Let $M$ be a finitely generated $R$-module, then for some multiplicative subset $D \subseteq R$ is $D^{-1}M$ finitely generated as a $D^{-1}R$-module?

Let $m_1, \dots, m_n$ be a set of generators of $M$, with some possible relations. Then every element $m \in M$ can be written as

$$ m=a_1m_1 + \dots + a_n m_n$$

Passing to the localization, each element $\frac{m}{d} \in D^{-1}M$ can then be written as $$ \frac{m}{d}= \frac{a_1m_1 + \dots + a_n m_n}{d} = \frac{a_1}{d}\frac{m_1}{1} + \dots + \frac{a_2}{d} \frac{m_n}{1}$$

This make me believe that $\pi: M \to D^{-1}M$ which maps $m \to \frac{m}{1}$ would map generators to generators. This would give that $M$ as an $R$-module is isomorphic to $D^{-1}M$ as a $D^{-1}R$-module.

What is wrong with my thinking?

  • $\begingroup$ Someone should tag this question with #algebraic-geometry because of its relation to finite morphisms of (affine) schemes. $\endgroup$ – Tanner Strunk Mar 30 '17 at 16:22

It's correct, except the last assertion for two reasons:

$1)$ What is an isomorphism between modules over different rings?

$2)$ Anyway you did not prove the map $M\to D^{-1}M$ is surjective (in general it is neither surjective not injective).

Example: suppose $R$ is an integral domain, $D=R-\{0 \}$, and $D^{-1}R=K$ its quotient field. If $M$ is a non-zero torsion module, then $D^{-1}M=0$, so $M$ and $D^{-1}M$ are certainly not isomorphic in whatever sense.


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.