Searching for EeV photons with Telescope Array Surface Detector and neural networks

TL;DR

This paper presents updated limits on ultra-high-energy photon flux using Telescope Array data, employing neural networks to distinguish photon-induced events from hadronic backgrounds over 14 years.

Contribution

It introduces a neural network classifier trained with experimental data to improve photon-hadron discrimination in ultra-high-energy cosmic ray observations.

Findings

No excess photon candidates found, consistent with background expectations.

Established upper limits on photon flux at energies above 10^{19} eV and 10^{20} eV.

Demonstrated effectiveness of neural networks in astrophysical particle identification.

Abstract

Ultra-high-energy photons play an important role in probing astrophysical models and beyond-Standard-Model scenarios. We report updated limits on the diffuse photon flux using Telescope Array's Surface Detector data collected over 14 years of operation. Our method employs a neural network classifier to effectively distinguish between proton-induced and photon-induced events. The input data include both reconstructed composition-sensitive parameters and raw time-resolved signals registered by the Surface Detector stations. To mitigate biases from Monte Carlo simulations, we fine-tune the network with a subset of experimental data. The number of observed photon candidates is found to be consistent with the expected hadronic background, yielding upper limits on photon flux $\Phi_\gamma(E_\gamma > 10^{19} \text{eV}) < 2.3 \cdot 10^{-3} $, and $\Phi_\gamma(E_\gamma > 10^{20} \text{eV}) < 3.0 \cdot 10^{-4} $ $ (\text{km}^2 \cdot \text{sr} \cdot \text{yr})^{-1} $.