1. Monte Carlo simulation of the imaging properties of a scintillator- coated X-ray pixel detector M. Hjelm * B. Norlin H-E. Nilsson C. Fröjdh X. Badel Department of Information Technology and Media Mid-Sweden University Sundsvall, Sweden Department of Microelectronics and Information Technology KTH, Stockholm, Sweden * Also affiliated to KTH

2. Outline Simulated devices Simulation method Results Conclusions

3. Detector top view Division 45  m Wall thickness 6  m

4. CCD detector, side view No transmission of X-rays into Si detector is assumed Wall: 2 x (1  m SiO 2 + 2  m Si) Poly-Si layer thickness: 0.6  m => large damping Real device also includes a fiber plate in order to avoid direct absorption in the CCD

5. Diode detector, side view Two wall designs simulated: –2 x (1  m SiO 2 + 2  m Si) –2 x (2  m SiO 2 + 1  m Si) Two layouts of diodes simulated: –On sides and bottom –On bottom

6. Simulation method Based on 3 MC simulations: 1.X-ray absorption –MCNP 2.Light transport –In-house ray-tracing code 3.Complete detector –Small special program for each detector type

7. Example of X-ray energy absorption data CsISi Absorbed energy in 15-20 keV range 238  m deep pore, walls: 2 x (1  m SiO 2 + 2  m Si)

8. Example of light transport data Light absorbed in 2  m bottom diode CsI pore, 238  m deep pore, walls: 2 x (1  m SiO 2 + 2  m Si)

9. SNR, CsI - CCD light detector 16*N defects with a damping of 5 % each are randomly distributed in the scintillator pores N=number of pixels=625

10. Fixed pattern image due to pore defects Defects as in previous picture Compensated with fixed- pattern noise correction, which is considered in SNR calculation

11. SNR, Gadox - CCD light detector Defects as in previous two pictures Gadox compares well to CsI due to longer wave length of light, which better passes the poly-Si layer This is very much dependent on the charac- teristics of the poly-Si layer

12. SNR, CsI - diode detector

13. SNR, Gadox - diode detector Gadox is poor for diode on 5 surfaces due to relatively low light emission

14. Thickness and contribution a)SNR for different thicknesses of CsI diode detector b)SNR for signal from X-ray direct absorption in Si diode and indirect CsI – light – light absorption in diode, 238  m thickness a) b) X-ray dose=25 mR diodes on 5 surfaces

15. The walls should not be completely depleted to permit collection of charge –Depletion controlled with bias Charge collection from walls can be switched off with high bias To suppress direct absorption from bottom: –Important to select suitable diffusion length Limiting lifetime and/or mobility in substrate –Alternatively: thick scintillator with high X-ray absorption Charge transport issues for diode detector

16. Conclusions To get high SNR, the signal from direct absorption of X-rays has to be minimized compared to the signal generated from scintillator light absorption  High light emission from the scintillator material is very important for designs with diodes on side surfaces From the point of view of SNR: –The designs based on diode light detectors at pore surfaces are not better than the CCD design Diode solutions have other advantages: –higher signal, less damage by high radiation dose

17. Conclusions For designs with diodes on side surfaces: –Increasing the SiO 2 layer thickness leads to less high- energy electrons emitted from scintillator into silicon –Should be balanced with less X-ray absorption in a smaller scintillator pore