Development of ATLAS Primary Vertex Reconstruction for LHC Run 3

TL;DR

This paper presents new algorithms for primary vertex reconstruction in the ATLAS experiment, designed to handle increased pile-up in LHC Run 3, improving efficiency, purity, and spatial resolution.

Contribution

Introduction of a Gaussian track density seed finder and an adaptive multi-vertex finder for enhanced vertex reconstruction in high pile-up conditions.

Findings

Improved vertex finding efficiency under high pile-up.

Enhanced vertex purity and spatial resolution.

Successful optimization of new reconstruction algorithms.

Abstract

Increasing luminosity at the Large Hadron Collider (LHC) poses a challenge for primary vertex reconstruction in the ATLAS experiment. A rate of 70 or more inelastic proton-proton collisions per beam crossing was observed during the recently-completed Run 2 and even higher vertex density, or pile-up, is expected in Run 3. To meet this challenge, ATLAS has developed new tools: a Gaussian track density seed finder and an adaptive multi-vertex finder. The former constructs a simple but powerful analytic model of the track density along the beam axis to locate candidate vertices, and the latter applies a global approach to vertex finding and fitting, allowing vertices to compete for nearby tracks. These proceedings document the strategy, optimization and preliminary performance of this new vertex reconstruction software, highlighting improvements in vertex finding efficiency, purity and spatial resolution under Run 3 pile-up conditions.