Triggering on Emerging Jets

TL;DR

This paper proposes new trigger strategies for detecting emerging jets from confining dark sectors at the LHC, improving detection efficiency using initial state radiation and machine learning-based hit count triggers.

Contribution

It introduces novel trigger methods leveraging initial state radiation and machine learning to enhance sensitivity to emerging jets from dark sectors at the LHC.

Findings

Trigger efficiency improves with additional jets simulation.

Machine learning triggers have low background rates.

New triggers are sensitive across various dark sector parameters.

Abstract

Confining dark sectors at the GeV scale can lead to novel collider signatures including those termed emerging jets with large numbers of displaced vertices. The triggers at the LHC experiments were not designed with this type of new physics in mind, and triggering can be challenging, especially if the mediator is relatively light and/or has quantum numbers such that additional jets are not automatically produced in each event. We show that the efficiency and the total event rate at current triggers can be significantly improved by considering initial state radiation of the events, with the largest increase in rate coming from simulation of two additional jets. We also explore possible new triggers that employ hit counts in different tracker layers as input into a machine learning algorithm. We show that these new triggers can have reasonably low background rates, and that they are sensitive to a wide range of new physics parameters even when trained on a single model.