Update from the UCSC/SLAC ESTB T-506 Irradiation Study LCWS 2016 Aiina Center & MALIOS Morioka, Iwate, Japan December 5-9, 2016 Bruce Schumm UC Santa Cruz Institute for Particle Physics 1

Main Points and Updates Summary: Exposures and progress in characterizing radiated samples Reprise: Results from 270 Mrad exposure of p-type float-zone sensor, 300 Mrad exposure of n-type float-zone sensors New Results: Industrial sapphire and LumiCal N-type Si Diode sensor (300 Mrad)

T-506 Motivation BeamCal maximum dose ~100 MRad/yr BeamCal is sizable: ~2 m2 of sensors. A number of ongoing studies with novel sensers: GaAs, Sapphire, SiC  Are these radiation tolerant?  Might mainstream Si sensors in fact be adequate?

Radiation Damage in Electromagnetic Showers Folk wisdom: Radiation damage proportional to non-ionizing component of energy loss in material (“NIEL” model) BeamCal sensors will be embedded in tungsten radiator Energy loss dominated by electromagnetic component but non-ionizing contribution may be dominated by hadronic processes

Hadronic Processes in EM Showers There seem to be three main processes for generating hadrons in EM showers (all induced by photons): Nuclear (“giant dipole”) resonances Resonance at 10-20 MeV (~Ecritical) Photoproduction Threshold seems to be about 200 MeV Nuclear Compton scattering Threshold at about 10 MeV;  resonance at 340 MeV  These are largely isotropic; must have most of hadronic component develop near sample 5 5

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

Irradiating the Sensors 7 7

Mauro Pivi SLAC, ESTB 2011 Workshop, Page 8 LCLS and ESA Use pulsed magnets in the beam switchyard to send beam in ESA. Mauro Pivi SLAC, ESTB 2011 Workshop, Page 8 8

Daughter Board Assembly Pitch adapter, bonds Sensor 1 inch 9 9

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

Recent T506 Exposures 11 11

Both irradiated to 300 Mrad Summer 2016: LumiCal Sensors and Industrial Sapphire LumiCal sensor fragment via Sahsa Borisov, Tel Aviv Industrial sapphire sensor courtesy of Sergej Schuwalow, DESY Zeuthen Both irradiated to 300 Mrad 12 12

Other Recent Exposures and Summary December 2015, Summer 2016 offered high-rate exposures: 10 24hr days of ~20 Mrad/hr Sensor Type Notable Exposures (Mrad) GaAs 20, 100 SiC 80 Si PF 270, 570 Si NF 300 Si PC 600 Si NC 290 Red indicates results available; others await evaluation Operated at 12-15 GeV at close to 1 nA 13 13

Assessing the Radiation Damage 14 14

Charge Collection Measurement For pad sensors use single-channel readout Daughter-board Low-noise amplifier circuit (~300 electrons) 15 15

Charge Collection Apparatus Readout: 300 ns 2.3 MeV e- through sensor into scintillator Sensor + FE ASIC DAQ FPGA with Ethernet 16 16

Pulse-height distribution for 150V bias Median pulse height vs. bias Measurement time Pulse-height distribution for 150V bias Mean Pulse Shape Single-channel readout example for, e.g., N-type float-zone sensor Readout noise: ~300 electrons (plus system noise we are still addressing) Median pulse height vs. bias 17 17

Reprised Results 18 18

Si Diode Results P-Type Float Zone (PF) to 270 Mrad N-Type Float Zone (NF) to 300 Mrad Just by way of reminder (reported at Santander in June) 19 19

PF Charge Collection after 270 Mrad @600 V, ~20% charge collection loss (60C annealing) 20 20

PF I-V after 270 Mrad Exposure (-10 C) At 600 V, about 80 A (0.05 W) per cm2 (sensor area ~ 0.025 cm2) Not much improvement with annealing 270 Mrad Exposure 21 21

NF Leakage Current after 300 Mrad After normalizing for sensor area, results are similar to those for the PF sensor. Sensor area = 0.1 cm2 Temperature  -15 C 22 22

NF Charge Collection after 300 Mrad @600 V, ~55% charge collection loss (32C annealing) 23 23

New Results 24 24

Industrial Sapphire Sensor provided by Sergej Schuwalow Short collection length for mobile carriers  intrinsically small signals Current low (< 10 nA) after irradiation 25 25

Industrial Sapphire Charge Collection Low pre-irradiation charge-collection and significant charge loss after irradiation Sensors via Sergej Schuwalow, DESY Zeuthen 500 m thick Al2O3 300 Mrad Exposure 26 26

Industrial Sapphire High-Bias Annealing Trends Annealing does not seem to significantly restore charge collection 500 m thick Al2O3 300 Mrad Exposure Sensors via Sergej Schuwalow, DESY Zeuthen 27 27

LumiCal Sensor (N-Type) Fragment cleaved from broken sensor provided by Tel Aviv After treatment (heat, UV light) were able to bias center strip to ~150V After 300 Mrad irradiation, could bias to > 300 V 28 28

N-Type LumiCal Prototype Fragment Typical (~20 A/cm2) post-irradiation currents 300 Mrad Exposure 29 29

N-Type LumiCal Prototype Fragment Sensor via Sasha Borisov, Tel Aviv Significant charge loss; annealing studies continue 300 Mrad Exposure Sensor via Sasha Borisov, Tel Aviv 30 30

Summary and Outlook For all tested technologies, so far only p-type Si seems to maintain good charge collection as dose  1 Grad (might have been expected…) LumiCal sensors might work (charge collection and currents typicalfor n-type), but not optimal Currents do develop, but preliminary estimates have suggested overall power draw remains quite tolerable; these estimates are being refined (FLUKA simulation) A number of highly-irradiated Si Diode samples (300, 600 Mrad) await characterization, as does a 100 Mrad GaAs exposure (but poor performance was observed for 20 Mrad for GaAs)

BackUp BackUp… Bruce Schumm

GaAs 21 Mrad Exposure 33 33

GaAs Dark Current (-100 C) for 21 Mrad 45C anneal Room temp anneal 21 Mrad Exposure Before annealing Dark current as a function of annealing temp 34 34

GaAs Charge Collection (21 Mrad Exposure) Collected Charge (fC) Vbias (V) Charge Collection v. Bias and Annealing Temp 35 35

GaAs Charge Collection (21 Mrad Exposure)  Try even higher annealing temperatures  Higher exposure in next T506 run 21 Mrad Exposure Vbias = 600 V Slice at VB=600 vs. function annealing temp 36 36

Compare to Direct Electron Radiation Results (no EM Shower) A bit better performance than direct result kGy Pre-anneal Post-anneal at room temp Georgy Shelkov, JINR 1000 kGy = 100 Mrad 37 37

SiC Results Bohumir Zatko, Slovak Institute of Science 4H-SiC crystal geometry Irradiated to 80 Mrad 38 38

SiC Dark Current Before/After Annealing 80 Mrad Exposure 39 39

SiC Charge Collection after 80 Mrad @600 V, ~25% charge collection loss (60C annealing) 80 Mrad Exposure 40 40

Luc d'Hauthuille Bruce Schumm University of California, Santa Cruz Power Draw of the Beam Calorimeter as a function of Temperature & Radiation Dosage Luc d'Hauthuille Bruce Schumm University of California, Santa Cruz

Assumptions Power modeled as a function of radiation and temperature Power drawn scales linearly with radiation dosage Temperature dependent IV data was taken at SCIPP for a Si sensor exposed to 270 MRads of radiation after a 60˚C annealing process, by Cesar Gonzalez & Wyatt Crockett. A 3rd degree polynomial was fit to this I vs. T data, for a Bias Voltage = 600V

Overview The LCSIM framework was used to compute the energy deposited from 10 simulated background events (bunch crossings at 500 GeV collision energy) Energy deposited was then extrapolated for 3 years of runtime, and converted to radiation dosage Temperature was input and combined with radiation dosage to compute the power draw for each mm^2 pixel, at each layer, for 600V bias (charge-collection about 90% after 270 Mrad) Power draw of these pixels was plotted on a heatmap for a range of temperatures (-7, 0,7, 15 ˚C)

IV Curves at various Temperatures

Polynomial Fit for Temperature Dependent Current (600 V)

P(R,T) = (R/270MRads)*(600V)*I(T) Model for Power Draw Using these assumptions, power drawn by a pixel is: P(R,T) = (R/270MRads)*(600V)*I(T) where R is radiation dosage, T is temperature, 600V is the Bias Voltage and I(T) is the current given by the fit.

Layers 2 & 10 of BeamCal at T = 0˚C P_max = 4.59 mW (for a single mm2 pixel)

Power Drawn(Watts) of BeamCal collapsed T = -7 ˚C T = 0 ˚C P_total = 4.467 W P_total = 11.01 W P_max = 1.86 mW P_max = 4.59 mW (for a single pixel) (for a single pixel)

Total Power Draw (3 Years of Running with Si Diode Sensors Operating Temperature (0C) Total Power Draw (W) Maximum for 1mm2 Pixel (mW) 15 56 23 7 25 10 11 4.6 -7 4.5 1.9

Summary GaAs after 20 Mrad exposure retains low current. Charge loss severe but recovers with annealing. Need to try higher annealing temp and then larger exposure. 4H-SiC after 80 Mrad exposure suffers ~25% CC loss; possible annealing gain at higher temperatures. Currents remain low.  Try higher temp and larger exposure also. PF and NF silicon diodes shows significant CC after 3-year equivalent dose. Significant currents but overall power draw still low. Additional diode sensors (up to ~600 Mrad) await evaluation. Sapphire sensors now characterized and mounted. Low-noise amplifier being commissioned. 50 50

BACKUP 51 51

Looking Forward Cointue GaAs, SiC annealing studies 300 Mrad exposures of PC, NC, NF silicon diode sensors awaiting evaluation PF sensor exposed to another 300 Mrad (total 550-600 Mrad); awaiting study Low-noise amlipfier (<300 electrons) under development for exporation of Sapphire sensors; initial probe evaluation underway. Ongoing offer for more beam time at SLAC, but large backlog of sensors to study at SCIPP 52 52

SiC CC Before/After Annealing 80 Mrad Exposure 53 53

Confirmed with RADFET to within 10% Dose Rates (Including 1 cm2 Rastering) Mean fluence (cm-2) per incident e- Confirmed with RADFET to within 10% Maximum dose rate (e.g. 10.6 GeV; 10 Hz; 150 pC per pulse): 20 Mrad per hour 54 54

Summer 2013: Initial Si Doses “P” = p-type “N” = n-type “F” = float zone “C” = Czochralski 55 55

T-506 Idea Embed sample sensors in tungsten: “Pre-radiator” (followed by ~50 cm air gap) spreads shower a bit before photonic component is generated “Post-radiator” brings shower to maximum just before sensor “Backstop” absorbs remaining power immediately downstream of sensor Realistic EM and hadronic doses in sensor, calibrated to EM dose

Charge Collection Measurement For strip sensors use multichannel readout Median Collected Charge Channel-over-threshold profile Efficiency vs. threshold 57 57

GaAs I-V after 21 Mrad Exposure (-10 C) At 600 V, about 0.7 A (0.0005 W) per cm2 GaAs IV GaAs Dose of 21 Mrad Post-anneal Pre-anneal 58 58

Results: NF Sensor to 90 Mrad, Plus Annealing Study Dose of 90 Mrad Limited beneficial annealing to 90oC (reverse annealing above 100oC?) 59 59

~15% charge loss at 300 ns shaping Results: NC sensors Dose of 220 Mrad Incidental annealing ~15% charge loss at 300 ns shaping 60 60

Results: PF sensors Doses of 5 and 20 Mrad No annealing 61 61

Results: PC sensors Dose of 20 Mrad No annealing 62 62

G.P. Summers et al., IEEE Trans Nucl Sci 40, 1372 (1993) Departure from NIEL (non-ionizing energy-loss) scaling observed for electron irradiation NIEL e- Energy 2x10-2 0.5 MeV 5x10-2 2 MeV 1x10-1 10 MeV 2x10-1 200 MeV G.P. Summers et al., IEEE Trans Nucl Sci 40, 1372 (1993) Also: for ~50 MRad illumination of 900 MeV electrons, little loss of charge collection seen for wide variety of sensors [S. Dittongo et al., NIM A 530, 110 (2004)] But what about the hadronic component of EM shower? 63 63

Results: NF sensor for low dose Doses of 5 and 20 Mrad No annealing 64 64