Toric Elliptic Pairs with Picard Number Three

TL;DR

This paper classifies certain elliptic pairs on rational surfaces with Picard number three, showing that Lang-Trotter quadrilaterals do not exist and analyzing the structure of these pairs.

Contribution

It proves the non-existence of Lang-Trotter quadrilaterals among elliptic pairs with Picard number three and studies the Zariski decomposition of related divisors.

Findings

Lang-Trotter quadrilaterals do not exist in this setting.

Classification of elliptic pairs with Picard number three.

Analysis of the Zariski decomposition of $K_X+C$.

Abstract

An elliptic pair $(X, C)$ is a generalization of a rational elliptic fibration $X \to \mathbb{P}^1$ with fiber $C,$ introduced in \cite{jenia_blowup}. Here, $X$ is a projective rational surface with log terminal singularities, and $C$ is an irreducible curve contained in the smooth locus of $X,$ with $p_a(C)=1$ and $C^2=0.$ These naturally arise as blowups $X:=\text{Bl}_e(\mathbb{P}_\Delta)$ of projective toric surfaces, whose Newton polygon is elliptic. The order of $\mathcal{O}(C)|_C$ in $\text{Pic}^0(C)$ gives a quantitative way to check if $X$ is an elliptic fibration, which is equivalent to finiteness of the order. We call $\Delta$ a Lang-Trotter polygon when this order is infinite, in which case $\overline{\text{Eff}(\text{Bl}_e(\mathbb{P}_\Delta))}$ is non-polyhedral. The paper \cite{lizzie} shows there are exactly $3$ elliptic triangles up to $\text{SL}_2(\mathbb{Z}),$ none of which is Lang-Trotter. The paper \cite{jenia_blowup} gives an infinite family of Lang-Trotter pentagons and heptagons, and various examples of other polygons when $\rho(\mathbb{P}_\Delta)>2.$ Remark 4.7 in the paper asks if any Lang-Trotter quadrilaterals exist, and we answer this in the negative by studying the curves in the Zariski Decomposition of $K_X+C.$