Existence of meromorphic connections

300 Views Asked by At

Suppose we start with the line bundle $ O(N) \rightarrow \mathbb{C} \mathbb{P}^m$. By a holomorphic connection, I mean an operator $\nabla : \Gamma(\mathbb{C} \mathbb{P}^m, O(N)) \rightarrow \Gamma(\mathbb{C} \mathbb{P}^m, O(N)) \otimes \Omega^1(\mathbb{C} \mathbb{P}^m)$ where $\Omega^1(\mathbb{C} \mathbb{P}^m)$ is the set of holomorphic 1-forms. By meromorphic connection, I mean that we replace $\Omega^1(\mathbb{C} \mathbb{P}^m)$ with $\Omega^1_M(D)$,the set of meromorphic one-forms with poles in some prescribed polar divisor $D$.

I have two questions about these things:

  1. From my understanding, there does not exist a holomorphic connection for our line bundle, since such a connection would be automatically flat, and thus our line bundle would be flat, which is false for $N \neq 0$. Am I correct in this assertion?

  2. Do there exists meromorphic connections? If so, can one provide an explicit example?

1

There are 1 best solutions below

0
On

The answer to both your questions is positive.

  1. As is explained in Huybrechts's "Complex Geometry – An Introduction" here, the existence of a holomorphic connection is equivalent to the Atiyah class being trivial.

    Huybrechts, Daniel. Complex geometry. An introduction. Universitext. Springer-Verlag, Berlin, 2005. xii+309 pp.

    The Atiyah class in this case is just the first Chern class which is nontrivial for $N \ne 0$.

  2. Just trivialize $\mathcal{O}(N)$ on a standard open set. Any connection there can be considered as a meromorphic connection in your sense.