2
$\begingroup$

This is a exercise in "Deformation Theory [Hartshorne]".

Let C be a local Artin ring with residue field k.

Let X be a scheme flat over C , and let$X_0=X\times _Ck$.

If F is a coherent sheaf on X which is flat over C,

and $F_0=F⊗_{O_X}O_{X_0}$ is locally free on $X_0$,show that F is locally free on X.

what 'local Artin ring' implies in this question?

$\endgroup$
3
$\begingroup$

It suffices to show that if $X$ is affine and $F_0$ is free, then $F$ is free. Since $F_0$ is free over $X_0$, we can find an isomorphism $\mathscr{O}_{X_0}^r\rightarrow F_0$.

Arbitrarily lift this to a map $\mathscr{O}_X^r\rightarrow F$. We claim that this lift is an isomorphism. Let $K$ be the kernel and $M$ be the cokernel.

First, we see that, since $\mathscr{O}_X^r\rightarrow F$ is a surjection after applying $\cdot\otimes_{\mathscr{O}_X}\mathscr{O}_{X_0}$, $M\otimes_{\mathscr{O}_X}\mathscr{O}_{X_0}=0\Leftrightarrow M=mM$, where $m$ is the maximal ideal of $C$. Since $C$ is artinian, $m^N=0$ for $N>>0$, so $M=0$.

Now, we take $0\rightarrow K\rightarrow \mathscr{O}_X^r\rightarrow F\rightarrow0$ and tensor to get $0\rightarrow K\otimes_{\mathscr{O}_X}\mathscr{O}_{X_0}\rightarrow \mathscr{O}_{X_0}^r\rightarrow F_0\rightarrow 0$, which is left exact since $F$ is flat over $C$.

Therefore, we again see that $K\otimes_C k=0$, which means $K=mK$, so $K=0$ from the same argument as above.

(So we used $C$ is artinian to apply a version of Nakayama's lemma in the case where our coherent sheaves over $X$ aren't finitely generated over $C$.)

$\endgroup$
  • $\begingroup$ (I copied the same trick Daniele A gave when I asked math.stackexchange.com/questions/1698112/…, if you want something clearer) $\endgroup$ – DCT Apr 17 '16 at 13:36
  • $\begingroup$ thank you for the really nice proof! But I still not understand why the lift of the first isomorphsm exists? $\endgroup$ – 8Frog Apr 17 '16 at 20:31
  • $\begingroup$ Okay, I made a small edit to the beginning of the proof, so we only need to consider the affine case. Then, the map $\mathscr{O}_{X_0}^r\rightarrow F_0$ is given by $r$ elements $f_1,\ldots,f_r$ of $F_0$. Since $F\rightarrow F_0$ is surjective, we can lift this map to $\mathscr{O}^r\rightarrow F$ by lifting those elements $\tilde{f}_1,\ldots,\tilde{f}_r$ to $F$ such that $\tilde{f}_i$ maps to $f_i$ under the surjection. $\endgroup$ – DCT Apr 18 '16 at 4:48
  • $\begingroup$ Before the edit, I think I tried to claim that we could lift the map $\mathscr{O}_{X_0}^r\rightarrow F_0$ to $\mathscr{O}_X^r\rightarrow F$ in general (even if $X$ is not affine), and I doubt this is true. We have the long exact sequence $H^{0}(F)\rightarrow H^0(F_0)\rightarrow H^1(mF)$, where $m$ is the maximal ideal of $C$, and the fact that $H^1$ can be nontrivial is evidence to me that we shouldn't expect this to be true. (If this wasn't what you were worried about then, don't worry about this.) $\endgroup$ – DCT Apr 18 '16 at 4:53

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.