1
$\begingroup$

I am looking for an example of a non-zero commutative ring $R $ with identity such that $R $ has no idempotent elements, and there exists a non-nilpotent element $s\in R $ such that for each $n\in\mathbb{ N }$ there exists $r\in R$ with $r (r-1)s=0$, $rs^n\not=0 $ and $(r-1)s^n\not=0$.

$\endgroup$
1
  • 1
    $\begingroup$ Please provide more context. Why are you interested in such an example? And what have you tried so far? Recently there was a similar question here, was this also asked by you under a different account? $\endgroup$ Mar 3, 2020 at 16:46

1 Answer 1

1
$\begingroup$

For this type of problem, I find that free algebras modulo the relations you want often suffice. That is the case here. Take $R=\mathbb{Z}\langle r,s : sr=rs,r^2s=rs\rangle$. We check that $$ r^2(sr)=r^3s=r^2s=rs \text{ and } (r^2s)r=rsr=r^2s=rs $$ as well as $$ s(r^2s)=srs=rs^2 \text{ and } (sr)rs=rsrs=r^2s^2=rs^2. $$ So the two overlaps between the two relations both resolve to the same expression, and so Bergman's Diamond Lemma tells us that we can write elements of $R$ in a unique reduced form as $$ t=f(r) + sg(s) + rsh(s) $$ where $f(x),g(x),h(x)\in \mathbb{Z}[x]$ are polynomials. We achieve this normal form by repeatedly applying the two rules $sr\mapsto rs$ and $r^2s\mapsto rs$ to simplify.

Let us now show that there are no nontrivial idempotents. If we were to have $t^2=t$, then $f(r)^2=f(r)$ (work modulo the ideal generated by $s$), and so $f(r)\in \{0,1\}$ (since those are the only idempotents in $\mathbb{Z}[r]$). After replacing $t$ by $1-t$ as necessary, we may assume $f(r)=0$. Now, $t^2=s^2(g(s)+rh(s))^2$ has degree (in $s$) larger than $t$, unless $g(s)=h(s)=0$. Thus the only idempotents are $0,1$.

It is also clear that $rs^n\neq 0$ and $rs^n\neq s^n$, from the normal form, as desired.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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