The CMS electromagnetic calorimeter: status, performance with cosmic and
first LHC data
The CMS electromagnetic calorimeter: status, performance with cosmic and
first LHC data
Cristina Biino* - INFN Torino
11th ICATPP Conference onAstroparticle, Particle, Space Physics, Detectors and Medical Physics
ApplicationsVilla Olmo, 5‐9 October, 2009
*On behalf of the CMS Electromagnetic Calorimeter Group
Cristina Biino* - INFN Torino
11th ICATPP Conference onAstroparticle, Particle, Space Physics, Detectors and Medical Physics
ApplicationsVilla Olmo, 5‐9 October, 2009
*On behalf of the CMS Electromagnetic Calorimeter Group
• Electromagnetic calorimeter at CMS– System description
• Results from cosmic ray running– Performances
– Response stability
• Results from LHC beams data• Current status• Conclusions
Outline Outline
PixelsTrackerECALHCAL
SolenoidMuons
• Benchmark channel: discovery of low mass Higgs in H channel
• Target energy resolution 0.5% at high energy for unconverted photons
ECAL BarrelECAL BarrelECAL EndcapECAL Endcap
The CMS detector requirements The CMS detector requirements
Scintillating PbWO4 crystals; Pb/Si preshower
E.M. Calorimeters:ECal barrel & endcap
• Compact & modular
• Hermetic
• Large Energy Range
• Fast & Stable
• Radiation Resistant
• Excellent Energy Resolution
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
Barrel (EB)36 SuperModules (18 per half barrel)
61,200 crystalsTotal crystal mass 67.4t
|| < 1.48 x = 0.0175 x 0.0175
Barrel (EB)36 SuperModules (18 per half barrel)
61,200 crystalsTotal crystal mass 67.4t
|| < 1.48 x = 0.0175 x 0.0175
Endcap Preshower (ES)Pb (2Xo) / Si (1Xo) 4 Dees (2 per endcap)
4,300 Si strips1.8mm x 63mm1.65< || < 2.6
Endcap Preshower (ES)Pb (2Xo) / Si (1Xo) 4 Dees (2 per endcap)
4,300 Si strips1.8mm x 63mm1.65< || < 2.6
Endcaps (EE)4 Dees (2 per endcap)
14,648 crystals Total crystal mass 22.9t
1.48< || < 3 x = 0.01752 ↔ 0.052
Endcaps (EE)4 Dees (2 per endcap)
14,648 crystals Total crystal mass 22.9t
1.48< || < 3 x = 0.01752 ↔ 0.052
Barrel crystals
Pb/Si EndcapPreshower
Endcap ‘Dee’ with ‘Supercrystals’
ECAL design and layout ECAL design and layout
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
Crystals are projective and positioned pointing slightly off the IP to avoid cracks.
Homogenous Lead Tungstate (PbWO4) Crystal Calorimeter & Pb-Si Preshower
Challenges:Crystal LY temperature dependence -2.2%/OC
Need excellent thermal stability (±0.05 OC)
Formation/decay of colour centres Need precise light monitoring system
Low light yield (1.3% NaI) Need photodetectors with gain in magnetic
field
Challenges:Crystal LY temperature dependence -2.2%/OC
Need excellent thermal stability (±0.05 OC)
Formation/decay of colour centres Need precise light monitoring system
Low light yield (1.3% NaI) Need photodetectors with gain in magnetic
field
Properties: Homogeneous medium
Fast light emission ~80% in 25 ns
Short radiation length X0 = 0.89 cm
Small Molière radius RM = 2.10 cm
Emission peak 425nmReasonable radiation resistance to very high doses
Light yield (23cm) 100 /Mev
Properties: Homogeneous medium
Fast light emission ~80% in 25 ns
Short radiation length X0 = 0.89 cm
Small Molière radius RM = 2.10 cm
Emission peak 425nmReasonable radiation resistance to very high doses
Light yield (23cm) 100 /Mev
23cm 25.8Xo
22cm 24.7Xo
• EB crystal, tapered34 types, ~(2.6x2.6 cm2 at rear)x23cm
• Two avalanche photodiodes (APD), 5x5 mm2 each, QE ~75%, Temperature coeff.: -2.4%/°C
• EB crystal, tapered34 types, ~(2.6x2.6 cm2 at rear)x23cm
• Two avalanche photodiodes (APD), 5x5 mm2 each, QE ~75%, Temperature coeff.: -2.4%/°C
• EE crystal, tapered 1 type, (3x3 cm2 at rear)x22cm
• Vacuum phototriodes (VPT), more rad hard than diodes; gain 8 -10 (B=3.8T), Q.E. ~20% at 420nm
• EE crystal, tapered 1 type, (3x3 cm2 at rear)x22cm
• Vacuum phototriodes (VPT), more rad hard than diodes; gain 8 -10 (B=3.8T), Q.E. ~20% at 420nm
Scintillating crystals and photodetectors Scintillating crystals and photodetectors
PbWO4 Producers:
BTCP (Bogoroditsk, Russia)
SIC (Shanghai, China)
PbWO4 Producers:
BTCP (Bogoroditsk, Russia)
SIC (Shanghai, China)
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
ECAL readout & pulse shape reconstruction ECAL readout & pulse shape reconstruction
• VFE card: 3-gain amplification, shaping, digitization & sampling every 25nsec
• From ten time samples reconstruct the signal amplitude A and Tmax using digital filtering technique, weights, fit and ratio methods.
• Subtract the pedestal P on event-to-event
digitization via ADC
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
Energy resolution goal is 0.5% at high energyTime resolution goal is 0.1 nsec
2007:2007:Individual signoff
of each SM during installation in P5; H4 EE Test Beam
2006 2007 2008
2006:2006:H4 Test Beam:
9 SM calibrated;H2 Combined Test Beam:
ECAL&HCAL
2006-20072006-2007::Commissioning & calibration of each SM with cosmics on
surface
2006:2006:2 SM tested with B-field on surface
(MTCC)
2008:2008:Endcap
Installation; Commissioning
with cosmics and first beam in-situ
2009:2009:Installation of
Preshower and commissioning
of Endcap trigger
2009
Highlights from ECAL project timeline Highlights from ECAL project timeline
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
Toyoko Orimoto, Caltech 8
EB Module: 400/500 crystals
EB @ P5EB SM with electronics
EE Dee
EE Dee 1 & 2 @ P5
SuperCrystal
ECAL construction ECAL construction
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
ECAL barrel installation in 2007 ECAL barrel installation in 2007
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
ECAL Endcaps assembly in 2008 ECAL Endcaps assembly in 2008
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
ECAL Endcaps installation in 2008 ECAL Endcaps installation in 2008
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
basic cluster
super-cluster
Particle energy reconstructionElectrons and Photons are essential in:• at least two of the Higgs decay channels • decay of a new heavy bosons• Supersymmetry• Standard electroweak and QCD processes
containement correction
CMS ECAL Performance ECAL Perfomance C. Biino – ICATPP 09
• Unconverted Photons– Best energy estimate: energy sum in fixed arrays of
crystals (97% of the shower contained in a 5x5 arrays)
– Containment corrections (position dependent) precisely measured at test beam with electrons
• Electrons/Converted Photons– Require recovery of Bremsstrahlung in tracker
material (1 X0)– Super-clusters of clusters along (bending direction)– In the endcaps, add also the energy deposited in the
preshower
ECAL energy resolution ECAL energy resolution Nine Barrel SuperModules were studied at test beam with electrons
in the energy range 15-230 GeV. • Achieved constant term in energy resolution better than 0.5%• Noise is at 40 MeV level per channel, as expected.
Each Barrel SuperModule was exposed to cosmics for at least one week with increased APD gain.• 5M triggers per SM (average of about
500 good events per crystal)
(σ/E)2 = (3.37%/√E) 2 + (108 MeV/E) 2 + (0.25) 2
Only 500 Endcap channels were calibrated with 120 GeV electrons.Light yield measurement for each
crystal; photocathode QE, gain and total photo-electron yield measured for each VPT
CMS ECAL Performance ECAL Performance C. Biino – ICATPP 09
Calibration & MonitoringSome Results
CMS ECAL Performance ECAL Perfomance C. Biino – ICATPP 09
ECAL Calibration & Monitoring
Uncalibrated Supermodule : 13%-25% spread in resolution among channels
Lab Pre-Calibration: 4% EB, 10% EE (all crystals)
Cosmic Pre-Calibration: 1.5-2.5% (all EB)
TestBeam Pre-Calibration: 0.3% (1/4 of EB & 500 EE xtals)
In-Situ Physics Calibration: 0.5% resolution
• Without inter-calibration, same signal would produce different outputs in different crystals.
• Also need overall energy scale
Calibration of ECAL crucial to maintain high energy resolution.
ECAL Stability (<< 0.5%):Monitored with Laser System
Transparency Change Correction:Signal Change under Irradiation, Measured with Laser Monitoring System
ECAL Monitoring (Monitor Stability and Measure Radiation Effects):
CMS ECAL Performance ECAL Calibration & Monitoring C. Biino – ICATPP 09
ECAL In-Situ CalibrationGoal: improve startup calibration as quickly as possible in-situ
Strategy Time Precision
symmetry: use invariance of mean energy deposited by jets at fixed Few hours ~ 2-3%
0 : mass peak @ low luminosity Few days <= 1%
Zee: absolute energy calibration 100 pb-1 < 1%
We: E/p measurement 5-10 fb-1 0.5%
CMS PreliminaryL=2x1030cm-2s-1
CMS PreliminaryL=2x1030cm-2s-1
CMS ECAL Performance ECAL Calibration & Monitoring C. Biino – ICATPP 09
• During LHC cycles the ECAL response will vary, depending on irradiation conditions and crystal characteristics:– Transparency changes and fast recovery
(a few hours)
• Damage and recovery are monitored by laser light injected into each crystal through optical fibres– Blue light (440 nm) tracks response– Infrared (796 nm) provides a check
• The laser is pulsed during the LHC ‘orbit gaps’
• An optical switch directs light to onehalf-supermodule or one quarter-Dee in turn. A complete cycle takes ~ 20 min.
• Small transparency changes at start-up (L = 1030–1031 cm-2s-1)
• Corrected response• Raw response
Time (hours)
ADC
2/ndf =73.9/68
Test beam data
Simulation of crystal transparency evolution
at LHC (L = 2 x 1033cm-2s-1)
- based on test beam irradiation results
0.2%
Stability of the crystal responseStability of the crystal response
Laser monitoring system Laser monitoring system
CMS ECAL Performance ECAL Calibration & Monitoring C. Biino – ICATPP 09
Performance with CosmicsSome Results
CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
Commissioning ECAL with cosmicsCommissioning ECAL with cosmics
CRAFT: Cosmic Run At Four Tesla
– continuous running for several weeks to gain operational experience
– > 300 M cosmic events collected
– magnetic field operated at 3.8T
– most of CMS subsystems participating
Minimum ionizing particles deposit 250 MeV in ECAL. Increase efficiency: signal/noise enhanced (x4) in EB to the value of 20, by increasing the gain of the APD.
CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
ECAL: timing and occupancy
SUSY09A. David
ϕ
Eseed > 100 MeV
Timing – bottom is late (t.o.f.)
Occupancy – top is busier (shaft side)
Top
Bottom
20
CRAFT
CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
(25 nsec units)
3x3 crystal energy deposit confirms absolute energy scale to few %
<260> MeV
3x3 crystal energy deposit confirms absolute energy scale to few %
<260> MeV
Energy deposits per ECAL cluster from cosmics.
Depend on track length inside the active ECAl volume
Cosmic rays
CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
Validate ECAL calibration with muons: measure energy deposition vs muon momentum
ECAL stopping power
Collision loss
Bre
mss
trahl
ung
Tracker momentum matches well with ECAL energy loss, energy scale is correct
CRAFT
momentum p measured in the CMS silicon tracker
dE: energy from ECAL cluster
dx: length traversed in ECAL crystals
dE/ρdx energy deposit matched to the track corrected for muon path length
Experimental data vs Expected stopping power for PbWO4 from literature
Not a fit !Not a fit !
CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
LHC Beam (Sept. 2008)Some Results
CMS ECAL Performance First LHC Beam – ECAL Perfomance C. Biino – ICATPP 09
First LHC Beam in 2008
“Splash” Event
Wed, 10 Sept. 2008“Splash” events observed when beam (450 GeV, 4.109 p) struck closed collimators 150m upstream of CMS Halo muons observed once beam (uncaptured and captured) started passing through CMS
Data-taking with LHC beam.
Data-taking with LHC beam.
High energy deposit in the calorimeters, particles travelling horizontally
useful to commission forward detectorsAll systems ON except Tracker and Solenoid
CMS ECAL Performance First LHC Beam – ECAL Perfomance C. Biino – ICATPP 09
Commissioning ECAL with first beamCommissioning ECAL with first beamBeam Splash Events: Single beam shots of 2109 protons onto
closed collimators 150m upstream of CMS
Longitudinal views
BEAMDebris
Transverse views
BEAM
Collimators
146m
CMSCMSDebris
Beam Splash Schematic
BEAM450 GeV
A “wave” or “splash” of secondary particles
passed through CMS, depositing a huge amount
of energy
ECAL ECAL EnergEnerg
yy
CMS ECAL Performance First LHC Beam – ECAL Perfomance C. Biino – ICATPP 09
Beam Splash: ECAL Energy
• More than 99% of ECAL channels fired
• Estimated hundreds of thousands of muons passing through CMS per event
• ~200 TeV energy deposited in EB+EE
• Inter-crystals timing established (< 1ns), inter-crystal calibration: EB (1.5-2.5% - test beam + cosmics), EE (~7% from splash events)
• White areas: channels masked from readout
ECAL Endcaps
crystal index ix
cry
sta
l in
dex
iy
crystal index ix
cry
sta
l in
dex
iy
En
erg
y
(GeV
)
TOP BOTTOM
ECAL BarrelECAL Barrel
cry
sta
l in
dex
i
crystal index i
En
erg
y
(GeV
)
CMS ECAL Performance First LHC Beam – ECAL Perfomance C. Biino – ICATPP 09
• Average energy per crystal over 50 splashes: 5-8 GeV
• Patterns:– shielding structures (square) and
floor of the LHC tunnel (bottom)
– lower energy at large radius of downstream EE, due to shielding effect of barrel
• EE pre-calibrations (spread 25%):
– Measurements from laboratory applied (precision of 9%): smoother and enhanced patterns
– New set being derived assuming local uniformity, to be combined with lab measurements for better startup values
CMS ECAL Performance First LHC Beam – ECAL Perfomance C. Biino – ICATPP 09
Beam Splash: ECAL Energy
CMSCMS
TANTAN
TCTV
TCTV
TCTH
TCTHTCLP
TCLP
BEAM
Correlation between Correlation between Energies in barrel HCAL Energies in barrel HCAL
and ECALand ECAL
Correlation between ECAL Correlation between ECAL & Beam Loss Monitors& Beam Loss Monitors
~150 TeV deposited in ECAL & ~1000 TeV deposited in HCAL
per splash event
Beam Splash Correlations
CMS ECAL Performance First LHC Beam – ECAL Perfomance C. Biino – ICATPP 09
Beam Splash: ECAL Timing
Muons
Muons
RED is a profile of the raw data, and BLUE is the nominal timing
according to the equation above.
• Observed pattern is due to pre-synchronization obtained with laser light
• Latency then adjusted w/ splashes: hardware allows steps of 1ns steps
• Further synchronization applied in offline reconstruction, better than 1 ns
• Synchronization from splashes will be start-up condition; better precision w/ LHC data
Beam splash events provide a source of synchronous hits throughout detector, allowing to internally synchronize ECAL
CMS ECAL Performance First LHC Beam – ECAL Perfomance C. Biino – ICATPP 09
Laser distributi
on modularit
y
Shutdown activities: 1st maintenance cycle
CMS ECAL status Shutdown Activities C. Biino – ICATPP 09
* Only a few/mille channels are not functioning
Preshower:– Installed in the months of
February-March 2009
– First data collected to check out components and connections same status of health as in the laboratory, prior to installation:
• 99.88% good channels (tot 137k)• MIP Signal/noise: 3.6 in low gain
(physics) and 9 in high gain (calib)
Crystal calorimeter:
EB and EE active through LHC beams and extended cosmic ray run in 2008/2009
More than 99.5% of the channels are in good health for physics
System routinely operated in CMS global exercises, collecting data to monitor the detector and consolidate data acquisition and procedures
Trigger commissioning in the endcaps: first data collected, being finalized
Beam pipe
preshower
HCAL
EE
ECAL status of the detector ECAL status of the detector
CMS ECAL status Shutdown Activities C. Biino – ICATPP 09
Closing of CMS: 2009CMS is now closed after a 7-months
long and successfulmaintenance period
and is moving again into “beam-ready” state
CMS ECAL status Shutdown Activities C. Biino – ICATPP 09
Conclusions
• Crystal part of CMS Electromagnetic calorimeter has collected data with LHC circulating beams and during cosmic ray test runs
• Preshower detector installed in feb-march 09– Optimal health
– Joined CMS global runs
• Beam splash events allowed to validate performance and improve:– Endcap startup calibrations
– Internal synchronization
• Long cosmic ray run has allowed to validate energy scale in the barrel and assess stability of temperature and transparency monitoring, both matching specifications
• CMS ECAL on track for first LHC collision data
CMS ECAL status ECAL Detector C. Biino – ICATPP 09
Spares
CMS ECAL Performance ECAL Calibration C. Biino – ICATPP 09
precision vs index (ring)
• Intercalibration precision at start-up• ECAL Barrel:
– 0.3% on 10 SM (electron beams)– 1.5-2.5% on 26 SM (cosmic
rays)
• ECAL Endcaps: – 1015% (LY measurement VPT gain)
• High energy electron test beam (0.3%) “Single crystal maximum response”
• Cosmic ray calibration (1.5-2.5%) Muons aligned to crystal axis Reference signal 250 MeV
• Crystal LY and transmission (4%) Co-60 gamma source
– Both validated against test beam data
36 Supermodules (100%) intercalibrated with cosmics
electron beam calibration
reproducibility (Aug - Sept)
/2 = 0.2%
10 Supermodules (25%) intercalibrated
with e-
Precalibration
1%
Light injected into each crystal using quartz fibres, via the front (Barrel) or
rear (Endcap)
Laser pulse to pulse variations followed with PN diodes to 0.1%
Normalise calorimeter data to the measured changes in transparency
F1 F2
PIN FE
LaserS
PWO
F1 F2
PIN FE
LaserS
PWO
Transparency and colour centres:These form in PbWO4 under irradiation
Partial recovery occurs in a few hours
Damage and recovery during LHC cycles tracked with a laser monitoring system; 2 wavelengths: 440 nm and 796 nm
Transparency and colour centres:These form in PbWO4 under irradiation
Partial recovery occurs in a few hours
Damage and recovery during LHC cycles tracked with a laser monitoring system; 2 wavelengths: 440 nm and 796 nm
Black: irradiation at test beamRed: after correction
1%
Reference diode
Transparency correction:Response to laser pulses relative to initial response provides correction
for loss of light yield loss PbWO4
Test beam irradiation exercises showed precision of correction of 0.15% on several channels
Transparency correction:Response to laser pulses relative to initial response provides correction
for loss of light yield loss PbWO4
Test beam irradiation exercises showed precision of correction of 0.15% on several channels
ECAL Laser monitoring system ECAL Laser monitoring system
CMS ECAL Performance ECAL Calibration & Monitoring C. Biino – ICATPP 09
ECAL: laser calibration chain ready
Data from a 300 h sequence
RMS (APD/PN) (RMS) VPt/PN
J.Malcles & laser team
% %
•Readout of PN ref working
•< 1 permill reproducibility
achieved•LED ‘stabilizer’
pulsing fully commissioned for
endcaps
Missing FEDs used for development
EE+
EE-
On EE- concurrent with LED load test
CMS ECAL Performance ECAL Calibration & Monitoring C. Biino – ICATPP 09CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
RMS (%)
Most of EB: stability better 1‰
In absence of transparency variation, the stability of the monitoring system can be assessed
• Laser data collected throughout CRAFT; laser sequence loops over all ECAL channels every 20 minutes;• For each channel and each sequence (600 events), the average <APD/APDref> is employed as monitoring variable• “Stability” is defined as the RMS over all laser sequences of normalized <APD/APDref>• Stabilities are computed for each channel on a period of 200 hours with stable laser conditions • APDref is chosen as a reference because of readout problems with PN reference diodes, which are being fixed• White regions lack statistics (2 supermodules were not readout for LV problems, now fixed)
CMS preliminary
Transparency monitoring stability (1)
CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
CRAFT
Mean = 0.3 ‰RMS = 0.2 ‰
99.6% of channels with RMS<1‰99.9% of channels with RMS<2‰
CMS preliminary
• 1-d projection of map in previous slideTransparency monitoring system stable in EB to better than than 2‰ in99.9% of the channels (Consistent with specifications needed to achieve the design resolution)
Transparency monitoring stability (2)
CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
CRAFT
°C
Average spread: 0.009
°C
CMS preliminary
• EB equipped with one precision temperature sensor every 10 channels, in good thermal contact with APD and crystal
• For each sensor, thermal stability is quantified with the RMS of the temperature measurements over one month of data taking
• The observed stability is 0.009°C on average and better than 0.05°C in all the channels.
1 month
Temperature stability during CRAFT
CMS ECAL Performance CRAFT – ECAL Perfomance with Cosmics C. Biino – ICATPP 09
CRAFT
• At the end of CRAFT fully readout ( albeit with some synch problems on few FEDs)
• Latency scan performed
• Data being analysed ES residuals: distribution dominated by track extrapolation error
Preshower fully part of CMS Preshower fully part of CMS
CMS ECAL status Shutdown Activities C. Biino – ICATPP 09
• Following a meeting with the LHC people, experiments and CERN management the plan to restart has been agreed.
• Once collisions at injection energy are established will move to collision at 7 TeV center-of-mass energy.
• In consultation with experiments and LHC operation will move to higher energy once some luminosity will be accumulated by the experiments and experience gained by the machine operations.
Prospects for 2009-2010 Run
CMS ECAL status Shutdown Activities C. Biino – ICATPP 09
Early Physics Programme Detector commissioning – much already done using cosmics/testbeam,..
Early beam: splash events, first collisions at injection energy, then at 7 TeV
Detector synchronization, alignment with beam-halo events, minimum-bias events. Earliest in-situ alignment and calibration
Early beam - collisions, up to 10-20 pb-1 @ 7 TeV Commission trigger, start “physics commissioning” – “rediscover SM”:
Physics objects; measure jet and lepton rates; observe W, Z, top And, of course, first look at possible extraordinary signatures…
7 TeV, up to 100 pb-1 measure Standard Model, start searches Per pb-1: 3000 W l (l = e,); 300 Z ll (l =e, ); 5 ttbar +X
Improved understanding of physics objects; jet energy scale from W j j’; extensive use (and understanding) of b-tagging
Measure/understand backgrounds to SUSY and Higgs searches Early look for excesses from SUSY & Z’ resonances.
Collisions at higher energy: extend searches; Explore large part of SUSY and resonances at ~ few TeV ~ 1000 pb-1 entering Higgs discovery era
CMS Physics Early Physics Prospects C. Biino – ICATPP 09