I'm reading an article and I have a question about one of the steps.

Setting: k is a field, K is the function field of the projective line $P^1$ over $k$. Let $A=k[t,t^{-1}]$ be the ring of all the rational functions having poles only at $0,\infty$ . $V$ is a 2-dimensional vector space over $K$ and $M$ is a free module over $A$ containing a basis of $V$.

The author claims and I quote "The A-module M is free and can be regarded as a (trivial) rank 2 vector bundle on $Spec A = P^1 - \{0,\infty\}$".

I'm not familiar with vector bundles (I do know the definition), and I don't understand the duality between free modules over $A$ and vector bundles over $spec(A)$.

Any explanation is appriciated!



$\newcommand{\Spec}{\operatorname{Spec}}$When we're talking about schemes, when you read "vector bundle" you should read "locally free sheaf". Here's what the latter term means precisely:

Let $X$ be a scheme and write $\mathcal O_U:=\mathcal O_X|_U$. We say an $\mathcal O_X$-module $\mathscr F$ is a locally free sheaf if for each $p\in X$, there is a set $I$ and an open neighborhood $U$ of $p$ such that $\mathscr F|_U\cong\mathcal O_U^{\oplus I}$ (note this is an isomorphism of $\mathcal O_U$-modules).

Usually we will be interested in when $I$ is finite and $\mathscr F$ has constant rank, so there is some integer $n$ such that $\mathscr F|_U\cong\mathcal O_U^{\oplus n}$ for our "trivial" neighborhoods $U$.

Now, here's a specific construction you should be aware of:

Let $A$ be a ring and $M$ an $A$-module. Let $X=\Spec A$ and $\mathcal O_X$ its structure sheaf. There is a natural $\mathcal O_X$-module $\widetilde M$ defined as follows: for $f\in A$, we let $\widetilde M(\Spec A_f)=M_f$, the localization of $M$ at $f$, which is an $A_f$-module. Since the open subsets $\Spec A_f$ form a base for $X$ as $f$ runs over all elements of $A$, this is enough to determine $\widetilde M$ as a sheaf on $X$, and you can double check any details to convince yourself that $\widetilde M$ is actually an $\mathcal O_X$-module.

Now you should practice unwinding definitions to convince yourself that the following is true:

The $\mathcal O_X$-module $\widetilde M$ is a locally free sheaf on $X=\Spec A$ if and only if $M$ is a locally free $A$-module, which means that there exists a set $\{f_1,\dots,f_n\}$ generating $A$ such that $M_{f_i}$ is a free $A_{f_i}$-module.

In particular, if $M$ is a free $A$-module, then our generating set above if just $\{f_1=1\}$, so $\widetilde M$ is actually a free $\mathcal O_X$-module, i.e. it is isomorphic to $\mathcal O_X^{\oplus 2}$.

This is why for a free $A$-module $M$, we can consider $M$ as a "trivial" locally free sheaf on $\Spec A$. As to why we identify "locally free sheaves" with "vector bundles", you can try to prove the following for yourself.

If $X$ is an $A$-scheme, then there is a one-to-one correspondence between rank $n$ locally free sheaves on $X$ and rank $n$ vector bundles on $X$, where a (rank $n$) vector bundle is an $A$-scheme $E$ along with a projection morphism $\pi:E\to X$ with the following property: for each $p\in X$ there is a neighborhood $U$ of $p$ and an isomorphism $\Phi:\pi^{-1}(U)\overset{\sim}\to U\times\Bbb A_A^n$ such that $\pi=\mathrm{pr}_2\circ\Phi$, where $\mathrm{pr}_2$ is the projection $\pi^{-1}(U)\times\Bbb A_k^n\to\Bbb A_A^n$.

If it's not immediately obvious, the "trivializable" neighborhoods $U$ in the above definition of vector bundle should correspond naturally to the "trivializable" neighborhoods $U$ where $\mathscr F|_U\cong\mathcal O_U^{\oplus n}$.


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.