Minimal resolutions, Chow forms and Ulrich bundles on \(K3\) surfaces (Q2405320)

Betti diagrams are an effective way to organize information concerning the minimal free resolutions of graded modules over a polynomial ring. Let \(X\subset \mathbb P^r\) be a projective variety. The Minimal Resolution Conjecture (MRC), as formulated in [\textit{G. Farkas} et al., Ann. Sci. Éc. Norm. Supér. (4) 36, No. 4, 553--581 (2003; Zbl 1063.14031)], predicts that, given a sufficiently general subset \(\Gamma\subset X\) consisting of sufficiently many points, the Betti diagrams of \(X\) and \(\Gamma\) differ as few as possible. More concretely, it is shown in [\textit{M. Mustaţǎ}, Matematiche 53, Suppl., 53--81 (1998; Zbl 0943.13010)] that the Betti diagram of \(\Gamma\) differs from the Betti diagram of \(X\) only on two rows \(u-1\) and \(u\), lying beneath all of the non-trivial rows of the Betti diagram of \(X\), in other words, \(u-1\) is greater than the regularity of \(X\). The MRC predicts that \[ b_{i+1,u-1}(\Gamma)\cdot b_{i,u}(\Gamma)=0 \] for every \(i\geq 0\), which means that each of the diagonals of the two new rows of the Betti diagram contains at most one non-zero entry. The Ideal Generation Conjecture (IGC) is the \(i=1\) case of the MRC, which is equivalent to the statement that the number of generators of \(I_\Gamma/I_X\) is minimal. Theorems 1.1 and 1.2 of the paper under review establish the MRC for a large class of projective curves with general moduli, under some numerical assumptions on the degree and the genus of the curve and the dimension of the ambient projective space. Theorem 1.3 establishes the IGC for every general curve. As pointed out in the paper, there are special curves for which MRC and IGC fail, so the results presented are already optimal in an appropriate sense. In another direction, the paper under review studies Ulrich bundles on \(K3\) surfaces. Ulrich bundles are a class of bundles on projective varieties with interesting geometric and homological properties which has been intensively studied in recent years; see, for example, the survey article [\textit{A. Beauville}, Eur. J. Math. 4, No. 1, 26--36 (2018; Zbl 1390.14130)]. One of the key questions is whether every smooth projective variety admits an Ulrich bundle. The authors prove that a large class of \(K3\) surfaces, including all \(K3\) surfaces of Picard rank 1, admit Ulrich bundles of every even degree; see Theorems 1.4 and 1.5. As a corollary, for such a \(K3\) surface \(S\), the Cayley--Chow hypersurface \(\mathscr Z(S)=\{L\in \mathbb G\mid L\cap S\neq \emptyset\}\) inside the Grassmannian \(\mathbb G\) of codimension 3 planes in \(\mathbb P(H^0(S,\mathcal O_S(1)))\) can be realised as the zero locus of a Pfaffian; see Theorem 1.6. A detailed account of the many results of the paper is given in a convenient form in the introduction. The two main topics of the paper, the Minimal Resolution Conjecture and Ulrich bundles on \(K3\) surfaces, come together in the final Section 4 of the paper. There, it is shown that the existence of a rank 2 Ulrich bundle on a \(K3\) surface implies a variant of the MRC for certain zero-cycles on the surface.