3
$\begingroup$

The ring of $p$-adic integer is a complete local ring.

I already know the ring of $p$-adic integer is complete. Problem is to show it's a local ring. I know this answer is already answered here.

In that answer, it uses one useful fact: If $P\subsetneq R$ is a proper ideal such that $x\in R\setminus P\Rightarrow x\in R^\times$, then $P$ is a unique maximal ideal i.e. $R$ is a local ring. Then it says in the case of power series $f\notin (X)\Rightarrow f\in R^\times$ so $(X)$ is the unique maximal ideal. Applying this to may case, $x\notin (p\Bbb Z_p)$ then $x\in (\Bbb Z_p)^\times$. This is not clear to me. Could you explain?

Edit: I already know $p\Bbb Z_p$ is a maximal ideal. But don't know why it's a unique maximal ideal.

Edit: I know there is some lecture notes that shows $\Bbb Z_p$ is a local ring using DVR argument but I don't want to do that. Or is that the simplest way?

$\endgroup$
11
  • $\begingroup$ The is a typo in the last sentence: the claim is that if $x \not\in (p\mathbb{Z}_p)$ then $x \in (\mathbb{Z}_p)^\times$. The much smaller ring $(\mathbb{Z}/p\mathbb{Z})$ does not play a role here $\endgroup$
    – Vincent
    Commented Jun 1, 2021 at 9:26
  • $\begingroup$ That said: what needs to be done is that for some abstract element $f$ of $\mathbb{Z}_p$ that is not in $p\mathbb{Z}_p$ you want to construct some element $h$ of $\mathbb{Z}_p$ such that $fh = 1$. This construction can follow all the steps in the other answer. $\endgroup$
    – Vincent
    Commented Jun 1, 2021 at 9:30
  • $\begingroup$ Ah wait... there is one thing. In general we have that $f$ is of the form $a + pg$ for some element $a \in \mathbb{Z}$ and $g \in \mathbb{Z}_p$. Now in the power series case we can easily assume without loss of generality that $a = 1$. Here that is less obvious, it seems we need a 'separate' proof for the special case that $f \in \mathbb{Z}$ first. For that step it seems it helps that $a$ does have in inverse in the smaller ring $(\mathbb{Z}/p\mathbb{Z})$, so it does play some role after all... $\endgroup$
    – Vincent
    Commented Jun 1, 2021 at 9:31
  • 1
    $\begingroup$ Viewing $\mathbb{Z}_p$ as the inverse limit of $\mathbb{Z}/p^n\mathbb{Z}$ it is immediately clear that some sequence of compatible residues is invertible in each component if and only if its first component is non-zero, hence the claim $\endgroup$
    – Mike Daas
    Commented Jun 1, 2021 at 9:41
  • 1
    $\begingroup$ I think @MikeDaas 's argument is the simplest and best argument $\endgroup$ Commented Jun 2, 2021 at 11:44

1 Answer 1

3
$\begingroup$

Let $f$ be an arbitrary element of $\mathbb{Z}_p$. We can expand it as $a_0 + a_1p + a_2p^2 + \ldots$ with the $a_i$ integers (which we can even take in $\{0, \ldots, p-1\}$ but that is irrelevant for the answer).

When $a_0 = 0$ we have $f \in p\mathbb{Z}_p$ and there is nothing to prove.

When $a_0 = 1$ we can copy the steps in the other answer to construct an element $f^{-1}$ (as an infinite series of the above form) satisfying $ff^{-1} = 1$.

What remains is the case that $a_0 \neq 0$ but we do not necessarily know that it equals $1$ either. Now in this case there is a $b \in \mathbb{Z}$ such that $a_0b \equiv 1 \mod p$. ($b$ is of course not unique, just pick one). This uses that $\mathbb{Z}/p\mathbb{Z}$ is a field.

Now it is easy to check that the element $fb \in \mathbb{Z}_p$ is of the form $1 + b_1p + b_2 p^2 + \ldots$. So by the earlier answer we can construct an element $(fb)^{-1}$ such that $(fb)(fb)^{-1} = 1$.

The last step is then observing that $f(b(fb)^{-1}) = 1$ as well, so we can take $f^{-1} = b(fb)^{-1}$.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .