I have a simple question about coherent sheaves and line bundles:

if I have a coherent sheaf $F$ on a good scheme $X$ and I know that $F_x \otimes k(x) = k(x)$ for alle points $x$ on $X$ (where $k(x)$ is the residue field of $x$ and the tensor product goes over the local ring $\mathcal O_{X,x}$), can I then say, that $F$ is already a line bundle?

My strategy would be:

the stalk of $F$ is finitely generated over the local ring; by Nakayama you can choose a generating element $s_x$ of $F_x$ for every $x$, so you have $F_x \simeq \mathcal O_{X,x}$ and now use the statement that if a coherent sheaf is free in a point, then it is in a neighborhood. Is this the right way to see it?


No. You need that the scheme is reduced.

It is certainly true that if $F_x$ is a free $\mathcal{O}_{X,x}$-module of rank $n$, then there exist an open neighborhood $U$ of $x$ such that $F \vert_U$ is a free $\mathcal{O}_U$- module of rank $n$.

But from $\dim_{k(x)} F_x \otimes_{\mathcal{O}_{X,x}} k(x) = 1$ you can deduce that $F_x$ is a cyclic module and not that $F_x$ is free of rank $1$.

However on a reduced scheme the statement is true: exercise II.5.8 of Hartshorne.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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