Give a prime number $q$, let be $\omega = \sqrt{q}$ is $q \equiv 2,3 ( \: \mod \: 4)$ and $\omega = \dfrac{1+\sqrt{q}}{2}$ if $q \equiv 1 ( \: \mod \: 4)$. Let $R= \mathbb{Z}[\omega]$. (I think that's the ring of integers of the quadratic field extension $\mathbb{Q}[\sqrt{p}]$).

If $\alpha \in R$ is such that $N(\alpha)=p$ with $p$ a prime integer, where $N$ is the complex norm restricted to $R$, then $\alpha$ is a prime.

I know how to prove that's irreducible, which is easy. If $\alpha = \beta \gamma$ then $N(\alpha) = N( \beta) N(\gamma)$ and as $N(\alpha)=p$ is a prime, then $N(\gamma)=1$ or $N(\beta)=1$. So $\gamma$ or $\beta$ is a unit.

But now I don't know how to proced. I know that $\mathbb{Z} \cap M = p \mathbb{Z}$ if $\alpha \in I$ and $I$ is an ideal, because $p= \alpha \overline{\alpha} \in I$ because $I$ is an ideal, and $p \mathbb{Z}$. I guess that the only ideals $I$ of $R$ such that $I \cap \mathbb{Z} = p \mathbb{Z}$ are $(\alpha)$ and $(-\alpha)$.

But I don't know how to prove it. Well, I am not even 100% sure this is true.

  • 3
    $\begingroup$ math.stackexchange.com/questions/3314752/… $\endgroup$ – Viktor Vaughn Aug 6 '19 at 23:24
  • $\begingroup$ Oh, I don't know anything about ring norms, but I will read that one. At least I should be able to understand the isomorphism. $\endgroup$ – P.Luis Aug 6 '19 at 23:46
  • $\begingroup$ How is it that two users came to ask the same question within a few hours of each other? Is there something you're not telling us, P.L? $\endgroup$ – Gerry Myerson Aug 7 '19 at 0:24
  • 1
    $\begingroup$ It was a homework question due to this monday. As the deadline has already ended and I already submitted my homework without this, I thought it would be fine to ask it as I spend 3 hours last night reading bibliography and I wake up thinking about this. I asked the TA and the professor and neither was able to give me an answer. I tried lots of approachs. For example, I think that it is enough to prove that the circle or radius sqrtp has p elements of R, at least for imaginary extensions. I don't know who the other user is, but I was surprised when I read that question made 22 hours ago. $\endgroup$ – P.Luis Aug 7 '19 at 0:53
  • $\begingroup$ Well, thats all. I haven't read the rules in a long time and I think that making homework questions is forbidden. But after toying with a diophantic equation solver trying to come out with a solution I got really curious of this. Moreover, most people I have talked about this problem with think it is false. Sorry if this was against the rules. I should read them to refresh it, this is my honest answer. $\endgroup$ – P.Luis Aug 7 '19 at 0:57

Let $f : Z[w] \to Z^2$ be an isomorphism of abelian group. The multiplication by $\alpha$ in $Z[w]$ becomes the multiplication by a matrix $A\in M_2(Z)$ in $Z^2$.

$Q[\sqrt{q}]=Q+\sqrt{q}Q$ is a $Q$-vector space isomorphic to $Q^2$ and the multiplication by $\alpha=a+ b \sqrt{q}$ in $Q+\sqrt{q}Q$ becomes the multiplication by the matrix $B = \pmatrix{a & b q \\ b & a}$ in $Q^2$.

The field norm of $\alpha$ is defined as $N_{K/Q}(\alpha)=\det(B) = a^2-b^2 q$. We also have $\det(B) = \det(A)$ since $B = M A M^{-1}$.

Finally $$\# Z[w] / \alpha Z[w] = \# f^{-1}(Z[w])/f^{-1}( \alpha Z[w]) = \# Z^2/ A Z^2 = |\det(A)| = |N_{K/Q}(\alpha)|$$

Thus if $|N_{K/Q}(\alpha)|=p$ is prime then $Z[w]/(\alpha) \cong Z/ p Z$ is the field with $p$ elements and $(\alpha)$ is a maximal ideal.

The exact same argument holds in larger number fields.


If $N(\alpha)=p$ then $(p) \subset (\alpha)$ as $\alpha \overline{\alpha} = p \subset ( p )$. Also, every element on $(p)$ can be written as a product $\beta p$, and $ N(\beta p)=N(\beta)N(p) = N( \beta ) p^{2}$, so $\alpha \notin (p)$ and the inclusion $(p) \subset (\alpha)$ is proper.

I will use the third isomorphism theorem for groups.

\begin{equation*} \dfrac{\mathbb{Z}[w]}{(\alpha)} \cong \dfrac{\dfrac{\mathbb{Z}[w]}{(p)}}{ \dfrac{(\alpha)}{(p)} } \end{equation*} As abelian groups: $$ \dfrac{\mathbb{Z}[w]}{(p)} = \dfrac{ \mathbb{Z} \oplus \mathbb{Z} \omega } {p \mathbb{Z} \oplus p \mathbb{Z} \omega } \cong \dfrac{\mathbb{Z}}{ p \mathbb{Z} } \oplus \dfrac{\mathbb{Z} \omega}{ p \mathbb{Z} \omega } \cong \dfrac{ \mathbb{Z} }{ p \mathbb{Z} } \oplus \dfrac{ \mathbb{Z} }{ p \mathbb{Z} } $$ So $\dfrac{\mathbb{Z}[w]}{(p)}$ is an abelian group of order $p^{2}$. Also $(\alpha) \subset (p)$ is a proper contention, that is $(\alpha) \neq (p)$. So $\dfrac{\alpha}{(p)}$ has order larger than $1$. As

\begin{equation*} 1 < \left| \dfrac{\mathbb{Z}[w]}{(\alpha)} \right| \cong \dfrac{\left| \dfrac{\mathbb{Z}[w]}{(p)} \right| }{ \left|\dfrac{(\alpha)}{(p)} \right|} < \left| \dfrac{\mathbb{Z}[w]}{(p)} \right| = p^{2} \end{equation*} (The $1<$ is because $\dfrac{\mathbb{Z}[w]}{(\alpha)}$ isn't the trivial group)

As $\left| \dfrac{\mathbb{Z}[w]}{(\alpha)} \right|$ divides $\left| \dfrac{\mathbb{Z}[w]}{(p)} \right|=p^{2}$, under these restriction it can only be $\left| \dfrac{\mathbb{Z}[w]}{(\alpha)} \right| = p$ as the other divisors of $p^{2}$ are $1$ and $p^{2}$ and we already ruled out those possilibities. Thus the abelian group structure of $\dfrac{\mathbb{Z}[w]}{(\alpha)}$ is $\mathbb{Z} / p \mathbb{Z}$. The only ring with unity of $p$ elements is $\mathbb{F}_{p}$, and as out quotient has unity, it must be isomorphic to $\mathbb{F}_{p}$. Thus our ring structure is $\dfrac{\mathbb{Z}[w]}{(\alpha)} \cong \mathbb{F}_{p}$, and as it's a field, $(\alpha)$ is maximal, thus a prime ideal.

Therefore, $p$ is prime.

PD: My professor meant irreducibles when he wrote the exercise. But this one was a very nice exercise and it motivated me to study a lot, so I liked it.

  • $\begingroup$ It's a easy result to deduce? Where can I read about that? It sounds very useful but I didn't know that. I should have taken the linear algebra approach like Galois groups does and use matrix representations of my rings. I still don't see why the determinant is that one. Sorry, this is my first course with ring theory and it barely started a 6 weeks ago. All I know about rings before I had learned it reading Dummit and Foote on my free time. It's a little embarrasing to me, but I hadn't have ring theory on my undergrad. $\endgroup$ – P.Luis Aug 8 '19 at 0:13

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.