Problem is:

Let $R$ be a commutative ring and let $P$ be a prime ideal.

(a) Prove that the set of non-units in $R_{P}$ is the ideal $P_{P}$.

(b) Deduce that $R_{P}$ has a unique maximal ideal.

I tried (a) as:

Let $u$ be a non-unit in $R_{p}$ and $p$ be any elements in $R_{p}.$ Then, since $u$ is not unit, $up\neq1$ and $pu\neq1.$ Suppose that $up$ or $pu$ is a unit in $R_{p}.$ Then there exists $q$ such that, say, $upq=1.$ Then $u(pq)=1.$ So, $u$ is a unit in $R_{p}.$ This is contradiction. So, the set of non-units in $R_{p}$ is an ideal in $R_{P}.$

So, I have to show that the set of non-units is equal to $P_P$ and (b)

But I am stuck at this point.


3 Answers 3


The ideal $P_P$ is $$ P_P=\left\{\frac{a}{s}: a\in P, s\in R\setminus P\right\} $$ Let's prove that no element of $P_P$ is a unit. Suppose $a\in P$, $s\in S=R\setminus P$, $x\in R$ and $t\in S$ be such that $$ \frac{a}{s}\frac{x}{t}=\frac{1}{1} $$ By definition there exists $u\in S$ with $u(ax-st)=0$ or $uax=ust$. This is a contradiction because $uax\in P$, but $ust\in S$.

On the other hand, every element of $R_P$ not in $P_P$ is a unit. Indeed, if $a/s\notin P_P$, we have $a\notin P$, so $(a/s)^{-1}=s/a$.

Therefore the set of nonunits is an ideal and the ring is local, by general results. The maximal ideal is the set of nonunits, which we proved to be $P_P$.

  • $\begingroup$ Which is the same argument in the picture I gave one hour ago :) $\endgroup$
    – user 1
    Commented Jun 13, 2016 at 14:06
  • $\begingroup$ @user1 Basically so; but I redid it without looking in a book. By the way, I'm unsure adding a picture of a book is not a copyright violation. Of course, the general argument is folklore and not subject to copyright. $\endgroup$
    – egreg
    Commented Jun 13, 2016 at 14:16
  • $\begingroup$ it is not copyright violation. it is from google books. $\endgroup$
    – user 1
    Commented Jun 13, 2016 at 14:17
  • $\begingroup$ @user1 Do you think that all books on that site are free? You're wrong. $\endgroup$
    – egreg
    Commented Jun 13, 2016 at 14:34
  • $\begingroup$ not the book, but parts of it. BTW do you think an snapshot of (half of) a page is copyright violation? $\endgroup$
    – user 1
    Commented Jun 13, 2016 at 14:36

Let $x\in R$, denote by $[x]$ the class of $x$ in $R_P$, let $[M]=\{[p],p\in P\}$,

$M$ is an ideal, if $[x],[y]\in \in M, [z]\in R_P$ $[x]+[y]=[x+y]\in M$, $[zx]\in M$ since $x+y, zx\in P$

$M$ is distinct of $R_P$, suppose $1\in M$, this implies $1=p\in P$ impossible since $P$ is an ideal.

$M$ is maximal. Let $N$ an ideal which contains strictly $M$, there exists $[s]\in N$, $[s]$ is not in $M$ this implies $s\in R-P$ thus $[s]$ is invertible and $1\in N$.


Part (a) is Lemma 5.20 of the Sharp's book, "Steps in Commutative Algebra" (page 89):

enter image description here

It is easy to see that $I$ is an ideal of $R_P$. Sharp proves that $R_p\setminus I$ is the set of units. One direction is easy: for any $a/s\notin I,$ one has $a\in S$, so $\dfrac{a}{s}\dfrac{s}{a}=\dfrac{1}{1}.$
On the other hand, if $b/t$ is a unit, say $\dfrac{b}{t}\dfrac{c}{v}=\dfrac{1}{1},$ then there exists $w\in S$ such that $wbc=wtv\in R\setminus P.$ Hence $b\notin P,$ which means $b/t\notin I.$

For part (b), note that a ring is local if and only if the set of non-units of the ring is an ideal (this is lemma 3.13 of the book).


You must log in to answer this question.

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