3
$\begingroup$

Let $S$ be a smooth projective surface with an ample divisor $X\subset S$. Consider the moduli stack of vector bundles $F$ on $S$ such that

1) $c_1(F)=0$

2) $c_2(F)=n$

3) The restriction of $F$ to $X$ is fixed (say, isomorphic to a bundle $M$).

$\mathbf{Questions:}$

1) Is it true that for any $n$ and $M$ the above stack is of finite type?

2) Is it true, that for any given $M$ the above stack is empty for $n$ sufficiently small?

$\endgroup$

1 Answer 1

2
$\begingroup$

Take Q 2). Assume that we are in characteristic zero, that $F$ has rank $2$ and that $X$ is smooth. If $n<0$ then $F$ is Bogomolov unstable: there is a short exact sequence $$0\to O(A)\to F\to I_Z.O(-A)\to 0$$ with $A^2>0$ and $A.X>0$ and $I_Z$ the ideal sheaf of a $0$-dimensional subscheme $Z$. In fact $c_2(F)=\deg Z-A^2$, so $A^2\ge -n$. Suppose that we can make $n$ arbitrarily small (i.e., large and negative). Then (index theorem) $A.X$ is arbitrarily large (and positive). On the other hand, restrict the inclusion $O(A)\to F$ to $X$; this gives a non-zero homomorphism $O_X(A)\to M$, which is impossible for fixed $M$.

So yes, your stack is empty in this case (rank $2$, char. $0$). In higher rank Bogomolov gives a destabilizing flag and maybe an elaboration of the argument above will then work. I don't know about char. $p$.

$\endgroup$
2
  • $\begingroup$ Where did you use char=0? By the way, I am also not sure what you mean by $O(A)$ - do you mean that you have chosen an embedding into ${\mathbb P}^n$? $\endgroup$ Sep 14, 2011 at 22:54
  • $\begingroup$ No I have not chosen $O(A)$, it is given by Bogomolov's theorem, which is this. Suppose that $S$ is a smooth projective surface over a field of char. zero and $F$ a rank $2$ vector bundle on $S$ with $c_1(F)^2>4c_2(F)$. Then there is a short exact sequence $$0\to O(A)\to F\to I_Z O(B)\to 0$$ where $O(A),\ O(B)$ are invertible sheaves on $S$ and $I_Z$ is the ideal sheaf of a $0$-dimensional subscheme $Z$ of $S$ such that $(A-B)^2>0$ and $(A-B).X>0$ for all ample $X$ on $S$. There is a version for bundles of higher rank and there are counter-examples in char. $p$. $\endgroup$
    – inkspot
    Sep 15, 2011 at 10:53

Your Answer

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

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