Cubic fourfolds containing highly singular hyperplane sections
Lisa Marquand, Sasha Viktorova · 2026
Read Cubic fourfolds containing highly singular hyperplane sections by Lisa Marquand, Sasha Viktorova (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-20778.txt.
Introduction
Lisa Marquand and Sasha Viktorova construct new families of a classic object in algebraic geometry. Their 2026 paper studies cubic fourfolds, shapes cut out by a cubic equation. The authors focus on smooth cubic fourfolds that contain highly singular hyperplane sections. A hyperplane section is the shape left when a flat cut passes through.
The pair construct five irreducible divisors in the moduli space of such fourfolds. Each divisor parametrizes fourfolds with one of these singular sections. They then prove that none of the five is a Noether-Lefschetz or Hassett divisor. The proof uses a computational method due to Addington and Auel. The result adds new landmarks to the map of how cubic fourfolds can vary.
Catalog record
| CREATOR | Lisa Marquand, Sasha Viktorova |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-20778
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Cubic fourfolds containing highly singular hyperplane sections Lisa Marquand, Sasha Viktorova Abstract . We construct five irreducible divisors in the moduli space of complex cubic fourfolds parametrising smooth cubic fourfolds that contain highly singular hyperplane sections. We prove that each is not a Noether-Lefschetz (or Hassett) divisor utilising the computational method developed by Addington–Auel. 1. Introduction Divisors in the moduli space of smooth cubic fourfolds, denoted \mathcal{C} , have received a lot of recent attention, due to their connection with rationality constructions. Indeed, the recent work of [ KKPY25 ] implies that the locus of rational cubic fourfolds is contained in the union of the Hassett divisors , a countable union of irreducible divisors that parametrise cubic fourfolds containing a surface non-homologous to a complete intersection. These divisors were introduced by Hassett in [ Has00 ] , and they parametrise so-called special cubic fourfolds. It is a complementary problem to find explicit equations for cubic fourfolds, away from these Hassett divisors, where the theory of Hodge atoms of [ KKPY25 ] guarantees irrationality. To our knowledge, there are few known non-Hassett divisors in the moduli space \mathcal{C} . Ranestand–Voisin constructed three divisors D_{V-ap},D_{rk3},D_{IR} parametrising cubic fourfolds apolar to certain surfaces, and a fourth divisor D_{copl} parametrising coplanar cubics. They proved that D_{V-ap} was non-Hassett by carefully analysing its singularities [ RV17 ] . On the other hand, Addington–Auel used computational techniques to prove that D_{rk3} and D_{IR} were non-Hassett as well [ AA18 ] . Further, Costa–Harvey–Kedlaya used p -adic methods to show the same conclusion holds for D_{copl} [ CHK19 ] . More recently, Sammarco used techniques similar to [ RV17 ] to identify another non-Hassett divisor [ Sam25 ] . It is shown in [ RV17 ] that D_{V-ap} is different from D_{rk3} . However, it is not known if the other non-Hassett divisors, including the ones that we define in this work, are different. In this paper, we consider smooth cubic fourfolds that have a hyperplane section that is more singular than expected. We define D_{E_{6}},D_{D_{6}}\subset\mathcal{C} to be the locus parametrising smooth cubic fourfolds containing a hyperplane section Y:=X\cap H with an E_{6} or D_{6} singularity, respectively. We prove: Theorem 1.1 . The loci D_{E_{6}},D_{D_{6}}\subset\mathcal{C} are irreducible divisors in \mathcal{C}. Further, both divisors are not Hassett divisors. In fact, one can write down the general form of an equation for a cubic contained in either D_{E_{6}},D_{D_{6}} (see Proposition 2.8 and Equation 3.1 ), and thus applying [ KKPY25 ] we can write down examples of irrational cubic fourfolds. If X is a general cubic fourfold, then necessarily all hyperplane sections have total Tjurina number (or total Milnor number) \leq 5 [ LSV17 , Lemma 3.8] . Thus it is natural to consider hyperplane sections with total Tjurina number 6. A hyperplane section with six nodes is such an example, however cubic fourfolds with such a hyperplane section are contained in either the Hassett divisor \mathcal{C}_{12} (containing a cubic scroll) by [ HT10 ] , or \mathcal{C}_{8} (containing a plane) if the singularities are not in general position. The condition for a cubic threefold Y (necessarily singular) to contain either a cubic scroll or a plane was further studied in [ MV25 ] , where we showed this condition is equivalent to the positivity of the defect : \sigma(Y):=b_{2}(Y)-b_{4}(Y)>0 . The defect \sigma(Y) is a global invariant of the cubic threefold, which can be computed using both the position of the singularities and the local numerical invariants. In particular, in order to obtain divisors away from \mathcal{C}_{8}\cup\mathcal{C}_{12}, we restrict our attention to cubic fourfolds containing singular hyperplane sections with \sigma(Y)=0. With that in mind, we can consider loci D_{K}\subset\mathcal{C} parametrising cubic fourfolds with a hyperplane section Y with multiple singularities of type K=\sum_{i=1}^{r}K_{i} , with each K_{i} of ADE type, and total Tjurina number 6. One would expect to find more non-Hassett divisors this way, as long as \sigma(Y)=0 . We prove this is the case: Theorem 1.2 . The loci D_{D_{5}+A_{1}},D_{D_{4}+A_{2}}\subset\mathcal{C} are irreducible divisors, and D_{D_{4}+2A_{1}} has two irreducible divisoral components D_{D_{4}+2A_{1}}^{0},D_{D_{4}+2A_{1}}^{1}, indexed by the value of \sigma(Y) . Further, the divisors D_{D_{5}+A_{1}},D_{D_{4}+A_{2}},D_{D_{4}+2A_{1}}^{0} are non-Hassett divisors, while D_{D_{4}+2A_{1}}^{1}=\mathcal{C}_{8}. We will also consider the locus D_{T_{333}}\subset\mathcal{C} that parametrises smooth cubic fourfolds containing a hyperplane section Y with a T_{333} singularity. This is a unimodular singularity of corank 3 and Tjurina number 8 . One then expects this locus to be codimension 2 in moduli, which we verify: Proposition 1.3 . The locus D_{T_{333}}\subset\mathcal{C} is irreducible of codimension 2. Further, it is not contained in a Hassett-divisor. Our interest in singularities of hyperplane sections of a cubic fourfold X\subset\mathbb{P}^{5} also stems from the study of compactified intermediate Jacobian fibrations \pi:\mathcal{J}_{X}\rightarrow(\mathbb{P}^{5})^{*} , first constructed in [ LSV17 ] for general X . These are examples of hyperkähler manifolds of OG10 type, and for a general H\in(\mathbb{P}^{5})^{*} the fiber of \pi is the intermediate Jacobian of the cubic threefold X\cap H . Although existence of such a hyperkähler compactification is guaranteed for any cubic by results of Saccá [ Sac23 ] , it is non constructive. Recently, [ DM26 ] extend the construction of [ LSV17 ] to very good cubic fourfolds - cubics not containing a plane, a cubic scroll, or admitting a hyperplane section with a corank 3 singularity. Since the locus of cubic fourfolds admitting a hyperplane section with a corank 3 singularity is the closure of D_{T_{333}} (see Remark 2.9 ), our results imply the following: Corollary 1.4 . A cubic fourfold X\in\mathcal{C}\setminus(\mathcal{C}_{8}\cup\mathcal{C}_{12}\cup D_{T_{333}}) is very good. In particular, for a cubic X\in\mathcal{C}\setminus(\mathcal{C}_{8}\cup\mathcal{C}_{12}\cup D_{T_{333}}) , the construction of [ DM26 ] produces a smooth projective OG10 -type hyperkähler \mathcal{J}_{X} with a Lagrangian fibration \pi:\mathcal{J}_{X}\rightarrow(\mathbb{P}^{5})^{*} whose fibers are compactified Prym varieties of reduced, irreducible planar curves. Outline We recall necessary background on cubic fourfolds, and singular cubic threefolds in Section 2 . In Section 3 , we prove that the divisors described above are indeed irreducible divisors in the moduli of smooth cubic fourfolds. We obtain results on the deformations of singular hyperplane sections which may be of independent interest (Proposition 3.3 and Lemma 3.5 ). We complete the proofs of Theorems 1.1 and 1.2 in Section 4 by utilising the computational methods of Addington–Auel. To implement their method to prove our divisors are non Hassett, we use code developed in [ AKPW25 , AKPW24 ] . Funding The first author was supported by NSF grant DMS-2503390. 2. Preliminaries Throughout, X\subset\mathbb{P}^{5} denotes a smooth cubic fourfold defined over \mathbb{C}. We denote by \mathcal{C} the moduli space of smooth cubic fourfolds, constructed using GIT [ Laz09 ] . Recall the following definition [ Has00 ] : Definition 2.1 . A cubic fourfold X is special if it admits an algebraic cycle of codimension 2 not homologous to a complete intersection. Equivalently, a smooth cubic fourfold X is not special if and only if the algebraic lattice A(X):=H^{2,2}(X)\cap H^{4}(X,\mathbb{Z}) is generated by the class h^{2} , where h\in H^{1,1}(X) is the hyperplane class. Our goal is to construct divisors in \mathcal{C} parametrising geometrically relevant but not special cubic fourfolds. We will do so by parametrising cubics containing a hyperplane section with specific singularities. For a general hyperplane section Y:=X\cap H