Bloch's conjecture for Catanese and Barlow surfaces (Q2509563)

In the paper under review the author proves the Bloch conjecture for two classes of surfaces of general type, namely, \textit{F. Catanese} [Invent. Math. 63, 433--465 (1981; Zbl 0472.14024)] and \textit{R. Barlow} [Invent. Math. 79, 293--301 (1985; Zbl 0561.14015)] surfaces. To do so she establishes an improved version, in the surface case, of the main theorem of [\textit{C. Voisin}, Ann. Sci. Éc. Norm. Supér. (4) 46, No. 3, 449--475 (2013; Zbl 1282.14015)]. The Bloch Conjecture for surfaces with \(p_g(S)=0\) predicts that \[ \mathrm{CH}^0(S)_{\mathrm{alb}} := \ker (\mathrm{CH}^0(S)_{\mathrm{hom}} \longrightarrow \mathrm{Alb}(S ))= 0. \] This statement is known to hold for surfaces which are not of general type, therefore it is interesting to prove it for surfaces of general type. Two classes of such surfeces are the so called Catanese surfaces and Barlow surfaces. The Catanese surfaces where introduced by \textit{F. Catanese} in [Invent. Math. 63, 433--465 (1981; Zbl 0472.14024)] to give a counterexample of Babbage's conjecture. Moreover, they specialize to double covers of Barlow surfaces. A Catanese surface is a numerical Campedelli surfaces which is constructed in the following way: start from a \(5 \times 5\) symmetric matrix \(M(a)\), \(a \in \mathbb{P}^{11}\), of linear forms on \(\mathbb{P}^3\) satisfying certain conditions making their discriminant invariant under the Godeaux action of \(\mathbb{Z}/5\mathbb{Z}\) on \(\mathbb{P}^3\). The general quintic surface \(V(a)\) defined by the determinant of \(M(a)\) has \(20\) nodes corresponding to the points \(x \in \mathbb{P}^3\) where the matrix \(M(a,x)\) has rank \(3\), and it admits a double cover \(S(a)\) which is étale away from the nodes, and to which the \(\mathbb{Z}/5\mathbb{Z}\)-action lifts. Then the Catanese surface \(\Sigma(a)\) is \(S(a)/(\mathbb{Z}/5\mathbb{Z})\) by this lifted action. The idea of the proof is the following. In a more general setting let \(\mathcal{S} \longrightarrow B\) be a smooth projective morphism with two dimensional connected fibers, where \(B\) is quasi-projective. Let \(\Gamma \in \mathrm{CH}^2(\mathcal{S} \times_B \mathcal{S})_{\mathbb{Q}}\) be a codimension 2 cycle, which provides a relative 0-self correspondence of \(\mathcal{S}\) over \(B\). For \(t \in B\), let \(\Gamma_t := \Gamma|_{\mathcal{S}_t\times \mathcal{S}_t}\) be the restricted cycle, with cohomology class \([\Gamma_t] \in H^4(\mathcal{S}_t \times \mathcal{S}_t, \mathbb{Q})\). We have the actions \[ \Gamma_{t*} \colon \mathrm{CH}^0(S_t)_{\mathbb{Q}} \longrightarrow \mathrm{CH}^0(S_t)_{\mathbb{Q}}, \quad [\Gamma_t]_* \colon H^{i,0}(\mathcal{S}_t) \longrightarrow H^{i,0}(\mathcal{S}_t). \] First the author proves the following theorem. Assume the following: {\parindent=6mm \begin{itemize}\item[1.] The fibers \(\mathcal{S}_t\) satisfy \(h^{1,0}(\mathcal{S}_t) = 0\) and \([\Gamma_t]_* \colon H^{2,0}(\mathcal{S}_t) \longrightarrow H^{2,0}(\mathcal{S}_t)\) is equal to zero. \item[2.] A non-singular projective (equivalently any non-singular projective) completion \(\overline{\mathcal{S} \times_B \mathcal{S}}\) of the fibered self-product \(\mathcal{S} \times_B \mathcal{S}\) is rationally connected. \end{itemize}} Then \[ \Gamma_{t*} \colon \mathrm{CH}^0(\mathcal{S}_t)_{\mathrm{hom}} \longrightarrow \mathrm{CH}^0(\mathcal{S}_t)_{\mathrm{hom}} \] is nilpotent for any \(t \in B\). Next, the author shows that the nilpotence property is sufficient to imply the vanishing (hence Bloch's conjecture) in a number of situations, in particular for Catanese surfaces. Indeed, she proves that, let \(\Sigma\) be a Catanese surface, then \(\mathrm{CH}^0(\Sigma) =\mathbb{Z}\). Finally, a Barlow surface \(\Sigma'\) has \(\mathrm{CH}^0\) equal to \(\mathbb{Z}\) too. Since the Barlow surface \(\Sigma'\) considered is a quotient of the Catanese surface \(\Sigma\) by an involution, hence \(\mathrm{CH}^0(\Sigma') \hookrightarrow \mathrm{CH}^0(\Sigma)\) since \(\mathrm{CH}^0(\Sigma')\) has no torsion. In the last section of the paper she will also give a conditional application (more precisely, assuming the variational Hodge conjecture) of the same method to the Chow motive of low degree \(K3\) surfaces.