5
$\begingroup$

Let $p$ be a prime and let $\mathbb{Z}_p$ denote the $p$-adic integers. One has a canonical inclusion of rings $$\mathbb{Z}_{(p)}\longrightarrow\mathbb{Z}_p$$ given by identifying the rational integers in $\mathbb{Z}_p$.

What is the quotient group $\mathbb{Z}_p/\mathbb{Z}_{(p)}$?

It was suggested to me, without proof, that the quotient is a $\mathbb{Q}$-vector space (of uncountable rank). I have been trying to verify this in detail.

My approach was the following:

  1. Show that the quotient is $p$-divisible
  2. Show that the quotient is torsion-free

For then the quotient is torsion-free and divisible, and the result follows.

However, I can neither prove 1 nor 2; my knowledge of the $p$-adic integers is a bit weak. How might I prove these statements, or should I try another approach?

$\endgroup$
2
  • $\begingroup$ But $\Bbb Z[1/p]$ is $p$-divisible and torsion-free, but not a $\Bbb Q$-space. Did you mean $q$-divisible for every prime $q$? $\endgroup$
    – Lubin
    Oct 22 '15 at 4:02
  • $\begingroup$ Well the $p$-adics are already $q$-divisible for every prime $q\not=p$, and this property is inherited by quotients. $\endgroup$
    – H.Durham
    Oct 22 '15 at 7:20
3
$\begingroup$

The approach works, and ends up being elementary.

Firstly: for an abelian group $A$, one has $A\otimes \mathbb{Z}/p\mathbb{Z}=A/pA$. Thus $p$-divisibility is equivalent to the tensor product with $\mathbb{Z}/p\mathbb{Z}$ vanishing.

In our case we have $$\frac{\mathbb{Z}_p}{\mathbb{Z}_{(p)}}\otimes \frac{\mathbb{Z}}{p\mathbb{Z}}\cong \frac{\mathbb{Z}_p}{\mathbb{Z}_{(p)}+p\mathbb{Z}_p},$$ but this latter quotient is $0$, since any $\gamma\in\mathbb{Z}_p$ may be written as

$$\gamma=a_0+p\sum_{i\geq1}a_ip^{i-1}\in \mathbb{Z}_{(p)}+p\mathbb{Z}_p.$$

To see why the quotient is torsion-free, suppose that $\gamma\in\mathbb{Z}_p$ has $n\gamma\in\mathbb{Z}_{(p)}$. If $p\!\!\not|\;n$ then $\frac1n$ exists and $\gamma\in\mathbb{Z}_{(p)}$. Thus we may reduce to the case that $n=p$. Write $$p\gamma=\frac km$$ for $p\!\!\not|\;m$. If now $p$ divides $k$, we are done. If not, we have that $p\gamma$ is a unit, which is impossible.

$\endgroup$

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.