Motives of moduli spaces on \(K3\) surfaces and of special cubic fourfolds (Q2285838)

The first part of the paper deals with smooth projective moduli spaces of Gieseker stable sheaves on a given complex projective \(K3\) surface \(S\). The goal is to prove that for such a space \(M\), of dim \(m\), it is the case that the associated Chow motive \(\mathfrak{h}(M)\) can be realized as a direct summand of a motive \(\bigoplus \mathfrak{h}(S^{k_i})(n_i)\) with \(k_i\le m \). For curves a similar result is due to \textit{S. del Baño} [J. Reine Angew. Math. 532, 105--132 (2001; Zbl 1044.14005)]. The proof depends on Grothendieck Riemann-Roch computations. One basic input for this task comes from \textit{E. Markman} [J. Reine Angew. Math. 544, 61--82 (2002; Zbl 0988.14019)], where the class of the diagonal on \(M \times M\) is identified as being \(c_m (W)\) of a certain natural complex \(W\), associated with the universal sheaf. The author notes that his method also applies to moduli spaces of twisted sheaves and to moduli spaces of stable objects in \(\text{D}^{\text{b}}(S,\alpha )\), for a Brauer class \(\alpha \in \text{Br}(S)\). Given a variety \(X\), one defines \(\mathrm{Mot}(X)\) to be the full pseudo-abelian tensor subcategory of motives generated by \(\mathfrak{h}(X)\) and by the Lefschetz motive \({\mathbb{L}}\). The result above then signifies that \(\mathfrak{h}(M)\) belongs to \(\mathrm{Mot}(S)\), one says that \(\mathfrak{h}(M)\) is motivated by \(\mathfrak{h}(S)\). The author expects that the converse is also true. The second part of the paper considers the special cubic fourfolds \(X\) which lie in those Noether Lefschetz loci \(\mathcal C_d\) which are treated in [\textit{D. Huybrechts}, Compos. Math. 153, No. 3, 586--620 (2017; Zbl 1440.14180)]. There it is proved, at least generically, that the Kuznetsov component \(\mathcal{A} (X)\) is exact equivalent to \( \mathrm{D}^{\mathrm{b}}(S)\) or to \(\mathrm{D}^{\mathrm{b}}(S, \alpha )\) for some \(K3\) surface \(S\). The main result here states that the transcendental motive of \(X\) is isomorphic to the transcendental motive of some \(K3\) surface \(S\) and that there is a decomposition \( \mathfrak{h}(X) \cong\mathrm{1} \oplus \mathfrak{h}(S)(1) \oplus {\mathbb{L}}^2 \oplus {\mathbb{L}}^4 \). The proof is given first under the hypothesis that the categorical condition on \(\mathcal{A} (X)\) holds, so for \(X\) generic in \(\mathcal C_d\). Specialization yields the result for every \(X\) in \(\mathcal C_d\). The author remarks that his expectation \(\mathrm{Mot}(S) =\mathrm{Mot}(M) \) is verified in this setting, because the Fano variety \(F\) turns out to be birational to a moduli space for \(S\), and therefore one has \(\mathrm{Mot}(M) =\mathrm{Mot}(F) =\mathrm{Mot}(X) = \mathrm{Mot}(S)\).