Status Report of The CMS Experiment Christos Leonidopoulos CERN-PH on behalf of the CMS...
-
Upload
bruno-thompson -
Category
Documents
-
view
222 -
download
5
Transcript of Status Report of The CMS Experiment Christos Leonidopoulos CERN-PH on behalf of the CMS...
Status Report of The CMS Experiment
Christos Leonidopoulos
CERN-PHon behalf of the CMS Collaboration
102nd LHCC Meeting, CERN7 July 2010
The Collaboration
2
AACHEN-1, AACHEN-3A, AACHEN-3B, ADANA-CUKUROVA, ANKARA-METU, ANTWERPEN, ATHENS, ATOMKI, AUCKLAND, BARI, BEIJING-IHEP, BOGAZICI, BOLOGNA, BOSTON-UNIV, BRISTOL, BROWN-UNIV, BRUNEL, BRUSSEL-VUB, BRUXELLES-ULB, BUDAPEST, CALTECH, CANTERBURY, CARNEGIE-MELLON, CATANIA, CCCS-UWE, CERN, CHANDIGARH, CHEJU, ILLINOIS-CHICAGO, CHONNAM, CHUNGBUK, CHUNGLI-NCU, COLORADO, CORNELL, DEBRECEN-IEP, DELHI-UNIV, DEMOKRITOS, DESY, DONGSHIN, DUBLIN-UCD, DUBNA, EINDHOVEN, FAIRFIELD, FERMILAB, FIRENZE, FLORIDA-FIU, FLORIDA-STATE, FLORIDA-TECH, FLORIDA-UNIV, FRASCATI, GENOVA, GHENT, HAMBURG-UNIV, HEFEI-USTC, HELSINKI-HIP, HELSINKI-UNIV, HEPHY, IOANNINA, IOWA, IPM, ISLAMABAD-NCP, JOHNS-HOPKINS, KANGWON, KANSAS-STATE, KANSAS-UNIV, KARLSRUHE-IEKP, KHARKOV-ISC, KHARKOV-KIPT, KHARKOV-KSU, KONKUK-UNIV, KOREA-UNIV, KYUNGPOOK, LAPP, LAPPEENRANTA-LUT, LIP, LIVERMORE, LONDON-IC, LOUVAIN, LYON, MADRID-CIEMAT, MADRID-UNIV, MARYLAND, MEXICO-IBEROAM, MEXICO-IPN, MEXICO-PUEBLA, MEXICO-UASLP, MILANO-BICOCCA, MINNESOTA, MINSK-INP, MINSK-NCPHEP, MINSK-RIAPP, MINSK-UNIV, MISSISSIPPI, MIT, MONS, MOSCOW-INR, MOSCOW-ITEP, MOSCOW-LEBEDEV, MOSCOW-MSU, MOSCOW-RDIPE, MUMBAI-BARC, MYASISHCHEV, NAPOLI, NEBRASKA, NICOSIA-UNIV, NORTHEASTERN, NORTHWESTERN, NOTRE DAME, NUST, OHIO-STATE, OVIEDO, PADOVA, PAVIA, PEKING-UNIV, PERUGIA, PISA, POLYTECHNIQUE, PRINCETON, PROTVINO, PSI, PUERTO RICO, PURDUE, PURDUE-CALUMET, RAL, RICE, RIE, RIO-CBPF, RIO-UERJ, ROCHESTER, ROCKEFELLER, ROMA-1, RUTGERS, SACLAY, SANTANDER, SAO PAULO, SEONAM, SEOUL-EDU, SEOUL-SNU, SHANGHAI-IC, SKK-UNIV, SOFIA-CLMI, SOFIA-INRNE, SOFIA-UNIV, SPLIT-FESB, SPLIT-UNIV, ST-PETERSBURG, STRASBOURG, SUNY-BUFFALO, TAIPEI-NTU, TALLINN, TASHKENT, TBILISI-IHEPI, TBILISI-IPAS, TENNESSEE, TEXAS-TAMU, TEXAS-TECH, TIFR-EHEP, TIFR-HECR, TORINO, TRIESTE, UCDAVIS, UCLA, UC RIVERSIDE, UC SANTA BARBARA, UC SAN DIEGO, UNIANDES, VANDERBILT, VILNIUS-ACADEMY, VILNIUS-UNIV, VINCA, VIRGINIA-TECH, VIRGINIA-UNIV, WARSAW-IEP, WARSAW-INS, WARSAW-ISE, WAYNE, WISCONSIN, WONKWANG, YEREVAN, ZAGREB-RUDJER, ZURICH-ETH, ZURICH-UNIV
• 182 Institutions • 3000 scientists and engineers• 2000 Authors
Reminder: we went from this…3.8T Superconducting Solenoid
All Silicon Tracker (Pixels and Microstrips)
Lead tungstate E/M Calorimeter (ECAL)
Redundant Muon System(RPCs, Drift Tubes,
Cathode Strip Chambers)
Hermetic (|η|<5.2) Hadron Calorimeter (HCAL)
[scintillators & brass]
3
First 7 TeV collisions in CMS – 30 March 2010
…to this
…and this, just three months later
Life did not begin in a vacuumwith the first collisions
One Billion Cosmic Muons
7
before collisions
23 JINST papers: March 2010 (Vol. 5)
8
Feedback into realistic simulationto help us prepare for collisions
Detector understanding• “Why should we believe that the simulation
correctly describes the detector performance?”• Excellent question!• TeVatron experience: it takes a long time to
commission & understand collider experimentsAccelerator, detector, trigger, background,
underlying event, software: very complicated problems
Claim:• Cosmic runs/beam tests have made a difference• First data distributions agree well with simulation
9
CMS is still in the commissioning phase
From data-taking to the plots
10
• Hard work, long hours
• Despite early phase and complexity of experiment Unprecedented levels of readiness Very encouraging first results
• But: Always problems seeking solutions Hardest part is ahead of us
Operations
Integrated luminosity
12
Since end of M
arch (7
TeV):
100 nb-1 delivere
d (*)
88 nb-1 re
corded (~88%)
• ~3/4 of data recorded arrived in last 10 days• Working hard to integrate full datasets for ICHEP• Most performance plots use only fraction of data
(*) Stable beams only
L≈ 1027-1029 cm-2s-1
L≈ 1030cm-2s-1
Subdetectors status
13
MUON-CSC
MUON-DT
MUON-RPC
HCAL BARREL
HCAL ENDCAP
HCAL FORWARD
ECAL BARREL
ECAL END-CAP
PRE-SHOWER
STRIP TRACKER
PIXEL TRACKER
MUON-CSC
MUON-DT
MUON-RPC
HCAL BARREL
HCAL ENDCAP
HCAL FORWARD
ECAL BARREL
ECAL END-CAP
PRE-SHOWER
STRIP TRACKER
PIXEL TRACKER
Series1
98.5
99.8
98.8
99.9
100
99.9
99.3
98.9
99.8
98.1
98.2
Alignment/calibration status, dead/masked channels mirrored in MC
“The Trigger does not determine which Physics Model is Right.
Only which Physics Model is Left.”
DAQ/Trigger
15
• L1/DAQ rate: 45 kHz, @<0.5 MB/evt• High-Level Trigger: have successfully deployed
online trigger menus spanning luminosities from 1E27 through 2E30o Very smooth running throughout (200-400 Hz)
• HLT CPU-performance: 49 ms/evto Primary contributors: commissioning
and early analysis triggerso Contingency: factor of 2o Constantly on watch list Overflows taken into
account in the mean
Run 138737
Trigger Performance
16
• HLT muon efficiency wrt L1• L1 objects matched to offline
objects• ~90% efficiency at the plateau
• Photon efficiency wrt offline “super clusters”• For barrel & endcaps• Nearly 100% efficient
Predicting trigger rates: MC vs. data
17
“Building trigger menus 101”
Predicting trigger rates: MC vs. data
18
Monte Carlo:• Only used as a cross-check
at this point• Some trigger paths have
significant cosmic or noise distributions that are not modeled with “baseline” MC
• Still, impressive agreement overall
Using MC to cross-check 4.6E29 rates
Predicting trigger rates: MC vs. data
19
Data:• Most triggers exhibit fairly
linear behavior vs. luminosity
• Extrapolation errors minimized by using most recent data to keep the rate non-linearities under control
• Rates of all main players are predicted within ~20%
Using 1.2E29 rates to predict 4.6E29 rates
Calibration Trigger Streams
20
• Calibration triggers have access to full L1 rate, and they output small fraction of event
• Feature unique to CMS HLT• Calibration starts online!
Trigger calibration streams
21
• Calibration triggers have access to full L1 rate, and they output small fraction of event
• π0 peak reconstructed offline 200 seconds into 7 TeV run
30 March 2010
LHC has delivered
Trigger has accepted
CMS will analyze22
Analysis Activity
23
10-20k analysis jobs running on Tier-2s continuously every day of June
Routinely delivering 100k jobs per dayOctober 09
MC Exercise
7 TeV data
Winter Break
Physics production
25
3+1 CMS papers since May
• Rise of the particle density at (2.36) 7 TeV steeper than in models• Careful tuning effort of the MC generators is ongoing
“Transverse Momentum and Pseudorapidity Distributions of Charged Hadrons in pp Collisions at √s=7TeV”, submitted to PRL
26
CMS paper at 7TeV
Detector & Physics PerformanceCalorimetry Jets
Tracking b-tagging
Muon EWK/Onia
Particle
Flow Electron EWK/Onia
Detector & Physics PerformanceCalorimetry Jets
Tracking b-tagging
Muon EWK/Onia
Particle
Flow Electron EWK/Onia
1.46M of π0 → γγPT(γ) > 0.4 GeV, PT(pair) > 1 GeV
MC based correction applied according to cluster η and energy
0.43 nb-1 1.46M π0
0.43 nb-1 25.5k η
25.5K η → γγPT(γ) > 0.5 GeV, PT(pair) > 2.5 GeV
Calorimetry: π0 and η → γγ
• Statistics refer to < 0.5 nb-1
• Very useful tool to intercalibrate the crystals• Good agreement in width and Signal/Background ratio• Masses agree with expectations to within 1% 29
DATA
DATA
MC
MC
Calorimetry: Missing ET
Calorimetric MET (GeV)
30
• Jets reconstructed with the anti-kT R=0.5 algorithm• Dijet selection : Jet PΤ > 25 GeV, Δφ > 2.1, |η| < 3• Loose ID cuts on number of components and neutral/charged energy fraction
Detector & Physics PerformanceCalorimetry Jets
Tracking b-tagging
Muon EWK/Onia
Particle
Flow Electron EWK/Onia
Calorimetric di-jet events
32
Dijet m
ass
Δφ(j1, j2)
# of Calo Towers
Fraction of EM energy in Calo-Jets
Detector & Physics PerformanceCalorimetry Jets
Tracking b-tagging
Muon EWK/Onia
Particle
Flow Electron EWK/Onia
Tracking distributions
Muon distributions
35
“Global Muons”: matched tracks from Muon system and Tracker
Global Muons
• η and pT distributions dominated by light hadron decay muons (red)• good agreement with MC prediction, including o heavy flavor decays (blue)o punch-through (black)o fakes (green)
Global Muons
pT spectrum
η distribution φ distribution
Tracking distributions
36
η distribution φ distribution
Tracker Material Budget
37
η distribution φ distribution
Tracker Material Budget
38
pixel cluster charge
Tomography
39
Detector & Physics PerformanceCalorimetry Jets
Tracking b-tagging
Muon EWK/Onia
Particle
Flow Electron EWK/Onia
3D impact parameter value and significance all tracks with pT> 1GeV belonging to jets with pT >
40 GeV and |η| < 1.5 - PFlow Jets anti-kT R=0.5)
Excellent alignment and general tracking performance
b-tagging 3D IP significance
41
42Two b-jets candidate
b-tagging example
42
CMS experiment at LHC, CERNRun 136100 / Event 2568584382010-25-5 03:43:48 CEDTB- → J/yK- candidate
CMS experiment at LHC, CERNRun 136100 / Event 2568584382010-25-5 03:43:48 CEDTB- → J/yK- candidate
All other tracks: pT > 1.0 GeV/c
2
2
GeV/ 135.3
GeV/ 268.5
cM
cKM
Detector & Physics PerformanceCalorimetry Jets
Tracking b-tagging
Muon EWK/Onia
Particle
Flow Electron EWK/Onia
Particle Flow• Particle Flow: Full Event reconstruction
Topological matching between charged particle momenta measured with tracker with clusters in calorimeter Corrects for energy loss along trajectories Better precision, full event info
• High-level object: requires holistic detector view Excellent tracker High E/M calorimeter granularity (0.017 × 0.017) Strong magnetic field to separate tracks
• CMS very well suited for P-Flow reconstruction 46
Particle Flow MET
47
Particle Flow MET
48
Laser forgotten on
Need cleaning strategies developed based on timing constraints
Particle Flow MET
49
Comparison between calorimetric and Particle-Flow MET(Minimum bias events)
Detector & Physics PerformanceCalorimetry Jets
Tracking b-tagging
Muon EWK/Onia
Particle
Flow Electron EWK/Onia
51
J/ψ → μ+μ−
Ongoing studies:• Momentum scale corrections by studying mass as a function of η, pT (material budget)• Efficiency studies with tag-n-probe• Flight distance with determination of prompt and b→J/ψ + Χ
terms
Signal events: 4150 ± 222Sigma: 43.1 ± 1.9 (stat.) MeVM
0 : 3.094 ± 0.001 (stat.) GeV
S/B= 5.2χ2/Νdof = 1.0
56 nb-1
52
J/ψ → μ+μ− : The Best Of
Selection of central (barrel), high-quality dimuons:• Resolution: 43.1 MeV → 21.0 MeV
56 nb-1
53
J/ψ → μ+μ− plus friends
60 nb-1
54
J/ψ → μ+μ− plus friends
Not enough statistics to disentangle all resonances (yet)
60 nb-1
• Event selection: Muon id cuts Isolation, pT and MET cuts
• Monte Carlo: Event count normalized to integrated luminosity
W→μν candidate
W± →μ±ν observation
55
# of candidate (MT > 50 GeV) = 137
# of expected signal (MT > 50 GeV) = 128
# of expected background (MT > 50 GeV) = 7
37 nb-1
• Event selection: Muon id cuts Loose isolation, pT cuts
• Monte Carlo: Event count normalized to integrated luminosity
Z →μ+μ− observation
Z→μμ candidate
56
#of candidate = 25#of expected signal = 24.7 #of expected background = 0.08
60 nb-1
Detector & Physics PerformanceCalorimetry Jets
Tracking b-tagging
Muon EWK/Onia
Particle
Flow Electron EWK/Onia
58
J/ψ → e+e−
• Higher background, tighter selection compared to muon channel• Challenging analysis, Particle-Flow selection crucial• Very promising preliminary results, signal clearly established
Signal events: 132 ± 14Sigma: 98 ± 12 (stat.) MeVM
0 : 3.070± 0.013 (stat.) GeV
37 nb-1
Two event selections: Basic electron ID, no MET cuts More advanced electron ID, cuts
on ET, MET, ΣET
W± →e±ν observation
59
173 candidates with MT > 50 GeVMC: Sig=163, Bkg=5
196 candidates with MT > 50 GeVMC: Sig =176, Bkg =11
51 nb-1
• Event selection: Two electrons with ET > 20 GeV
• Monte Carlo: Event count normalized to integrated luminosity
Z →e+e− observation
Z→ee candidate
60
52 nb-1
#of candidate = 18#of expected signal = 19 #of expected background = 0.8
Summary
7 TeV collisions: a very exciting run!
• The CMS detector is working according to design First performance results are very encouraging Its behavior can be reproduced in Monte Carlo
simulation Our level of understanding for this early
commissioning phase is very advanced
• The “rediscovery” of the SM has begun • We are setting the grounds for challenging it
as early as the end of 2010
62
Epilogue
• The technology of the LHC accelerator and experiments is unprecedented
• Massive amount of work and preparation invested in building and commissioning hardware & software
• But: we do not forget that the real challenges are still ahead (for all of us)
• We should consider this truly exciting period as the beginning of a marathon
63
The Beginning of The Journey
Credit for “Da Vinci” drawings: Sergio Cittolin
Credit for material used in this talk: LHC, CMS
Backup
65
The CMS Detector
MUON BARREL
CALORIMETERS
PixelsSilicon Microstrips210 m2 of silicon sensors9.6M (Strip) & 66M (Pixel) channels
ECAL76k scintillating PbWO4 crystals
Cathode Strip Chambers (CSC)Resistive Plate Chambers (RPC)
Drift Tube Chambers (DT)
Resistive Plate Chambers (RPC)
Superconducting Coil, 4 Tesla
Steel YOKE
TRACKER
MUONENDCAPS
HCALPlastic scintillator/brasssandwich
66
Muon pT resolution with cosmics1B events of (mostly muon) cosmic events collected make muons the best understood reconstructed object in CMS
Compare muon pT in upper, lower detector halves to evaluate resolution
67
12% resolutionat 1 TeV
π0 → γγ η, Φ distributions
Relative calibrationprecision ~ 2%target ~ 0.5% at 10pb-1
π0s and ECAL calibration
π0 and Φ symmetry
ECAL Barrel ECAL Barrel
HLT: CPU performance & pile-up
69
• First look at impact of pileup on CPU-performance
1 coll/bunch
2 coll/bunch
• Have deployed “multiple-vertex” trigger to
facilitate pile-up studies with real data
70
J/ψ → μ+μ−
• Run range: 132440-139370• Common selection:
No scraping Tracker Muons of opposite charge Pixel layers >= 2 Tracker hits >= 12 Tracker chi2 < 3 Mu pT > 2.5 Mu segments >=2 Matched L1DoubleMuOpen vertex Prob > 0.05
Run 136100, Event 256858438 • Measured Parameters:
18
mm 11.0
mm 93.1
844.0Prob
GeV/ 3.5
GeV/ 4.3
GeV/ 1.10
GeV/ 6.18)(
GeV/ 135.3
GeV/ 268.5
2
2
2
xyxy
xy
xy
T
T
T
T
LL
L
L
cKp
cp
cp
cBp
cM
cKM
3-trk vertex that is displaced from the PV by 2mm (18s).
Our background is dominated by real J/yDimuon mass in data (points) compared to MC (hist)