부산대학교 도서관

The electromagnetic calorimeter of the PANDA detector at the future FAIR facility, will be one of the central components to achieve the physical goals in studying the interaction of cooled antiprotons with a fixed target. The barrel part of the target electromagnetic calorimeter will consist of 11 crystal geometries with a different degree of tapering. Due to tapering the crystals show a non-uniformity in light collection, which is resulting from an interplay between the focusing and the intrinsic absorption of light in the crystal. For the most tapered crystals the light detected by the photo sensor is enhanced by a factor of $ > 1.4$, if the scintillation light is created in the front part of the crystal. Due to the spread of the electromagnetic shower within the crystal and due to its fluctuations, this effect leads to a smearing of the response, resulting in a reduction of the energy resolution. Therefore, one lateral crystal surface has been de-polished for 9 crystals to a roughness of $0.3 ~\mu \hbox{m}$ , which decreases the non-uniformity from up to 40% to less than 5%, with a tolerable decrease of the light yield. This paper will compare the response of a $3\times 3$ array of crystals with one de-polished side face with an identical matrix of completely polished crystals using high energy photons from 56 MeV up to 767 MeV, respectively. The results are compared to GEANT4 simulations and show a significant improvement of the energy resolution at energies above $\sim 200~\hbox{MeV}$ with no deterioration down to 50 MeV.