[论文解读] Noether-Lefschetz general complete intersection K3 surfaces over the rationals

Asher Auel, Henry Scheible|arXiv (Cornell University)|Mar 3, 2026

Algebraic Geometry and Number Theory被引用 0

一句话总结

本文证明，在度数分别为 4、6、8 的 Noether–Lefschetz 一般极化 K3 曲面在其模空间中是 Zariski 稠密的，使用度数特定的策略与 Mukai 的 Hodge 同构。

ABSTRACT

We prove that the locus of Noether-Lefschetz general polarized K3 surfaces of degree at most 8 defined over the rational numbers is Zariski dense in the moduli space. Previously, this was proved by van Luijk in the quartic case, and it follows from work of Elsenhans and Jahnel in the degree 2 case. Innovations on their methods, and employing Mukai's Hodge isogeny, suffices to handle the degree 8 case. New methods allow us to deal with the case of degree 6.

研究动机与目标

• Motivate and address whether primitively polarized K3 surfaces of even degree d (d = 4, 6, 8) over Q can have Picard rank 1 over C, i.e., rational points on the Noether–Lefschetz general locus.
• Establish Zariski density of Noether–Lefschetz general K3 surfaces over Q within the moduli spaces K_d for small even degrees (d ≤ 8).
• Extend techniques from quartic (d=4) and degree-2/8 connections to handle degree-6 cases and unify the density results across these complete intersections.

提出的方法

• Utilize specialization of the Néron–Severi group under good reduction and Tate conjecture inputs to bound and determine geometric Picard ranks.
• For degree 8: relate X (three quadrics in P^5) to its discriminant K3 surface Y of degree 2 via Mukai’s Hodge isogeny, showing ρ( X ) = ρ( Y ) in favorable cases.
• For degree 6: exploit projection from a line on a sextic K3 in P^4 to obtain a degree-2 model X → P^2, and compute the branch sextic via a specialized linear-algebra/Gröbner-basis approach to verify Picard rank and lift to characteristic 0.
• For degree 4 and 2: build on van Luijk and Elsenhans–Jahnel strategies by constructing reductions with compatible Picard lattices and employing discriminant/line-tritangent techniques to deduce liftability to Q.
• Provide explicit constructions and verifications (including Gröbner-basis checks) to produce Q-points with Picard rank 1 and use these to deduce density.

实验结果

研究问题

• RQ1 Do primitively polarized K3 surfaces of even degree d (d = 4, 6, 8) defined over Q have Noether–Lefschetz general loci that contain rational points?
• RQ2 Is the locus of Noether–Lefschetz general K3 surfaces over Q Zariski dense in the moduli spaces K_d for d ≤ 8?
• RQ3 Can degree-specific techniques (quartic, degree-2/8 Mukai isogenies, and degree-6 projections) be integrated to establish density across all small even degrees?
• RQ4 How can discriminant K3 surfaces and Mukai’s isogeny be used to detect and preserve Picard rank 1 when lifting from reductions modulo p to characteristic 0?
• RQ5 What explicit constructions over Q yield geometric Picard rank 1 K3 surfaces that drive density arguments?

主要发现

• The set of Noether–Lefschetz general K3 surfaces over Q is Zariski dense in K_d for d = 4, 6, 8.
• Degree 4: density is achieved via lifting and explicit quartic constructions following van Luijk’s method.
• Degree 2/8: Mukai’s Hodge isogeny shows that the degree-8 discriminant K3 surface Y is isogenous to the X’s discriminant, preserving Picard rank under suitable reductions, enabling density arguments.
• Degree 6: a specialized projection-from-a-line method yields a degree-2 model X → P^2 and, with Gröbner-basis verification, constructs X with ρ( X̄ ) = 1, which then lifts to yield density via Prop. 2.3 and the discriminant-dominated approach.
• The combination of explicit mod p constructions, Picard-rank analysis via Frobenius eigenvalues, and Mukai-type isogenies underpins the density results across the three degrees.
• Explicit examples are provided (including a 47-step example for degree 8) to demonstrate the Picard rank computations and lifting behavior.