3
$\begingroup$

I want a reference that treated the proof of this proposition :

Proposition : Suppose that $X$ is a finitely generated, torsion $\Lambda-$module. Then there are uniquely determined $\Lambda-$submodules $Z$ and $Y$ of $X$ with the following properties :

  1. $Z$ is finite and $X/Z$ has no nonzero, finite $\Lambda-$submodules.

  2. $Y$ is annihilated by a power of $p$ and $X/Y$ is a free $\mathbb Z_p-$module of finite rank.

Or someone help me to prove.This proposition allows us to define certain Iwasawa's invariants ($\lambda$ and $\mu$)

$\endgroup$
1
  • 2
    $\begingroup$ I think this question is better suited for Math SE but anyway: the canonical historical reference is numdam.org/item?id=SB_1958-1960__5__83_0 Note also that the Wikipedia page on Iwasawa algebras contains what you look for. $\endgroup$
    – Olivier
    Dec 17, 2013 at 18:07

1 Answer 1

3
$\begingroup$

I'm assuming $\Lambda=\mathbf{Z}_p[[\mathbf{Z}_p]]\cong\mathbf{Z}_p[[T]]$ (with a topological generator $\gamma$ of $\Gamma=\mathbf{Z}_p$ going to $1+T$).

Let $Z$ be the $\mathbf{Z}_p$-torsion submodule of the $\Lambda$-submodule $X^\prime=\bigcup_{n\geq 0}X^{\Gamma^{p^n}}$. Since $X^\prime$ is finitely generated over $\Lambda$, $X^\prime=X^{\Gamma^{p^n}}$ for some $n\geq 0$. This means that $X^\prime$ is annihilated by $(1+T)^{p^n}-1$, and hence is a finitely generated module over $\Lambda/((1+T)^{p^n}-1)$, a finitely generated $\mathbf{Z}_p$-module. So $X^\prime$ is finitely generated over $\mathbf{Z}_p$, and thus its $\mathbf{Z}_p$-torsion submodule $Z$ is finite. Conversely, if $Z^\prime$ is any finite $\Lambda$-submodule of $X$, then $\Gamma$ must act discretely on $Z^\prime$, so $Z^\prime$ is fixed element-wise by some open subgroup of $\Gamma$, i.e., $Z^\prime\subseteq X^\prime$. Since $Z^\prime$ is of finite cardinality, in fact $Z^\prime\subseteq Z$. Thus $Z$ is the unique maximal finite $\Lambda$-submodule of $X$.

Now let $Y$ be the $\mathbf{Z}_p$-torsion submodule of $X$, i.e., $Y=\bigcup_{n\geq 1}X[p^n]$. This is a $\Lambda$-submodule, hence finitely generated, and it follows that, for some sufficiently large $n$, $Y=X[p^n]$, so $Y$ has finite exponent. Because $X$ is torsion, the structure theory for finitely generated $\Lambda$-modules shows that $X/Y$ is finitely generated over $\mathbf{Z}_p$, and since it is torsion-free by construction, it is finite free over $\mathbf{Z}_p$. Conversely, if $Y^\prime\subseteq X$ is a $\Lambda$-submodule killed by some power of $p$, then $Y^\prime\subseteq Y$, and if $X/Y^\prime$ is torsion-free over $\mathbf{Z}_p$, we must have $Y\subseteq Y^\prime$, whence the equality. So $Y$ is unique.

$\endgroup$
1
  • $\begingroup$ thank you Keenan , I don't have this article, you can prove the proposition ? if it is possible of course $\endgroup$
    – Med
    Dec 17, 2013 at 19:30

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.