6
$\begingroup$

Let $A$ be an abelian group. Are there an algebra $\mathfrak{X}(A)$ s.t. the multiplication group is isomorphic to A ? i.e. $$ \mathfrak{X}(A)^{\times} \simeq A. $$

For example, for $A=\mathbb{Z}/4\mathbb{Z}$, $\mathfrak{X}(A)=\mathbb{F}_{5}$. I want to know the sufficient conditions of $A$ for existence of $\mathfrak{X}(A)$. Is it well-known ?

$\endgroup$
  • 2
    $\begingroup$ Did you want to restrict to finite abelian groups? If not, then I suspect that this questions gets a lot more complicated. $\endgroup$ – Joe Silverman Oct 8 at 11:54
  • 1
    $\begingroup$ If anything I am interested in the case of infinity abelian groups. $\endgroup$ – M masa Oct 9 at 9:03
14
$\begingroup$

I am going to assume that by "algebra" you simply mean a ring.

The answer is "no", in general. For example $\mathbb{Z}/5\mathbb{Z}$ is not the unit group of a ring. Indeed, suppose it was the unit group of a ring $R$, let $u\in R$ denote a generator of the unit group. Now $-1$ is always a unit, and has order dividing $2$. But it is supposed to live in a cyclic group of order $5$, which forces its order to be $1$, i.e. $1=-1$ in $R$, so that $R$ has characteristic $2$. But then $u$ generates a subring of $R$ that is a quotient of $\mathbb{F}_2[\mathbb{Z}/5\mathbb{Z}]\cong \mathbb{F}_2[x]/(x^5-1)\cong \mathbb{F}_2\times \mathbb{F}_{16}$. Since that quotient has at least $5$ distinct elements, it must be at least all of $\mathbb{F}_{16}$, so that $R^\times$ contains $\mathbb{F}_{16}^\times\cong \mathbb{Z}/15\mathbb{Z}$ — contradiction.

In general, this argument shows that a group of prime order $p$ is isomorphic to the unit group of a ring if and only if $p=2$ or of the form $2^n-1$ for some $n$.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Why does $-u \in R^\times$ force $1 = -1$? $\endgroup$ – LSpice Oct 8 at 12:48
  • 3
    $\begingroup$ @LSpice: we are assuming that $R^\times=\{1,u,\ldots,u^4\}$, so we must have $-u=u^i$ for some $i$. Since $u$ is a unit, this means that $-1=u^{i-1}$ for some $i$. But $-1$ has order dividing $2$, and we are in a cyclic group of order $5$, so $-1$ must have order $1$, i.e. be equal to $1$. I guess, I could have bypassed the whole $-u$ business. Let me edit. $\endgroup$ – Alex B. Oct 8 at 13:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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