5
$\begingroup$

If $f \colon X \to Y$ is a proper monomorphism of schemes [monomorphism in the sense of category theory], then the underlying map of topological spaces is continuous, closed, and injective, hence a homeomorphism onto its image. Properness also implies finite type, which eliminates some of the more egregious monomorphisms of schemes. But I don't see any obvious argument (or counterexample) for why the corresponding morphism of sheaves $\mathcal{O}_Y \to f_* \mathcal{O}_X$ should be surjective.

Motivation: $X \to Y$ is a monomorphism iff the corresponding morphism of functors $\mathrm{Hom}(-,X) \to \mathrm{Hom}(-,Y)$ is injective. Thus, a positive answer to my question here would, in nice situations, allow for a valuative criterion as to whether a subfunctor is closed.

$\endgroup$
3
  • 3
    $\begingroup$ If $f$ is of finite presentation then the answer is yes; cf. EGA IV${}_3$ 8.11.5. $\endgroup$ Nov 10, 2010 at 18:22
  • 6
    $\begingroup$ Without assuming finite presentation, it is EGA IV$_4$, 18.12.6 (and 18.12 has plenty of other nifty results removing finite presentation restrictions). $\endgroup$
    – BCnrd
    Nov 10, 2010 at 18:59
  • $\begingroup$ A reference in english is the Stacks project Tag 04XV $\endgroup$
    – Mathmop
    Jan 2, 2022 at 14:50

0

Your Answer

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

Browse other questions tagged or ask your own question.