The production of radiation tolerant vacuum phototriodes and their HV filters for the CMS endcap electromagnetic calorimeter

Transcription

1 The production of radiation tolerant vacuum phototriodes and their HV filters for the CMS endcap electromagnetic calorimeter Peter R Hobson* On behalf of the Group *School of Engineering & Design, Brunel University, UK 1

2 Overview Introduction Endcap calorimeter construction Radiation Environment VPT and HV filter: description and status: Photo-detector HV filter Other components Summary 2

3 Compact Muon Solenoid HCAL Total weight: 12,500 t Overall diameter: 15 m Overall length: 21.6 m Magnetic field: 4 T Muon chambers Tracker ECAL 4T solenoid Iron yoke 3

4 Endcap at LHC/SLHC η=1.48 Supercrystals and their internal components are inaccessible and cannot be replaced. Components: VPTs, HV pcbs, capacitors, resistors Signal & HV cable, quartz monitoring fibres η=3 Repair of SC array would require the dismounting of EE readout electronics on rear of backplate. High activation levels, (5 msv/hour at inner edge) thus access time limited. Qualify SC components for SLHC before EE build 4

5 Challenges & Choices Challenges: High radiation doses and neutron fluences (assume SLHC happens) (10 year doses: n/cm 2, 1kGy at η=0 2x10 14 n/cm 2, 50kGy at η =2.6) Cost, cannot use expensive materials, processes or components No realistic option to replace during CMS lifetime Like a space mission in this respect Long term stability Choices: Faceplates of Vacuum Phototriode (VPT) COTS HV components (capacitors and resistors) Cable insulation (HV and signal) Connectors, labels, thermal insulation etc. 5

6 Endcap Radiation Environment LHC 10 year normal operation assumed 6

7 Photodetectors: end caps Vacuum Phototriodes (VPT) Developed by CMS in conjunction with industry (Western, Japanese & Russian) φ =26.5 mm Vacuum devices offer greater radiation hardness than Si diodes Gain 8-10 at B = 4 T Active area of ~ 280 mm 2 /crystal Q.E. ~ 20% at 420 nm UV glass window much less expensive than quartz and no increase in tube length - much more radiation resistant than borosilicate glass MESH ANODE Order placed with RIE (Russia): devices delivered and tested 7

8 VPT production status 8

9 VPT and HV filter Two approaches All components must be have extreme radiation tolerance. Use vacuum photodetectors rather than the silicon Avalanche PD used for the barrel ECAL 1. Design key components using new materials 2. Extensively evaluate existing COTS components We have used a combination of these approaches New types of glass for photocathode faceplates Extensive testing of COTS components: cables, capacitors & resistors, structural elements, glues, thermal conducting compounds etc. 9

10 VPT faceplate glass NO! Surely it is easy just use quartz faceplates? Significant additional cost (approximately doubles price) Adds about 10 mm in length to the VPT (graded glass seals) and we seek to minimise space behind the calorimeter. Develop or find a glass that is radiation tolerant, and thermally matched to tube glass (a borosilicate). We use a Russian glass US49 (similar to Schott Technical glass 8337B) 10

11 VPT faceplate testing Samples of cut and polished faceplates are irradiated to 20 kgy 60 Co gamma ray dose in our facility at Brunel. At least 3 faceplate samples per glass batch are tested. Samples are irradiated in the dark, at a temperature around 20 C, at a dose rate of about 0.2 kgy per hour. Induced absorption is measured within two hours of irradiation. Irradiated samples are stored in the dark at 20±1 C. Spectrophotometer is recalibrated every 12 months. 11

12 VPT faceplates All batches of VPT faceplates are tested before VPT are manufactured. Only batches which result in < 10% induced loss (weighted by PbWO 4 scintillation spectrum) after 20 kgy are accepted. Induced Absorption Glass sample "a" after 19.3 kgy PbWO 4 Absorbance (OD) 0.15 Unirradiated Irradiated Wavelength (nm) Wavelength (nm) This batch was QA passed with an average (n=4) of 92.5% internal transmission of PbWO 4 spectrum after 19.3 kgy Irradiation causes the growth of an absorption band centred around 315 nm. 12

13 VPT faceplates Gamma induced damage in glasses typically saturates with dose (analogous to capacitor charging). Most damage in our region of interest ( nm) occurs after a few kgy. Glass sample "b" after 20.6 kgy Full dose Half dose 0.2 Induced Absorption Wavelength (nm) 13

14 Irradiation of complete VPT Samples of fully QA accepted VPT devices are also tested for radiation tolerance. Standard set of tests at 0T,1.8T and 4T are carried out. 95.0% 0T data 4T data 90.0% Relative response 85.0% 80.0% Tube ID Relative VPT response to pulsed blue LED after 10 kgy of gamma irradiation 14

15 HV filter components COTS passive components on card: 1 nf and 470 pf HV capacitors 22MΩ, 10M Ω and 100 Ω thin-film resistors 15

16 HV filter components 1nF 2.2kV rated capacitor 1 kv leakage current - unirradiated 60 Co irradiation tests About 10% decrease in capacitance with recovery over a few weeks. No change in leakage current. Some recent batches show essentially no change in capacitance. Resistors are completely unaffected Post-irradiation recovery of capacitance 1 kv leakage current after 345 kgy 16

17 Supercrystal items, Co 60 Eta (SLHC equivalent) Dose (kgy) VPT faceplates VPTs DC 3145 VPT-xtal glue Irradiation tests HT cable, 2KV Co-axial signal cable Capacitors (HV, unbiased10 Capacitors (HV, biased) 10 Resistors (HV,LV) 20 2 to 362kGy Thermal compound Carbon fibre alveolar composite strength tests to 5MGy!! All tests so far OK capacitors (unbiased) 10% change in value at worst Dynamic leakage in capacitors negligible Batch testing of HV filter components for QA during assembly period is now underway 17

18 Supercrystal items, Neutron Irradiation tests Eta (SLHC equivalent) neutrons/cm2*10** VPT faceplates 7.10**14 (PNPI, Apr 2003) VPTs **14 (PNPI, Sep 1999) Capacitors (biased) 5 Resistors (HV, biased) 4 All neutron irradiation tests so far OK 1 capacitor, measured under irradiation, long cables, -17% Tests carried out at Minnesota, 252 Cf source, 2.14 MeV neutrons Neutron rate 10 7 cm -2 s -1 rate at η = 3 at cm -2 s -1 Noise induced in VPT from local activation ~ 3200e e - at Compton electrons, from β s γ s, enter VPT faceplate Light, from electrons above Cerenkov threshold, yield VPT photo-electrons 18

19 Summary All faceplate glass samples were QA assured before VPT batches produced and complete VPT have been tested to check our procedures Neutron activation induced noise is acceptable and comparable to electronic only at inner edge of endcap All VPT have now been made for CMS endcap calorimeter HV cards use standard COTS components A set of capacitors, resistors, connectors and cables have been found that meet the electrical and radiation tolerance requirements No significant dynamic effects on capacitor leakage currents Samples of all other components used near VPT have been tested for tolerance to gamma irradiation We have achieved low-cost volume production of VPT and HV filter that meet the radiation environmental challenge of LHC and SLHC. 19

20 Spares 20

21 VPT dynamic currents under irradiation Complete tubes evaluated in a high-rate gamma ray cell (60-Co) Signal due to Cerenkov light in the UV glass faceplate RIE 51 RIE Anode current (na) Source OFF Dose (Gy) 21

22 VPT dynamic currents under irradiation Dynamic leakage current for capacitor alone (effect of leads subtracted). Irradiation rate about X10 greater than the maximum we expect to see in practice. Source OFF 22

23 ECAL Layout Pb/Si Preshowers: 4 Dees (2/endcap) 4300 Si strips (~ 63 x 1.9 mm 2 ) Tapered crystals Pointing ~ 3 o from vertex Barrel: 36 Supermodules (18 per half-barrel) Crystals (34 types) total mass 67.4 t Dimensions: ~ 25 x 25 x 230 mm 3 (25.8 X 0 ) η x ϕ = x Endcaps: 4 Dees (2 per endcap) Crystals (1 type) total mass 22.9 t Dimensions: ~ 30 x 30 x 220 mm 3 (24.7 X 0 ) η x ϕ = x x 0.05