MathOverflow is a question and answer site for professional mathematicians. 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

Can anyone suggest a good thorough reference for $p$-localization and $p$-completion of the integers? I'm an algebraic topologist who's found himself washed up without any intuition.

In particular, here are the two questions I'm trying to answer right now:

Question 1: what is $\mathbb{Q}/\mathbb{Z}_p$ tensored with itself?

I mean $\mathbb{Z}_p$ to be the $p$-local integers. The tensor is over $\mathbb{Z}_p$ or $\mathbb{Z}$ - the answer should be the same.

I know $\mathbb{Q}/\mathbb{Z}_p$ is an injective $\mathbb{Z}_p$-module, and $\mathbb{Z}_p$ is a nice local ring, so maybe something can be said about a flat resolution of $\mathbb{Q}/\mathbb{Z}_p$?

Question 2: What is known about the cokernel of the map $\mathbb{Z}_p \rightarrow \mathbb{\hat{Z}}_p$?

I mean the map that $p$-completes the $p$-local integers. I think the cokernel is a rational vector space, but of finite or infinite dimension? Does it have any nice description?

share|cite|improve this question
Note that number theorists normally notate $\mathbb{Z}_{(p)}$ to be the "$p$-local integers", and leave $\mathbb{Z}_p$ to mean the $p$-adic numbers, that is the completion of $\mathbb{Z}_{(p)}$. – Daniel Loughran Apr 25 '11 at 13:08
Remember that $\mathbb{Q} \otimes_{\mathbb{Z}} \mathbb{Q} \cong \mathbb{Q}$. (At one time I found this incredibly counterintuitive, and at slightly earlier times I thought the answer was a vector space of countably infinite dimension. But I have gotten used to it.) I think that should help you with Question 1. As for Question 2: well, the cokernel is big: continuum cardinality, anyway. To check whether it's a $\mathbb{Q}$-vector space, it's enough to check whether it is a uniquely divisible abelian group. – Pete L. Clark Apr 25 '11 at 14:08
What are $p$-local integers? Those with powers of $p$ in the denominators, or those with denominators coprime to $p$? – darij grinberg Apr 25 '11 at 14:45
@darij I think Luke means the localization of $\mathbb{Z}$ at the prime ideal $(p)$, so no $p$ in the denominator. – Keenan Kidwell Apr 25 '11 at 15:21
up vote 7 down vote accepted

I'm going to switch to a more standard notation, where the localization is $\mathbb{Z}_{(p)}$ and the completion is $\mathbb{Z}_p$.

For your first question, the general rule that concerns us is that the tensor product of a $p$-divisible group with a $p$-torsion group is zero. The reason is that if $a$ is a $p^n$th power and $p^n b = 0$, then $a \otimes b = p^{-n}a \otimes p^n b = 0$. Since $\mathbb{Q}/\mathbb{Z}_{(p)}$ is both $p$-divisible and $p$-torsion, its tensor square is zero.

The cokernel of completion is a $\mathbb{Q}$-vector space, because it admits a $\mathbb{Z}_{(p)}$-action by multiplication, and it is uniquely $p$-divisible (Proof: lift any element to $\mathbb{Z}_p$, subtract the constant term of its $p$-adic expansion, divide by $p$, and check that different lifts yield the same answer). Its dimension over $\mathbb{Q}$ is the cardinality of the continuum, since that is its cardinality as a set.

I think most number theory texts have some discussion of $p$-adic numbers, and Gouvea even wrote a whole book about them.

share|cite|improve this answer

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.