17
$\begingroup$

It is true that over a contractible manifold all differentiable vector bundles are trivial. However the method of proof does not apply in the holomorphic category.

It is also true that a contractible one dimensional complex manifold has no non-trivial line bundles. (Shown using the exponential sequence and a bit of tinkering.)

Moreover, I have read that the topological and holomorphic classification of bundles are the same over Stein manifolds. (No adequate reference was given.)

Hence my question:

Is it true that all holomorphic vector bundles are trivial over a contractible complex manifold?

Probably this is false. In this case, could you give me counter-examples?

$\endgroup$
1

1 Answer 1

20
$\begingroup$

No, even for line bundles. We have the short exact sequence of sheaves $$0 \to \underline{\mathbb{Z}} \overset{2 \pi i}{\longrightarrow} \mathcal{O} \overset{\exp}{\longrightarrow} \mathcal{O}^{\ast} \to 0$$ where $\underline{\mathbb{Z}}$ is locally constant $\mathbb{Z}$ valued functions, $\mathcal{O}$ is holomorphic functions and $\mathcal{O}^{\ast}$ is nonzero holomorphic functions. If $X$ is contractible, then $H^1(X,\underline{\mathbb{Z}}) = H^2(X,\underline{\mathbb{Z}}) = 0$. So $H^1(X, \mathcal{O}) \cong H^1(X, \mathcal{O}^{\ast}) \cong Pic(X)$.

Now use the standard example of a contractible manifold with nonzero $H^1(X,\mathcal{O})$:

$$X = \{ (z,w) \in \mathbb{C}^2 : (|z|, |w|) \in [0,1) \times [0,2) \cup [0,2) \times [0,1) \}.$$

$\endgroup$
4
  • 5
    $\begingroup$ How do you show that $H^1(X, \mathcal{O}) \neq 0$? $\endgroup$ Nov 8, 2014 at 7:10
  • $\begingroup$ @MichaelAlbanese I just noticed this comment. In case you still care: Let $U_1 = \{ (z,w) : (|z|, |w|) \in [0,1) \times [0,2) \}$ and $U_2 = \{ (z,w) : (|z|, |w|) \in [0,2) \times [0,1) \}$. Then $U_1$, $U_2$ and $U_1 \cap U_2$ are polydiscs, so their $H^1(\mathcal{O})$ vanishes, and we can compute $H^1(X, \mathcal{O})$ using the Cech complex on $U_1$ and $U_2$. So $H^1(X, \mathcal{O})$ is the cokernel of $\mathcal{O}(U_1) \oplus \mathcal{O}(U_2) \to \mathcal{O}(U_1 \cap U_2)$. $\endgroup$ Feb 19, 2019 at 14:12
  • $\begingroup$ An explicit nonzero class in the cokernel is $\sum z^i w^i$, since the coefficient of $z^i w^i$ is $O(2^{-i})$ in any function which extends to $U_1$ or $U_2$. $\endgroup$ Feb 19, 2019 at 14:15
  • $\begingroup$ Thanks! I asked a question on MSE about this and it remains unanswered. Would you mind writing an answer for it? $\endgroup$ Feb 19, 2019 at 16:01

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.