On lattice-polarized K3 surfaces
Valery Alexeev, Philip Engel
TL;DR
The paper corrects and refines the framework for lattice-polarized and lattice-quasipolarized K3 surfaces by replacing Weyl-chamber data with generalized 'small cones' to stabilize period-map fibers. It extends the theory to ADE K3 surfaces, proving separated moduli stacks and identifying coarse moduli spaces with period-domain quotients D_Λ/Γ, while showing independence from the chosen very irrational vector within a small cone. The approach unifies smooth and ADE cases, clarifies non-separation phenomena, and provides a robust moduli-theoretic foundation likely adaptable to K-trivial and hyperkähler varieties. Key contributions include corrected definitions (mqpol), a detailed small-cone analysis, and a separated moduli-theoretic description for ADE K3 surfaces. The results have broad implications for applications and future generalizations in the realm of K3-type and hyperkähler moduli.
Abstract
We propose modifications to the commonly used definitions of lattice-polarized and lattice-quasipolarized smooth K3 surfaces, collecting various versions of the definition, and determining the effects of these choices on the resulting moduli space. We fill a gap in the theory, by replacing Weyl chambers with the new notion of a ``small cone'': the true datum in the definition of lattice quasipolarized K3 surfaces. In addition, we describe the separated moduli stack and moduli space for lattice-polarized K3 surfaces with $ADE$ singularities, an important notion for applications.
