3

Show that $R = \bigcap_mR_m$ whenever $R$ is an integral domain, where the intersection is indexed by all maximal ideals of $R$.

$R \subset \bigcap_mR_m$ is clear since $R \subset R_m$ for all $m$ maximal because $R$ is an integral domain which tells us that $R \twoheadrightarrow R_m $ is injective.

Now I am not sure how the other directions works. My idea:

Choose $x \in(\bigcap_mR_m)\setminus R$. Then for all $m$ maximal x must be a unit in $R_m$ since otherwise $x \in m$ which is a contradiction to $x \notin R$. But for $x\notin R$ to be a unit in $R_m $ for all $m$ maximal, $x^{-1} \in R \setminus m$ for all $m$ maximal. But that means $x^{-1}$ is a unit in R, because otherwise there exist a maximal ideal containing m. We conclude that $x$ must already be in $R$ which is a contradiction to the coice of x. Therefore $(\bigcap_mR_m)\setminus R = \emptyset$. Thus the equality $R = \bigcap_mR_m$ holds, whenever $R$ is an integral domain.

Is this correct?

1 Answers1

1

The first claim: "$x$ must be a unit in $R_m$ since otherwise $x\in m$..." it's wrong. What means that $x$ is not a unit in $R_m$? Obviously $x\in mR_m$ which is equivalent to $\exists s\in R-m$ such that $sx\in m$, and this is not exactly what you get.

For the proof you start with $x\in R_m$ for all $m$, so there is $s_m\in R-m$ such that $s_mx\in R$, and then notice that the ideal generated by all the $s_m$ equals $R$.

user26857
  • 53,190