On the abelian fivefolds attached to cubic surfaces (Q2510089)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | On the abelian fivefolds attached to cubic surfaces |
scientific article |
Statements
On the abelian fivefolds attached to cubic surfaces (English)
0 references
31 July 2014
0 references
In this paper, the author manages to say something new about the twenty-seven lines on a cubic surface. The main object of interest is the abelian fivefold \(Q(Y)\) associated with a cubic surface \(Y\): essentially this is the intermediate Jacobian of the cubic threefold \(Z\) obtained by taking the triple cover of \({\mathbb P}^3\) ramified exactly along \(Y\). By means of a Prym construction one may do this not just complex-analytically but over \({\text{Spec}}\, {\mathbb Z}[1/6]\), as shown by the author in [Zbl 1325.14053]. The abelian fivefolds that arise have an action of the Eisenstein integers of signature \((4,1)\) and are principally polarised. The idea here is to study the moduli space of stable cubic surfaces via this correspondence, thus exploiting the theory of Shimura varieties. Two distinct kinds of result are shown here. Over the complex numbers, the author calculates all the possible Hodge groups (and Mumford-Tate groups) of the abelian fivefolds \(Q(Y)\), showing that they do all occur. In fact the Mumford-Tate groups are determined by the endomorphism algebras, and it is these that are classified. It is quite delicate to show that one may obtain the desired endomorphism algebra (which is an isogeny invariant) exactly while still retaining principal polarisation. On the other hand, one can ask about the monodromy groups of the universal abelian fivefold pulled back to the moduli of cubic surfaces. Over the complex numbers this is quite well understood, but here, by careful use of some rather general monodromy calculations for unitary Shimura varieties, the calculation is carried out (\(\ell\)-adically) over more or less arbitrary fields (with some limitations at \(2\) and \(3\)). In particular, it is shown that the Galois group of the field of definition of the \(27\) lines on a cubic surface over a field \(K\) of characteristic prime to \(6\) is usually \(W(E_6)\): for the complex numbers, this is a modern rephrasing of a result of Jordan from 1870.
0 references
cubic surface
0 references
cubic threefold
0 references
abelian variety
0 references
Shimura variety
0 references
Mumford-Tate group
0 references
monodromy
0 references