Speaker
Description
Dual-readout calorimeters are designed to achieve precise hadronic energy reconstruction by simultaneously measuring scintillation and Cherenkov signals, providing sensitivity to the electromagnetic and hadronic components of particle showers. To ensure a stable detector response, high-granularity readout is employed, enabling uniform sampling of the shower and access to its spatial structure. This spatial information supports particle identification (PID), including topological discrimination such as the separation of neutral pions from single photons. While relying on an unsegmented longitudinal structure, the depth development of showers is only indirectly reflected in the spatial information. In this context, timing information measured at the readout level provides an additional handle, offering indirect sensitivity to the longitudinal structure of particle showers.
In this work, we investigate the role of timing information for particle identification in a dual-readout calorimeter using simulated events over a range of energies and incident conditions. Using a point cloud representation of calorimeter hits and deep learning models as a probe, we study the impact of timing through controlled input variations. We find that timing information provides complementary discriminative power in certain regimes and contributes to improved particle identification performance, indicating its potential to enhance the reliability of dual-readout calorimetry.