Proton induced radiation damage studies on plastic scintillators for the Tile calorimeter of the atlas detector

Figure 5-8: The effect of the experimental systematics on the measured light yield signal.
In Figure 5-8, the sample signal is the integrated signal over the sample region. 
The fiber signal is the signal integrated over the fiber region. Since the position 
of the fibers and holder are kept fixed, this measurement can act as a control for 
systematics such as variation in the particle flux of the source, fluctuations in the 
high voltage gain applied over the PMT, etc. The edge length is the length of the 
scintillator sample edge that makes contact with the two fibers. In this 
comparison, it is evident that these systematic v ariations have a large impact on 
the results of the experiments conducted.
In order to analyse the light yield loss as a function of irradiation dose, the sample 
signals were normalised by correcting for variations in thickness, contact length 
and fiber signal. The ratio between the corrected signals for the irradiated and 
un-irradiated samples was then computed. These results are shown in Figure 5-9. 
Whilst the correction factors slightly improved the precision of the three 
repetitive measurements, large uncertainties in the accuracy of the ratio computed 
still remain.

---

46 
The results of the corrected light loss ratio however, do correlate with the 
expected performance of the scintillators. EJ260 exhibits the least loss, however 
the actual signal obtained for its un-irradiated sample was a factor of 3 smaller 
than that obtained for the un-irradiated EJ200 sample. The blue scintillators 
performed very similarly once again, with EJ208 exhibiting the most radiation 
resistance against loss to light yield.
Figure 5-9: Relative light yield as a function of radiation damage dose.
The light box set-up used for these experiments typically measure much larger 
scintillator samples. Although a measurement could be made for the small 
samples in our study, the actual signals were small and sensitive to systematic 
errors. The UPS923A data particularly showed a large variation, but these 
samples also had large geometric variations. Other factors that may have led to 
errors in the data include the surface conditions of the scintillators. Since the 
edges were handled with metal tweezers, slight scratches and cracks could ha ve 
formed which would influence the contact made between these edges and the 
read-out fibers.