11
$\begingroup$

Is the group $\mathbb Q^*$ (rationals without $0$ under multiplication) a direct product or a direct sum of nontrivial subgroups?

My thoughts:

Consider subgroups $\langle p\rangle=\{p^k\mid k\in \mathbb Z\}$ generated by a positive prime $p$ and $\langle -1\rangle=\{-1,1\}$.

They are normal (because $\mathbb Q^*$ is abelian), intersects in $\{1\}$ and any $q\in \mathbb Q^*$ is uniquely written as quotient of primes' powers (finitely many).

So, I think $\mathbb Q^*\cong \langle -1\rangle\times \bigoplus_p\langle p\rangle\,$ where $\bigoplus$ is the direct sum.

And simply we can write $\mathbb Q^*\cong \Bbb Z_2\times \bigoplus_{i=1}^\infty \Bbb Z$.

Am I right?

$\endgroup$
1
  • 4
    $\begingroup$ Of course, each $\langle p\rangle\cong\mathbb Z$, so we might write this isomorphism using familiar groups: $$\mathbb Q^*\cong\mathbb Z_2\times\bigoplus_{i\in\mathbb N}\mathbb Z$$ $\endgroup$ – Karl Kroningfeld May 10 '13 at 13:56
14
$\begingroup$

Yes, you're right.

Your statement can be generalized to the multiplicative group $K^*$ of the fraction field $K$ of a unique factorization domain $R$. Can you see how?

In fact, if I'm not mistaken it follows from this that for any number field $K$, the group $K^*$ is the product of a finite cyclic group (the group of roots of unity in $K$) with a free abelian group of countable rank, so of the form

$K^* \cong \newcommand{\Z}{\mathbb{Z}}$ $\Z/n\Z \oplus \bigoplus_{i=1}^{\infty} \Z.$

Here it is not enough to take the most obvious choice of $R$, namely the full ring of integers in $K$, because this might not be a UFD. But one can always choose an $S$-integer ring (obtained from $R$ by inverting finitely many prime ideals) with this property and then apply Dirichlet's S-Unit Theorem.

$\endgroup$
2
  • $\begingroup$ Would you explain why can we always choose an $S$-integer ring of number field $K$ to be UFD ? $\endgroup$ – user682141 Apr 25 '20 at 16:13
  • $\begingroup$ Since the ring of integers $\mathbb{Z}_K$ has finite ideal class group (a basic result in algebraic number theory), it suffices to take $S$ to be the set of prime ideals divididing all elements of a set of generators for the class group. See e.g. p. 61 of math.uga.edu/~pete/8430Elliptic_Curves.pdf for more details. $\endgroup$ – Pete L. Clark Apr 29 '20 at 18:45

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.