Let be an abelian variety over an algebraically closed field . If , then corresponds to a *complex torus*: that is, can be expressed complex analytically as where is a complex vector space of dimension and is a lattice (i.e., a -free, discrete submodule of rank ). In this case, one can form the *dual abelian variety*

At least, as defined is a complex torus, but it turns out to admit the structure of an abelian variety.

The purpose of the next few posts is to describe an algebraic version of this duality: it turns out that can be constructed as a scheme, purely algebraically. I’d like to start with a couple of posts on Picard schemes. A useful reference here is this article of Kleiman.

**1. The Picard scheme analytically**

Let be a smooth projective variety over the complex numbers . The collection of line bundles is a very interesting invariant of . Usually, it splits into two pieces: the “topological” piece and the “analytic” piece. For instance, there is a first Chern class map

which picks out the topological type of a line bundle. (Topologically, line bundles on a space are classified by their first Chern class.) The admissible topological types are precisely the classes in which project to -classes in under the Hodge decomposition.

But there is more to the classification of (algebraic or equivalently holomorphic) line bundles on than the first Chern class: a line bundle may be topologically trivial without being holomorphically trivial. A good way to see this is to use the exponential sequence

where is the sheaf of analytic functions on the complex points of . The exponential sequence yields after applying cohomology an exact sequence

Here is precisely and we get an exact sequence

This shows that the group (which is a complex torus) precisely picks out the topologically trivial holomorphic line bundles on : they fit into a connected complex manifold called (the **Picard variety**) of dimension .

**Example 1 **Let be a complex abelian variety. Then is identified with while is identified (using Hodge theory) with the translation-invariant -forms on . (These are the harmonic -forms with respect to a translation-invariant metric.) Such a form is given by an element in the -cotangent space to , i.e. by an element of

We find that the Picard variety is exactly , i.e. the complex torus we encountered before.

In particular, we conclude from this analysis that the Picard group is the extension of a finitely generated abelian group (the so-called **Néron-Severi group**) by the complex torus , i.e. the Picard variety. We also conclude that the tangent space to the Picard variety is precisely : this is something that holds in the algebraic context as well, but via deformation theory.

**Example 2** On a rational variety (or even a unirational variety), the group vanishes. More generally, this is true for any simply connected variety (over ) by the Hodge theoretic equality , and consequently the Picard group is discrete for such varieties. (Recall from the previous post that unirational varieties are simply connected by a theorem of Serre; as Jason Starr points out in the comments, this result has been extended significantly, e.g. to rationally connected varieties.)

**Example 3 **Since we are over the complex numbers, we have and the latter is a birational invariant. Therefore, the dimension of the Picard variety is a birational invariant.

**2. The Picard scheme**

Let’s now move to an algebraic setting. If is a projective variety over the algebraically closed field , we would like to put on the group the structure of an algebraic variety. Or rather, it should be a scheme, which is the disjoint union of countably many algebraic varieties.

In order to turn into a scheme, we’ll need to say what it means for a scheme to map into . Roughly, this should be the same as a “family” of line bundles on parametrized by . Here is a reasonable definition of that:

Definition 1A family of line bundles on parametrized by is simply a line bundle on .

We could thus try to hope that maps from would be isomorphism classes of such line bundles. Unfortunately, this does not work: the functor thus defined fails to satisfy flat descent, because line bundles admit automorphisms. For example, if , then the functor thus defined is

which is clearly not a sheaf in the flat topology (or even Zariski topology!). The reason it fails to be a sheaf is precisely that in order to do flat descent, you need to keep track of automorphisms rather than quotient by them.

But one could instead define a functor

sending any to the *groupoid* of line bundles on (rather than the set of isomorphism classes). This functor is a sheaf in the flat topology: this is equivalent to the theory of flat descent for quasi-coherent sheaves. We can use this to get a Picard **stack** of line bundles.

However, in this case, there’s no real need to form the Picard stack, because there aren’t any interesting automorphisms to keep track of. The only automorphisms come from multiplication by invertible functions on . In fact, given a line bundle on , we have

because is reduced and projective. In particular, the automorphisms of are exactly .

This suggests that if we rigidify the moduli problem, we can eliminate the need for stacks without losing any interesting information. In other words, we can define the improved **Picard functor** via

where is a fixed basepoint.

The main deep result is the following:

Theorem 2 (Grothendieck)is representable by the countable union of quasi-projective schemes.

Moreover, the **Picard stack** described previously is simply (*mistaken assertion corrected and recorrected here)*: it’s what you get by adding a ‘s worth of automorphisms to each object in (and stackifying). Observe that is a group scheme by tensor product of line bundles.

We can define to be the connected component at the identity of : this represents line bundles which are **algebraically equivalent to zero** (i.e., which can be connected to the zero bundle by a connected family). The quotient is the **Néron-Severi group**, which over picks out the topological type of a line bundle. In characteristic , it is a deep theorem that the group is still finitely generated.

March 19, 2013 at 12:46 pm

One might say the Picard stack is a -gerbe over the Picard scheme!

March 19, 2013 at 2:09 pm

Actually, I think my assertion was mistaken: there’s no natural map from the stack to the scheme. It’s a -quotient rather than a product.

March 19, 2013 at 4:08 pm

Hmm…I think what you had is OK if interpreted properly. The sheafification of the Picard stack is the Picard scheme, so as you said this is just a rigidification by killing off the -automorphisms. Unless I have something backwards ….

March 20, 2013 at 7:00 am

I was thinking about this a bit more, and as you say I now think the original assertion was correct. That is, the Picard stack *does* map to the Picard scheme, it’s just that you have to twist the line bundle on by the pull-back of the line bundle on given by restricting to so that you can make the trivialization. In other words, I believe it is the product stack as claimed.