5th International Workshop on Radiation Imaging Detectors

Riga11th September 2003

Paula Collins, CERN

Silicon based detectorsVertexing for Linear Collider Silicon Pixel sensors –

MAPs, HAPs, CCDs, DEPFETsTracking for Linear Collider Silicon drift

High radiation environments Lazarus, 3d detectors, edgeless

devices, diamond Large system operation DELPHI, CDF, CMS

2

Silicon for vertexing @ the LC

Use a silicon based pixel detector Confine the e+e- background with

a high solenoidal field Keep occupancy reasonable by

reading inner layer in 50 sec

(IP) < 5 m 10 m/(p sin3/2 )

best SLD 8 m 33 m/(p sin3/2 ))

Vertex detector characteristics

point resolution 1-5 m

Thickness ~ 0.1 % X0

5 layers

Inner radius ~ 1.5 cm

required performance

ttee disentangle complexevents

discriminate b from cand from background

,,, ggccbbH

qqbqqb Train Time

LEP0.2 TeV

NLC/JLC

TESLA0.5 TeV

bunches perTrain

4 190 2820

Rep rate, Hz 45500 100 5

3

Silicon TrendsBasic idea Start with high resistivity silicon

More elaborate ideas:•n+ side strips – 2d readout

•Integrate routing lines on detector•Floating strips for precision

Hybrid Active Pixel SensorsChip (low resistivity silicon)

bump bonded to sensorFloating pixels for precision

CCD: charge collected in thin layerand transferred through silicon

MAPS: standard CMOS waferIntegrates all functions

chip chip

chip

n+

n+

p

DEPFET:Fully depleted sensor

with integrated preamp

Al strip

amplifier

SiO2/Si3N4

+ Vbias

+

+

+

+--

- n bulk

p+

n+

chip

4

+

+

+

+

•Amplifying transistor integrated into high resistivity silicon detector

•Low noise operation possible at room temperature

•Thinning possible to 50 m

0V

15 V

0V

DEPFET sensorsKemmer, Lutz, 1987

R&D for tracking ~ 2000

Image of toothed watch wheelResolution: 9 m (50 m pixels)

NIMA 465 (2001) 247-252

R&D: pixel size, power, thinning, speed

30

0

m

5

Steuerchips

Auslesechips

520 x 4000 pixelDEPFET-Matrix

(25 x 25µm Pixel)

Auslesechips

n x mpixel

IDRAIN

DEPFET-matrix

VGATE, OFF

off

off

on

off

VGATE, ON

gate

drain VCLEAR, OFF

off

off

clear

off

VCLEAR, ON

clear,clearcontrol

output

0 suppression

VCLEAR-Control

3 steps for each matrix row

•Read all signal currents

•Reset pixel row via clear contacts

•Read pedestal currents

Proposed TESLA Layout

DEPFET readout scheme

R&D issues

•Mimimise pixel size•Thin from backside•Minimise power consumption•Speed up readout cycle

6

Recent DEPFET R&DThinning technology

Aim for 0.1-0.15% X0 per

layer (including chips)

NIM A 511 2003 (250-256)Also:Move from circularly shaped JFETs to linear structures to access smaller pixel sizes – currently achieved 30 x 20 m2

Development of new readout chipNIM A 511 2003 (257-264)

1.5 mm

4 mm

7

Same unique substrate for detector and electronics

No connections (e.g. bumps)

Radiation hardness (no bulk charge transfer)

Advantages of CMOS process: Easy Design/good yield/low power/Rad hard

Very small pixel sizes achieveable

1999 – small scale prototypes

1999-2000 first beam tests

2001 – large prototypes

2002 – first circuit with on-line processing

Mimosa I II III IV V VI

Process 0.6 m 0.35 m 0.25 m 0.35 m 0.6 m 0.35 m

Epi layer 14 m 4.2 m 2.3 m 0 (!!!) 14 m 4.2 m

# pixels 64x64x4 64x64x6 128x128x2 64x64x4 1,000,000x7 24x128x..

Monolithic Active Pixel Sensors (MAPS)~

15

m

8

Irradiation of MIMOSA I,II & V up to 1013 1 MeV/c neutrons

MAPS Beam Test Results

Gain constantNoise constantLeakage current moderate rise collected charge 50-70% of initial value

(smooth decrease after 1012 or few x 1011 )

MIMOSA I MIMOSA II MIMOSA V (120 m)

Resolution 1.4 m 2.2 m 1.7 m

Efficiency at S/N > 5 99.5 % 98.5 % 99.3%

Signal to Noise Resolution [m]

NIMA 478 (2002) 311-315

MIMOSA I

MIMOSA V

9

MAPS: Next R&D steps

MIMOSA VI: First prototype with column parallel readout

30 columns (128 Pixels each) with parallel readout30 MHz clock, 5 MHz readout (6 clocks per pixel)New pixel design allows on-pixel signal amplification and double sampling operation

Also:MIMOSA VII, <50 um thinning, explore fabrication processes, simulate irradiation, etc.

10

Charged Coupled Devices - CCDs

CCDs invented in 1970 – widely used in cameras, telescopes etc.

Tracking applications for HEP:

1980-1985 NA32 120 kpixels

1992-1995 SLD 120 Mpixels

1996-1998 SLD upgrade 307 Mpixels

TESLA 799 Mpixels

~1000 signal electrons are collected by a combination of drift and diffusion over a ~20m region just below surface

• Small pixel size – 20 x 20 m

• Possibility of very thin detectors

•Column parallel readout: serial register -> direct bump bonding to chip

10V

2V

11

• Speed up readout 5 MHz readout -> 50 MHz Reduce clock amplitudes 10V->3V -> 1.5V Build with high resistivity epitaxial material

CCD R&D for LC requirements ICCD 55Fe spectrum: 50MHz 3V

clocks

noise=2.8 counts

280.125 MHz/250 ms6.9125 x 22605

7614.350 MHz/50 ms3.3100 x 13151

Integrated background(kHits/train)

Background (Hits/mm2)

Clock / readout timeCCD Size (Mpix)

CCD lxw(mm x mm)

Radius (mm)Layer

More info at: http://hepwww.rl.ac.uk/lcfi/

First ever CPCCD and chip (CPR1) now available!

Bonded assembliesfor testing thisyear (watch VERTEX03)

12

Other steps:

• Study radiation resistance to LC doses of 100Krad ionising radiation + 5 x 109 neutrons Temperature dependence Sacrificial charge

• Minimise material in fiducial volume Unsupported “stretched silicon” option Highly thinned silicon glued to substrate

CCD R&D for LC requirements II

S. Hansen, VERTEX02

Be substrate

~300µm(0.12% X

0)

CCD

Recessed NuSil adhesive pads

CCDBe

substrate

1mm

CCD 20 μm CCD bonded with adhesive pads to 250 μm Be substrate

On cooling adhesive contracts more than Be pulls Si down on to Be surface

Layer thickness 0.12% Xo

IEE Trans Nucl Sci, Vol 47, No. 6

13

Integration into LC design

• Baseline design 5 layers R_min=15 mm 3 layer coverage

to |cos |=0.96 5 layer coverage

to |cos |=0.9

Cryostat for CCD option

flavour tagging performance at s=91 GeV

b jets roughly equal SLD performancec jets improved by factor 2-3 in efficiency

Efficiency

Puri

ty

14

Silicon for tracking: Silicon Drift Detectors

• Principle of sideways depletion – as for DEPFET sensors

• p+ segmentation on both sides of silicon

• Complete depletion of wafer from segmented n+ anodes on one side

y

x!! Drift velocity must be predictable

Temperature control resistivity control Calibration techniques

• SDD fully functioning in STAR SVT since 2001

• 216 wafers, 0.7 m2

• 10 m in anode direction• 20 m in drift direction• Particle ID

15

Silicon for tracking: Drift detectors

• SDD are a mature technology – attractive for LC

• 5 precise silicon layers to replace TPC

• 56 m2 silicon• R&D needed:

Improve resolution to 5 m Improve radiation length Improve rad hardness

• Track stamping possible at nanosecond level

2 separation for out-of-time tracks for different drift direction configurations

16

Irradiation

LHC upgrade/VLHC will be WORSE!

The LHC environment will be FIERCE

L = 1034 cm –2 s-1

8 x 108 pp collisions / s

Hadron fluences to 1015 cm-2

1014

1013

neq/c

m2

per

year

LHCb vertex detector

radius [cm]

Other collider upgrades are hot tooe.g. Super Belle/Babar

L = 1036 cm-2 s-1

Dose ~ 5-10 MRad

Improved semiconductor designs/materials are well worth considering

17

Irradiation

• focussing on Defect engineering New detector materials Cryo/forward operation 3d and thin devices

Bulk Damage Effects in SiliconIncreased Leakage Current

Noise

Hard to bias

Effective Doping Changes

Depletion grows from n+ side

will coverthese points

Annealing effects

Buildup of negative space charge worsens in time

Strongly temperature dependent

Dep

leti

on

volt

ag

e [

V]

time [years]

Dep

leti

on

volt

ag

e [

V]

Fluence

rd50.web.cern.ch/rd50

18

Irradiation: strips for LHCb

•Reminder from Ramo (1939)

1 – 3 x 1014 n/cm2

Irradiated detectors

Vbias

Q = e * d/w Vbias

Underdepletion has two bad consequences

Charge loss Charge spread:

A killer for fine pitch detectors!

NIM A 412 (1998) 238 Similar story for trapping…

19

Mr. RamoI co-invented the electron microscope

I pioneered microwave technology

I founded TRW

I had a theorem

20

Irradiation: strips for LHCb•Reminder from

Ramo (1939)

1 – 3 x 1014 n/cm2

Irradiated detectors

Vbias

Q = e * d/w Vbias

Underdepletion has two bad consequences

Charge loss Charge spread:

A killer for fine pitch detectors!

NIM A 412 (1998) 238 Similar story for trapping…

21

Irradiation: strips for LHCb

Effi

cie

ncy

LHCb

ATLAS

For LHCb

n-on-n detectors are the technology choice

Vbias

Charge spread causes problems on the p side only

Up to ~1014 underdepletion is still more important than

trapping

NIM A 450 (2000) 297

NIM A 440 (2000) 17

LHCb

ATLAS

n side

p side

p side

n side

Re

solu

tion

[m

]

Vbias

VbiasVbias

22

Irradiation: the Lazarus effect• Cool detectors have little

leakage current• Cool detectors don’t reverse

anneal• Possible to control doping –

hence underdepleted detectors magically become depleted

COLD is COOL – as Lazarus knew

VdeplNeff [cm-3]

positive space charge“n”

Negative space charge“p”

h-trapping

e-trapping

Irradiation

It’s all about space

charge!

Vdepl | space charge (Neff) |

23

Irradiation: the Lazarus efect

Resurrection of a microstrip detector

Recovery is temporary – but this can be solvedNIM A 440 (2000) 17

NIM A 440 (2000) 5

Reverse biasForward bias

0 min

5 min

15 min

30 min

charg

e c

olle

ctio

n

EfficiencyHOT

EfficiencyCOLD

V bias

24

Irradiation: alternative materials

Defect Engineering

Oxygen enriched siliconOxygen dimer in silicon

New Sensor Materials

Silicon CarbideAmorphous siliconCompound semiconductorsDiamondCzochralski silicon

Rule of thumb: Below 1015 neq cunning manipulation of space charge can often save you

Above this value, new materials and/or device engineering is better

Will discuss these

25

Alternative Materials: DiamondsBand gap 5 x Si•Cohesive crystal•Room temperature operation

Ionisation energy high: MIPgives 2x less signal for same X0

Low low capacitance

Low Ileak low noise

Fast signal collection time

Collection efficiency < 100%

FWHM/MP ~ 0.95 (for silicon is 0.5)

Diamonds traditionally grown by CVDNew: homoepitaxic monocrystalline films

Figure of merit: Collection distance (proportional to mobility and electric field)

26Residual shifts measured

Alternative materials: Diamonds

RD42 in collaboration with Element Six have achieved impressive improvements in collection distances

pCVD structures show good radiation hardness

52% loss of S/N at 2.9 1015 p/cm2

23% improvement in resolution

+ Big step forward in understanding non-uniform response (A. Oh, to be published)

Lateral field component simulated

27

Alternative materials: Diamonds

Mono-crystalline diamond showsspectacular performance

Future R&D stepsContinue monocrystalline diamond studiesPush pCVD to 300m40 MHz readout speedsApplicationsBeam monitoring for BaBar, CMS

Charg

e C

olle

ctio

n D

ista

nce

(m

m)

E field (V/micron)

28

Alternative materials: CzSingle crystal pulled out of melt (High resistivity Czochralski silicon now available)

Oxygen content:

Standard silicon ~1e15cm-3

DOFZ silicon ~1e17cm-3

Cz silicon ~1e18cm-3

First Cz measurements show•Depletion voltage changes less drastically after irradition•No type inversion up to 1015 / cm2 190 MeV pions

See Lindstroem et al: http://rd50.web.cern.ch

29

Alternative materials: Cz

167V Bias

First full scale detector prototypesirradiated and operated in test beam at 40 MHz Summer 2003

243V Bias 371V Bias

30

Irradiation: 3d detectors

• Maximum drift and depletion distance governed by electrode spacing

Lower depletion voltages Radiation hardness Fast response At the price of more complex processing Narrow dead regions on edges

Proposed by Parker, Kenney

1995

Unit cell defined by e.g. hexagonal array of electrodes

20

0m

Electrode

Planar technology 3-D technology

20

0m

50m 10 m

p+ p+ n+n+

Comprehensive summary: CERN Courier Vol 43, Number 1, Jan 2003

31

3d detectors: characteristics

Performance after irradiation~1015 p/cm2

Low leakage currents Low depletion voltages Gaussian X ray lines fast charge collection

Am241 spectrum

14 KeV

17 KeV

59.5 KeV

v50 V 100 V 150 V

Charg

e

Colle

ctio

n

good depletionvoltage

Fast charge collectionrise time ~ 3.5 ns

32

Active Edge Planar Devices

GUARD RINGSinks surface leakage current

E-fieldp + Al

n + Al

E-fieldp + + Al

n ++ Al

n ++ Al

PLANAR PLANAR-3D = PLANAR DETECTOR + DOPANT DIFFUSED IN FROM DEEP ETCHED EDGE THEN FILLED WITH POLYSILICON

Microcracks, chips etc..

A TRENCH IS ETCHED AND DOPED TO TERMINATE THE E-FIELD LINES

AFTER THE FULL PROCESS IS COMPLETED THE MATERIAL SURROUNDINGTHE DETECTORS IS ETCHED AWAY AND THE SUPPORTWAFER REMOVED : NO SAWING NEEDED!!! (NO CHIPS, NO CRACKS)

33

Active Edge Planar Devices

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 50 100 150 200 250 300 350 400

Position (microns)

Sig

nal

(a.u

.)

Strip 1

Strip 2

20 Volts BiasJuly 2003

X ray microbeam signal on 1st and 2nd strips of edgeless device

MEASURED DEADEDGE ~ 5 m

34

Silicon for tracking: Large Systems

DELPHI 1990

DELPHI 1994

DELPHI 1996

CDF 2001CMS 2006

2 !

35

Silicon for tracking: Large Systems

LEP Tevatron LHC

Whoops…

36

CDF & D0 are demonstrating the possible

A huge system is up and running:D0 K

ND=5632019 micron resolution (before alignment)9 micron resolution (after alignment)S/N > 10 - as expectedSilicon participating in triggerSilicon is used for physics!

37

CDF silicon sensitive to !

Radiation profile measured with high precision (~1 mrad) by 1000 TLD’s

~300 rad/pb at r=3cm (~1010 mips/pb)

1/Rwhere 1.5 < < 2.1

This can be compared to risesin leakage current…

… thanks to accurate temperature and current monitoring

38

CDF silicon sensitive to !C

urr

en

t (u

A)

Delivered Luminosity (pb)

Corrected data give beautiful

fits

And very good agreement with TLD’s

Fluence

(1

010 c

m-2 p

b-1

)

(rad)

Also compatible with very small noise increase

39

Construction: philosopy shift

Taken from: DELPHI (888 detectors, 8 geometries) CDF (8000 sensors, 8 geometries) CMS (25000 sensors, 15 geometries)

Each sensor treated individually, nurtured into

life in many hours of careful handling

Systematic assembly line production with decent QA systems

40

Construction: Tales of the unexpected

Range from the glamarous

•resonating wire bonds

•Endoscopic operations on cooling tubes

•super long kapton problems

•Chemically active packing materials

Through to the less glamarous

(Vendors lie)

41

Cautionary tales IDELPHI “sticky plastic saga”

Received sensors from vendor, tested and distributed to assembly labs. All = OKAssembly labs got worse results – confirmed at CERNUS TO VENDOR: YOUR SENSORS AGE!VENDOR: YOU ARE RUINING THEM!

Finally found “flakes”

Zoom on flake thru packing

Zoom on packing

Vendor had changed anti static packing plastic – 60 sensors affected, big delay

A story repeated withvariations elsewhere

42

Jumper bond wires route signal from Rto

Rz side of module

Cautionary tales IICDF “resonating bond saga”

Under a very particular set of conditions:

•L2 “torture test” or SVT trigger•Bonds orthogonal to 1.4T field•Large current swing (100 mA – only on one bond)

Ipp ~ 160 mA

If pulsed at the right frequency the tiny Lorentz force (10-50 mg) can excite resonances which fatigue the heel of the wire bond.

Eventually cracks are induced and electrical continuity lost

43

Cautionary Tales III CMS are confronting the enormous challenge of tracker construction with a sophisticated QA scheme

2.8 m

1.3

m

15 different designs24244 different sensors!To be completed within 2 years!

Full testing on pre-series (5% of sensors): IV, CV, strips, optical, irradIrradiation of 5% of test structures and 1% of stripsCMS process control on test structures: (12 measurements)

Scheme has weeded out many problems at an early stage and gives confidence in sensor performance

44

Cautionary Tales IIISample measurements:

Depletion Voltage Leakage current

Note however that even CMS are not totally immune to theoccassional broken bond! Commercial transportation of some modules caused 20% of bonds to break (these modules were fixed in ~1 day)

Vibration tests show that transportation can give > 3.4 g force

45

Cautionary Tales IIINASA style vibration test Used laser to identify cantilever

resonances at 88 Hz

and at 120 Hz

Reinforcement glue bedstotally solved the problem

46

Conclusions

Detector R&D making a huge impact on the future of HEP

Bear in mind:

“Even if you think it will take a long time, it will still take longer than you think”

H. Bieler, this conference

And what about the cost?

47

Acknowledgements

LC Keith Riles, Paul Dauncey, Ron Settles Silicon Hans Dijkstra MAPS Marc Winter, Grzegorz Deptuch, Wojtek Dulinski, CCD Chris Damerell, Stefania Hansen DEPFET Marcel Trimpl Johannes Ulrici SDD Rene Bellweid, Vladimir Rykov HAPS Massimo Caccia, Wojtek Kucewicz, Peter Chochula, Peter Rosinsky Irradiation Sherwood Parker, Cinzia da Via, Angela Kok, Mahfazur Rahman, Michael Moll, Mika Huhtinen, William Trischuk, Zheng Li Tevatron Alan Sill, Regina Demina, Gino Bolla, Rainer Wallny, Chris Hill, Rick Tesarek CMS Manfred Krammer, b factory David Leith, David Hitlin Dark Matter Hans Kraus, Fabrice Feinstein, Harry Nelson Overview Tejinder Virdee, Guy Wilkinson

Thanks to all people who provided material, including

and not forgetting

the organisers of IWORID, Riga 2003

48

Challenges of Large Systems

LHC GPD?

ack: G. CharpakThe Tower of Babel by Pieter Breugel

49

More Challenges of Large Systems…

mass production, QA, electronics, computing, long time scales(technology, aging), risk factors, mechanics, etc. etc.

Example: Calibration systemsWill assume huge importance: from cross check to full integration

STAR calibration system CMS calorimeter80,000 crystals

Intercalibration ~ 0.4%Will use data and monitoring

200-400 UV laser beamsMeasures drift velocity

Next step: embed in data stream

W eZ ee

Huge effort for database

50

From now on backup transparencies

51

Hybrid Active Pixel Sensors (HAPS)

Hybrid pixels

(some) Tracking applications for HEP:

1995 WA97 0.5 Mpixels 75 x 500 m2

1996 DELPHI 1.2 Mpixels 330 x 330 m2

NA60 0.7 Mpixels 50 x 400 m2

ATLAS 100 Mpixels 50 x 400 m2

CMS 23 Mpixels 150 x 150 m2 ALICE 100 Mpixels 50 x 400 m2 BTeV 23 Mpixels 50 x 400 m2

HAPS are fast and tolerant of LC radiation

R&D to cover precision and material issues

Solder bump

52

R&D for HAPS

R&D programmeFine implant pitch + coarse readout pitchUse capacative charge sharing between pixels to improve resolution – use good S/N pixel performancePrototype results: 100 m implant + 200 m readout gives 3 – 10 m resolutionExpected to scale with pitch: 20-25 m prototypes in production in 2002

Test structure with

interleaved pixels

hep-ex/0101012

Resolution Resolution vs position

53

LEP vs LC At LC: “x sections are tiny”“No radiation issues”“Triggerless operation possible”“Modest rates”

Why not use a LEP/SLD detector?

LC physics demandsexcellent

Vertexing (b,c,t)and Tracking

in a high B field with energy flow

54

Irradiation: NIEL

NIEL allows us to look into the future and predict what will happen in complex environments

(!) Has been known to fail for neutrons/charged hadrons in some cases

A common language:

“1 MeV neutron equivalent”

Use the NIEL scaling factors

55

LEP0.2 TeV

NLC/JLC TESLA0.5 TeV

Train length, s 0.750 0.265 950

Number of bunches/Train 4 190 2820

Bunch separation, ns 200 1.4 337

Repetition rate, Hz 45500 100 5

Timing @ the LC

TimeTrain/rf pulse

at TESLA, keep occupancy reasonable by reading out innermost layer in 50 sec

56

MPGD: GEM detectorsGas Electron Multiplier – F. Sauli 1996

• Copper clad kapton foil• Perforated by many

(~104/cm2) holes• U between electrodes

creates amplification region inside holes

• Gain is a property of foil

• Can cascade several GEMS Higher gain

• Amplification and readout stages are separate

• Many readout schemes possible, e.g. 3d, Sauli 2002

u + v + w = 0

57

Beamstrahlungbackgrounds

r=12 mmr=15mm

Hit

s /

mm

2

z [cm]

Silicon for vertexing @ the LC

Use a silicon based pixel detector Confine the background with a big solenoidal

field

(IP) < 5 m 10 m/(p sin3/2 )

best SLD 8 m 33 m/(p sin3/2 ))

Vertex detector characteristics

point resolution 1-5 m

Thickness ~ 0.1 % X0

5 layers

Inner radius ~ 1.5 cm

Required Vertexing performance

Flavour tagging: beauty + charmttee discriminate b

from background

discriminate b from c ,,, ggccbbH

disentangle complex events

ttttAHee

12 jets

qqbqqb

58

R&D for photodetectors

•Pixel HPD photodetector developed

Cross focused image intensifier tubeDemagnification x 5

Silicon pixel array bump bonded to binary readout chip

pixels 500 m x 500 m

75

mm

Electronics must be compatible with bakeout cycles!

PID is a fundamental requirement for LHCb, BTeV, BaBar, Belle, etc.

LHCb (e.g.) requirements:•Total image surface ~ 2.6 m2

•Granularity 2.5 x 2.5 mm2

•High fill factor•LHC speed readout

Key R&D effort towardsdevelopment of suitablephotodetector system

59

Extra Requirements: Time capability ~200ps B field operation

..and for future DIRC/TOP?

At L=1036 cm-2 sec-1 new readout

scheme needed for the BaBar DIRC

Similar ideas inplace for Belle TOP

Currently evaluating MaPMTs, Pixel HPDs, Pad HPDs, HAPDs

NIMA 460 p326, 2001, NIMA 463 p220, 2001

Single photon peak obtained with

Hybrid Avalanche Photo Diodes

60

Backup transparency - oxygenation

M Moll

61

Backup transparency - oxygenation

62

Backup transparency - oxygenation

63

MPGDs: micropixelsTowards true gaseous pixel detectors

• Micro-dot chamber (Biagi, 1993)• Micro-pin structure (Rehak, 1999)• Micro-pixel chamber (-PIC) (Ochi et al, 2001)

3 x 3 cm device: Pitch 400 mAnode thickness 50 m5 days continuous operation without

discharge at gain=103

64

R&D for LHCb RICHPID is a fundamental requirement for LHCb

LHCb requirements:•Total image surface ~ 2.6 m2

•Granularity 2.5 x 2.5 mm2

•High fill factor•LHC speed readout

Key R&D effort towardsdevelopment of suitablephotodetector system

65

3d Sensors: Speed

Cell WidthCell Width Rise timeRise time T T FluenceFluence

200 200 mm 1.5 1.5 0.25ns0.25ns

130 130 KK

zerozero

200 200 mm 3.5 3.5 0.25ns0.25ns

300 300 KK

zerozero

100 100 mm 3.5 3.5 0.25ns0.25ns

300 300 KK

10101515p/cmp/cm22200m

Vbias VbiasVsig

100 mp

n

n

= 40VVbiasVbias Vsig

134 m

n

n

p

100 m

=40V

1 x 1015 p/cm2, 300 K

No Oxygen No Oxygen DiffusionDiffusion

Reverse Reverse annealedannealed

Non-Irradiated, 130 K

66

3D detectors: Excavating the holes

Hole depth/diameter: ~ 25Hole depth/diameter: ~ 40(but..)

Hole depth/diameter ~ 26

Electrochemical etching

Laser DrillingDry Etching

Any material No photolithography

Standard photolithography process No sidewall damage

Slow process for big arraysSidewall damage TaperingRepeatability

Sidewall damageSi and GaAs only

Si only (GaAs and SiC?)Complex photolithography

1 hole / 3-5sec. 0.6m / min.1m / min.

67

Lazarus efect: Beamscope for NA60

Requirements:• determine primary vertex • tracking in the beam (p … Pb)

• high dose - ~Grad per day• fast readout (beam intensity)

Solution:• Lazarus effect (RD39, ‘98)

• cryogenic operation • CCE restored

Detector: • single sided Si strip• 50 m pitch, 24 strips• cryogenic operation ~130 K

• steel cryostat, LN• tens of Grad !

• 2 sensors = 1 space point

Discovery: 1998 Application: 2001!

68

3D detectors: Excavating the holes

Hole depth/diameter: ~ 25Hole depth/diameter: ~ 40(but..)

Hole depth/diameter ~ 26

Electrochemical etching

Laser DrillingDry Etching

Any material No photolithography

Standard photolithography process No sidewall damage

Slow process for big arraysSidewall damage TaperingRepeatability

Sidewall damageSi and GaAs only

Si only (GaAs and SiC?)Complex photolithography

1 hole / 3-5sec. 0.6m / min.1m / min.

69

• Other collider upgrades are hot tooe.g. Super Belle/Babar

Irradiation

• Even an LC is enough to fry an SLD• LHC detectors will be upgraded

where possibleLHCb vertex detectorPixel layers of GPDs

Improved semiconductor designs/materials are well worth considering

L = 1036 cm-2 s-1

Dose ~ 5-10 MRad

70

MPGDs: Micro Strip Gas Chamber

• S/N=31 (98% hit efficiency for CMS)

• Resolution 35 m• No aging for an “LHC lifetime”• Robust for broken strips

HEP applications HERA-B 184 GEM-MSGC ~30*30 cm HERMES 48 MSGC 14*14 cm

MSGC: mature technology

Proposed by Oed, Bellazini, Udo (1998-89)

71

Silicon for tracking: Drift detectors• Principle of sideways depletion – as for DEPFET sensors• p+ segmentation on both sides of silicon• Complete depletion of wafer from segmented n+

anodes on one side

• Electrons drift along potential trough in detector mid plane skewed towards anodes at the end

• X coordinate measured with drift time (~8 m/ns)• Y coordinate measured from anode c.o.g.

y

x

Reduction of channels vs pixel detectors

Multi track capability dE/dx capability Small anode capacitance

Drift velocity must be predictable Temperature control resistivity control Calibration techniques

72

RPC developmentResistive Plate Chamber:

Parallel plate chamberwith resistive electrodes

Used for TOF, Trigger, TrackerStreamer mode Avalanche mode

L3 300 m2

BaBar 2000 m2

BELLE 2200 m2

ALICE 144 m2

TOF mode

ATLAS ?? m2

CMS ?? m2

LHCb ?? m2

Trigger-tracker modeRate capability as amplitude ”Lower HV hits time resolutionSmall signal amplitude: noise, x talk issues

Big worldwide activityPlenty of accumulated

experience NA49 0.5 m2

HARP 8 m2

STAR 60 m2

ALICE 144 m2

73

RPC development: MRPCs

•Multigap Resistive Plate Chamber : good time resolution

Internal plateselectrically floating

Single anode/cathodereadout

Recent tests (ALICE,STAR)demonstrate feasibility ofLarge area devicesInsensitivity to variations in gaps(à la micromega)

STAR prototypeSix 220 m gas gaps

6.5 x 19.5 cm2

NIMA 478 (2002) 176

Time to Amplitude correction = 55 ps Eff=98%

74

Silicon for tracking: Drift Detectors

• SDD fully functioning in STAR SVT since 2001

• 216 wafers, 0.7 m2

• 10 mm in anode direction• 20 mm in drift direction• Particle ID

SDD also chosen by ALICE (1.3 m2)

Similar requirements:

-high multiplicity

-dE/dx

Strips

Silicon DriftPixels

ALICE

STAR

75

RPCs for the LC HCAL

Ability to finely segment: O(1cm2)Works in magnetic fieldFits in gap of < 8 mmEfficient (>95%)ReliableAffordable

All requirementsmet by RPCs

HV

Glass/Bakelite Gas Graphite

Mylar Copper

Analog 10 GeV π’s Digital

Eflow driven simulations will study detector parameters e.g.•Transverse segmentation•Cell depth absorber density•Layer geometry

NhitsEnergy

76

Micromegas for COMPASS

• Compass requirements matched to MPGD performance High rate capability ~100 kHz per

strip Good spatial resolution Large active area No aging/discharges

•COMPASS: fixed targed experiment at the CERN SPS: •measure gluon spin contribution to nucleon spin

• Compass is largest MGPD application 12 planes, 40 x 40 cm2

Spatial resolution ~ 70 m Time resolution ~ 10 ns 0.45% radiation length

•PHYSICS DATA WITH FULL DETECTOR SINCE JUNE 24 2002

77

Size of silicon detectors

78

Impact parameter resolution

79

Rubbia (I)

80

Rubbia (II)

81

Rubbia (III)

82

MPGDs: Micro-Pattern Gas Detectors

MWPC The “daddy” of proportional counters

MSGCSpatial resolutionRate capability

Photolithographic techniques

advanced PCB technology

Micro WELL

Micro Groove

GEMSand GlassMicroMegaMicroCAT

Hybrid devices

83

MPGDs: MicroMegas Micromesh gaseous structureG. Charpak, Y. Giomatraris, 1992

• Thin gap || plate structure Good energy resolution High rate capability

• Conversion space separated from amplification gap by micromesh

• Ion feedback suppressed• U~500V -> 50 kV/cm• Saturation of Townsend

coefficient stable against gap variations

Variation on a theme: MicroCompteurATrousTwo dimensional interpolating readout StructureLarge active area with new spacer concept (e.g. Wagner, 2002)

gap sizeg

ain

84

Micromegas

New fabrication technique•Integrate manufacture of mesh and spacers

•NIMA 461 (2001) 84

pillars

mesh

85

MPGD’s for TESLA TPC

TPC design• 200 3d points per track

with r~150 m

• d (1/pT) = 1.4x10-4 GeV-1 (barrel)

• Drift time 50 s 150 bx• 20 MHz readout

Ar-CO2-CH4

230V/cm-60kV

R&D focus: new gas amplification system

MGPD readout (GEM/Micromega)

(1/pT) < 5x10-5 GeV-1

Precision on MH from Zl+l-

Tracking requirements for TESLA:

86

MPGD’s for TESLA TPCnew gas amplification systemMGPD readout (GEM/Micromega)

• Ion feedback suppressed to 2% level• Good dE/dx• E x B effect reduced to 50 m• Gating @ 2cm possible between

trains• Chevron pads compensate low

induction signalsAnother option: use silicon pixel readout a

la “x ray polarimeter”• Each GEM hole has its own pixel(s)• GEM gain 1000-3000 + pixel noise (200

e) = single electron detection• 1.7 GpixelsA digital TPC limited only by diffusion

87

MGPDs : Advantages

• Common advantages of MGPD family: Efficiency 99% plateau

at G=6500 in COMPASS Spatial resolution

micromega best 12 m Rate capability Double GEM 105 Hz/mm2

Energy resolution and gain stability

Time resolution ~10 ns for micromega Adaptable gain

88

MPGD’s by Albert Cuyp

Watch out for discharges

89

MPGD’s by Albert Cuyp

Watch out for discharges

•Triple GEM solution:•Reduced spark probability for equal gain

GEM + SGC:Tune the voltages to optimisedischarge rate -> factor 100

90

MPGD’s by Rembrandt

Ageing can be a problem

1629

1640

1669

91

MPGD’s: AgeingMSGC ageingDependence on materialsSensitive to pollutants, strip and substrate material

Long term ageing test for GEM+SGC:After 500 LHC days (5 kHz/mm2 at gain=2800)•5% anode loss•25% gain loss

•NIMA 485 (2002) 322

Long term ageing test for triple GEM

No degradation after 27 mC/mm2

NIMA 478 (2002) 263

92

Calorimetry for the LC• LC physics all about jets:

tth 8 jets hZ 2l + 2 jets, 4 jets hhZ 2l + 4 jets, 8 jets

+ SUSY, quark, tagging, lepton/hadron idHigh magnetic field demands compact design

Calorimetry at the LC will bedemanding

ENERGY FLOW is the name of the game

Ejet = Ech + E + Eneutrals

Identification and reconstruction of all eflow objects

Charged tracks from tracking systemPhotons from ECALNeutral hadrons from ECAL and HCAL

93

Calorimetry for the LCECAL: silicon-tungsten (Si-W)

sampling electromagnetic calorimeter:

• 40 layers, between

0.4 and 1.2X0

(radiation lengths).

• 24X0 total thickness.

• 32 million channels.

•1 x 1 cm2 pad size

HCAL: digital calorimeter concept

•count 1cm2 cells

•16 million channels

•Stainless steel absorber plates

•Readout with RPCs/wire chambers

E/E = 60%/E E/E = 30%/E

For hadronic jets, TDR calorimeters give E/E = 33%/E

DistinguishesW+W– and ZZfinal states Higgs gives m ~ 2 GeV

94

Applications in Dark Matter Searches

Example: Direct detection of WIMPsformidable challenge to detector technologyfrom nature rather than an accelerator

Search for a tiny signal with low rates

Nuclear recoil fromWIMP elastic scatter

~ 10 keV ~ 0.1 count/kg/day

phononsionisation

scintillation

CRESST I, ROSEBUD, Tokyo LiF, Milano Te02

HD Moscow, GENIUS

UKDMC NaI, DAMA NaI

EDELWEISS,CDMS

UKDMC Xe

CRESST IICaWO2

95

Applications in Dark Matter Searches

next generation detectors 10 kg

this side collects

phonon energy

this side collects

ionisation energy

to SQUID array

Si or Ge crystal zip or blip

CDMS:

Array of superconductingTransition Edge Sensors

CRESST II : Simultaneous detection of phonons + scintillation light

with CaW04 crystals

Low pressure gas Xe TPCCharge carried by negative ionsReadout with MWPC/CCD/MPGD

DRIFT :