Let $X$ be a variety. Let $D^b(Coh(X))$ be the derived category of bounded complexes of coherent sheaves on $X$, and $D^b_{coh}(X)$ be the derived category of bounded complexes of sheaves of $\mathcal{O}_X$-modules with coherent sheaves as cohomologies. Similarly, we have $D^b_{qc}(X)$ and $D^b(Qco(X))$ by replacing $coherent$ sheaves to $quasi$-$coherent$ sheaves.

It is proved in "Residues and Duality" by Hartshorne (Chapt. II Corollary 7.19) that $D^b_{qc}(X)$ and $D^b(Qco(X))$ are derived equivalent. I was wondering if the same thing is still true for $D^b_{Coh}(X)$ and $D^b(coh(X))$?

Hartshorne's proof seems could not be generalized to this case because I feel that any quasi-coherent module might not be embedded to a quasi-coherent injective module.

  • 3
    $\begingroup$ I think that noetherianity is necessary. Otherwise the cohomology of a complex of coherent sheaves need not be coherent. $\endgroup$ Feb 28, 2014 at 15:59
  • $\begingroup$ I assume $X$ to be a variety, isn't that enough? $\endgroup$
    – Li Yutong
    Feb 28, 2014 at 19:39
  • 1
    $\begingroup$ varieties are by definition at least of finite type over a field. these are automatically noetherian, so yes. $\endgroup$ Mar 1, 2014 at 9:35

1 Answer 1


For X noetherian this is still true. (Proposition 3.5 in Daniel Huybrechts' book)

  • $\begingroup$ Not exactly. Huybrechts prove that $D^b(Coh(X))$ is equivalent to $D^b_{coh}(Qcoh(X))$, the derived category of quasi-coherent complexes with coherent cohomology. I do not think this is equivalent to $D^b_{coh}(X)$. $\endgroup$
    – abx
    Feb 28, 2014 at 15:59
  • 1
    $\begingroup$ I guess you are right - can we not bootstrap the fact that $D(qc(X)) = D_{qc}(Sh(X))$ to get what you want? $\endgroup$ Feb 28, 2014 at 16:14
  • $\begingroup$ I don't see how to do that. $\endgroup$
    – abx
    Feb 28, 2014 at 16:32
  • 11
    $\begingroup$ We have the embedding $D(coh(X)) \to D_{coh}(qc(X)) \to D_{coh}(Sh(X))$ and we want to show it's essentially surjective. Take a complex in $D_{coh}(Sh(X))$. By Corollary 3.4 in H this is quasi-isomorphic to a complex of quasi-coherent sheaves and by Prop 3.5 such a thing is quasi-isomorphic to a complex of coherent sheaves. In the whole process we never changed cohomology, so that ought to do it. Or am I misunderstanding something fundamental? $\endgroup$ Feb 28, 2014 at 16:58
  • 1
    $\begingroup$ Seems OK to me. I just wonder why Huybrechts did not state it that way. $\endgroup$
    – abx
    Feb 28, 2014 at 17:22

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.