[Paper Review] On elliptic K3 surfaces

Ichiro Shimada|ArXiv.org|May 9, 2005

Algebraic Geometry and Number Theory2 references6 citations

TL;DR

This paper presents a complete classification of all possible ADE-types of singular fibers and torsion parts of Mordell-Weil groups for complex elliptic K3 surfaces using lattice-theoretic methods, particularly Nikulin's theory of discriminant forms. It generates a list of 3,693 realizable pairs $(\Sigma, G)$, with explicit rules for deriving all configurations from a minimal set via elementary transformations, and establishes that nontrivial torsion arises only under restricted transformation rules.

ABSTRACT

We classify all the possible configurations of singular fibers and the torsion parts of Mordell-Weil groups of complex elliptic K3 surfaces. The complete list of 3279 configurations is attached.

Motivation & Objective

• To determine all possible ADE-types of singular fibers and torsion parts of Mordell-Weil groups for complex elliptic K3 surfaces.
• To provide a complete, algorithmically generated list of 3,693 realizable pairs $(\Sigma, G)$, where $\Sigma$ is an ADE-type and $G$ is a finite abelian group.
• To establish transformation rules that generate all such realizable pairs from a minimal set $\mathcal{S}$, enabling systematic construction.
• To clarify the geometric distinction between configurations with trivial and nontrivial Mordell-Weil torsion via elementary transformations.

Proposed method

• Leverages Nikulin’s theory of discriminant forms of even integral lattices to analyze the Néron-Severi lattice of elliptic K3 surfaces.
• Applies Kondo-Nishiyama’s lemma to relate the Néron-Severi lattice structure to the singular fiber configuration and Mordell-Weil torsion.
• Uses computer-aided computation in Maple V to systematically enumerate all valid combinations of ADE-types and torsion groups.
• Employs a criterion for the existence of even lattices with given signature and discriminant form, adapted for machine computation.
• Defines a generating set $\mathcal{S}$ of 336 pairs corresponding to extremal K3 surfaces and applies transformation rules to generate the full list $\mathcal{P}$.
• Applies elementary transformations (vertex deletions in Dynkin diagrams) to generate all non-extremal configurations, with restrictions for nontrivial torsion.

Experimental results

Research questions

• RQ1 Which ADE-types of singular fibers can occur on complex elliptic K3 surfaces with a given torsion part of the Mordell-Weil group?
• RQ2 What is the complete set of all possible pairs $(\Sigma, G)$, where $\Sigma$ is an ADE-type and $G$ is the torsion subgroup of the Mordell-Weil group, realizable on an elliptic K3 surface?
• RQ3 How can all such realizable configurations be generated from a minimal set using simple lattice-theoretic transformations?
• RQ4 What are the restrictions on elementary transformations (vertex deletions in Dynkin diagrams) that preserve realizability when the Mordell-Weil torsion is nontrivial?
• RQ5 Why is the set of ADE-types with trivial Mordell-Weil torsion precisely the closure of extremal types under elementary transformations?

Key findings

• The complete list $\mathcal{P}$ contains 3,693 realizable pairs $(\Sigma, G)$ of ADE-types and Mordell-Weil torsion groups for elliptic K3 surfaces.
• An ADE-type $\Sigma$ with trivial Mordell-Weil torsion arises if and only if it is obtained from an extremal elliptic K3 surface with trivial torsion via elementary transformations (vertex deletions in Dynkin diagrams).
• For nontrivial torsion, certain elementary transformations are forbidden; the list $\mathcal{P}$ is generated from a minimal set $\mathcal{S}$ of 336 extremal pairs using transformation rules that exclude these forbidden cases.
• The list includes configurations with torsion groups $\mathbb{Z}/n\mathbb{Z}$ for $n=4,5,6,7,8$ and $\mathbb{Z}/2\mathbb{Z} \times \mathbb{Z}/2\mathbb{Z}$, $\mathbb{Z}/4\mathbb{Z} \times \mathbb{Z}/2\mathbb{Z}$, $\mathbb{Z}/6\mathbb{Z} \times \mathbb{Z}/2\mathbb{Z}$, and $\mathbb{Z}/3\mathbb{Z} \times \mathbb{Z}/3\mathbb{Z}$, with specific ADE-type realizations.
• The paper provides an algorithmic framework to generate all 3,693 pairs from a small seed set using explicit transformation rules, accessible via the author’s website.
• The study confirms that the geometric realizability of a pair $(\Sigma, G)$ is decidable via purely lattice-theoretic computation, using discriminant form invariants and signature constraints.