On the existence of Ulrich vector bundles on some surfaces of maximal Albanese dimension (Q2326974)

Given a polarized algebraic variety \((X,\mathcal{O}_X(1))\) of dimension \(d\) with \(\mathcal{O}_X(1)\) a very ample line bundle, it becomes a central question to understand those vector bundles \(\mathcal{E}\) supported on \(X\) with the simplest cohomology with respect to multiples of \(\mathcal{O}_X(1)\).They are the so-called Ulrich bundles defined by the following conditions: \(H^i(\mathcal{E}(-j))=0\) for \(i=0,\dots,d\) and \(j=1,\dots d\). Equivalently, they are arithmetically Cohen-Macaulay bundles such that \(H^0(\mathcal{E}(-1))=0\) and \(H^0(\mathcal{E})=d.rank(\mathcal{E})\), the maximal permitted number of global sections. \textit{D. Eisenbud} and \textit{F.-O. Schreyer} [J. Am. Math. Soc. 16, No. 3, 537--575, appendix 576--579 (2003; Zbl 1069.14019)] whether any polarized variety \((X,\mathcal{O}_X(1))\) supports an Ulrich bundle. It is a classical result, due to Horrocks, that \(\mathcal{O}_{\mathbb{P}^d}\) is the only indecomposable Ulrich bundle on \(\mathbb{P}^d\). Already in the aforementioned paper, it was proved that any polarized curve supports Ulrich line bundles. So the interest shifted to the case of algebraic surfaces. Thanks to the collaboration of many mathematicians, a quite large list of polarized surfaces has been shown to support Ulrich bundles: rational surfaces, \(K3\) surfaces, abelian surfaces, etc (see for instance [\textit{A. Beauville}, Eur. J. Math. 4, No. 1, 26--36 (2018; Zbl 1390.14130)] for an account). However, for large families of surfaces, specially for those of general type, the answer is still unknown. The paper under review contributes to this circle of ideas showing that a polarize surface \((X,\mathcal{O}_X(1))\) with maximal albanese dimension, \(\chi(\mathcal{O}_X)=0\) and \(h^1(\mathcal{O}_X(1))=0\) (a non-special surface) supports rank two simple (i.e., the only endomorphisms are the homotheties) Ulrich bundles.