Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I wanted to come up with a few examples of rings of fractions $S^{-1}R$. Can you tell me if these are correct:

1.Let $R = \mathbb Z$, $S = (2 \mathbb Z \setminus \{0\}) \cup \{1\}$. Then every $[x] = \frac{r}{s} \in S^{-1}R$ consists of the elements: $[x] = \{ 2x, \frac12 x\}$. The ring homomorphism $f: R \to S^{-1}R$ is injective since if $f(r) = [\frac{r}{1}] = [\frac{r^\prime}{1}] = f(r^\prime)$ we have $2k r = 2k r^\prime$ for $2k \in \mathbb Z$ and since $\mathbb Z$ is an integral domain, $r=r^{\prime}$.

2.Now let $R=\mathbb Z / 12 \mathbb Z$ and $S = \{1,2,4,6,8,10\} = (2 R \setminus \{0\}) \cup \{1\}$. Then $f: R \to S^{-1}R$ is not injective since $f(6) = f(3)$.

Are there more interesting examples where $f: R \to S^{-1}R$ is not injective?

share|cite|improve this question
Note: $R$ is a commutative unital ring and $S$ is a multiplicative subset of $R$. – Rudy the Reindeer Jun 12 '12 at 11:51
I believe that you have non-injectivity when the ring in question has zero divisors. – user38268 Jun 12 '12 at 11:55
@BenjaminLim Thank you, that's right. In fact, we have $$ f(r) = \frac{r}{1} = 0 \iff \exists s \in S: sr = 0$$ – Rudy the Reindeer Jun 12 '12 at 19:09
@BenjaminLim: Not necessarily; e.g., if $S$ consists only of elements that are not zero divisors. Since $\frac{as}{s} = \frac{0}{s}$ holds if and only if there exists $s'\in S$ such that $s'(sas - 0s)=0$, if neither $s$ nor $s'$ are zero divisors, then $sas$ cannot be zero, and therefore $s'(sas)$ cannot be zero. So even if $R$ has zero divisors, you still may get injectivity. E.g., taking $S$ to be the set of all nonzero divisors gives you the total ring of fractions of $R$, and every commutative ring embeds into its total ring of fractions. – Arturo Magidin Jun 12 '12 at 19:09
@ArturoMagidin Thank you for this comment! – Rudy the Reindeer Jun 13 '12 at 6:49
up vote 2 down vote accepted

It's difficult to find a place to begin, because there are so many examples. For $R=\mathbb{Z}$,

  1. there is of course the choice $S=R\setminus\{0\}$, producing $\mathbb{Q}$.

  2. You could also just choose a prime ideal, say $P=2\mathbb{Z}$, and use $S=R\setminus P$. This yields the subring of $\mathbb{Q}$ consisting of all fractions of odd denominator.

  3. You could fix a nonzero element, say $a\in R$ and use $S=\{a^i\mid i\in\mathbb{Z}^+\}$. This produces the subring of $\mathbb{Q}$ whose denominators are a power of $a$.

  4. To break from $\mathbb{Z}$ for a second, I'd like to apply the last point to $R=F[x]$ and choose $a=x$. The ring of fractions based on the powers of $x$ then yield $F[x,x^{-1}]$, the Laurent polynomials.

In your second example you put $0$ in your multiplicative set... are you aware that will make $S^{-1}R=\{0\}$?

If the map is not injective, and you exclude $0$ from $S$, then there must be a nonzero zero divisor in $S$. If $a\neq b$ in $R$, but $(a,1)=(b,1)$ in $S^{-1}R$, then that says there exists $s\in S$ such that $as=bs$. This means $(a-b)s=0$.

Using that idea, let $as=0$ for some nonzero $a\in R$, nonzero $s\in S$. Let $b$ be anything nonzero in $R$. Note that $0=as=(a+b-b)s=(a+b)s-bs$. Then $(a+b,1)=(b,1)$.

A simple way to produce a nontrivial ring of fractions such that the canonical map is not an injection would be to find a nonzero zero divisor $x$ in a ring, such that $x$ is not nilpotent. Then localize at the nonnegative powers of $x$.

share|cite|improve this answer

I don't follow what you mean here: if by $\,[x]\,$ you mean the equivalence class of $\,x\,$ in the localization then I can't see how $\,2x\,,\,\frac{1}{2}x\,$ are equivalent in $\,(1)$...?

If you're talking of localization of $\,R\,$ wrt $\,S\,$, then an element of $\,S^{-1}R\,$ is of the form $\,\displaystyle{\frac{r}{s}\,\,,\,r\in R\,,\,s\in S}$, and of course $\,S\,$ cannot contain zero and has to be multiplicatively closed...

So in $\,(1)\,$ , the elements we have there are of the form $$\,\displaystyle{\frac{m}{r}\,,\,m\in\mathbb{Z}\,,\,r\in 2\mathbb{Z}\cup \{1\}}$$ and $$\frac{m}{r}=\frac{n}{s}\Longleftrightarrow ms=nr\,\,\text{in}\,\,R$$

In your second example $\,S\,$ is not valid because it contains zero

share|cite|improve this answer

For the first of your examples, I don't think that $S^{-1}R$ consists of just those two elements. For example you have $\frac{13}{4}$ in the localisation that is not of the form $2x$ or $\frac{x}{2}$. On second thoughts I don't get what you mean by $[x]$ is just those two elements. Do you mean to talk about the equivalence class of a fraction? In that case there are infinitely many elements in an equivalence class. For example

$$\frac{13}{4} = \frac{26}{8} = \frac{39}{12} = \ldots $$

As can be seen $\frac{13}{4}$ is not even of the form you claimed that an element is in an equivalence class. Other important examples of localisation:

  1. Let $P$ be a prime ideal in a ring $R$. Then $S = R - P$ is a multiplicative set, and $S^{-1}R$ is a local ring. We call this "localisation at the prime ideal $P$". For example if $P$ is all the multiples of $5$ in $\Bbb{Z}$, then

$$R_P =\{q \in \Bbb{Q} | q = \frac{a}{b} \hspace{1mm} \text{with} \hspace{1mm} a,b\in \Bbb{Z}\}$$

  1. Let $f$ be a non-zero non-nilpotent element. Then consider $S = \{1,f,f^2,\ldots \}$. When localising by this multiplicative subset, we sometimes write $R_f$. This example crops up a lot in algebraic geometry I believe.

  2. More advanced example: Let $R$ be a $\Bbb{Z}$ - graded ring. Let $f$ be a non-zero homogeneous element of degree one. Notice that $R_f$ is naturally a graded ring. Then $R_f \cong (R_f)_0[x,x^{-1}]$ where $x$ is an indeterminate and $(R_f)_0$ we denote by the degree zero component of $R_f$. More generally, let $P$ be a homogeneous prime ideal in $R$ and $U$ the multiplicative subset of all homogeneous elements not in $P$. Then it is surprising that for a more general $U$ we than before we still have

$$U^{-1}R \cong (U^{-1}R)_0[x,x^{-1}]$$

where $(U^{-1}R)_0$ is the degree zero component of $U^{-1}R$.

  1. Interesting example. Let $R$ be a PID and $K$ its field of fractions. Let $A$ be a subring of $K$ containing $R$. Then one can show that $A$ is actually isomorphic to $S^{-1}R$ where $S$ is the multiplicative set of all elements in $R$ that are units in $A$. You can use this to prove that $A$ must always be a PID!

By the way in your last example you have included zero in the multiplicative subset $S$. We claim that $S^{-1}R$ is the zero ring. Indeed, any

$$\frac{a}{b} = \frac{0}{1}$$

because there exists $0 \in S$ such that $(a\cdot 1- b\cdot 0)\cdot a = 0$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.