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

Suppose that $R$ is a commutative ring (with unity) of characteristic $0$, and that the set $Z$ of zero divisors in $R$ forms an ideal. Does it follow that the characteristic of $R/Z$ is $0$?

Equivalently (with less verbiage), suppose that the ring homomorphism $\iota:\mathbb{Z} \to R$ injects. Does it follows that $\iota(m)$ is not a zero divisor, for all $m$?

This proposition holds when all zero divisors are nilpotent:

Example: Suppose that $Z = \mathrm{nil}(R)$, the nilradical of $R$. Then $Z$ is an ideal, and if $\iota(m) \in Z$, then $\iota(m^n)=0$. Here $\mathrm{char}(R)=0$ forces $m^n=0$, so $m=0$ and $\mathrm{char}(R/Z)=0$.

If this proposition does not hold under my given hypotheses, I would appreciate a counter-example or a new hypothesis on $Z$; something more restrictive than $Z$ forming an ideal and less restrictive than $Z =\mathrm{nil}(R)$.

share|cite|improve this question
Is R commutative – Amr Nov 9 '12 at 9:27
@Amr Yes; and to be clear I've also written that $R$ should contain a unit element. – A Walker Nov 9 '12 at 9:31
Why is the set of divisors of R a subring of R – Amr Nov 9 '12 at 9:44
up vote 1 down vote accepted

No. Consider $R = \mathbb{Z}[t]/(m t)$ for any integer $m > 1$ together with the natural inclusion $\mathbb{Z} \hookrightarrow R$.

share|cite|improve this answer
Unfortuantely, it seems to me that the image of $m+t$ is not a zero divisor of $R$, so that $R$ does not satisfy the hypothesis that its set of zero divisors forms an ideal (even if you throw in $0$). It looks like this can be repaired by replacing $R$ by its quotient $R = \mathbb{Z}[t]/(mt,t^2)$; now the image of $(km+lt)t$ for $k,l\in\Bbb Z$ is zero, and the zero divisors are the ideal generated by the images of $m$ and of $t$. – Marc van Leeuwen Nov 9 '12 at 11:59
For $R=\mathbb{Z}[t]/(4t,t^2)$, the image of $2$ is a zero divisor, as $2 \cdot 2t =0$. So here shouldn't $Z=(2,t)$? In general, one should expect $Z$ to have generators with divisors of your $m$. Of course, this too can create a problem: Consider $R=\mathbb{Z}[t]/(6t,t^2)$. Both $2$ and $3$ are zero divisors, but $1$ is not. It seems to me that your example should work exactly when $m$ is a prime power. – A Walker Nov 9 '12 at 17:23
I see your point, Marc. Thanks for mentioning it. A Walker is also right that the ideal generated by $m$ and $t$ does not contain all zero divisors in general. I am very tired now, so I cannot tell if one can fix this example to make it work. – Lennart Nov 9 '12 at 21:25

The set of divisors of R does not contain 0, thus it is not a subring of R

share|cite|improve this answer
It looks to me that you can improve your answer. What about if $Z:= \{\text{zero divisors}\}\cup \{0\}$? What about the "equivalent" statement suggested by A Walker? Moreover you forgot a "zero" in your answer. – Giovanni De Gaetano Nov 9 '12 at 9:52
Even if this is the case, one can it guarantee that {zero divisors} U {0} is closed under addition. – Amr Nov 9 '12 at 9:55
You cannot guarantee it but you can suppose it. As A Walker specifically stated. – Giovanni De Gaetano Nov 9 '12 at 9:55
I didnt undertand that he imposed this condition when I read his question. I will think about it again. – Amr Nov 9 '12 at 9:56
As a side comment, some authors allow 0 as a zero divisor. – rschwieb Nov 9 '12 at 12:03

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.