Scintillator-based Hadron Calorimetry at the...

ScintillatorScintillator--basedbasedHadronHadron CalorimetryCalorimetry

for the ILC/for the ILC/SiDSiDDhiman Chakraborty

International Linear Collider WorkshopsSnowmass, 14-27 Aug, 2005

Why (not) Why (not) scintillatorsscintillators??• Tested and true, well understood & optimized,• New developments in cell fabrication & photo-

detection help meet ILC/PFA demands – Fine segmentation at a reasonable cost– Photodetection and digitization inside detector ⇒min. signal loss/distortion, superior hermeticity

• Can operate in both analog and digital modes– Measures energy, unlike RPC & GEM that only offer

hit-counting (“tracking” or “imaging” calorimetry)– Remains a viable option if digital PFA fails to deliver.

ILCW2005, Snowmass Scintillator-based HCal for ILC Dhiman Chakraborty

2

Design Considerations: PFADesign Considerations: PFA• Need ≲10 cm2 lateral segmentation.• At least ~35 layers and ~4λ must fit in ~1 m along R.

– Min. Rin driven by tracker performance.– Max. Rout limited by magnet and material costs.– Min. absorber fraction limited by the need for

shower containment.• ⇒2 cm thick absorber layers if SS (less if W).• ⇒0.6-0.8 cm sensitive layers must respond to MIPs

with good efficiency and low noise.(cont’d...)


3

Design Considerations: PFADesign Considerations: PFA(…cont’d)

• Good lateral containment of showers is important for minimizing the confusion term.

• W absorber in ECal ⇒ e/π compensation is not built in ⇒ must be achieved in software ⇒ particle id (inside calorimeter by shower shape?) may be important.


4

Design Considerations: OthersDesign Considerations: Others• The technology must be

– Reliable,– Mechanically sound, – Operable inside strong (~4.5T) magnetic field,– Capable of 15+ years of running,– Tolerant to ~5σ fluctuations in T, P, humidity,

purity of gas (if any). Monitoring will be necessary if response depends strongly on any of these,

– Suited for mass-production and assembly of millions of cells in ~40 layers, (cont’d…)


5

Design Considerations: OthersDesign Considerations: Others• The technology must be (…cont’d)

– Allow hermetic construction (minimum cracks/gaps) – Safe (HV, gas, …),– Compatible with other subsystems (MDI?),– Amenable to periodic calibration,– Able to handle the rates (deadtime < 0.1 s?)– Cost-effective (including construction, electronics,

operation).


6

Hardware testsHardware testsCells made of cast (Bicron) and extruded scintillators

(NICADD/FNAL) have been extensively tested with many variations of– Shape (hexagonal, square),– Size, thickness– Surface treatment (polishing, coating),– Fibers (manufacturer, diameter, end-treatment)– Grooves (σ− shaped, straight)– Photodetectors (PMT, APD, SiPM, MRS)


7

Hardware testsHardware tests• Different cell and groove shapes with extruded and

cast scintillators


8

Hardware Prototypes (DESY Hardware Prototypes (DESY ““MiniCalMiniCal””))•DESY 6 GeV e beam 2003-2004•108 scintillator tiles (5x5cm)•Readout with Silicon PMs on tile, APDs or PMTs via fibres

2 cm steel

0.5 cm active

DES Y, Hamburg U, ITEP, LPI, MEPHI, Prague


9

DESY DESY ““MiniCalMiniCal”” Test Beam resultsTest Beam results• Resolution as good as

with PM or APD*• Non-linearity can be

corrected (at tile level)– Does not deteriorate

resolution – Need to observe

single photon signals for calibration

• Well understood in MC • Stability not yet

challenged

NIM A (2005)


10

Choosing the Optimum ThresholdChoosing the Optimum ThresholdEfficiency and Noise Rejection

0%

20%

40%

60%

80%

100%

120%

-0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

Number of MIPs

Perc

ent

EfficiencyNoise Rejection

0.25 MIP threshold: efficient, quietILCW2005, Snowmass Scintillator-based HCal for ILC

Dhiman Chakraborty11

Miscellaneous MeasurementsMiscellaneous Measurements

Response ratios between types, glues, fibers,…• Scintillator type: extruded/cast ≈ 0.7• Optical glue: EJ500/BC600 ≈ 1.0• Fiber: Y11/BCF92 ≈ 3.2

– Y11 = 1 mm round Kuraray,– BCF92 = 0.8 mm square Bicron

Extruded/cast (cost) ≈ 0.05


12

Optimum parametersOptimum parameters• Shape: Hexagonal or Square• Thickness: 5 mm• Lateral area: 4 - 9 cm2

• Groove: straight• Fiber: Kuraray 1 mm round (or similar)• Fiber glued, surface painted• Scintillator type: Extruded

But a bigger question is the photodetector:PMTs are costly, bulky, won’t operate in B field.We have been investigating APDs, MRS, Si-PM…

Based on what we have learnt so far


13

The Metal/Resistor/ The Metal/Resistor/ Semiconductor Photodiode (MRS)Semiconductor Photodiode (MRS)• From the Center of Perspective Technologies & Apparatus (CPTA),

• Multi-pixel APDs with every pixel operating in the limited Geiger

multiplication mode & sensitive to single photon,

• 1000+ pixels on 1 mm x 1 mm sensor,

• Avalanche quenching achieved by resistive layer on sensor,

• Detective QE of up to 25% at 500 nm,

• Good linearity (within 5% up to 2200 photons)

• Immune to magnetic field,

• Radiation-tolerant.


14

Study of MRS/Study of MRS/SiPMSiPM• Determination of Working point:

– bias voltage, – threshold, – temperature

• Linearity of response• Stress tests: magnetic field, exposure to

radiation.• Tests with scintillator using cosmic rays

and radioactive source.


15

Metal Resistive Semiconductors (MRS)Metal Resistive Semiconductors (MRS)

LED signal

0

200

400

600

800

1000

1200

1 33 65 97 129

161

193

225

257

289

321

353

385

417

449

481

513

ADC counts

Even

ts

Typical pulseheight spectrum

Poduced by the Center of Perspective

Technologies & Apparatus (CPTA)

B. Dolgoshein


16

SiPMSiPM SummarySummaryWe have conducted a set of measurements to illustrate the potential

use of Si photodetectors in High Energy Collider experiments in general, and for hadron calorimetry at the ILC in particular.

• Good MIP sensitivity, strong signal (gain ~O(106)),• Fast: Rise time ≈ 8 ns, Fall time < 50 ns, FWHM ≈ 12 ns (w/ amp)• Very compact, simple operation (HV, T, B,…),• Each sensor requires determination of optimal working point,• Noise is dominated by single photoelectron: a threshold to reject

1 PE reduces the noise by a factor of ~2500,• The devices operate satisfactorily at room temperature (~22 ˚C).

Cooling reduces noise and improves gain, • Not affected by magnetic field (tested in up to 4.4 T + quench),• No deterioration of performance from 1 Mrad of γ irradiation.


17

SiPMSiPM prospects on the horizonprospects on the horizon• Bigger SiPMs are under development

– 3mm x 3mm made, but require cooling to -40 C– 5mm x 5mm thought possible– cost increase: insignificant (CPTA), linear (H’matsu)?

• 5mm x 5mm may help us eliminate fibers– put the SiPM directly on the cells– wavelength matching by n-on-p (sensitive to blue

scint. light) or WLS film– hugely simplifies assembly

• Better uniformity across sample with purer Si.ILCW2005, Snowmass Scintillator-based HCal for ILC

Dhiman Chakraborty18

Simulation StudiesSimulation StudiesGeometries considered

Scint HCal

Steel 20mmPolystyrene 5mm

Gas Geom1

Steel 20mm

Gas 5mm

Gas Geom2

G10

Glass 1mm

Gas 1mm


19

ππ++ energy resolution as function of energy resolution as function of energy for different (linear) cell sizesenergy for different (linear) cell sizes


20

CompensationCompensation• Cell counting has its own version of the

compensation problem (in scintillators).• With multiple thresholds this can be

overcome by weighting cells differently (according to the thresholds they passed).

• In MC, 3 thresholds seem to be adequate.


21

ππ++ energy resolution vs. energyenergy resolution vs. energy

Two-bit (“semi-digital”) rendition offers better resolution than analog


22

NhitNhit vs. fraction of vs. fraction of ππ++ E in cells with E>10 MIP:E in cells with E>10 MIP:Gas vs. scintillatorGas vs. scintillator

2-bit readout affords significant resolution improvement over 1-bit for scintillator, but not for gas


23

ππ++ energy resolution vs. energyenergy resolution vs. energyMultiple thresholds not used


24

NonNon--linearitylinearity• Nhit/GeV varies with energy.• This will introduce additional

pressure on the “constant” term.• For scintillator, the non-linearity

can be effectively removed by “semi-digital” treatment.


25

ππ±± angular width: density weightedangular width: density weighted


26

Simulation SummarySimulation Summary• Large parameter space in the nbit-

segmentation-medium plane for hadron calorimetry. Optimization through cost-benefit analysis?

• Scintillator and Gas-based ‘digital’ HCalsbehave differently.

• Need to simulate detector effects (noise, x-talk, non-linearities, etc.)

• Need verification in test-beam data.• More studies underway.


27

TB: TB: ScintScint HCalHCal layer assemblylayer assembly


28

SummarySummary• Simulations indicate (semi-) digital approach to be

competitive with analog calorimetry• Prototypes indicate scintillator offers sufficient

sensitivity (light x efficiency) & uniformity.• Now optimizing materials & construction to minimize

cost with required sensitivity.• SiPM and MRS photodetectors look very promising.• Preparations for Test Beam (Analog tile HCal and Strip

tail-catcher/muon tracker) are in full swing.

All-in-all scint looks like a competitive option.We are moving toward the next prototype.


29

Thank you!Thank you!For further details, see talks given by DC at• The LC study group mtg on 26 May ’05,• The Beaune Photodetection Conference,

19-24 June ’05. Links athttp://www.fnal.gov/~dhiman/talks.html


30

Backup slidesBackup slides


31

Working point determination with LEDWorking point determination with LED

• The MRS is to able to separate single photoelectrons• Different response under identical setup ⇒ working point must be determined for each channel individually

0

500

1000

1500

2000

1 61 121 181 241 301 361 421 481

ADC Channels

Cou

nts

Channel 2

0

250

500

750

1000

1 61 121 181 241 301 361 421 481

ADC Channels

Cou

nts

Channel 4

0

250

500

750

1000

1 61 121 181 241 301 361 421 481

ADC Channels

Cou

nts

Channel 5


32

Cosmic MIP detection with Cosmic MIP detection with SiPMSiPM

Comparable to PMTComparable to PMT

Light output vs bias

0

10

20

30

40

50

51 52 53 54 55

Bias (Volts)


33

Noise Rate vs. Bias Voltage & ThresholdNoise Rate vs. Bias Voltage & Threshold

1

10

100

1000

10000

100000

1000000

10000000

48.5 50.5 52.5 54.5 56.5Bias (V)

Freq

uenc

y (H

z)

70mV

80mV

90mV

1

10

100

1000

10000

100000

1000000

10000000

0 50 100 150Threshold (mV)

Noi

se (H

z)

52V

50.6V

49.6V

• The right end of the plateau region in the Figure on left is optimal for our purpose.

• For thresholds in the range of 80±10 mV and bias voltage in

50.0±0.5 V, the dark noise is well under control.


34

Signal & Noise Amplitudes vs. Bias VoltageSignal & Noise Amplitudes vs. Bias Voltage

0

100

200

300

400

500

600

700

48 50 52 54 56Bias (V)

Am

plitu

de (m

V)

020406080

100120140160180

48 50 52 54 56 58

Bias (V)

Noi

se A

mpl

itude

(mV)

0

5

10

15

20

48 50 52 54 56 58

Bias (V)

S/N

• For this particular device S/N peaks at Vbias≈ 52 V • Sharp peaking in S/N⇒ working point must be found for each piece.


35

Temperature DependenceTemperature Dependence

0

100

200

300

400

500

19.5 20.5 21.5 22.5 23.5 24.5 25.5Temperature (oC)

Noi

se (H

z)

y = -14.369x + 667.86χ2/DoF=0.89

330

340

350

360

370

380

390

19.5 20.5 21.5 22.5 23.5Temperature (oC)

Sign

al A

mpl

itude

(mV)

• Bias = 51.3 mV, threshold = 80 mV

• Loss in signal amplitude with increase in T ≈ 3.5%/˚C


36

Fiber Positioning on MRSFiber Positioning on MRS

Optimal fiber-sensor

mating is crucial.


37

0

0.2

0.4

0.6

0.8

1

1.2

2.5 3 3.5 4 4.5Position, mm

Sign

al

0

0.2

0.4

0.6

0.8

1

1.2

-1 0 1 2 3 4

Distance from Sensor, mm

Sign

al

0

0.2

0.4

0.6

0.8

1

1.2

-3.5 -1.75 0 1.75 3.5Angle, o

Sign

al

Linearity of ResponseLinearity of ResponseSince the response of an individual pixel is not proportional to nγ,

(unless it has had time in between to recover), non-linearity is

expected when the detector receives a large number of photons.


38

0.8

0.85

0.9

0.95

1

1.05

1.1

0 1500 3000 4500

Number of Photons

Res

pons

e: m

easu

red/

"bes

t"Deviation reaches 5% (10%)

at nγ ≈ 2200 (3000) or,

nPE ≈ 550 (750).

One MIP ≈17 PE

⇒ up to 32 MIPs can be

measured within 5%

linearity.

Stress Tests: Effect of Stress Tests: Effect of MagMag. field. fieldNo significant effect of fields up to 4.4 T and quenching at 4.5T:


39

Stress Tests: Effect of IrradiationStress Tests: Effect of Irradiation


40

• No detectable damage from 1 Mrad of γ:y = 6E-05x + 1χ2/DoF=0.857

0.92

0.94

0.96

0.98

1

1.02

1.04

1.06

1.08

48 50 52 54 56 58Bias (V)

Rat

io o

f Fre

quen

cies

y = -1E-05x + 1χ2/DoF=0.851

0.95

0.97

0.99

1.01

1.03

1.05

0 40 80 120

Threshold (mV)

Fatio

of N

oise

Fre

quen

cies

y = 3E-05x + 1χ2/DoF=0.873

0.96

0.98

1

1.02

1.04

48 50 52 54 56 58

Bias (V)

Rat

io o

f Sig

nal A

mpl

itude

s

Hit timingHit timing


41

Hit timing (contd.)Hit timing (contd.)


42

Hit timing (contd.)Hit timing (contd.)


43

Hit timing (contd.)Hit timing (contd.)Scintillator

RPC

•Same Z→jj event at pole

•Same cell size (1cm x 1 cm)

•Same threshold of 0.25 MIP


44

Hit timing (contd.)Hit timing (contd.)ECal hits in the same events as on the last slide


45

Time of flightTime of flight

t (ns)

dNhi

t/dt


46

TimeTime--ofof--flight dependence of resolutionflight dependence of resolution

100ns/5µs


47

Avalanche PhotoAvalanche Photo--DiodesDiodesHamamatsu APD gain vs V @ diff wavelengths (T= 18 ºC)

1.0

10.0

100.0

1000.0

100 150 200 250 300 350 400

Bias Voltage, V

Gai

n

486nm 565nm for 587nm 660nm


48

Scintillator-based Hadron Calorimetry at the...

Documents

Transcript of Scintillator-based Hadron Calorimetry at the...