We describe a method to show that certain elliptic surfaces do not admit purely inseparable multisections (equivalently, that genus one curves over function fields admit no points over the perfect closure of the base field) and use it to show that an
y non-Jacobian elliptic structure on a very general supersingular K3 surface has no purely inseparable multisections. We also describe specific examples of such fibrations without purely inseparable multisections. Finally, we discuss the consequences for the claimed proof of the Artin conjecture on unirationality of supersingular K3 surfaces.
We develop a theory of twistor spaces for supersingular K3 surfaces, extending the analogy between supersingular K3 surfaces and complex analytic K3 surfaces. Our twistor spaces are obtained as relative moduli spaces of twisted sheaves on universal g
erbes associated to the Brauer groups of supersingular K3 surfaces. In rank 0, this is a geometric incarnation of the Artin-Tate isomorphism. Twistor spaces give rise to curves in moduli spaces of twisted supersingular K3 surfaces, analogous to the analytic moduli space of marked K3 surfaces. We describe a theory of crystals for twisted supersingular K3 surfaces and a twisted period morphism from the moduli space of twisted supersingular K3 surfaces to this space of crystals. As applications of this theory, we give a new proof of the Ogus-Torelli theorem modeled on Verbitskys proof in the complex analytic setting and a new proof of the result of Rudakov-Shafarevich that supersingular K3 surfaces have potentially good reduction. These proofs work in characteristic 3, filling in the last remaining gaps in the theory. As a further application, we show that each component of the supersingular locus in each moduli space of polarized K3 surfaces is unirational.
We prove that a twisted Enriques (respectively, untwisted bielliptic) surface over an algebraically closed field of positive characteristic at least 3 (respectively, at least 5) has no non-trivial Fourier-Mukai partners.
