Exploring High-Energy Proton-Pion Separation in Granular Calorimeters

The development of highly granular calorimeters has opened up new possibilities for particle identification in high-energy physics experiments. These calorimeters provide detailed information about hadronic-shower development, which can be used to distinguish between different types of particles. A recent study has explored the limits of high-energy proton-pion separation in granular calorimeters, using Geant4 simulations and a Deep Sets model to analyze the data.

## Introduction to Granular Calorimeters Granular calorimeters are designed to detect and measure the energy of particles produced in high-energy collisions. They consist of a large number of small cells, each of which is capable of detecting the energy deposited by a particle. The high granularity of these calorimeters allows for detailed studies of hadronic-shower development, which can provide valuable information for particle identification.

## Achieving Accurate Particle Identification The study used Geant4 simulations of isolated particles with energies from 10 to 100 GeV in a homogeneous lead-tungstate calorimeter. The researchers compared the performance of a Deep Sets model operating directly on cell positions and detected energy and time with a boosted decision tree based on reconstructed shower observables. The results showed that the Deep Sets model outperformed the boosted decision tree, achieving an accuracy of 93.8% at 10 GeV and 67.2% at 100 GeV.

## Factors Affecting Particle Identification The study also investigated the factors that affect particle identification in granular calorimeters. The results showed that shower topology is independently informative, deposited energy provides the largest additional contribution, and timing supplies complementary information. The researchers also found that coarser segmentation reduces discrimination, with performance more sensitive to longitudinal than transverse granularity. The following table summarizes the accuracy of the Deep Sets model at different energies: | Energy | Accuracy | | --- | --- | | 10 GeV | 93.8% | | 100 GeV | 67.2% |

## Conclusion and Future Directions The results of this study provide an encouraging benchmark for calorimeter-based hadron identification and motivate its inclusion among the optimization targets for future highly granular calorimeters. The use of Deep Sets models and other machine learning techniques may enable more accurate particle identification, which could have a significant impact on high-energy physics experiments. Further research is needed to fully explore the potential of granular calorimeters for particle identification and to develop new technologies that can take advantage of their capabilities.