Challenging Technology: Detectors
Beyond the LHC
Craig Buttar
Higgs-Maxwell Meeting Feb ‘06
Examples – ATLAS SCT
Future of Particle Physics in a nutshell
• Luminosity upgrade of LHC (sLHC)
– H self couplings  precision tracking in high multiplicity environment
– Radiation hardness x10 cf LHC
– Large scale production
• LHCb upgrade
– Velo replacement required after 3 years
– LHCb data taking limited to 1/50th of design luminosity
– Make better use of LHC luminosity  improve rare decay limits test
new physics
• Precision physics at ILC
– Higgs couplings, SUSY (Higgs) mass spectrum……..
– High resolution vertexing for flavour tagging
– High resolution tracking calorimetry based on energy flow
• The neutrino sector
– Unravelling the neutrino mass matrix
– Large scale detectors
– Beam systematics
sLHC
Luminosity upgrade for ATLAS-SCT
Tracking and b-tagging
Design issues
• layout
• occupancy
• radiation damage
50mm pitch
Neutron 65%
Pion
29%
Proton 6%
Neutron 18%
Pion
74%
Proton 8%
Neutron 66%
Pion
25%
Proton 7%
Pixels: r=6cm, 15cm, 24cm
Ministrips: r=35cm, 48cm, 62cm
Microstrips: r=84cm, 105cm
Si technologies for sLHC
• Si Traditionally use p-in-n detectors but
cannot operate underdeplented
• Use n-in-p as no type inversion
n-in-p
7000e-
P-type 1cm detector
after 7.5×1015p cm-2
≈4.5×1015 1MeV neqcm-2
(850V)
– Production issues
– HV required, careful design to avoid
breakdown
• Use Cz based n-type Si
– Does not invert (?)
– Now a common material in Si
manufacture
– Harder to charged particles best for
intermediate radii or pixels
Ionisation in oxidehigh field regions
+ P
++ -current from
+
+ -cut edges
Displacement damage cluster
np; increased dark current
guard region
Detector region
SLHC
R=40cm
SLHC
R=20cm
P-type Si
ionisationsignal
CCE is now an issue
W3D~20mm
W2D~300mm
SiO
2
+ve -ve +ve
3D detectors
p+
h+
h+
Bulk
n
e-
e-
W3D
E
n+
+ve
E
W2D
• 3D uses MEMS
technology to engineering
Si structures
– Smaller detector ‘cells’
 Lower depletion voltages
 Faster and more efficient
charge collection
BUT small-scale suitable for
LHCb upgrade
For 4.5 x 1014 24GeV/c p/cm2
(2.7 x 1014 1MeV n/cm2)
2 years LHCb Velo inner radius
Depletion voltage = 19V
Type inversion observed
Other issues for sLHC
• Cooling
– Thermal runaway:
I  Power dissipation  T
• Large scale production
– SCT area 60m2 210m2
– Electronics 6M  60M channels
– Need close links with industry for mass
production of components and assembly?
Flux 2E15, bias 800, width 9cm
-20
Temperature (degC)
– To avoid thermal runaway need to have
edges a -30oC to dissipate heat and keep
temp low
– Current technology has a coolant
temperature of ~-30oC to give –7oC at LHC
– Low mass cooling with robust minimum
mass pipes
-44 degC
-25
-42 degC
-30
-40 degC
-35
-38 degC
-36 degC
-40
-34 degC
-45
-32 degC
0
2
4
x (cm)
6
8
-30 degC
LHCb upgrade triggering at LHC lumi
• Factor five higher lumi
Cope 3-4 interactions per beam crossing
• Muon trigger fine
– 4 of 10 benchmark channels have m+m- in final state
• Hadron Trigger bandwidth saturates
– Need displaced track trigger at first trigger level
8
FPGA based Triggering system
– Pattern recognition / tracking
– Primary Vertex Identification
– Displaced Track Trigger
– 4ms latency
Vertex detectors for ILC
• Physics requirements
– Identify b and c-jets and jet-charge
– Momentum resolution s(1/pT) at 100GeV ~4-5TeV-1 (ATLAS 100TeV-1) important
for energy flow—see later
• Detector requirements
– Very low mass
– High resolution sip~3mm  ~20mm pixels  ~109 channels  reduces
occupancy
– Beam structure  readout every 50ms ~20 times during 1ms train
• Data readout – possible em interference
• Data stored -- readout during ~0.2 between bunches
– Radiation hardness less of an issue
• CCDs
ILC vertex detectors
– Used in linear collider at SLAC -- SLD
– Precision but readout speed is slow as
reading out each column and then row
– Readout each column – CPCCD
• First prototypes CPC1 by LCFI
– Readout at 25MHz
– RF pickup from beam is a worry!
CPC1
RO chip
(Bump bonding
Packaging and
interconnects)
Detectors with storage ISIS and MAPS
• ISIS – In-situ Storage Image
Sensor
– Store charge in 20-element CCD
– Readout during 0.2s between
bunches
– Solves potential RF problem and is
more radiation hard
– But processing has to be
demonstrated
• Monolithic Active Pixel sensors
– CMOS imaging technology
– Smaller active volume
– Requires development of on-chip
memory
Slab
Slab
x5000
Slab
Slab
30%
best achieved at
E
Slab
s
– Requires 
LEP 60% E
Slab
• Need to resolve W and Zs
Slab
Calorimetry for the ILC
5000
fibres
1Gb
E
– Need to measure energy flow in the
event
• Match charged tracks to clusters
• Measure neutral clusters in calorimeter
•  good spatial resolution – more
important than energy resolution!
Large Network Switch/s
5Tb
10Gb
Event Event Event
Builder Builder Builder
PC
PC
PC
x250
Target
Control
Event
Builder
PC Busy
– Leads to a number of technical issues
• Readout density and getting data out of
the calorimeter: 0.3-3GBytes/s per ASIC,
200TByte/s total ECAL
• Achieving the required spatial
resolution -- MAPS
• Managing the thermal load
VFE chip
Cooling
Si Wafers
PCB
Tungsten
8.5mm
Neutrino factory
• Golden channel
muon appearance
• Requires MINOS like
detectors but x10 larger
• The large volume leads to
problems with reading out
scintillator (~10m)
m
50%
50%
e
detector 
m

m
e

e
m



m


wrong sign muon
Large Magnetic Detector
20 m
 beam
• But need to understand the
beam!

40KT
B=1 T
20 m
iron (4 cm) + scintillators (1cm)
Neutrino factory
• Near detector
– Need to measure flux and charm background
rates for measurements at far detector
– Use active target
– Large area coverage required (18 layers covering
50x50cm2
– Use MAPS
• 2D readout
• Cheap large area coverage
Si tracker in NOMAD
Not covered
• Development of technologies for PP is as active as ever
• Too many to cover!
• Some topics not covered (still only a selection)
–
–
–
–
Rad-hard readout chips for sLHC
Development of ionisation cooling for nufact – MICE @ RAL
Fast feedback systems for ILC to optimise luminosity
Low background detectors for Dark Matter searches and
neutrinoless double beta decay
– Precision physics at LHC (FP420): Small scale tracking
detectors able probe the edge of the beam
– …..
Other applications
• Particle physics technology
has found applications in
– Radiotherapy
– Imaging
Medipix1
2 mm
Imaging
plate
Medipix2
Summary and conclusions
• Future Particle Physics experiments has many
challenges for detectors and systems
–
–
–
–
–
–
–
Rad-hardness
Speed
Precision
Large scale production
Connections
System building detectors+readoutmodulessubsystems
Data handling
Backup
LHCb Upgrade – Why ?
• LHCb uses only 1/50th of LHC design luminosity
– Average of one interaction per beam crossing
– Limit Radiation damage
• Many physics results would benefit from higher luminosity, e.g.
– Rare B decays
• Clear Benefit
• e.g. Bsm+m– SM BR 3.5 x 10-9 , 3.7 s after 3 (107 second) years
– 
• Not theoretically limited after 3 (107 second) years
• More (clean) events would help !
LHCb Upgrade -Technically Feasible?
• Radiation Hard Silicon Technology Developed
•Czochralski Silicon
•n-on-p
•3D detectors
•Hybrid Pixels
Construct a Vertex Detector with
•better proper time resolution
•withstand 10 times more radiation damage
First Velo need replacement ~ 2011
VEtex LOcator
Priority only if
large Dms
1015 1 MeV
neutron equiv. /cm2
VELO Superior Performance Apparatus
•New Triggering Strategy
• Factor five higher lumi
Cope 3-4 interactions per beam crossing
• Muon trigger fine
– 4 of 10 benchmark channels have m+m- in final state
• Hadron Trigger bandwidth saturates
– Need displaced track trigger at first trigger level
8
FPGA based Triggering system
– Pattern recognition / tracking
– Primary Vertex Identification
– Displaced Track Trigger
– 4ms latency
Plan
• Rad. Hard Sensors being produced
•6” mask designed
•MCz, n-on-p, pixels
• Prototype VELO modules and Trigger 2009
• Upgraded VELO & displaced vertex trigger 2011
Physics case for future neutrino facilities
• Neutrino oscillations for atmospheric,
solar, reactor neutrinos provide fit to
q23, q12, Dm122 and Dm232
• Need more experiments for q13, mass
hierarchy and CP violation phase d
Super-beams
• Super-beams (ie. JPARC beam for T2K) provide monochromatic
Flux • s (arbitrary unit)
off-axis neutrino beams for e appearance
2.5o
q =2o
 2 

Flux  
2 2
1  q 
3o
0
q=0o
1
T2K
2
2
E (GeV)
• Proposals for super-beams
with ~ Mton Water Cherenkov
detectors (Memphys in
Frejus, HyperKamiokande
Japan, UNO in USA)
Goal T2K: down to
q13~2-30
• Beta-beam facility
6
2
He    36Li     e e 
18
10
Ne  F νe e
18
9

• Neutrino factory:
CERN layout: Emax~ 600 MeV
Beta-beams and neutrino factories
High-energy part
Low-energy part
Neutrino source
Acceleration
Ion production
Proton Driver
6He:
18Ne:
Ion production
ISOL target &
Ion source
 = 100
 = 100
Neutrino
Source
Beam preparation
ECR pulsed
SPS
Ion acceleration
Linac
Acceleration to
medium energy
RCS
PS
m   e   m  e
m   e   m  e
Decay
Ring
Oscillation signatures
• Golden channel at a NuFact: “wrong-sign” muons
detector
m
50%
e
50%
m

m


m
e
m
wrong sign muon
• Silver channel: tau appearance


Pb
Emulsion layers
 beam
40KT

1 mm
20 m


Emulsion (4 kton) Plastic base
Large Magnetic Detector

e

e m
B=1 T
20 m
iron (4 cm) + scintillators (1cm)
 e  
Liquid Argon (100 kton)
(kinematic selection )
Near Detector
• Near detector: control flux, systematics, maeasure crosssections, charm backgrounds for oscillation signals, ….
m CC event
• Use silicon detectors for vertex
(prototype in NOMAD)
• 109  interactions per year in 50 kg!!!
• Monolithic Active Pixel (MAPS): fully
active neutrino target
Physics reach neutrino factory
• Sensitivity to d-q13: best for neutrino factory (except at high q13,
in which matter effects dominate).
P. Huber et al.
High- b beam
not included
Performance
comparable to
Nufact
(Burget et al.)
First oscillation
maximum only
Introduction to 3D detectors
• Co-axial detector
– Arrayed together
• Micron scale
– USE Latest MEM techniques
• Pixel device
– Readout each p+ column
• Strip device
– Connect columns together
Proposed by S.Parker NIMA
395 pp. 328-343(1997).
Operation
SiO2
+ve
-ve
-ve
+ve
-ve
-ve
p+
h+
h+
Bulk n
-
e
W3D
E
e-
W2D
n+
Equal detectors
thickness
W2D>>W3D
+ve
E
Carriers drift total
Carriers swept horizontally
thickness of material
Travers short distance between electrodes
• Low full depletion bias
• Low collection distances  High CCE
• Fast
Two methods to form pores
DRIE
Electro-chemical etching
•So far, maximum aspect
ratio: 18:1 (depth 183μm)
•Modification to standard
equipment to obtain deep narrow
parallel walled pores
•In conjunction with STS
• maximum aspect ratio: 30:1
(depth 440 μm, =14 μm)
• 24 hours per wafer
• Cheap
Results - IV of devices
• Good rectifying np junction
formed
• Oxide as diffusing barrier
isolates individual cells
•
RIE removal of top surface caused
increase in current after -10V
Glasgow 3D detector
Results - Proton irradiation
High res n-type silicon, 85mm pitch, close-packed hexagonal pixels
Irradiation with 24 GeV/c protons at CERN
7 fluences from 5 x 1012 to 4.5 x 1014 24GeV/c p /cm2
For 4.5 x 1014 24GeV/c p/cm2
(2.7 x 1014 1MeV n/cm2)
2 years Velo inner radius
Depletion voltage = 19V
Type inversion observed
3D-stc detectors proposed at ITC-irst [2]
For fab simplification
n+ electrodes
50 mm
p-type substrate
n + electrodes
0V
+
potential distribution
-5 V
-5V
-
-
+
-
-
+
vertical cross-section
between two electrodes
--patterned
Uniform
p+ layer
0V
-8 V
-8V
Uniform/grid
-
-20 V
electrons are
swept away by
the transversal
field
holes drift in the
central region and
diffuse towards p+
contact
ionizing particle
Recently, Semi-3D radiation detectors with p+ columns in n-type substrates
were proposed by Eränen et al. [3]
C. Piemonte, M. Boscardin, G.-F. Dalla Betta, S. Ronchin, N. Zorzi, Nucl. Instr. Meth. Phys. Res. A 541 (2005) 441
[3] S. Eränen, T. Virolainen, I. Luusua, J. Kalliopuska, K.Kurvinen, M. Eräluoto, J. Härkönen, K. Leinonen, M.
Palviainen and M. Koski, 2004 IEEE Nuclear Science Symposium, Conference Record, paper N28-3, Rome (Italy),
October 16-22, 2004
[2]
Mask Layout-Test structures
Standard (planar)
test structures
10x10 matrix
Ø hole 10 µm
44 holes GR
p-stop 20 µm
Ø implant 44 µm
3D-Diode
Pitch 80 µm
Backplane full-depletion-voltage
Preliminary “3d-diode”/back
capacitance measurements
D4_4
10
Cdiode [pF]
Lateral depletion contribution to measured
capacitance at low voltages
12
C-V
8
6
4
2
Linear 1/C2 vs V region corresponding to
the same doping level of planar diodes
0
0
2.50
 Column depth ~ 150µm
2.00
C
-2
-2
[pF ]
Saturation capacitance corresponding to a
depleted width of ~ 150µm)
10
20
30
Vrev [V]
40
50
60
1.50
1.00
1/C2-V
0.50
0.00
0
10
20
30
Vrev [V]
40
50
60
~ 40V full depletion voltage (300µm
wafer)
S/N for HEPAPS2
Test beam at HERA
FAPS: RAL group
• FAPS could be extended to a full 20 samples per train, stored in pixel
• If this doesn’t fit with 0.25 mm CMOS, will surely be OK with 0.13 mm
• Idea is to relax the requirement for fast, precise, signal transmission to chip
periphery during train, and so render long columns feasible, with all processing logic
outside the detector active volume, as for the CCD architecture
• Test devices implemented using a 0.25 mm process – TSMC(imaging)
FAPS resolution
13um hit resolution