Radiation hardness

At the KEK B factory, a large number of soft $\gamma$-rays having an energy of up to a few MeV are generated by spent beam electrons and positrons hitting the beam pipes and radiation shielding masks. The dose of the beam induced background in the barrel region has been estimated to be up to 5 rad/y in the first 2 cm of depth under steady operation at an integrated luminosity of $10^{41} cm^{-2}$/y. With a large safety margin, the requirement for radiation hardness is set such that the light-output decreases to less than 3 %, 10 %, and 20 % at radiation dose of 10, 100, and 1000 rad, respectively. According to the previous measurements, it was not self-evident whether our CsI($Tl$) crystals would satisfy this requirement.
In order to test the radiation hardness, several full-size crystals made by each producer were irradiated up to 1000 rad by the $^{60}$Co source of 51 TBq at the irradiation facility of Tokyo Institute of Technology [61]. The irradiation rates were about 3, 3, and 25 rad/min for 10, 100, and 1000 rad irradiation, respectively. The attenuation length of 1.17 and 1.33 MeV $\gamma$-rays from the $^{60}$Co source is about 5.4 cm in a CsI crystal.
Irradiation was done in two ways: (a) uniform irradiation in which crystals were irradiated from the sides of the crystal over the whole region and (b) front face irradiation in which the sides of the crystals were shielded and the irradiation was made from the front surface simulating the actual case. The radiation damage is restricted in the front region, and approximately 90 % of the total dose is absorbed in the first 10 cm region out of the total length of 30 cm. The light output of irradiated crystals was measured in two methods: one with a photomultiplier tube (PM) (Hamamatsu R1847-S) and a $^{137}$Cs source, and the other with two photodiodes (PDs) (Hamamatsu S2744-08, the sensitive area of 10 mm $\times$ 20 mm) and cosmic-ray muons. The latter arrangement is the same as in the actual Belle ECL.
Two kinds of recoveries of damage after irradiation were observed: (1) a decrease in the phosphorescence intensity and (2) a partial restoration of the light output. The intensity of the stray phosphorescence was too strong to measure the photoelectric peak of 662 keV $\gamma$-rays immediately after irradiation. After a period of a few hours to one day the intensity decreased and weakened to the level at which measurements were possible. In general, the loss of light-output of a CsI($Tl$) crystal after radiation damage partially recovered. Fig. shows a typical history of light output measured using a PMT. The light output recovered partially during a period of about 1, 2, and 4 weeks after irradiation of 10, 100, and 1000 rad, respectively. Unless specifically mentioned, we use the saturated value of the light output at each dose.

Figure: Time dependence of the light output for uniform irradiation (PMT readout).

The light output measured with a PMT and PDs as a function of dose for front-face irradiation is shown in Fig. . The results with a PMT and with PDs agree at a level of a few %. The figure shows that all of the crystals tested, except for the prototype made by Shanghai Institute of Ceramics (SIC), satisfy the requirement of radiation hardness up to 1000 rad. The radiation hardness of SIC crystals was improved to fulfill the requirement by doping with a material with the property to cancel the suspicious impurity. Fig. shows the results of light output for new SIC crystals and the original prototype SIC crystals measured using the PMT readout after front-face irradiation.

Figure: Light output as a function of radiation dose for front-face irradiation. The upper and lower figures are the data with the PMT readout and the PD readout, respectively. The light output values are normalized to those before irradiation. The thick line represents the requirement for the radiation hardness of ECL. All crystals, except for SIC-prototypes, satisfy the requirement.

Figure: Light output of new SIC crystals versus radiation dose (front-face irradiation).

The position dependence of the light output measured with PMT and PDs readouts for uniform and front-face irradiation indicates that the change in the light output depends little on the position along a crystal, even for front-face irradiation. This result itself is a good evidence that the mechanism of scintillation emission in CsI($Tl$) crystals is little affected by irradiation. We conclude that only the overall gain factors need to be corrected for the effect of radiation damage in the running conditions.
The results of studies of the radiation-damage mechanism indicate that the main cause of radiation damage is due to the degradation in the attenuation length caused by the formation of color centers, rather than a deterioration of the mechanism of scintillation light emission. The coloration of CsI($Tl$) to red or brown was observed after irradiation. The degree of coloration was correlated with the change in the light output. The prototype crystals by Shanghai Institute of Ceramics (SIC) were colored at 100 rad irradiation while those by Novosibirsk (NOV) and Crismatec were at 1000 rad or beyond. In the case of the front-face irradiation, coloration was observed in the front region and the front part of SIC prototype crystals became red after 100 rad irradiation.
In order to investigate the cause of radiation damage we measured the spectra of the transmittance, excitation, and emission at 0, 10, 100, and 1000 rad irradiation with spectrophotometers. Sliced pieces of 5.5 cm $\times$ 5.5 cm $\times$ 2.7 cm in size were used for these studies. The transmittance was measured across a thickness of 5.5 cm with the reproducibity of $\pm$1 %. The measurements of the excitation and emission were made on the surface of each crystal with somewhat poorer reproducibilities of $\pm$5 %. The results for the SIC prototype and NOV are shown in Fig. . Clear absorption bands at around 430 and 520 nm on the transmittance curves can be seen for the SIC prototype, indicating a formation of color centers. On the other hand, the deterioration in the transmittance was small for NOV. A Monte Carlo ray-tracing simulation also supported the idea that the main result of radiation damage is the formation of color centers.

Figure: Spectra of the transmittance across a thickness of 5.5 cm and of excitation and emission at a surface of each crystal after 0, 10, 100 and 1000 rad irradiation for SIC prototype and NOV. The transmittance data were corrected for the reflection loss at the surfaces. Clear absorption bands are observed on the transmittance curve for the SIC prototype. The spectra of excitation and emission were corrected for wavelength dependence of the quantum efficiencies of PMT.