}M$ is already right exact as "tensoring is a right exact functor". To conclude, one just has to show that $M\mapsto S^{-1}M$ sends injective maps to injective maps. So let $\varphi : M\to N$ be injective and let $\frac{m}{s}$ be in the kernel of $S^{-1}\varphi$, which means that $\frac{\varphi(m)}{\varphi(s)} = 0$ in $S^{-1}N$, which means that you can find an $s'\in S$ such that $s'\varphi(m) = 0$, which means $\varphi(s'm) = 0$, and $\varphi$'s injectivity implies that $s'm = 0$, which implies that $\frac{m}{s} = \frac{1}{ss'} s'm = 0$ in $S^{-1} M$. $\square$

Remark. The proposition answers to your question 2.


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.