0

Here is the question I want to answer:

Let $R = \mathbb Z[\sqrt{-5}]$ and $ I = (2, 1 + \sqrt{-5})$ be an ideal of $R$.

Show that $I$ is not a free $R-$module.

But here in this question How to show that an $\mathbb{Z}[\sqrt{-5}]$-module is not free? in its answer the author said that: to prove that it is free of rank 1, we should try to prove that no element of $R$ divides both $2$ and $1 + \sqrt{-5}.$ could anyone explain to me why this will prove that it is not free of rank $1$?

Here is the answer I am speaking about:

It is generated by two elements so if it were free, it would be free of rank $1$ or $2$.

If it were free of rank $1$, it would be generated by a single element: you can then try to prove that no element of $R$ divides both $2$ and $1+\sqrt{-5}$.

If it were free of rank $2$, then $2$ and $1+\sqrt{-5}$ would have to be free generators (can you see why ?). You can then try to prove that this is not the case, that is, that $2$ and $1+\sqrt{-5}$ are not linearly independent.

EDIT:

I know that we are using the contrapositive of this statement for the proof:

$I$ is free $R$ module $\implies$ I is a principal ideal.

And I know that we are trying to show that $I$ is not principal.

  • 1
    "could anyone explain to me why this will prove that it is free of rank 1"? No, the fact that no element of $R$ divides both $1+\sqrt{-5}$ and $2$ shows it is not free of rank 1. – user10354138 Feb 09 '21 at 14:36
  • @user10354138 I am sorry I will edit my question .... I meant to say not –  Feb 09 '21 at 14:44
  • My question here is different from the suggested duplicate. –  Feb 09 '21 at 16:40

1 Answers1

1

If $I$ were free of rank $1$ there would be an element $b \in I$ such that the $R$-linear map $\varphi: R \to I, x \mapsto xb$ would be an isomorphism. This would mean that there exist elements $r_1,r_2 \in R$ s.t. $r_1b=2$ and $r_2b=1+\sqrt{-5}$. If you are able to show that no $b$ in $R$ divides both elements this is a contradiction.

Edit. Let $b=\alpha+\beta \sqrt{-5}$, $\alpha, \beta \in \mathbb{Z}$. It holds that $N(2)=4, N(b)=\alpha^2+5\beta^2$. Since the norm is multiplicative $N(b)$ has to divide $N(2)=4$ in $\mathbb{Z}$ and thus $N(b) \in \{1,2,4\}$. Note that in all of these cases $\beta$ has to be zero and that the norm can never be two. Assume that the norm is one, then $\alpha=1$ or $\alpha=-1$ and $\beta=0$. On the other hand $N(b)$ has to divide (in $\mathbb{Z}$) $N(1+\sqrt{-5})=6$. This means that $b=1$ or $b=-1$ are in fact the only options. This however implies that $I=R$ which is a contradiction.

adh.
  • 198
  • 1
  • 4
  • Can I use the norm function of $\mathbb Z[\sqrt{-5}]$ to prove this also ? if so, how? is it easier? –  Feb 09 '21 at 15:12
  • why we are sure from the existence of this isomorphism ... is it because $I$ will generate the whole module in that case? –  Feb 09 '21 at 15:14
  • @hard I didn't calculate it yet so I can't say it for sure but I think one can use the norm here. As for your second question, it's one way to define free module of rank 1. What is the definition you are familiar with? – adh. Feb 09 '21 at 15:25
  • I = Rb this is our definition –  Feb 09 '21 at 15:28
  • 1
    @hard I edited my post which hopefully helps. Are you sure that this is your definition without further requirements like $R$ being an integral domain? I struggle to believe that is correct, but maybe I am just wrong. – adh. Feb 09 '21 at 16:20
  • Yeah my definition does not require $R$ to be an integral domain unfortunately. May be I am also wrong. –  Feb 09 '21 at 16:37