calorimetry r&d for the ilc
Post on 05-Jan-2016
26 Views
Preview:
DESCRIPTION
TRANSCRIPT
Part III
Calorimetry R&D
for the
ILC
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
2
The PFLOW paradigm• The confusion term dominates
• Each particle should be reconstructed and measured separately
• For the jet energy measurement spatial resolution / particle separation power is more important than energy resolution
2confusion
2neut.had.
2photons
2charged
2jet
had. neut.photonschargedjet
σσσσσ +++=
++=
EEEE
EEEE
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
3
Imaging calorimetry
ZHH qqbbbb
red: track based
green:calorimeter based
Calorimeters sorroundTrackers
Electromagnetic Calorimeter - Ecal
Hadronic Calorimeter - Hcal
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
4
A Particle Flow Algorithm (Extraction)Pandora Algorithm by M. Thompson Uni Cambridge
Cannot do justice to full complexity of algorithm but …
Two main steps:
Tracing a particle through the detector
- Find Individual Clusters created by one particle
- Merge clusters to reconstruct shower of the particle
Algorithms can only be qualified in MC simulationNeed extremely good knowledge in particular of hadronic cascadesMajor task for testbeam efforts with (physics) prototypes
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
5
Calorimeter concept
• large radius and length– to separate the particles
• large magnetic field– to sweep out charged tracks
• “no” material in front– stay inside coil
• small Molière radius– to minimize shower overlap
• high granularity– to separate overlapping showers
• figure of merit: B R2calo / (r2
M+r2cell)
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
6
Ecal - Main Task Photon measurement and photon/hadron separation
“Known” basic tools: Large R and B If small R: Force created by Large B-Field might compromise detector stabilityLimit: BR2 < 60 Tm2
Separation gets difficult if hadron andphoton are within RM
Photon Energy gets assigned to close-byHadron and vice versa
“Calorimetric” tools to improve photon-hadron separation?
“SiD”
TESLA TDR( LDC)
2 RMDistance Photon-Hadron[cm]
Erec./Ereall
(SD: R=1.27 m, here with 6T,TESLA TDR: R=1.68m, B=4T)
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
7
Choice of Absorber Material - Tungsten vs. Iron
IronTungsten(images courtesy H.Videau)
• elm./had separation: keep X0 / λI smallX0 = 1.8cm, λI=17cm
X0 = 0.35cm, λI=9.6cm
• Molière Radius for W: RM = 0.9cm
• Cellzise need to match RM !
• effectively a factor ( 1 + Gap / 2.5mm ) more
• technology challenge: thin readout gap
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
8
International effort
• Linear collider detector R&D is partially organized in (open) proto-collaborations, e.g. CALICE:
~200 Physicists, ~40 Institutes, 10 Countries: 3 Regions
• CALICE performs large scale testbeam 2005-2008(9) whith ‘physics’ protoypes
• ECAL and HCAL together, different options• First large scale module (‘technological’ prototype for 2008)
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
9
Ecal Prototype - CALICE Collaboration
- Sampling Technique (see Part I)
- W as absorber material
- Signal extraction by “Silicon Wafers”
- Extreme high granularity 1x1 cm2 cell size
- Detector is optimized for particle separation
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
10
Ecal in Testbeam @ CERN
Shower by 6 GeV e-
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
11
CALICE-ECal - results
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
12
Hadron CalorimeterSame imaging requirements as for EcalHigh granularity for single particle identificationMost important task: Measure neutral hadrons !
Two Options Sampling Technique
Digital Approach:
- Exploit statistical nature of (hadronic) shower- Extreme small cell size 1x1 cm2
1 signal/(particle and cell) Ncells with signal ~ Energy of primary Particle- Fe or W as Absorber (W very expensive !!!)- Gaseous Detector (RPC see later) as active Element
Analogue Approach:
- More “classical” Approach- Measure energy deposition in cell- Still small cell size O(3x3 cm2)- Fe or W as Absorber- Scintillator as active Element Amount of light ~ deposited energy Appetizer: Novel photosensitive devices will be employed (see later)- RPC also considered as active element
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
13
Hcal -Comparison of GranularitiesCell Grids
Digital Option Analogue Option
Cell Grid for a Hcal Prototypewith finely granulated inner core
Challenging to assembleKeep in mind: Need alsor/o devices
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
14
Analog vs. Digital
• Digital: pad size 1cm asymptotic value• suppress Landau fluctuations: at low E superior to analogue
• need ideas for high E, e.g. multiple thresholds (semi-digital)
Energy (GEV)
Num
ber
of c
ells
hit
σ(E)/E
0.3
0.2
0.1
10 20 30 40 50E/GeV
AnalogDigital (0.5x0.5)Digital (1.4x1.4)Digital (2.5x2.5)Digital (3.0x3.0)
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
15
Gas vs. Scintillator
• width of shower pattern appears larger in scintillator
• will be recovered using amplitude or density information
N
ri∑Δ2
∑∑ Δ
i
ii
E
rE 2
RPC
Scint.
digitaldistance fromshower axis
analogueenergy weighteddistance fromshower axis
(L.Xia)
Regard lateral extension of shower: Want to have narrow showers
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
16
Hcal - Active Media (not only) fot Digital Concept
Pad arrayMylar sheet
Mylar sheet Aluminum foil
1.1mm Glass sheet
1.1mm Glass sheet
Resistive paint
Resistive paint
(On-board amplifiers)
1.2mm gas gap
-HV
GND
Resistive Plate Chamber (RPC): - Gaseous Detector
Freon as chamber gas
- Spark by gas ionisation generates current in Pad Array
- RPC’s can be built in small units at low cost
(need to equip a huge calorimeter with 1x1 cm2 cells)
typ.1kV
Alternative Option GEM’s:
Absorber
Gas 3-foldGEM structure
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
17
Hcal - Active Media (not only) for Analogue ConceptScintillating Tiles with wavelength shifting fibres (also employed in alternative Ecal Concepts)
- Light production in scintillating tiles ~ 200 photons/MeV- Light transport in fibres- Measurement in photosensitive device Silicon Photomultipliers (details later)
Tile Sizes ~ 3x3 cm2
Optimized for separation ofhadronic showers
Reconstructionof neutralparticlesin analogueHcal
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
18
Silicon Photomultipliers SiPM
• A pixilated solid state Geiger counter (Semi Conducting)– 1000 pixels on 1mm2
– Gain 10**6, efficiency 10..15%– At 50 V typical bias voltage
Signal - analog sum
R 50
h
pixel
Ubias
Al
Depletion Region2 m substrat
e
ResistorRn=400 k
20m
42m
Single Photon Signals
0 100 200 300 4000
2000
4000
6000
8000
10000
Counts
Channel
0 pe 1 pe
2 pe
3 pe
4 pe
= 1 Photon
= 2 Photons
= 3 Photons
= 4 Photons
Scintillatingphotons createphoto-electrons
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
19
Silicon Photomultipliers cont’d
• Insensitive to magnetic fields Has to operate in O(4 T) magnetic Field Big advantage w.r.t. traditional PMTs)
• Dark rate ~ 2 MHz• Interpixel Crosstalk 20-30%
• Gain, efficiency, Xtalk (thus noise rate) depend on bias and temperature– Overall 7% / 0.1V, -4% / K
• Need to carefully optimize individual working point
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
20
SiPMs at work I
• High gain, low bias, small size:• Mount directly, no fiber routing• Coax cable readout, no preamp
Spectra from SiPM
Single Photon
MiPs
Saturation Curvecaused by limited number of pixels
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
21
SiPMs at work II - A view in the lab• tests with SiPM minical cassette
very fast26ns peaking time
1 photon
2 photons3 photons
Intensive R&D to understand properties of SiPMs
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
22
First Testbeam Experience - MiniCal
2 cm steel
0.5 cm active
• 108 scintillator tiles (5x5cm)
• Readout with
– Hamamatsu APD S8664 – 55 spl– 16 ch. MAPM (H6568) for ref.
– Silicon Photomutipliers• Joint development
– Moscow Engineering Physics Inst.
– PULSAR, Moscow– DESY
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
23
First Testbeam Experience - MiniCal
• DESY 6 GeV e beam• Non-linearity can be corrected
– Need to observe single photon signals for calibration– Does not affect resolution
• Resolution as good as with PM or APD
• Stability not yet challenged
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
24
Cubic Meter Prototype for large scale Testbeam
• 8000 channels
• Hadron test beam in 2006-2008
• Includes scintillator strip muon detector with SiPMs
Tail Catcher
Tile Hcal
Ecal (see above)
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
25
Testbeam Setup at CERN 2006/07
ECAL
HCAL
TCMT
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
26
Layer Composition of Prototype
29.73 mm
Approx. 1.14 X0
39 Layers
Approx. 4.7 Needed Thickness
for final Detector
- 16 mm Steel, S235 as Main Absorber
- 1mm Air Gap
- 2mm Scintillator Housing – Front Plate
- 0.115 mm 3M Foil
- 5 mm Scintillator
- 0.115 mm 3M Foil
- 1mm FR4
- 1mm Cable Fibre Mixture (PVC. Fibre)
- 2 mm Scintillator Housing – Rear Plate
- 1mm Air Gap
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
27
Testing Hadronic Models• 10’000 particles, compare Geant 3 (histo) vs. Geant 4 (points)
• differences vary with energy, particle type, detector material,…
1 GeV
50 GeV
RPC Scintillator
5 GeV
longitud.(ECAL+HCAL)
transversein HCAL
(study byD.Ward)
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
28
Overview on Hadronic ModelsLarge Variety of hadronic models
Citation from GEANT4webpage:“For very fine grain calorimeter like in the case of several LC designs the physics lists … might not be adequate”
Need testbeam data to tune hadronic models for ILC purposesE.g. for optimization of particle flow algorithms
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
29
The Step towards the Real Detector - EUDET Module
Module ECAL
1700 (1500)
1000(840)
2
10
(1
70
)
HCAL
ECAL
Technological Prototype (2006-09) - Large Scale Integration - Layout come from Detector Studies - Development of Front-End Electronics
Physics Prototype
- Demonstrate Physics Performance - Validate Simulation Tools - Running at DESY since Jan. ‘05 - Large Scale Test Beam in ‘06
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
30
Production at industrial level
Eeasier for mechanical constructionEeasier for mechanical construction Smaller Molière radiusSmaller Molière radius Industruial way of assembling Industruial way of assembling DAQ based on FPGA DAQ based on FPGA VFE ASICs more integratedVFE ASICs more integrated (ADC, more channel, zero suppress …) (ADC, more channel, zero suppress …) Smaller silicon pixelSmaller silicon pixel (glue, ac/dc coupling, thickness) (glue, ac/dc coupling, thickness) 2x thinner than protoype2x thinner than protoype
Eeasier for mechanical constructionEeasier for mechanical construction Smaller Molière radiusSmaller Molière radius Industruial way of assembling Industruial way of assembling DAQ based on FPGA DAQ based on FPGA VFE ASICs more integratedVFE ASICs more integrated (ADC, more channel, zero suppress …) (ADC, more channel, zero suppress …) Smaller silicon pixelSmaller silicon pixel (glue, ac/dc coupling, thickness) (glue, ac/dc coupling, thickness) 2x thinner than protoype2x thinner than protoype
R&D sampleR&D sample
FE Electronics has tofit into ~3cm gap betweenHcal and Ecal …
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
31
…and the electronics has to consume no power
The goal is 25µW/ch. (with Power Pulsing) Remember Ecal+Hcal O(5x108) Channels
time
Time between two trains: 200ms (5 Hz)
Time between two bunch crossing: 337 ns Train length 2820 bunch X
(950 us)
Electronicsactive
Electronics switched off
Electronicsactive
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
32
The 4th concept
Calorimeter
VertexDetector
TPCSolenoids
Muon System
Novel Features:muon system embedded in solenoidal field
‘traditional’ calorimeter but:multiple calorimeter readout to overcome typical deficiencies of hadron calorimetrybased on Results of DREAM calorimeter testbeam
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
33
The DREAM Calorimeter - Wigmans et al.
Dual REAdout Module A Dream Cell
Quartz Fibres: To measureCerenkov Light fromrelativistic particle in shower ~ electromagnetic Fraction
Scintillator Fibres: To measure scintillation light in showerIntegral Signal from all particles €
C = [ fem +1 − femη c
] • EC = Cerenkov SignalC=S(e)/S(h) for Cerenkov SignalE = Shower Energy
S = Scintillator SignalS=S(e)/S(h) for Scintillator Signal
€
S = [ fem +1 − femη S
] • E
Two Equations to determineunknowns fem and E=>Full control over fluctuatingfem
2.5mm
4mm
DREAM Calorimeter - Front View
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
34
DREAM - Testbeam Results Response to 200 GeV Pions
- Scintillator only 200 photoelectrons/GeV
-Cerenkov and Scintillator
lateral leakage
€
fem = [C
E shower−
1
η c]
-Cerenkov and Scintillator
‘no’ lateral leakagefem constrained by Beam Energy
€
fem = [C
Ebeam−
1
η c]
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
35
Summary
• ILC opens a new frontier for calorimetry Challenges are addressed in worldwide collaborations like Calice
• Need sophisticated hardware and software development to achieve the goals envisaged for the ILC
• Calorimeters are tailored for particle Flow algorithms rather than on energy resolution but 4th concept follows alternative approach
• Testbeam are maybe more than ever vital to reach precision and therefore physics goals envisaged for the ILC
Roman Pöschl IRTG Fall School Heidelberg Germany Oct. 2006
36
These people kindly let me use their material
Dietrich Wegener, University of Dortmund … for Part1Lutz Feld, RWTH Aachen … Introduction to semi-conductorsCALICE Collaboration … on calorimetry Victoria Field, Uni of Edinburgh for material on 4th Calorimeter and slides from her Presentation at Ambleside+ many other people whose talks I have exploited for this lecture
top related