Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $X$ be an scheme and $F$ be a quasi coherent sheaf on $X$. There are three ways to consider flatness in the category of quasi coherent sheaves Qco($X$). 1- $F$ is called flat if $F_x$ is flat for each $x$ in $X$. (This definition is more usual) 2- $F$ is flat if $F(u)$ is a flat $O(U)$-module for each open affine subset $U$ of $X$. 3-(categorical definition) In the case that $X$ is semi-separated quasi-compact scheme, Qco($X$) is a locally finitely presented category. So one can think of flat objects via purity. $F$ is called flat if for each exact sequence $$0\to M\to N\to F\to 0$$ of quasi coherent sheaves, the induced sequence $$ 0\to Hom(P,M)\to Hom(P,N) \to Hom(P,F)\to 0$$ of abelian groups is exact for each finitely presentedquasi coherent sheaf $P$.

\remark: Note that when $X$ is semi separated and quasi compact, $P$ is finitely presented if and only if it is coherent.(Murfet notes, therisingsea.org)

IS there any relation between these classes of quasi-coherent sheaves?

share|improve this question
add comment

1 Answer

It is well-known that flatness of usual modules over rings can be checked stalkwise. This easily implies that (1) and (2) are equivalent. Another quite natural condition concerns the structure of $\mathsf{Qcoh}(X)$ as an abelian symmetric monoidal category: (4) $F$ is flat if $F \otimes -$ is an exact functor. Using the compatibility between stalks and tensor products it is easily seen that (1) implies (4).

I claim that (4) implies (2) if $X$ is a quasi-separated scheme $X$. In fact, assume (4) and let $U \subseteq X$ be an open affine subset and $M \to N$ a monomorphism of quasi-coherent modules on $U$. Let $j : U \to X$ denote the open immersion. It is quasi-compact (since $X$ is quasi-separated) and quasi-separated, hence $j_{*} M \to j_{*} N$ is a monomorphism of quasi-coherent modules on $X$. It follows that $F \otimes j_* M \to F \otimes j_* N$ is a monomorphism. Applying $j^*$, we find that $F|_U \otimes M \to F|_U \otimes N$ is a monomorphism. Hence, $F|_U$ is flat (i.e. $F(U)$ is flat since $U$ is affine).

By the way, if $X$ is a quasi-compact quasi-separated scheme, then $\mathsf{Qcoh}(X)$ is locally finitely presentable, and "finitely presentable" defined categorically coincides with "of finite presentation" (this follows from results in EGA I (new), 6.9, I have proven this in my diploma thesis, should I make this public?). So we don't need that $X$ is semi-separated.

Probably one can also show that (2) and (4) coincide in this setting. Namely, (2) => (4) is trivial from the affine case and the converse uses, again, extension of quasi-coherent modules along open immersions.

share|improve this answer
1  
In murfets note, it is shown that over a quasi compact quasi separated, $F$ is locally finitely presented if and only if it is categorically finitely presented. –  hamid Jan 8 '13 at 13:30
1  
Thank you Martin for your answer. I don't know why (1) and (2) are equivalent yet. –  hamid Jan 8 '13 at 13:34
1  
@hamid - for (1)<=>(2). Firstly, it is enough to work on an affine scheme. Then you could use the usual definition of flatness for $F(U)$ and $F_x$ (i.e. tensor product exact functor) coupled with the fact that for a module $M$ over a commutative ring if $M_{\mathfrak{m}}$ = 0 for all maximal ideals $\mathfrak{m}$ then $M=0$. –  Dragos Fratila Jan 8 '13 at 14:09
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.