Beam tests on the ATLAS tile calorimeter demonstrator module

Abstract The Large Hadron Collider (LHC) Phase-II upgrade aims to increase the accelerator luminosity by a factor of 5–10. Due to the expected higher radiation levels and the aging of the current electronics, a new read-out system of the ATLAS experiment hadronic calorimeter (TileCal) is needed. A Demonstrator prototype of the electronics has been tested during different testbeam campaigns at the Super Proton Synchrotron (SPS) accelerator of CERN with the purpose of checking the calibration and determining the performance of the detector by exploiting the features of the interactions of different particles with matter. We present the current status and results where the Demonstrator new electronics were situated in calorimeter modules and data were collected by exposing it to beams of muons, electrons and hadrons, at various incident energies and impact angles.


Introduction
The High-Luminosity LHC (HL-LHC) will have an instantaneous luminosity of five times the nominal design value. ATLAS TileCal [1] (figure 1c) is a sampling calorimeter composed of steel plates and plastic scintillator tiles that is divided in four cylindrical barrels: the two central long barrel segments (LBA, LBC) and shorter extended barrel segments (EBA, EBC) at each end (figure 1b). Each cylindrical section is composed of 64 wedge-shaped modules where scintillators are grouped in pseudo-projective cells. Light from two sides of a cell is collected by wavelength shifting fibers and read out by two photomultiplier tubes (PMTs) (figure 1a). The Phase-II upgrade program of the ATLAS TileCal [2] will compensate for the higher rates of pileup, higher radiation levels and the aging of the current electronics. The read-out electronics will be replaced to adopt a fully-digital trigger system that will provide full-granularity digital data to the off-detector systems. The new design will provide better timing and energy resolution, and less sensitivity to out-of-time pileup. Protypes of the new electronics were coupled to calorimeter modules and tested in the Super Proton Synchrotron (SPS) at CERN. In order to check the calibration and to determine the performance of the detector, data were collected with beams of muons, electrons and hadrons of various incident energies and impact angles.

The Phase-II and read-out system.
The Phase II read-out system [3] is modularized in Superdrawers (SD), each composed of 4 Minidrawers (MDs). Frontend boards (FEB) interfaced with a Mainboard (MB) shape and condition the PMT signals. In every MD, a MB transfers to a Daughterboard (DB) the digitized 2 gains of 12 PMT signals at 40 Mhz. The DB interfaces the MB and the off-detector systems by receiving timing and slow control commands and transmitting data over redundant fiber-optic links (figure 2a). A Tile Preprocessor receives and stores two gains of PMT data from the DB in pipelines until a trigger decision event, while providing reconstructed data to the trigger. Given a trigger event, the data are read out by the FELIX system (figure 2b).

Testbeam setup and module layout on the scanning table
The SPS H8 beamline is equipped with trigger scintillators in coincidence, Cherenkov counters for beam particle identification and wire chambers for measuring the lateral coordinates of the beam particles (figure 3).

Testbeam Results.
Data taken from the Demonstrator module (LBC65) during the test-beam campaigns (2015-2017) was analyzed [4] [5]. The deposited muon energy followed a Landau distribution (figure 5a). Studies were done for the "Fit" and "Optimal Filter" reconstruction methods with different noise thresholds (figure 5b). Cell responses were analyzed applying amplitude and timing cuts to the sums of the corresponding two PMTs signals (figure 5c). The e − energy distributions agree for both the Demonstrator and Legacy system (6a). The simulation and experimental data agree nicely for all studied energies (figures 6b and 6c).
The combination of the quantities C long (fraction of energy deposited in the first two longitudinal layers) and C total (spread of the energy deposited in the cell c) were used for e − /hadron separation (figure 6d). The hadron response was estimated using a Gaussian fit in the range 2 σ around peak value (figure 7a). The analysis of the calorimeter response as a function of the beam energy for hadrons had good agreement with the simulation and resulted in higher response for pions and kaons than for protons, improving with increasing beam energy (figures 7b and 7c).

Conclusions
Five testbeam campaigns of two weeks each took place between 2015 and 2017 with three detector modules resulting in a good performance of the new electronics, and agreement with calibration and simulated data. Two additional testbeams are scheduled for May and November 2018 following further integration with new revisions of the Phase-II upgrade hardware, mechanics and electronics.