Canonical surfaces with big cotangent bundle (Q2454274): Difference between revisions

Let \(X\) be a smooth projective surface defined over the complex numbers, and let \(\Omega_X\) be its cotangent bundle. We say that the cotangent bundle \(\Omega_X\) is big if the tautological bundle on the projectivised bundle \(\mathbb P(\Omega_X)\) is a big line bundle. Manifolds with big cotangent bundle play an interesting role for hyperbolicity problems since they can contain entire curves but the Green-Griffiths-Lang conjecture holds [\textit{M. McQuillan}, Publ. Math., Inst. Hautes Étud. Sci. 87, 121--174 (1998; Zbl 1006.32020)]. A sufficient condition for the bigness of the cotangent bundle is the positivity of the second Segre class \(s_2 := c_1^2-c_2\), however this condition is far from being necessary. In the paper under review, the authors give a new sufficient condition for the cotangent bundle to be big and show that this leads to a large number of new examples. More precisely let \(X\) be a normal projective surface with ample canonical divisor having at most du Val singularities. Since du Val singularities are quotient singularities, one can consider the orbifold (stack) \(\mathcal X \rightarrow X\). Moreover we can consider the minimal resolution \(Y \rightarrow X\) which is a smooth projective surface with nef and big canonical bundle. The main theorem states that if \(s_2(Y)+s_2(\mathcal X)>0\), then the cotangent bundle of \(Y\) is big. The authors then proceed to constructing several series of surfaces where \(s_2(Y)\) is negative, but \(s_2(Y)+s_2(\mathcal X)>0\). For example they prove that if \(X \subset \mathbb P^3\) is a hypersurface of degree \(d\) with \(l\) singularities of type \(A_k\) such that \[ l > \frac{4(k+1)}{k(k+2)} (2d^2-5d), \] then the minimal resolution of \(X\) has a big cotangent bundle. This generalises and corrects an earlier paper of \textit{F. Bogomolov} and \textit{B. De Oliveira} [J. Reine Angew. Math. 596, 89--101 (2006; Zbl 1108.14013)]. The proof of the main theorem is based on Riemann-Roch calculations on orbifolds and extensions results for symmetric differentials with logarithmic poles.