We prove that the general quartic double solid with \(k\le 7\) nodes does not admit a Chow theoretic decomposition of the diagonal, (or equivalently has a nontrivial universal \(\mathrm{CH}_0\) group,) and the same holds if we replace in this statement “Chow theoretic” by “cohomological”. In particular, it is not stably rational. We also deduce that the general quartic double solid with seven nodes does not admit a universal codimension \(2\) cycle parameterized by its intermediate Jacobian, and even does not admit a parametrization with rationally connected fibers of its Jacobian by a family of \(1\)-cycles. This finally implies that its third unramified cohomology group is not universally trivial.

/pdf/unirational-threefolds-with-no-universal-codimension-2-cycle-3ei6jm6l9m.pdf

Unirational threefolds with no universal codimension \(2\) cycle

Preface vii 1Introduction 1 1.1 Decomposition of the diagonal and spread 3 1.2 The generalized Bloch conjecture 7 1.3 Decomposition of the small diagonal and application to the topology of families 9 1.4 Integral coefficients and birational invariants 11 1.5 Organization of the text 13 2Review of Hodge theory and algebraic cycles 15 2.1 Chow groups 15 2.2 Hodge structures 24 3Decomposition of the diagonal 36 3.1 A general principle 36 3.2 Varieties with small Chow groups 44 4Chow groups of large coniveau complete intersections 55 4.1 Hodge coniveau of complete intersections 55 4.2 Coniveau 2 complete intersections 64 4.3 Equivalence of generalized Bloch and Hodge conjectures for general complete intersections 67 4.4 Further applications to the Bloch conjecture on 0-cycles on surfaces 86 5On the Chow ring of K3 surfaces and hyper-Kahler manifolds 88 5.1 Tautological ring of a K3 surface 88 5.2 A decomposition of the small diagonal 96 5.3 Deligne's decomposition theorem for families of K3 surfaces 106 6Integral coefficients 123 6.1 Integral Hodge classes and birational invariants 123 6.2 Rationally connected varieties and the rationality problem 127 6.3 Integral decomposition of the diagonal and the structure of the Abel-Jacobi map 139 Bibliography 155 Index 163

/pdf/chow-rings-decomposition-of-the-diagonal-and-the-topology-of-523pv4m0e3.pdf

Chow Rings, Decomposition of the Diagonal, and the Topology of Families

I will discuss positive and negative results on the Hodge conjecture. The negative aspects come on one side from the study of the Hodge conjecture for integral Hodge classes, and on the other side from the study of possible extensions of the conjecture to the general Kahler setting. The positive aspects come from algebraic geometry. They concern the structure of the so-called locus of Hodge classes, and of the Hodge loci.

/pdf/some-aspects-of-the-hodge-conjecture-5ym9lzk3ek.pdf

Some aspects of the Hodge conjecture

After recalling the definition and basic properties of Ulrich bundles, we focus on the existence problem: does every smooth projective variety carry an Ulrich bundle? We show that the Serre construction provides a positive answer on certain surfaces and threefolds.

An introduction to Ulrich bundles

The existence of stable ACM vector bundles of high rank on algebraic varieties is a challenging problem. In this paper, we study stable Ulrich bundles (that is, stable ACM bundles whose corresponding module has the maximum number of generators) on nonsingular cubic surfaces X ⊂ ℙ3. We give necessary and sufficient conditions on the first Chern class D for the existence of stable Ulrich bundles on X of rank r and c1 = D. When such bundles exist, we prove that the corresponding moduli space of stable bundles is smooth and irreducible of dimension D2 - 2r2 + 1 and consists entirely of stable Ulrich bundles (see Theorem 1.1). We are also able to prove the existence of stable Ulrich bundles of any rank on nonsingular cubic threefolds in ℙ4, and we show that the restriction map from bundles on the threefold to bundles on the surface is generically etale and dominant.

/pdf/stable-ulrich-bundles-3irtupu12o.pdf

Stable ulrich bundles

The Abel-Jacobi map of the family of elliptic quintics lying on a general cubic threefold is studied. It is proved that it factors through a moduli component of stable rank 2 vector bundles on the cubic threefold with Chern numbers c_1=0, c_2=2, whose general point represents a vector bundle obtained by Serre's construction from an elliptic quintic. The elliptic quintics mapped to a point of the moduli space vary in a 5-dimensional projective space inside the Hilbert scheme of curves, and the map from the moduli space to the intermediate Jacobian is etale. As auxiliary results, the irreducibility of families of elliptic normal quintics and of rational normal quartics on a general cubic threefold is proved. This implies the uniqueness of the moduli component under consideration. The techniques of Clemens-Griffiths and Welters are used for the calculation of the infinitesimal Abel-Jacobi map.

The Abel-Jacobi map of a moduli component of vector bundles on the cubic threefold

Let M(2;0,3) be the moduli space of rank-2 stable vector bundles with Chern classes c_1=0, c_2=3 on the Fano threefold X, the double solid of index two. We prove that the vector bundles obtained by Serre's construction from smooth elliptic quintic curves on X form an open part of an irreducible component M' of M(2;0,3) and that the Abel-Jacobi map F:M'-->J(X) into the intermediate Jacobian J(X) defined by the second Chern class is generically finite of degree 84 onto a translate of the theta divisor. We also prove that the family of elliptic quintics on a general X is irreducible and of dimension 10.

A. S. Tikhomirov

Papers

The Abel-Jacobi map of a moduli component of vector bundles on the cubic threefold

The Abel-Jacobi map of a moduli component of vector bundles on the cubic threefold

A parametrization of the theta divisor of the quartic double solid

A parametrization of the theta divisor of the quartic double solid