Speaker
Description
Summary
The PANDA detector will be one of the central experiments of the future Facility for Antiproton and Ion Research (FAIR) which is currently under construction near Darmstadt in Germany. PANDA will be a fixed target experiment using a cooled antiproton beam of up to 15 GeV to study new aspects in the fields of hadron spectroscopy, nucleon structure investigations, the modification of hadrons in matter and the search and characterization of Hypernuclei.
The electromagnetic calorimeter (EMC) of the PANDA detector, which will be one of the central components to achieve the physical goals, will consist of more than 15,000 lead tungstate (PWO) crystals operated at -25 °C to increase the light yield [PA09]. To reach the physics goals, a detection of photons down to 10 MeV is mandatory. The barrel part of the target EMC will be composed of 11 crystal geometries with a different degree of tapering read out by two rectangular large area avalanche photo diodes with an active area of 1 cm2 each. Due to the tapering the crystals show a non-uniformity in light collection as a result of an interplay between the focusing and the internal absorption within the crystal (see Fig. 1 left). The non-uniformity has been determined using low energy gamma-rays, cosmic muons and a 80 MeV proton beam (see Fig. 1 right) leading to comparable results [DB14].
While crystals with an average degree of tapering (type 6) show a light yield difference between creation in the front and the rear part of the crystal of 20 %, this value increases for the most tapered crystals to more than 40 % with a slope in the front part of almost 3 %/cm. Due to the spread of the electromagnetic shower within the crystal and due to its fluctuations, this effect causes a smearing of the response leading to a reduction of the energy resolution, in particular the constant term. This problem has already been considered by the CMS-ECAL collaboration and was solved by de-polishing crystals on one lateral side [EA02]. The PANDA barrel EMC is focusing on much lower energies ranging from 10 MeV up to a few GeV. Based on the experiences from CMS 3 one lateral side face of a set of 25 crystals has been de-polished at the CMS setup to a roughness of Ra = 0.3 µm. Fig. 2 shows the absolute and relative light yield of six typical type 2 crystals wrapped with mirror reflective VM2000 before and after the de-polishing procedure. In the rear part of the crystal a significant increase of the absolute light yield can be observed combined with a slight decrease in the front part. As a direct result, the non-uniformity is reduced down to a typical value of 5 % with a nearly homogeneous response in the front and central part of the crystal.
The observed behaviour can be reproduced with the ray tracing model of GEANT4 (see Fig. 3).
The increase of the light yield in the rear part of the crystal can be explained by photons which change their direction on the way to the front face of the crystal, which is illustrated by the significant increase of photons with path lengths between 1 cm and 39 cm in a crystal with a de-polished side face in Fig. 4.
For the first time a 3x3 sub-matrix of de-polished crystals has been implemented in the current close to final barrel EMC prototype (Fig. 5) and compared with an identical array of completely polished crystals. The response to tagged photons in the energy range between 50 MeV and 800 MeV, respectively, has been measured at the MAMI facility in Mainz (Germany).
Fig. 6 shows the obtained energy resolution for the 3x3 sub-array of de-polished crystals in comparison to a similar matrix composed of polished crystals.
The experimental results show a significant improvement of the energy resolution for energies down to 200 MeV if crystals with one de-polished side face are used, while the energy resolution is comparable for both arrays in the region between 50 MeV and 200 MeV, respectively. As a consequence, the constant term of the relative energy resolution is significantly reduced from above 2 % down to 0.5 % while the statistical term increases only slightly. The shown results can be well reproduced with a specially developed GEANT4 model, including the non-uniformity and other empirical properties of the crystals and their readout [SD15].
Acknowledgements:
The authors would like to thank Etiennette Auffray from CERN for providing the de-polishing of the crystals.
References:
[PA09]PANDA collaboration, Technical Design report for: The PANDA Electromagnetic Calorimeter, Darmstadt 2009, arXiv:0820.1216.[DB14]D. Bremer, Measurement and Simulations on Position Dependencies in the Response of Single PWO Crystals and a Prototype of the PAND EMC, PhD thesis (Giessen, 2014).[EA02]E. Auffray et al., Crystal Conditioning for High Energy Physics Detectors, Nucl. Instrum. Meth. in Phys. Res. A 486, 22-34 (2002).[SD15]S. Diehl, Optimization of the Influence of Longitudinal and Lateral Non-Uniformity on the Performance of an Electromagnetic Calorimeter, PhD thesis (Giessen, 2015).
