1. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota1 Radiation hardness of the Avalanche Photodiodes for ECAL CMS detector at CERN A. Singovski The University of Minnesota

2. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota2 APD’s in the CMS detector PbWO 4 crystal

3. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota3 Radiation level after 10 years Radiation doses are in red, 10 4 Gy. Neutron fluence in green 10 13 neutrons/cm 2 with E > 100 keV. Levels outside of the detector are down by a factor of 100 or more. Crystal calorimeter 100

4. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota4 APD’s for CMS Manufacturer: Hamamatsu Photonics, Japan. Quantity: Two APD’s per crystal– 124,000 APD’s with spares. Accessibility during operation: None. Radiation levels: Maximum expected dose 200 kGy and 2 10 13 neutrons/cm 2.

5. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota5 Basic APD Structure: APD is grown epitaxially on an n ++ wafer. JunctionSi 2 N 4 AR coating Groove to minimize surface leakage current. 5  5 mm 2 active area

6. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota6 APD radiation damage Radiation damage of APD can influence ECAL performance by essentially two effects: -rise of the bulk current  increase APD noise contribution to the energy resolution -early breakdown  breakdown happens before APD can reach operation point at Gain=50

7. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota7 APD contribution to the ECAL resolution Resolution: where, a : due to intrinsic shower fluctuations & photo statistics b : related to stability and reproducibility c : noise contributions CMS design goal : a ~3%, b~0.5%, c~200 MeV APD contributions: a - photo statistics (area, QE) & excess noise factor b - gain variation with bias voltage and temperature c - capacitance as series noise and dark current as parallel noise

8. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota8 Irradiation Tests. Irradiation with protons: 70 MeV protons beam at PSI – Switzerland. 1  10 13 hadrons/cm 2 in ~ 2 hours. Irradiation with gammas. All irradiation with 60 C0 source. Irradiation with neutrons. Californium source ( 252 Cf) for irradiation at the University of Minnesota. 2 10 13 neutrons/cm 2 in ~ 2 days.

9. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota9 Device failure Irradiation in a 70 MeV proton beam.

10. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota10 Neutron irradiation facility Draws for irradiation Use old tandem laboratory facility in Minneapolis to store and operate two 7 mg sources for irradiation samples. High and low flux areas 10 13 and 10 11 n/cm 2 in 2 – 4 days. Must provide biases to components during irradiation. Return sources after 4 years of operation.

11. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota11 Neutron irradiation results

12. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota12 Durk current 70 MeV protons 1 MeV neutrons from 252 Cf

13. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota13 Neutron irradiation summary 1.All APD tested so far survived -> no significant shift in breakdown voltage. 2.The mean bulk current after 2x10 13 neutrons/cm 2 is Id  280nA (non-amplified value). 3.It corresponds to 14  A at Gain=50 and ~ 80MeV noise contribution (no-recovery case, CMS TDR). Acceptable for CMS ECAL detector

14. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota14 Gamma irradiation APDs 32 wires containing 60 Co surround the probe and give a very uniform irradiation field. Present activity is 2.5 kGy/h 60 Co irradiation facility at PSI

15. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota15 Gamma irradiation results

16. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota16 Gamma irradiation results 2 APD with a significant shift of Vb after 60 Co irradiation (vs. good one) Id/Gain Noise

17. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota17 Screening Method: - irradiate 100% of APD`s (0.5 Mrad) with Co-60 gamma source (at PSI); - measure VB and Id(V) of all irradiated APD`s 1 day after irradiation (at PSI); - measure noise at M=1, 50, 150, 300 before annealing (at CERN APD Lab) - anneal all APD`s in the oven (for 4 weeks at T=80C, at CERN APD Lab); - measure VB, Id(V) after annealing/ageing (at CERN APD Lab); - reject potentially non-reliable APD`s: ones showing - Shift of Vb more than 5V - high Id - high noise

18. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota18 APD rejection High dark currentHigh noise

19. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota19 Screening efficiency 225 APD`s which passed 1st irradiation and annealing were irradiated the 2nd time. No change of VB>2V was found for all APD`s !!!

20. RESMDD02 July 10-12 2002, Florence, A.Singovski, University of Minnesota20 Gamma irradiation summary 1.APDs are sensitive to the gamma irradiation. 2.Several percents “die” after irradiation -> get a breakdown close to the operation point. 3.Screening, applied to 100% of APDs make them 99.9% rad. Hard.