3
$\begingroup$

(All rings are commutative)

Let $A$ be a noetherian ring. Let $B$ be a noetherian $A$-algebra (not nessecerily f.g!)

Suppose $M$ and $N$ are finitely generated projective $B$-modules (for my application I can assume that $M=N$ are of rank $1$).

Consider the $B\otimes_A B$-module $M\otimes_A N$. Is it projective?

Note that it is flat: Given a $B\otimes_A B$-module $L$, there is a canonical isomorphism $(M\otimes_A N) \otimes_{B\otimes_A B} L \cong M \otimes_B L \otimes_B N$ which clearly shows that if $M$ and $N$ are flat over $B$ then $M\otimes_A N$ is flat over $B\otimes_A B$, but I am not sure about projectivity. In my application, the ring $B\otimes_A B$ is not noetherian.

Any ideas?

$\endgroup$
  • $\begingroup$ Isn't it finitely generated ? $\endgroup$ – user18119 Sep 12 '12 at 17:35
  • $\begingroup$ @QiL, I believe it is... why? $\endgroup$ – the L Sep 12 '12 at 17:42
  • $\begingroup$ This seems to be straightforward. Write a finite presentation of $M$ as $B$-module, tensor by $N$ over $A$ and see that $M\otimes_A N$ is cokernel of a map of f.p. $B\otimes_A B$-modules. $\endgroup$ – user18119 Sep 12 '12 at 21:09
  • $\begingroup$ Can you elaborate? how do you deduce projectivity? $\endgroup$ – the L Sep 12 '12 at 21:24
1
$\begingroup$

Consider an exact sequence of $B$-modules $$ B^n \to B^m \to M\to 0$$ Tensoring by $N$ gives rise to an exact sequence of $B$-modules $$ B^n\otimes_A N\to B^m\otimes_A N\to M\otimes_A N\to 0.$$ The above maps are also $C$-linear, where $C=B\otimes_A B$.

Similarly we have an exact sequence of $C$-modules $$C^{mr}=B^m\otimes_A B^r\to C^{ms}=B^m\otimes_A B^s \to B^m\otimes_A N\to 0$$ which shows that $B^m\otimes_A N$, and similarly, $B^n\otimes_A N$ are finitely presented $C$-modules. Therefore $M\otimes_A N$ is finitely presented over $C$.

Finally notice that a finitely presented module is projective if and only if it is flat.

$\endgroup$
  • $\begingroup$ Thanks! I actually knew that, I just didn't know that a finitely presented flat module is projective :) $\endgroup$ – the L Sep 13 '12 at 7:57

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.