Radiation Damage Effects

Radiation damage is an important issue for the particle detectors operated in a hostile environment where radiations from various sources are expected. This is particularly important for high energy physics detectors designed for the energy and intensity frontiers. This chapter describes the radiation damage effects in scintillating crystals, including the scintillation-mechanism damage, the radiation-induced phosphorescence, and the radiation-induced absorption. The radiation damage mechanism in crystal scintillators is also discussed. While the damage in halides is attributed to the oxygen/hydroxyl contamination, it is the structure defects, such as the oxygen vacancies, which cause the damage in oxides. Various material analysis methods used in investigations of the radiation damage effects as well as the improvement of crystal quality through systematic R&D are also presented.  Introduction Total-absorption shower counters made of inorganic crystal scintillators have been known for decades for their superb energy resolutions and detection efficiencies (Gratta et al. ). In high energy and nuclear physics experiments, large arrays of scintillating crystals of up to m have been assembled for precision measurement of photons and electrons. These crystals are working in a radiation environment, where various particles, such as γ rays, neutrons, and charged hadrons, are expected. >Table  (Mao et al. ) lists the basic properties of the heavycrystal scintillators commonly used in high energy physics detectors.They are NaI(Tl), CsI(Tl), undoped CsI, BaF, bismuth germanate (BGO), lead tungstate (PWO), and Ce-doped lutetium oxyorthosilicate (Lu(SiO)O or LSO(Ce)) (Melcher and Schweitzer ). All have either been used in, or actively pursued for, high energy and nuclear physics experiments. Some of them, such as NaI(Tl), CsI(Tl), BGO, LSO(Ce), and cerium-doped lutetium–yttrium oxyorthosilicate (Lu(−x)YxSiO:Ce, LYSO) (Cooke et al. ; Kimble et al. ) are also widely used in the medical industry. All known crystal scintillators suffer from radiationdamage (Zhu ).There are three possible radiation damage effects in crystal scintillators: () the scintillation-mechanism damage, () the radiation-induced phosphorescence (afterglow), and () the radiation-induced absorption (color centers). A damaged scintillation mechanism would reduce the scintillation light yield and cause a degradation of the light output. It may also change the light-response uniformity along the crystal length since the radiation dose profile is usually not uniform. The radiation-induced phosphorescence, commonly called afterglow, causes an increase of the dark current in the photodetectors, and thus an increase of the readout noise.The radiation-induced absorption reduces the light attenuation length (Ma and Zhu ), and thus the light output and possibly also the light-response uniformity. > Table  summarizes γ-ray-induced radiation damage effect for various crystal scintillators. There is no experimental data supporting a scintillation-mechanism damage. All crystal scintillators studies so far, however, suffer from the radiation-induced phosphorescence and the radiation-induced absorption. The radiation-induced absorption is caused by a process called color-center formation, which may recover spontaneously under the application temperature through a process called color-center annihilation. If so, the damage would be dose-rate dependent (Ma and Zhu , ; Zhu ). If the radiation-induced absorption does not recover, or the recovery speed is Radiation Damage Effects   ⊡ Table  Properties of some heavy-crystal scintillators Crystal NaI(Tl) CsI(Tl) CsI BaF BGO PWO LSO(Ce) Density (g/cm) . . . . . . . Melting point (C)        Radiation length (cm) . . . . . . . Molière radius (cm) . . . . . . . Interaction length (cm) . . . . . . . Refractive indexa . . . . . . . Hygroscopicity Yes Slight Slight No No No No Luminescenceb (nm)        (at Peak)    Decay timeb (ns)         .  Light yieldb,c   .   .  . . . d(LY)/dTb,d (%/○C) −. . −. −. −. −. −. . Experiment Crystal CLEO KTeV TAPS L CMS KLOE Ball BaBar BELLE ALICE SuperB BELLE PrimEx BES III Panda At the wavelength of the emission maximum Top line: slow component, bottom line: fast component Relative light yield of samples of . X and with the PMT QE taken out At room temperature ⊡ Table  Radiation damage in crystal scintillators Item CsI(Tl) CsI BaF BGO PWO LSO/LYSO Scintillationmechanism No No No No No No Phosphorescence (afterglow) Yes Yes Yes Yes Yes Yes Absorption (color centers) Yes Yes Yes Yes Yes Yes Recover at room temperature Slow Slow No Yes Yes No Dose-rate dependence No No No Yes Yes No Thermally annealing No No Yes Yes Yes Yes Optical bleaching No No Yes Yes Yes Yes very low, the color-center densitywould increase continuously under irradiations until all defect traps are fully filled. In this case, the corresponding radiation damage effect is not dose-rate dependent. Color centersmay also be annihilated thermally by heating the crystal to a high temperature through a process called thermal annealing, or optically by injecting light of variouswavelengths to the crystal through a process called optical bleaching (Ma and Zhu , ). The recovery process, either spontaneous or manual through thermal annealing or optical bleaching,   Radiation Damage Effects reduces the color-center density or the radiation-induced absorption. At the same time, it also introduces an additional instability for the crystal’s light output because of the variation of the crystal’s transparency. In this case, a precision monitoring system is mandatory to follow the variations of the crystal’s transparency. The radiation damage caused by neutrons and charged hadrons may differ from that caused by γ rays. Studies (Huhtinen et al. , , ) on proton-induced radiation damages in PWO crystals, for example, show a very slow (or no) recovery at room temperature, contrary to the radiation damage caused by γ rays. This leads to a cumulative damage in PWO with no dose-rate dependence for hadrons. The radiation damage level is also different at different temperatures for crystals with dose-rate-dependent damage since the spontaneous recovery speed is temperature dependent. PWO crystals used at low temperature, for example, suffer more damage than that at high temperature (Semenov et al. , , ). Commercially available mass-produced crystals usually do not meet the quality required for high energy physics detectors. The quality of mass-produced crystals, however, may be improved by removing harmful impurities and defects in the crystal. The rest of this chapter discusses γ-ray-induced radiation damage phenomena in scintillating crystals, the origin of the radiation damage in halides and oxides, as well as the improvement of crystal quality through systematic R&D. All data presented in this chapter are measured for full-size crystals adequate for calorimeter construction, which is typically – X long. Since both the radiation-induced phosphorescence and absorption are a bulk effect, it is important that only the full-size crystals are used in such studies.  Scintillation-MechanismDamage Experimental facts show that the crystal’s scintillation mechanism is not damaged. This is observed for irradiations of γ rays, neutrons, as well as charged hadrons (Batarin et al. ; Huhtinen et al. , , ; Batarin et al. , ). A common approach is to compare the shape of the emission spectra measured before and after irradiations. Direct comparison of the overall intensity of the emission spectra suffers from a large systematic uncertainty caused by the sample position and orientation, the surface quality, and the internal absorption which may be induced by the radiation. The top plots of > Fig.  show the photoluminescence spectra measured before (blue) and after (red) γ-ray irradiations for a PWO sample (left) and an LYSO sample (right).These spectra are normalized to the integration around the emission peaks as shown in the figure.The relative difference between these normalized spectra (green) is shown in the bottom plots. Also shown in the bottom plots are the averages of the absolute values of the relative difference.The numerical values are .% and .%, respectively, for PWO and LYSO, which are much less than the systematic uncertainty of these measurements, indicating that no statistically significant difference is observed between the photoluminescence spectra taken before and after irradiations for both PWOand LYSO.This observation consists with no damage to the scintillationmechanism. Similar studies show that there is no scintillation-mechanism damage observed for BGO (Wei et al. ; Zhu et al. ), BaF (Zhu ), and CsI(Tl) (Zhu et al. ) as well. This conclusion is also supported by more complicated measurements of the light-response uniformity before and after irradiations with a nonuniform dose profile (Batarin et al. , , ). Radiation Damage Effects   0 10 20 30 40 BTCP-PWO-2467 Before irradiation After 400 rad/h irradiation Before irradiation After 8,500 rad/h irradiation In te ns ity ( a. u. ) Normalized area 360–500 nm