Vector bundles on the moduli stack of elliptic curves (Q2257284): Difference between revisions

For \(R\) a commutative ring one denotes by \({\mathcal M}_R\) the moduli stack of elliptic curves over it. The \(\mathrm{Pic}({\mathcal M}_R)\) was determined by \textit{W. Fulton} and \textit{M. Olsson} [Algebra Number Theory 4, No. 1, 87--104 (2010; Zbl 1198.14010)]. The first main result of this paper is the following\newline``Theorem A. Let \(R\) be a discrete valuation ring or a field with (residue) characteristic not equal \(2\) or \(3\). Then every vector bundle on \({\mathcal M}_R\) is a sum of line bundles.''\newline The theorem is proved in a more general context and the author explains why the extension to the case \(R\) a regular local ring ``does not seem feasible in the moment'', sending to the open Bass-Quillen conjecture.\newline The next important result of the paper is the classification of those vector bundles on \({\mathcal M}_R\) for \(R={\mathbb Z}_{(3)}\) which can be obtained by iterated extensions of line bundles.\newline Using representation arguments, namely constructing a functor \(I:{\mathbb Z}[\mathrm{GL}_2({\mathbb F}_3)] \to \mathrm{QCoh}({\mathcal M}_{{\mathbb Z}_{(3)}})\), one proves the third important result of the paper, which is the assertion that there are infinitely many, of arbitrary high rank, indecomposable vector bundles on \({\mathcal M}_{{\mathbb Z}_{(3)}}\).\newline The paper is very well written and contains seven facts and details which can not be mentioned in this short review.