2
$\begingroup$

In a category $\mathcal C$, we will say that $(\mathcal T,\mathcal F)$ is a torsion theory if it satisfies:

(1) $Hom(T,F)=0$ for all $T\in \mathcal T$ and $F\in \mathcal F$.

(2) If $Hom(T,F)=0$ for all $F\in \mathcal F$, then $T\in \mathcal T$.

(3) If $Hom(T,F)=0$ for all $T\in \mathcal T$, then $F\in \mathcal F$.

For modules over a commutative integral domain $A$, torsion modules and torsion free modules form such a pair.

For sheaves of $\mathcal O_X$-modules over an integral scheme, we can take locally torsion and locally torsion free $\mathcal O_X$-modules, i.e., torsion and torsion free respectively on each affine.

My question is : does this still hold for quasi-coherent sheaves on an integral scheme? In other words, do locally torsion and locally torsion free quasi-coherent sheaves still give a torsion theory on $QCoh(X)$?

Improve this question
$\endgroup$
3
  • $\begingroup$ I guess that $\mathcal{T}$ and $\mathcal{F}$ are meant to be sub-collections of objects. $\endgroup$
    – YCor
    May 18, 2015 at 14:55
  • $\begingroup$ for coherent sheaves on a noetherian scheme this is pretty easy to see: take any coherent module M and consider all its torsion subsheaves. By noetherianness there will be a maximal one Mt and the quotient M/Mt will be torsion free. (this is a general fact: for a noetherian abelian category, whenever you have a subcategory closed under quotients and extensions it gives rise to a torsion pair) $\endgroup$
    – bananastack
    May 18, 2015 at 14:55
  • $\begingroup$ You need at least that $C$ has zero morphisms or else I don't know what $\text{Hom}(T, F) = 0$ means. Really you should restrict to something like abelian categories or else I doubt the notion is useful for much. $\endgroup$
    – Qiaochu Yuan
    May 18, 2015 at 19:32

1 Answer 1

Reset to default
2
$\begingroup$

Yes, it is still a torsion theory (by the way, you forget to mention the condition that each object $A$ can be represented as an extension $0 \to T(A) \to A \to F(A) \to 0$). To show you just take $T(A)$ to be the union of all torsion subsheaves of $A$. A similar technique is used in the paper Tarrío, Leovigildo Alonso, Ana Jeremías López, and María José Salorio. "Construction of 𝑡-structures and equivalences of derived categories." Transactions of the American Mathematical Society 355, no. 6 (2003): 2523-2543.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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