$q$-bic hypersurfaces and their Fano schemes
Raymond Cheng
TL;DR
$q$-bic hypersurfaces in positive characteristic are studied by treating them as moduli spaces of isotropic lines for a canonical $q$-bic form $eta$, enabling a moduli-theoretic approach to their Fano schemes. The paper develops a Hermitian-structure framework and relates the Fano schemes to Deligne–Lusztig theory for finite unitary groups, yielding explicit Betti numbers, canonical bundles, and a rich stratification of the $m$-plane Fano variety; in the odd-dimensional case these schemes are irreducible, smooth, of general type, and admit a purely inseparable complete-intersection cover. For $m=1$ a Clemens–Griffiths–type analogue emerges: the Albanese of the Fano surface of lines is essentially the intermediate Jacobian of $X$, via a purely inseparable Abel–Jacobi map, with deeper links to Prym-type structures in positive characteristic. Altogether, the work forges deep connections between $q$-bic geometry, Deligne–Lusztig theory, unitary groups, and algebraic representatives, providing explicit structural results and guiding principles for further study in characteristic $p>0$.
Abstract
A $q$-bic hypersurface is a hypersurface in projective space of degree $q+1$, where $q$ is a power of the positive ground field characteristic, whose equation consists of monomials which are products of a $q$-power and a linear power; the Fermat hypersurface is an example. I identify $q$-bics as moduli spaces of isotropic vectors for an intrinsically defined bilinear form, and use this to study their Fano schemes of linear spaces. Amongst other things, I prove that the scheme of $m$-planes in a smooth $(2m+1)$-dimensional $q$-bic hypersurface is an $(m+1)$-dimensional smooth projective variety of general type which admits a purely inseparable covering by a complete intersection; I compute its Betti numbers by relating it to Deligne--Lusztig varieties for the finite unitary group; and I prove that its Albanese variety is purely inseparably isogenous via an Abel--Jacobi map to a certain conjectural intermediate Jacobian of the hypersurface. The case $m = 1$ may be viewed as an analogue of results of Clemens and Griffiths regarding cubic threefolds.
