The RICH system of LHCbThe RICH system of LHCb
Roger Forty (CERN)on behalf of the LHCb collaboration
1. RICH detector layout
2. Photon detectors
3 Performance and first data3. Performance and first data
1st International Conference on Technology and Instrumentation in Particle PhysicsTIPP09, Tsukuba, 13 March 2009
The LHCb experimentp• Dedicated B physics experiment at the LHC
searching for new physics in complementary way to ATLAS & CMS through precision measurement of CP violation and rare decays
• B hadrons are predominantly produced in the forward directionsingle arm spectrometer (working in pp collider mode)→ single-arm spectrometer (working in pp collider mode)including two RICH detectors for charged hadron identification
10 – 300 mradRICH1RICH2
p p
Roger Forty The RICH system of LHCb 2
From design… to reality!
RICH2 RICH1
RICH2RICH1
Roger Forty The RICH system of LHCb 3The experiment it fully installed and commissioned, ready for first collisions
Requirementsq• Momentum range ~ 1 < p < 100 GeV
to match spectrum of B decay products
Without RICH
to match spectrum of B decay products
• Hadron identification is crucial for some modes (eg B → hh)
With RICH
Roger Forty The RICH system of LHCb 4
Radiators• Correlation between polar angle of tracks and p → two RICH detectors used
eθC max
250
200
Aerogel
eμ
p
K
π
242 mradC max
RICH1
θ C (
mra
d) 150
100
pRICH1
100
50
C4F10 gas
CF gas
53 mrad
32 mrad
RICH2
01 10 100
Momentum (GeV/c)
CF4 gasKπ
Momentum (GeV/c)
• RICH1 for low and intermediate momenta (before spectrometer magnet) combines use of two radiators: C4F10 gas (n = 1.0014) and aerogel (n = 1.03)RICH2 f hi h t i CF di t ( 1 0005)
Roger Forty The RICH system of LHCb 5
• RICH2 for high momentum, using CF4 gas as radiator (n = 1.0005)
RICH1 layouty
Magnetic
Cross section (side view) 3D model
Photon Detectors
Magneticshielding Vertex detector
vacuum tank
250 mrad
Beam pipe
Aerogel SphericalMirror
C4F10
Track
Beam pipe
VELO exit window
C4F10
PlaneMirror
0 100 200 z (cm)Photon detectors
Roger Forty The RICH system of LHCb 6
0 100 200 z (cm)
RICH1 construction• Challenge: to minimize material budget (since within tracking volume)
– “No entrance window” — attached directly to vertex detector tank– Inner acceptance defined by conical beryllium beam pipe– Carbon-fibre composite construction of spherical mirror (~ 1% X0)
Spherical mirror
Aerogel
Pl iRoger Forty The RICH system of LHCb 7
Plane mirror
RICH2 layouty• Much larger structure
High precision required ~ 0.7 mradInternal view
Glass mirror substrate, segmentedFlat mirrors
Spherical MirrorsSpherical Mirrors
Support Structure
7.2 m
Photon Detectors + Shielding
Central Tube
m
Roger Forty The RICH system of LHCb 8
Photon Detectors + Shielding
RICH2 installation• RICH2 constructed on surface, including alignment of optical elements
Delicate transport (~ 1 kph) to the LHCb cavern, 100 m underground
Roger Forty The RICH system of LHCb 9
Photon detectors• Hybrid Photon Detectors (HPD) developed in collaboration with industry
Principal partner: DEP-Photonis for encapsulation of pixel anode in tube
(−18 kV)
8192-channel pixel chip 8× OR → 1024 pixels ( )( ) (500 × 500 μm square)→ Presentation by K. Wyllie
(cm)
5× demagnification from electron optics → 2 5 mm at photocathode as requiredRoger Forty The RICH system of LHCb 10
5× demagnification from electron optics → 2.5 mm at photocathode, as required
HPD productionp• 484 HPDs required in LHCb (196 in RICH1, 288 in RICH2)
550 produced including spares over 18 months up to July 2007
• Multialkali (S20) photocathode, quantum efficiency improved during production, significantly better than the specification
Average & running average/batch
Window cutoff Sensitivity in visible range for aerogel
33
35
37<QE> per batchrunning <QE> (batch 0-25)
Average & running average/batch
25
27
29
31
QE
[%]
19
21
23
0 2 4 6 8 10 12 14 16 18 20 22 24batch no.
Batch number
Roger Forty The RICH system of LHCb 11
Batch number
HPD installation• HPDs mounted in columns, to cover detector plane
M l i hi ld b d h HPDMumetal magnetic shield tube around each HPDServices for HV, LV, and readout electronicsmounted in frame
Roger Forty The RICH system of LHCb 12
Fully equipped planey q pp p
Roger Forty The RICH system of LHCb 13
Test patternp• Regular array of light spots projected
over HPD plane in situ in RICH2
0 100 200 300 4000
500
1000
1500
2000
2500
3000
3500
4000
Pixel Hit Map - All Hits
over HPD plane in situ in RICH2 (similar results achieved in RICH1 using motorized stage with LEDs)
0 100 200 300 4000
500
1000
1500
2000
2500
3000
3500
4000
Pixel Hit Map - All Hits
• Nicely uniform response and very low noise for almost all the HPDsA few show noisy behaviour, peaking
0 100 200 300 4000
500
1000
1500
2000
2500
3000
3500
4000
Pixel Hit Map - All Hits
around the central axis of the tube
• Due to degraded vacuum quality in th t b
0 100 200 300 4000
500
1000
1500
2000
2500
3000
3500
4000
Pixel Hit Map - All Hits
those tubes
• Photoelectrons ionize residual gas Ions then accelerated back to the
0 100 200 300 4000
500
1000
1500
2000
2500
3000
3500
4000
Pixel Hit Map - All Hits
photocathode producing further p.e. → ion feedback 0 100 200 300 4000
500
1000
1500
2000
2500
3000
3500
4000
Pixel Hit Map - All Hits
Hits/pixel
Roger Forty The RICH system of LHCb 14
Ion feedback• Ion feedback rate determined using fraction of large clusters (≥ 5 hits)
Rate measured regularly for all HPDs over the last 18 months
• Most show a linear increase of ion feedback with time, with low gradientThe noisy tubes have a higher gradientThe noisy tubes have a higher gradient
• The bad tubes eventually start to glow, but only after ion feedback rate > 5%
Glow light
Roger Forty The RICH system of LHCb 15
HPD repairp• 24 HPDs in the RICH detectors
that started to glow have been l d i hreplaced with spares
• Failure of tubes can be accurately predicted from theaccurately predicted from the ion feedback measurements
~ 11 HPDs / year predicted to require replacement over the lifetime of the experiment(i.e. 2% / year)( y )
• Removed tubes are successfully repaired by DEP-Photonis O d l d h h dOpened, cleaned, photocathode reapplied: ion feedback gradient is low after repair
Roger Forty The RICH system of LHCb 16Age (days)
Beam test• Before final installation in the
RICH detectors, production HPDsRICH detectors, production HPDs used in a beam test with 25 ns bunch structure and C4F10 radiator
• Clean reconstructed ring spread over a number of HPDs →exercised alignment procedure
Number of detected photoelectrons
X (mm)
Number of detected photoelectronsmatches expectation from the GEANT-based full simulation
N b f hit i l
Roger Forty The RICH system of LHCb 17
Number of hit pixels
Magnetic distortionsg• Minor distortions to HPD image expected from fringe field of magnet
Test pattern: B = 0 B axial (30 G) B transverse (50 G)
• Calibration of distortion from LHCb spectrometer dipole made in situ, using light spot pattern, g g p p
• Raw data superimposed from runs taken with (B = 0, B +ve, B –ve)di l d i lldisplayed sequentially
• Distortion can be parameterizedwith a few constants per HPD
Roger Forty The RICH system of LHCb 18
with a few constants per HPD
Alignment & calibrationg• After all corrections, reconstructed
light spot positions match precisely
-300 -200 -100 0 100 200 300
-600
-400
-200
0
200
400
600
HPD array
with regular ~ 10 mm grid pattern
• Residual contribution to resolution σ 0 6 mm << pi el si e (2 5 mm)
-300 -200 -100 0 100 200 300
-600
-400
-200
0
200
400
600
HPD array
σ ~ 0.6 mm << pixel size (2.5 mm)→ correction of magnetic distortion and misalignment is under control
-300 -200 -100 0 100 200 300
-600
-400
-200
0
200
400
600
HPD arraym
m)
-300 -200 -100 0 100 200 300
-600
-400
-200
0
200
400
600
HPD arrayy
(m
xResEntries 3400
Mean 0.03103
RMS 0.7814
Constant 5.9± 240.9
Mean 0.01234± 0.06125 Sigma 0.0108± 0.6329
-6 -4 -2 0 2 4 60
50
100
150
200
250
300
xResEntries 3400
Mean 0.03103
RMS 0.7814
Constant 5.9± 240.9
Mean 0.01234± 0.06125 Sigma 0.0108± 0.6329
xRes yResEntries 3400
Mean -0.01304
RMS 0.7214
Constant 6.4± 257.6
Mean 0.010556± 0.007389
Sigma 0.0105± 0.5977
-6 -4 -2 0 2 4 60
50
100
150
200
250
300
350
yResEntries 3400
Mean -0.01304
RMS 0.7214
Constant 6.4± 257.6
Mean 0.010556± 0.007389
Sigma 0.0105± 0.5977
yResResiduals in both dimensions
-300 -200 -100 0 100 200 300
-600
-400
-200
0
200
400
600
HPD array
xResEntries 3400
Mean 0.03103
RMS 0.7814
Constant 5.9± 240.9
Mean 0.01234± 0.06125 Sigma 0.0108± 0.6329
-6 -4 -2 0 2 4 60
50
100
150
200
250
300
xResEntries 3400
Mean 0.03103
RMS 0.7814
Constant 5.9± 240.9
Mean 0.01234± 0.06125 Sigma 0.0108± 0.6329
xRes yResEntries 3400
Mean -0.01304
RMS 0.7214
Constant 6.4± 257.6
Mean 0.010556± 0.007389
Sigma 0.0105± 0.5977
-6 -4 -2 0 2 4 60
50
100
150
200
250
300
350
yResEntries 3400
Mean -0.01304
RMS 0.7214
Constant 6.4± 257.6
Mean 0.010556± 0.007389
Sigma 0.0105± 0.5977
yRes
-300 -200 -100 0 100 200 300
-600
-400
-200
0
200
400
600
HPD array
xResEntries 3400
Mean 0.03103
RMS 0.7814
Constant 5.9± 240.9
Mean 0.01234± 0.06125 Sigma 0.0108± 0.6329
-6 -4 -2 0 2 4 60
50
100
150
200
250
300
xResEntries 3400
Mean 0.03103
RMS 0.7814
Constant 5.9± 240.9
Mean 0.01234± 0.06125 Sigma 0.0108± 0.6329
xRes yResEntries 3400
Mean -0.01304
RMS 0.7214
Constant 6.4± 257.6
Mean 0.010556± 0.007389
Sigma 0.0105± 0.5977
-6 -4 -2 0 2 4 60
50
100
150
200
250
300
350
yResEntries 3400
Mean -0.01304
RMS 0.7214
Constant 6.4± 257.6
Mean 0.010556± 0.007389
Sigma 0.0105± 0.5977
yRes
Δx (mm) Δy (mm)
Roger Forty The RICH system of LHCb 19
-300 -200 -100 0 100 200 300
-600
-400
-200
0
200
400
600
HPD array
x (mm)Δx (mm) Δy (mm)
Pattern recognitiong• Simulated event from 14 TeV collision: high multiplicity, overlapping rings
60 60 RICH 2Radiator (Npe/saturated track)
40
60 RICH 1
40
60 RICH 2Aerogel (6)( pe )
0
20
y (c
m)
0
20
y (c
m)C4F10 gas (24)
-40
-20
y
-40
-20
y
CF gas (18)
-60
-60 -40 -20 0 20 40 60
x (cm)
-60
-60 -40 -20 0 20 40 60
x (cm)
CF4 gas (18)
• Pattern recognition based on reconstructed tracks: θC of each p.e. from each track calculated under (e, μ, π , K, p) hypotheses Vary chosen hypotheses of tracks to maximize the global event likelihood
x (cm)
Roger Forty The RICH system of LHCb 20
Vary chosen hypotheses of tracks to maximize the global event likelihood
Expected performancep p• Determined using bb events
from PYTHIA passed through the full simulation
• Efficiency to identify kaons and misidentification rate of
K → K ⟨ε⟩ = 96 %and misidentification rate of pions plotted vs momentum
• Performance fed into the physics studies recently undertaken for six of the key measurements of LHCb:
π → K ⟨ε⟩ = 4 %
measurements of LHCb:
B (Bs → μμ), AFB (B0 → K*μμ) ACP (Bs → φγ), ACP (Bs → J/ψφ) Momentum (GeV/c)CP ( s φγ), CP ( s ψφ)ACP (B → DK), ACP (B → hh)
• Particle ID performance matches well the requirements from the physics
( )
Roger Forty The RICH system of LHCb 21
First LHC data• No collisions yet, but data from injection
tests last September:HPD data from RICH2
tests last September:
Particle flux (plan view)
Beamstop
LHCb
• Unfortunately tracks passed through LHCb in “wrong” direction, high density ~ 10/cm2
→ Cherenkov light in the HPD windows→ Cherenkov light in the HPD windows
• Sufficient to measure relative timingReady for collisions!
Roger Forty The RICH system of LHCb 22
Ready for collisions!
Upgrade planspg p• Upgrade of LHCb under discussion to follow first data-taking phase
Expected at ~ 2015, but need to consider now to launch any R&D required
• Plan to increase B rate by running at higher luminosity (~ 1033 cm-2s-1)and read out full experiment at 40 MHz (instead of 1 MHz at present)→ upgraded HPD or a different photon detector will be required→ upgraded HPD or a different photon detector will be required
• Detector concept also under study to use of time-of-flight for p < 10 GeVQuartz plate with focusing system (PANDA-style) to correct for photon time-of-propagation in plate: p p p g p
Roger Forty The RICH system of LHCb 23σ(t) ~ 100 ps / photon required
ConclusionsConclusions• The LHCb RICH system is fully installed and commissioned
• The HPD photon detectors have been thoroughly tested in beam tests and with the first LHC beams, and most perform excellently
A f ff f i d t d d d lit d b iA few suffer from noise due to degraded vacuum quality and are being replaced as necessary (predict 2% of the HPDs / year)
• Alignment and calibration of magnetic distortions has been performedAlignment and calibration of magnetic distortions has been performed using light spot data, and will be refined with data from collisions
Particle ID performance determined from simulation matches the requirements of the LHCb physics programme
• First ideas for an upgrade are being considered, to be ready in case the (new) physics from the first phase of LHC running demands itthe (new) physics from the first phase of LHC running demands it
• We are ready for the first LHC collisions expected later this year
Roger Forty The RICH system of LHCb 24