The CDF Silicon Detector Performance and Longevity Alexandre

The CDF Silicon Detector: Performance and Longevity Alexandre Pronko Fermilab On behalf of the CDF Collaboration Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007

Tevatron in Run II • Record instantaneous luminosity ~284. 5 1030 cm-2 s-1 • Delivered luminosity: – Current: ~2. 5 fb-1 – Goal by 2009: 5 -8 fb-1 • Proton-antiproton collider • Center-of-mass energy 1. 96 Te. V • 36 36 bunches • Collisions every 396 ns • Two experiments: CDF & D 0 Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 2

CDF in Run II • Multi-purpose detector Muon Chambers • Physics program Silicon Tracker – Top Quark – CKM and Bphysics – Electroweak – QCD – Higgs & New Physics Wire Chamber (COT) Calorimeters 1. 4 T Superconducting Solenoid Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 3

Silicon Detector in Run II • Silicon Run II – 7 -8 concentric layers of silicon sensors (~6 m 2) – 722, 432 channels read out by 5456 chips – Dead-timeless readout • Three sub-detectors – Layer 00 (L 00) – SVX-II – Intermediate Silicon Layers (ISL) • Purpose – Precision tracking – Displaced vertex trigger SVX+L 00 Alexandre Pronko ISL Vienna Conference on Instrumentation, Feb. 18 -24, 2007 4

SVX-II • Mechanical design: – 3 barrels; 6 bulkheads; – 12 wedges per bulkhead 5 layers of double-sided sensors: 2. 5 cm<R<10. 6 cm y 2. 5 cm m – Layers 0, 1, 3 (Hamamatsu): axial and 90 – Layers 2, 4 (Micron): axial and 1. 2 stereo – Strip pitch: 60— 140 m – AC coupled readout: microdischarges limit bias voltage to 170 V (Layers 0, 1, 3) and 80 V (Layers 2, 4) ≈1 • z • x The core of the CDF Silicon – Displaced vertex trigger • Tight alignment • Fast wedge-wide readout (30 k. Hz) Alexandre Pronko 360 sensors; 3168 chips Vienna Conference on Instrumentation, Feb. 18 -24, 2007 5

Intermediate Silicon Layer (ISL) • 1 central layer x, y – | |<1, links tracks from SVX to Wire. Chamber (COT) • 2 forward layers – 1<| |<2 – Forward tracking z SVX+L 00 ISL • Double-sided sensors: – – 112 m strip pitch 1. 2 stereo angle Hamamatsu: Layer 5 Micron: Layer 6 1. 9 m 296 sensors; 2368 chips Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 6

Layer-00 (L 00) • Precise position measurement Hamamatsu – Mounted on Be beam pipe – Low material badget: ~0. 01 X 0 – 50 m effective strip pitch SVXII Inner bore Cooling channels 2. 3 cm Be Beam-pipe 4. 2 cm Sensors • Radiation hard single-sided (LHC style) sensors SCS Thomson/Micron – Non-oxygenated silicon (Hamamatsu, SGS Thomson) – Oxygenated silicon (Micron) – Actively cooled (<5 C) – Can be biased to 500 V 300 m installation clearance 72 sensors; 108 chips Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 7

Physics Performance • CDF Run II physics highlights: – – Tight constraint Mt= 172. 4 2. 7 Ge. V/c 2 on MH MW = 80413 48 Me. V/c 2 Observation of WZ production Measurement of BS oscillation • Silicon Detector played crucial role in all measurements A CDF BS Oscillation Sensitivity L=1 fb-1 95% CL sensitivity No L 00 3 sensitivity 5 sensitivity With L 00 Observed ms value of m. S Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 8

Silicon Detector Performance • Hard work to maintain system pays off… • Very stable after commissioning – ~92. 5% of sensor units integrated – ~86% of sensor units return data with <1% rate of digital errors – ~1/2 of digital errors can be corrected offline • Tracking with Silicon – – – Run number Alexandre Pronko Average efficiency ~94% Monitored using J/ Very stable since 2003 d 0~40 m for p. T>2 Ge. V Continuous work on improvement Vienna Conference on Instrumentation, Feb. 18 -24, 2007 9

Performance and Longevity Overview • The CDF Silicon detector must remain efficient till 2009 – SVX-II designed for 2 -3 fb-1 – Delivered to date: ~2. 5 fb-1 – Run II goal: 5 -8 fb-1 • Major concerns – Long term radiation damage – Beam incidents – Other component failures SVX+L 00 Components in Radiation field Longevity Status Optical transmitters 10% degradation at 200 k. Rad (or ~8 fb-1), no change in wave form Readout chips 17% noise increase at ~3 Mrad (or ~8 fb-1) for SVX Layer-0 Silicon Alexandre Pronko This talk ISL As far as we can get to Silicon… Vienna Conference on Instrumentation, Feb. 18 -24, 2007 10

Radiation Effects on Sensors • Effects of Radiation Damage – Effect-1: noise increase • Bulk & surface damage to sensors increase in leakage current, inter-strip capacitance • Damage to readout chips – Effect-2: signal degradation • Increased charge trapping in crystal defects loss in charge collection efficiency • Change in effective doping concentration higher depletion voltage inability to fully deplete due to bias voltage limitations • Lifetime is evaluated by 3 independent methods – Sensor depletion voltage (Vdep) monitoring • “Signal” method • “Noise” method – Direct Signal/Noise measurements – Sensor bias current monitoring Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 11

Radiation Environment • Measured by >1000 thermoluminescent dosimeters (TLD) – Ionizing radiation; low energy neutrons • Accurate radiation map • Dose dominated by collisions (>90%) – remainder from beam losses • Radiation scaling in Silicon volume – Radial scaling as ~1/r • Z-dependent: 1. 5< <2. 1 – Dose at r=3 cm & |Z|<45 cm: • 300 60 k. Rad/fb-1 – Integrated dose up to now (~2. 5 fb-1) • L 00 2. 5 MRad • SVX-II Layer-0 1 MRad Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 12

Depletion Voltage Basics • Evolution of depletion voltage, Vdep, due to radiation damage – Effective number of charge carriers, Neff, decreases until type inversion, decreasing Vdep – Increase of Vdep after type inversion, reaching maximum allowed bias voltage (170 V for Layer-0) • Predictions: – modified Hamburg model – Vdep~ Neff=NA+NC+NY • • • Alexandre Pronko NA( , T, t) beneficial annealing NC( ) stable component NY( , T, t) reverse annealing , T, t fluence, temperature, time study of leakage current T estimated from coolant temperature; largest source of uncertainty S. Worm, NIM A 549, 126 (2005) Vienna Conference on Instrumentation, Feb. 18 -24, 2007 13

Depletion Voltage: Signal vs. Bias Vdep Peak charge (ADC) • Study collected charge of silicon hits during beam • Find peak of ADC distribution as function of bias voltage (fit: Landau Gaussian) • Determine Vdep as 95% of amplitude Bias voltage (V) Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 14

Depletion Voltage: Noise vs. Bias • Use double-sided sensors • P-stops do not provide isolation unless fully depleted Þ large leakage current Þ inter-strip n-side thermal noise is minimal when bulk fully depleted • Vary Vbias & measure n-side noise • Pros… – Requires no beam – independent measurement of Vdep • Cons… – Only works for double-sided sensors (SVX-II, ISL) – not sure if it works after inversion Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 15

Depletion Voltage: Lifetime Estimate Data points are from Noise-Bias scans • Layer-0 of SVX-II is of most concern: Vdep<170 V – compare with 500 V for L 00 • • Layer-0 not inverted yet, expected after 2 fb-1 Assuming same slope after inversion, Layer-0 should outlast Run II – Vdep evolution follows “optimistic” prediction Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 16

Signal/Noise Ratio Basics • Ability to deplete sensors is important, but S/N ratio is really what matters – Limit-1: • S/N > 8 for displaced vertex trigger – Limit-2: • S/N > 3 for b-tagging • Effects of radiation damage – Signal • Decreases with L (i. e. , dose) • changes doping concentration & charge trapping – Noise • Increases with L (i. e. , dose) • Bulk & surface damage increased leakage current increased shot noise • Chip damage Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 17

Signal/Noise Ratio Measurement • Signal – Use tracks from unbiased J/ trigger sample – Most probable value of fit to ADC spectrum • Noise – Mean strip noise from calibration runs without beam – Runs taken every two weeks Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 18

Signal/Noise ratio: Lifetime Projection • S/N decreases with L – Dominated by leakage current shot noise – Micron sensors degrade faster than Hamamatsu • Physically inspired model L integrated luminosity a, b, c, d free parameters • Model extrapolation: SVX-II lifetime is not limited by S/N degradation • Continuing work on refining model Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 19

Summary and Conclusion • Silicon detectors in CDF: SVX-II, ISL, L 00 – Large and complex system: 6 m 2 of silicon; 722 K channels – Very stable performance – Crucial for success of physics program • Observed evolution of Vdep and S/N in agreement with estimates – Inner most layers of silicon are near type inversion • Studies suggest that the CDF Silicon Detector should outlast Run II without major degradation in performance • We continue to monitor detector performance – How will the detector behave after inversion? – Will the S/N predictions hold true? Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 20

Backup Slides Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 21

Silicon Detector as Dosimeter • Linear increase of bulk leakage current Ileak with fluence : – Ileak= V • ”damage” constant, V sensor volume • Assume Ileak Ibias • Obtain “damage” constant by comparing with TLD measurements • Obtain fluence as function of radius – e. g. , for SVX-II Layer-0: • Large uncertainty – temperature model: 13 C (no direct measurement) Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 22

Temperature of SVX-II Sensors • Fluence scale depends on temperature • Temperature is measured at bulkheads • Use finite element analysis to estimate temperature of sensors 15 Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 23

Lifetime Projections • Based on 1. 6 fb-1 of data • Work on updating results with 2. 5 fb-1 is in progress Alexandre Pronko Vienna Conference on Instrumentation, Feb. 18 -24, 2007 24