10 th international conference on instrumentation for colliding beam physics budker institute of...

1010thth INTERNATIONAL CONFERENCE INTERNATIONAL CONFERENCE

ON INSTRUMENTATION FOR COLLIDING BEAM PHYSICS ON INSTRUMENTATION FOR COLLIDING BEAM PHYSICS

Budker Institute of Nuclear Physics,Budker Institute of Nuclear Physics,Siberian Branch of Russian Academy of Science,Siberian Branch of Russian Academy of Science,

Novosibirsk, RussiaNovosibirsk, Russia

February 28 - March 5, 2008February 28 - March 5, 2008

Operation of the CDF Silicon DetectorOperation of the CDF Silicon Detector

Feb 29th, 2008 Ricardo Eusebi - INSTR 08, Novosibirsk, Russia 2

At the core of the CDF detector

Largest operating silicon detector 7-8 concentric layers of silicon 7 m2 of silicon with 1.2 cm < r < 32 cm 722,432 cha., 5644 chips, 704 sensors

Silicon Detectors at CDFSilicon Detectors at CDF

Silicon will have to survive through Run IIb (6/8 fb-1)

Designed only for Run IIa (~2/3fb-1) Upgrade for Run IIb was cancelled!


Silicon Sub-detectorsSilicon Sub-detectors Three Sub-detectors

SVX II: 5 double sided layers Intermediate Silicon Layers (ISL): 3 double sided layers Layer 00 (L00): Single sided, LHC-style sensors


Sub-detector: SVX IISub-detector: SVX II SVX II: The core of the silicon systems

Overall dimensions: 1 meter along beam directionRadii from 2.5 to 10.6 cm

Structure: three identical barrels 2 bulkheads12 wedges 5 concentric silicon layers

Silicon layersStrip pitch: 60 to 140 mLayers 0,1 and 3

(Axial and 90º strips)Layers 2 and 4

(axial and 1.2º strips)SVX II barrel

SVX II before installation


Sub-detector: L00Sub-detector: L00 Layer 00: Right onto the beam pipe

Overall dimensions: 1 meter along beam directionRadii from 1.2 to 2.1 cm

Structure: one single layer 2 bulkheads

– Three consecutive sensors

6 wedges

Silicon layersStrip pitch: 25 mAxial stripsRadiation tolerant (LHC style)

L00 during installation

1.2 cm2.1 cm


Sub-detector: ISLSub-detector: ISL ISL: Intermediate Silicon Layers

Overall dimensions: 1.9 m along beam directionRadii from 20.5 to 29 cm

Structure: three different barrelsCentral:

Single layer,14 wedgesForward:

Two layers,12 and 18 wedges Each barrel two bulkheads

Silicon layersStrip pitch: 55 mAxial and 1.2º strips

ISL sensors


Operational IssuesOperational IssuesDuring commissioning:During commissioning:

Blocked Cooling lines Blocked by glue, well inside the detector Solution: open them up with a powerful laser

Resonances Wire bonds to the magnetic field Synch. Readout wire oscillate and break Solution: Stop high frequency synchronous

readouts.

Beam Incidents High dose accidentally delivered to the detector Solution:

Collimators in key parts of the Tevatron New Diamond based BLM system.

After commissioning: Infrastructure & Aging

Jumper

B

groove left by beam


Infrastructure & Aging: Power SuppliesInfrastructure & Aging: Power Supplies Common failure modes of CAEN SY527

Communication loss Corrupted read back of voltages/currents Spontaneous switch off

Failure mode of power supply modules: Voltages in Analog, Digital and Port-Card

supply start slowly dropping. Up to 47 Power supplies started to show this.

Problem: aging of one type of capacitor 36 capacitors per power supply Can result in bit errors

Solution: Wait for the shutdown of September 2007 and …

take all faulty power supplies out replace all 36 capacitors (on FNAL site) put them back in and test them on location.

Time intensive effort, lasted about 2 months. Not enough time to change all Still expect to replace others as failure appears

All power supplies with this failure were replaced!


Infrastructure & Aging: Cooling LinesInfrastructure & Aging: Cooling Lines Cooling Lines

Symptoms: electronic-valves start failing.

Problem: ISL cooling line (10% glycol in water) became ACIDIC (ph=2) during the 2006 shutdown

Solution: coolant neutralized by draining and larger use of de-ionizing resin bed

Welds of the aluminum rings that cool optical transmitter had already been corroded One meter from the closest accessible point Why there ?

Corrosion-resistance: is alloy-dependent Heat affected zone around junctions

manifold most sensitive (alloy: 6061-Al).

Ion chromatography analysis showed carboxylic acids, mostly formic acid.

Likely came from the oxidation of glycol


Infrastructure & Aging: Cooling Lines RepairInfrastructure & Aging: Cooling Lines Repair Started shutdown of 2007:

Keep the silicon cold and dry at all times A plastic tent was setup to work.

A custom made air dryer changed the volume every 2 minutes. Dew Point was always kept below -10 Cº.

Basic Idea: Cover holes with epoxy from

the inside of the pipe using borescopes and catheters.

Repairs took a month 4 shifts of people

Current tests: Tight vacuum in the repaired lines

Hole

Repaired cooling system has been running stable for months !


Environmental Effects on Silicon SensorsEnvironmental Effects on Silicon Sensors

Radiation effects: Modifies the crystal structure of the sensor bulk.

Intrinsic parameters change with time.

P-N junction evolves, and eventually disappears with time.

Annealing effects: Due to temperature change of the sensors.

Defects created by radiation are strongly affected.

Performance of the sensor degrading with time: AGING Depletion voltage increases with time.

Sensor has a maximum breakdown voltage.

Can we fully deplete the sensor until the last day of operation ?

Signal decreases with time and noise increases with time Can we keep good signal and noise levels until the last day of operation ?


Measured using more than 1000 thermo-luminescent dosimeters (TLDs) Two different data-taking periods allowed for distinction between fields:

Radiation field is collision-dominated and scales with

Radiation FieldRadiation Field

(See R. J. Tesarek et al., IEEE NSS 2003)

due to beam losses due to pp collisions

p p

2.1(z)1.5 with ,)( zr

How this field affects the silicon sensors ?


Depletion Voltage: Signal Vs Bias Depletion Voltage: Signal Vs Bias Look at the charge collection distribution

Reconstruct a track w/o using the studied sensor If track points to hit in sensor record its charge

Charge collection distribution Follows a landau distribution Distribution is smeared by intrinsic noise Fit the curve to a Landau convoluted by Gaussian

(4 parameter fit)

Depletion Voltage: Maximal for a fully depleted sensor Study charge collection as function of VBIAS

Identify charge of Most Probably Value (MPV)in each distribution


Depletion Voltage: Signal Vs. BiasDepletion Voltage: Signal Vs. Bias Plot charge’s Most Probable Value for different bias voltages Fit to a sigmoid (parameters include the plateau of maximum charge) Define depletion voltage Vd

Our criteria: voltage that collects 95% of the charge at the plateau

Depletion Voltage as a function of luminosity 3rd order polynomial fit around the inversion point Linear fit to extrapolate to the future


Depletion Voltage: Noise Vs BiasDepletion Voltage: Noise Vs Bias Take advantage of double sided sensors, that have strips on the back side Depletion zone grows from the p+ side Noise on the other side’s strips (n+) reduce when the depletion zone reaches them.

Need a criteria for defining Vd

We use 95% reduction in noise between the two plateaus No beam required

no interference with data-taking Does not work after the sensor underwent inversion.

Depletion zone generated differently


Depletion Voltage: ResultsDepletion Voltage: Results Prediction for L00

Depends on type of sensor Oxygenated ladders invert much later

Prediction for SVX-L0

We should be able to deplete sensors until the end of Run II


Signal to Noise Ratio Signal to Noise Ratio

Signal Use J/ +- tracks Get total charge of cluster Decrease linearly with Lum.

Mean Strip Noise Average over strips in charge cluster Obtained from calibrations taken

every two week. Square root increase with Lum.

The figure of merit of the performance is the Signal to Noise Ratio (S/N) Signal: charge collected when a charged particle crossed the sensor Noise: intrinsic noise of the detector


Signal to Noise RatioSignal to Noise Ratio Fit of S/N

Limit I: S/N=8 (SVT eff.)S/N = 6, ~5% loss in SVT eff.

Limit II: S/N=3 (B tag eff.)

Sensor-type behavior Layers 2,4 (Micron) Layers 0,1,3 (Hamamatsu)

First layer careful monitoring to see if it is

going to be useful at 4/5 fb-1.

Most of the silicon layers will be fully operational until the end of Run II


ConclusionsConclusions

Operational Issues: Advice : expect the unexpected. We have recovered cooling to the full detector subsystems. Power supply modules with unstable voltages fixed at FNAL

Sensor Aging Data indicates that we will be able to fully deplete the sensors

The innermost layers of the detector have passed type inversion

Studies indicate we will remain with a very effective S/N ratio.

The CDF Run II Silicon Detector will continue successful operation for the rest of Run II !!

Backup Slides


Bias current increases with integrated radiation Theory:

Can be used to obtain flux

Comparison with TLD’s prediction

Radiation Effects on Silicon SensorRadiation Effects on Silicon Sensor

V Φ αIleak sensor’s volume

damage factor

fluence

dLαV

I

dL

ΦFlux leak

(See R. J. Tesarek et al., IEEE NSS 2003)


Fermilab (1967) Large number of H.E.P. projects

Tevatron Run II (2001–2009) Proton-antiproton collider Two multi-purpose detectors

CDF & DØ

2 km

Tevatron

[Fermilab Visual Media Service]

Fermi National Accelerator Laboratory – Aerial View

Proton-antiproton collider √s = 1.96 TeV, 36×36 bunches Record instant. peak luminosity

292 µb–1 s–1 (1 µb–1 s–1 10–30 cm–2 s–1)

Expect 6–8 fb–1 by end of Run II

Fermilab and the Tevatron Fermilab and the Tevatron

10 th international conference on instrumentation for colliding beam physics budker institute of...

Documents

cdf silicon detector

sensors silicon

m2 of silicon

beamricardo eusebi instr

cmricardo eusebi instr

concentric layers of

beam directionradii

detector solution