7
$\begingroup$

Let $R$ be a (right noetherian) ring. Is there always a right $R$-module which is both flat and injective? If $R$ is an integral domain, then the answer is indeed yes, as the quotient field is such.

Sorry if this is too silly, but so far I'm stuck.

$\endgroup$
1
  • 2
    $\begingroup$ The zero module is both flat and injective. $\endgroup$ Commented Jun 18, 2014 at 18:21

1 Answer 1

11
$\begingroup$

Here's a counterexample. Let $k$ be a field and $R=k[x,y]/(x^2,xy,y^2)$; this is a local ring with maximal ideal $m=(x,y)$. Suppose $F$ is a flat $R$-module. Then there is a short exact sequence $$0\to m\otimes F\to F\to F/m\to 0$$

But $m\cong R/m\oplus R/m$ (with $x$ and $y$ generating the summands), so this sequence can be rewritten as $$0\to F/m\oplus F/m\stackrel{(x,y)}\to F\to F/m\to 0$$

Choose a basis for $F/m$ over $R/m=k$ and lift it to $F$. By exactness of the sequence above, this lift will be a basis for $F$ over $R$.

Thus any flat $R$-module is free. Since $R$ is not self-injective, no flat $R$-module can be injective. More generally, a similar argument shows that flat modules over any Artinian local ring are free (replace this short exact sequence by the filtration by powers of the maximal ideal), so any such ring that is not self-injective is a counterexample.

$\endgroup$
1
  • $\begingroup$ "By exactness of the sequence above, this lift will be a basis for $F$ over $R$." Can you elaborate on this? $\endgroup$
    – Smiley1000
    Commented Sep 29 at 12:21

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .