1. Solid-State Radiation Damage Studies
FCAL Workshop
Sept 4-5, 2017
Bruce Schumm
UC Santa Cruz Institute
for Particle Physics 1

2. 2 X0 pre-radiator; introduces a little divergence in shower
Sensor sample
Not shown: 4 X0 “post radiator” and 8 X0 “backstop”

3. Silicon Diode Sensors n-bulk (N) and p-bulk (P)
Both float-zone (F) and Magnetic Czochralski (M) for each of N,P
m thick bulk
Various manufacturers
Heaviest doses for pad (not strip) sensors 3 3

4. NF Type Charge Collection for 300 Mrad
@600 V, ~40% charge collection loss (58C annealing)
1-hour annealing steps
300 Mrad Exposure
NF Si Diode Sensor 4 4

5. NP Type Charge Collection for 300 Mrad
Breakdown (probably not fundamental) limited VB
@600 V, charge collection loss likely less than 30%
300 Mrad Exposure
NP Si Diode Sensor
Annealing vs. time (at ~250 C) rather than vs. temperature 5

6. N-Type LumiCal Prototype Fragment
After annealing, charge collection at 600V likely well above 50% after 300 Mrad exposure
Sensor via Sasha Borisov, Tel Aviv
300 Mrad Exposure
“LumiCal” N-Type Diode Sensor 6 6

7. PF Type Charge Collection for 270 Mrad
@600 V, ~20% charge collection loss (60C annealing)
270 Mrad Exposure
PF Si Diode Sensor 7 7

8. PF Type Charge Collection for 570 Mrad
Currents roughly x2 that for 270 Mrad
570 Mrad Exposure
PF Si Diode Sensor 8 8

9. Silicon Diode Currents
Appear to be similar from one technology to the other
Appear to scale roughly linearly with dose
Not affected by high-temp annealing
These are expected
Thus, focus on one sensor: the 270 Mrad exposure of sensor WSI-P4 (PF-type) 9 9

10. PF Type I vs. Temperature; 270 Mrad
Define exposure in “T506” units. One “T506” equals
270 Mrad of ionizing energy loss
3.7x1011 “RAND” (see below) of non-ionizing loss
270 Mrad Exposure
Current doubling for event ~70 C (expected)
Detector area is about cm2 10 10

11. Comparison to Neutron Irradiation Results
Based on results from numerous neutron-irradiation studies, Lindstrom et al. NIMA 466(2),308 [2001] provide a damage proportionality factor  that relates neutron fluence to current density at T = -100 C.
Using FLUKA to estimate the T506 neutron fluence (see below), and the expected temperature dependence for a 2.50 C extrapolation, we converted this to an expectation for the T506 current density
The numbers agree within ~30%
Supports (but doesn’t prove) the notion that T506 leakage current is due primarily to non-ionizing energy loss (NIEL) from neutrons
Predicted Current Density at -100C, 300µm sensor for T506
94 A/cm2
Measured T506 Current Density at -100 C
65 A/cm2 11 11

12. Gallium Arsenide Sensor provided by Georgy Shelkov, JINR
Sn-doped Liquid-Encapsulated Czochralski fabrication
300 m thick 12 12

13. GaAs Charge Collection for 21 Mrad
Significant charge collection loss
21 Mrad Exposure
(0.078 “T506”)
GaAs Sensor 13 13

14. Industrial Sapphire Sensor provided by Sergej Schuwalow
Fabricated by Crystal GmbH, Berlin
Layered Al-Pt-Au contact structure
Current low (< 10 nA) after irradiation 14 14

15. Sapphire Charge Collection for 300 Mrad
Low pre-irradiation charge-collection and
significant charge loss after irradiation
Sensors via Sergej Schuwalow, DESY Zeuthen
500 m thick Al2O3
300 Mrad Exposure
(1.1 “T506”) 15 15

16. Silicon Carbide Sensor provided by Bohumir Zatko, Bratislava
Schottky-barrier contacts mounted on 4H-SiC structure
Epitaxial (active) layer thickness 70 m 16 16

17. SiC Charge Collection for 77 Mrad
4H SiC Sensor
98C anneal
77 Mrad Exposure
(0.29 “T506”)
Charge collection mostly above 50% 17 17

18. BeamCal Neutrons from FLUKA
Many thanks to Ben Smithers, UCSC undergraduate 18 18

19. BeamCal Simulation in FLUKA (Ben Smithers, SCIPP)
BeamCal absorbs about 10 TeV per crossing, resulting in electromagnetic doses as high as 100 Mrad/year
Associated neutrons can damage sensors and generate backgrounds in the central detector
GEANT not adequate for simulation of neutron field  implement FLUKA simulation
Design parameters from detailed baseline description (DBD)
Primaries sourced from single Guinea Pig simulation of e+- pairs associated with one bunch crossing
Proceeding into the simulations of the Beamcal, I would like to note that the geometric parameters of the beamcal were acquired from the ILC detailed baseline description. Additionally, we sourced our primaries from a Guinea Pig simulation of a bunch crossing.
After running the FLUKA simulations, I produced several cutouts of the resulting fluences of electrons and positrons, and neutrons, at various layers in the BeamCal. It gives a good idea of what is going on inside the BeamCal. I also used the raw data to find the layers of highest fluence. Layer 8 saw the highest electron positron fluence and 12 saw the highest neutron fluence.

20. Layer 2 Detector - Fluence
E+&E-
Neutrons

21. Layer 4 Detector - Fluence
E+&E-
Neutrons

22. Layer 6 Detector - Fluence
E+&E-
Neutrons

23. Layer 8 Detector - Fluence
E+&E-
Neutrons

24. Layer 10 Detector - Fluence
E+&E-
Neutrons

25. Layer 12 Detector - Fluence
Neutrons

26. Layer 14 Detector - Fluence
Neutrons

27. Layer 16 Detector - Fluence
Neutrons

28. FLUKA Simulation: T506 Baseline
51 C of 13.3 GeV SLAC ESA electrons onto target
Raster over 1 cm2 area
Realistic mix of e± and neutrons (Giant Dipole Resonance)
Project: Assuming this baseline damage is due entirely to neutron dose, use FLUKA to estimate damage effects throughout BeamCal 28 28

29. T506 Neutron Fluence from FLUKA
Mean number of neutrons per cm2 per 13.3 GeV primary 29 29

30. T506 Neutron Dose (Step 1/3) 30 30

31. T506 Neutron Dose (Step 2/3): NIEL Scaling31 31

32. NIEL in Silicon
N(E) 32 32

33. T506 Neutron Energy Spectrum (FLUKA)
In range where N(E) is slowly varying
Note that N(E) is for Si only; caveat (small?) for drawing assumptions about GaAs, Sapphire, SiC
Peaks around Ecrit for Tungsten 33 33

34. “T506 Unit” of Neutron Dose
T506 Neutron Dose (Step 3/3)
“T506 Unit” of Neutron Dose
Primaries
51 (3.1 x 1014)
C (electrons)
Neutron Fluence
7.4 x 1013
Neutrons/cm2
Rand/primary
1.2 x 10-3
Rand (MeV/cm3 of NIEL)
Total Dose (“T506”)
3.7 x 1011 34 34

35. T506 Neutron Non-Ionozing Energy Deposition
Mean MeV of non-ionizing energy deposition per cm3 per 13.3 GeV primary 35 35

36. Baseline is 270 Mrad T506 run (to connect to our measurements)36 36

37. Baseline is 270 Mrad T506 run (to connect to our measurements)37 37

38. Baseline is 270 Mrad T506 run (to connect to our measurements)38 38

39. Baseline is 270 Mrad T506 run (to connect to our measurements)39 39

40. Baseline is 270 Mrad T506 run (to connect to our measurements)40 40

41. Baseline is 270 Mrad T506 run (to connect to our measurements)41 41

42. Baseline is 270 Mrad T506 run (to connect to our measurements)42 42

43. Baseline is 270 Mrad T506 run (to connect to our measurements)43 43

44. Baseline is 270 Mrad T506 run (to connect to our measurements)44 44

45. Baseline is 270 Mrad T506 run (to connect to our measurements)45 45

46. Baseline is 270 Mrad T506 run (to connect to our measurements)46 46

47. Baseline is 270 Mrad T506 run (to connect to our measurements)47 47

48. Baseline is 270 Mrad T506 run (to connect to our measurements)48 48

49. Neutron Flux and BeamCal Sensor Radiation Damage
SLAC Experiment T506: prototype sensor
placed at shower max of electromagnetic
shower induced by tungsten  shower has
realistic hadronic component
Explored radiation-hardness properties of
several different Si diode and bulk solid-state
(GaAs, Sapphire, SiC) sensor technologies
T506 exposures in the Mrad range
(recall that maximum BeamCal dose is 100 Mrad
of electromagnetic radiation)
Recent discovery: FLUKA simulations suggest
that damaging (non-ionizing) component of neutron energy deposition, per MeV of dEdX from e+-, is much higher in the BeamCal than at the T506 exposure point
May have important implications for BeamCal sensor choice, given varying degrees and types (charge loss, leakage current) of radiation damage observed in T506
SLAC e- beam
Sensor prototype 49 49

50. Summary
Need to do precision calorimetry in high-dose, high speed environment is driving a lot of R&D and design work
Work reasonably advanced, even at systems level
Significant use of test beams for prototype evaluation and radiation damage studies
Picture continuing to clarify (LHCal perhaps a bit behind) 50 50

51. Backup 51 51