Kodaira dimension of moduli of special cubic fourfolds (Q2001378)

A cubic fourfold is a smooth complex hypersurface \(X \subset \mathbb P^5\) of degree four. Two such are isomorphic exactly when projectively equivalent, so the coarse moduli space \(\mathcal C\) of isomorphism classes of cubic fourfolds can be constructed as a GIT quotient and is irreducible of dimension 20. \textit{B. Hassett} pioneered a systematic study of \textit{special} cubic fourfolds \(X\), those containing an algebraic surface \(T\) not homologous to a complete intersection [Compos. Math. 120, No. 1, 1--23 (2000; Zbl 0956.14031)]. He showed that isomorphism classes of special cubic fourfolds are parametrized by a countable union of irreducible divisors \(\mathcal C_d \subset \mathcal C\), each determined by the discriminant \(d\) of a saturated lattice \(K \subset A(X) \cong H^{2,2} (X) \cap H(X, \mathbb Z)\) containing \(h^2\), where \(h\) is the hyperplane section; furthermore, \(\mathcal C_d\) is nonempty if and only if \(d > 6\) and \(d \equiv 0\) or \(2\) modulo \(6\). The authors investigate the Kodaria dimension \(\kappa (\mathcal C_d)\) of the spaces \(\mathcal C_d\). When \(d=6n+2\), the main theorem states that \(\kappa (\mathcal C_d) > 0\) if \(n > 18\) and \(n \neq 20,21,25\) and \(\kappa (\mathcal C_d) \geq 0\) if \(n>13\) and \(n \neq 15\). Likewise when \(d=6n\), \(\kappa (\mathcal C_d) > 0\) if \(n > 18\) and \(n \neq 20,22,23,25,30,32\) and \(\kappa (\mathcal C_d) \geq 0\) if \(n > 13\) and \(n \neq 18,20,22,30\). Their method uses the ``low weight cusp form trick'' developed by \textit{V. A. Gritsenko} et al. [Invent. Math. 169, No. 3, 519--567 (2007; Zbl 1128.14027)] to reduce the problem to finding a low weight cusp form on an associated orthogonal modular variety. Using ideas of \textit{S. Kondō} [Compos. Math. 116, No. 2, 111--117 (1999; Zbl 0948.14007)] and \textit{V. Gritsenko} et al. [in: Handbook of moduli. Volume I. Somerville, MA: International Press; Beijing: Higher Education Press. 459--526 (2015; Zbl 1322.14004)], this problem is then reduced to the existence of primitive lattice embeddings into the lattice introduced by \textit{R. E. Borcherds} [Invent. Math. 120, No. 1, 161--213 (1995; Zbl 0932.11028)]. A computer is used for smaller values of \(d\). Combined with earlier work and recent results of \textit{H. Nuer} [Lect. Notes Math. 2172, 161--167 (2016; Zbl 1371.14052)] and \textit{K.-W. Lai} [``New cubic fourfolds with odd degree unirational parametrization'', Preprint, \url{arXiv:1606.03853}] showing that \(\kappa (\mathcal C_d) < 0\) for \(d \leq 44\), there are 20 values of \(d\) for which we have no information about \(\kappa (\mathcal C_d)\).