Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
1
ATLAS Status ReportLHCC 17 February 2010
D Charlton (Birmingham), J Boyd (CERN)
Technical Aspects were Covered in Pre-Meeting with Referees Yesterday
Data Taken, Detector & TriggerCalorimetry, Jets & EM objects
Muons and Inner Detector TrackingMinimum Bias Analysis
2
3
First collisions 23 NovemberFirst stable beams 6 December2009 data complete 16 December
Total collision candidate events: 9.2 x 105, ~ 21 μb-1 In stable beam conditions: 5.4 x 105, ~ 12 μb-1 With ID+solenoid on, good data quality: 3.8 x 105, ~ 9 μb-1
Short periods at 2.36 TeV ~ 34k collision events(without stable beam conditions: ID not fully on)
Data-taking efficiency ~90%
Only high pT tracks shown
2009 Collision Data
4
Luminosity
Instantaneous luminosity ℒ derived from:Main 2009 measures:
● MBTS two-side coincidence trigger rate● LAr offline event selection (coincidence of
in-time endcap E deposits)Relative measurements also from:
● Dedicated LUCID forward detectors● FCal total energy measurement
Overall ℒ scale uncertainty ~ 30%
Specific luminosity ~ constant during stable fills, as expected
LUCID detector, ±17m from IPLuminosity Cerenkov Integrating Detector
5
Luminosity
Instantaneous luminosity ℒ derived from:Main measures:
● MBTS two-side coincidence trigger rate● LAr offline event selection (coincidence of
in-time endcap E deposits)Relative measurements also from:
● Dedicated LUCID forward detectors● FCal total energy measurement
Overall ℒ scale uncertainty ~ 30%
Specific luminosity ~ constant during stable fills, as expected
LUCID detector, ±17m from IPLuminosity Cerenkov Integrating Detector
Overlay: Fcal total E measurement vs time
6
Level-1 Trigger
Main 2009 collision trigger: hit in either arm of minimum-bias trigger scintillators (MBTS) 2.09<|η|<3.84
Higher ET triggers also active, eg:
● Lowest threshold jet trigger in 0.4x0.4 ηxф
● Lowest threshold EM trigger, on pairs of 0.1x0.1 ηxф EM cells
L1calo algorithms use 1 GeV bit-step and in 2009 a tight noise suppression
L1 muon trigger active, statistics very low
Offline: anti-kT(0.4)
7
High-Level Trigger
Rapid commissioning on collision data:● Run systematically offline on data from 20
November (splashes) onwards● HLT infrastructure run online from the start to
provide event streaming● Real HLT menu run online from 25 November in
pass-through mode, e.g. online beamspot determination available from level-2
● Active event rejection started on 6 December for specific physics analysis requirement
8
Active event rejection introduced at the HLT from 6 December● beam pickup level-1 trigger (prescaled by 20)● ID activity required at HLT (~104 rejection)
Provides a sample to measure MBTS efficiency
Further HLT commissioning will be required in 2010 for high-ET object trigger signatures
High-Level Trigger Rejection
Stable beamsL1 BPTX activatedHLT activated
L2 rejection
Sampling 5%
Level-1 collision trigger (MBTS)
9
Detector Readiness - Summary
High operational fractions of all detector systems
Minor repairs/maintenance during technical stop:
● Barrel toroid current lead repairs
● Solenoid cooling circuit cleaned
● Maintenance of cooling and gas systems
● Minor detector cabling fixes
● Additional supports installed for future install'n of more small EE muon chambers
CSC muon chamber readout continuing its commissioning ATLAS is ready for 2010 data-taking
● Taking combined cosmic data since 2 February
10
Prompt Data Processing and DistributionMB/sper day
Total data throughput through the Grid (Tier0, Tier-1s, Tier-2s)
Beam splashes
First collisions
Nov. Dec.
Cosmics
End data-taking
In steady-state 2010 operation, “calibration loop” will be introduced● express-stream (~5%) reconstructed promptly at Tier-0● wait ~36h for semi-automatic calibration updates before starting bulk reconstruction
● e.g. beamspot, channel maps derived from express processing● reconstructed bulk data available after ~48h, improved quality
Data reconstructed & exported promptly
Typically ~4h from DAQ to Tier-2, including reconstruction time
Excellent performance of Tier sites for data processing and analysis
11
Reprocessing
“Fast reprocessing” campaign started just after data-taking ended in December → most plots shown today
● consistent reconstruction code● updated ID alignment constants…
Monte Carlo samples also reprocessed with same reconstruction version
Reprocessing performed at Tier-1 sites according to the ATLAS computing model
A second reprocessing was started on Friday last week → now largely done
● Uses software version for first 2010 data-taking Nine Tier-1s participating
Flexible re-allocation between Tier-1 sites possible
Jobs/day
First collision data reprocessing campaign
12
Calorimetry
13
LAr CalorimeterRaw cell energy distributions
● simulation quality is very good● full angular range including FCal
14
Tile Calorimeter
Raw cell energy distributionsIncludes 2.36 TeV data
15
events with2 jets pT> 7 GeV
JetsMBTS trigger (rather open)
Simulation normalised to data
Uncalibrated energies(EM scale)
16
Calorimeter Response to Isolated Tracks
Sample of isolated hadrons● 0.5<p
T<10 GeV, |η|<0.8
● No track within ΔR=0.4
E(ΔR<0.1)/pa powerful tool to test
simulation of calorimeter response to single hadrons
Remarkable quality of simulation: very promising for detailed understanding of jet energy calibration
A pay-off from the many years of test-beam studies and detailed comparisons to G4 simulation (models, material, etc)
Uncalibrated energies(EM scale)
17
Missing-ET
METyMETx
900 GeV and 2.36 TeV dataResolution of the two components
vs. total ET sum
Uncalibrated energies(EM scale)
EM scaleEM scale EM scale
18
Photon candidates with pT>2.5 GeV
● Fraction of energy in first layer, f1
(longitudinal shower development)● Lateral shower width in middle EM
layer, w2
● Average η profile in front-layer strips● Shower width computed over three
cells in first layer ws3
Good description of the data both in depth and transverse profiles
Calorimeter shower shape variables
19
Photon candidates with pT>2.5 GeV
● Fraction of energy in first layer, f1
(longitudinal shower development)● Lateral shower width in middle EM
layer, w2
● Average η profile in front-layer strips● Shower width computed over three
cells in first layer ws3
Good description of the data both in depth and transverse profiles
Calorimeter shower shape variables
20
Diphoton Mass Distributions
Two photon candidates● p
T(γ)>0.4 GeV, p
T(γγ)>0.9 GeV
● Calibrated cluster energies
Tighter kinematic cuts, including:● p
T(γ)>0.6 GeV, p
T(γγ)>1.5 GeV
Remove clusters with matched tracks
Widths and positions well described by simulation
21
Conversions
Reconstructed conversion radius● clean conversion sample● Dalitz decays separable● pay-off for detailed assay of
ID material
Reconstruct here using tracks with hits in silicon detectors
22
Conversions
Reconstructed conversion radius● clean conversion sample● Dalitz decays separable● pay-off for detailed assay of
ID material
Reconstruct here using tracks with hits in silicon detectors
π0 mass peak with one converted photon
23
Electron Identification
Electron candidates reconstructed from EM clusters with a matched track p
T>0.5 GeV
Track-based distributions well described by sum of hadron and conversion-electrons:
● few pixel hits on tracks● Strong discrimination even at these
low-pT for transition radiation hits
24
Electron Identification
Electron candidates reconstructed from EM clusters with a matched track p
T>0.5 GeV
Track-based distributions well described by sum of hadron and conversion-electrons:
● few pixel hits on tracks● Strong discrimination even at these
low-pT for transition radiation hits
Conversions with particle ID on other leg: see onset of transition radiation, as expected
25
Electron Identification
Electron candidates reconstructed from EM clusters with a matched track p
T>0.5 GeV
Track-based distributions well described by sum of hadron and conversion-electrons:
● few pixel hits on tracks● Strong discrimination even at these
low-pT for transition radition hits
Looking to the future: starting to test forward electron selection
Beyond tracking acceptance:
2.5 < |η| < 4.9(EMEC, HEC, FCal)
Fraction of energy, fem
,
deposited in EM layers
Mainly photons at high fem
Towards an extended acceptance for Z→ee and
H→eeee…
26
EM shower lateral radius
and evolution of its mean value with
track incident angle
Towards Tau IdentificationPure background
sample useful to test modelling of τ ID
variables at low pT
ET
cone (0.1<ΔR<0.2)
ET
cone (ΔR<0.4)
Muon Spectrometer
27
Muon candidates – data / MC comparisons
28
Muoncandidatesveryforwardandlowmomentumat900GeV–thresholdPT>2.5GeV.
Comparisonofcombinedmuoncandidates(requiringtrackingandmuonspectrometerhits)betweendataandMC.
NiceagreementwithinthelimitedstaKsKcs(50muoncandidates).
The Inner Detector
29
The900GeVdataprovideslotsoflowPTtracksoriginaKngfromthecentreofthedetectortocommissiontheInnerDetector.
(overlayofmanyevents)
Tracking - data / MC comparisons ExcellentdescripKonofthedatabytheMC
Numberofpixelhitsontrackversusη/ϕ
NumberofSCThitsontrackversusη/ϕ
30
‐CorrectMCforbeamspotposiKonandsize‐SimulaKngthedetectorconfiguraKonasusedindatataking
31
x residual [mm]-0.2 -0.1 0 0.1 0.2
nu
mb
er
of
hits o
n t
racks
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000MC perfect alignment
=65%core
m, f!=23core
!m, !=0core!
Collision data alignment
=66%core
m, f!=28core
!m, !=0core!
ATLASPreliminaryPixel Barrel
x residual [mm]-0.2 -0.1 0 0.1 0.2
nu
mb
er
of
hits o
n t
racks
0
500
1000
1500
2000
2500
3000
3500
4000 MC perfect alignment
=83%core
m, f!=20core
!m, !=0core!
Collision data alignment
=79%core
m, f!=22core
!m, !=0core!
ATLASPreliminaryPixel EC
x residual [mm]-0.2 -0.1 0 0.1 0.2
nu
mb
er
of
hits o
n t
racks
0
5000
10000
15000
20000
25000 MC perfect alignment
=75%core
m, f!=36core
!m, !=0core!
Collision data alignment
=73%core
m, f!=41core
!m, !=0core!
ATLASPreliminarySCT Barrel
x residual [mm]-0.2 -0.1 0 0.1 0.2
nu
mb
er
of
hits o
n t
racks
0
2000
4000
6000
8000
10000
12000
14000
16000MC perfect alignment
=77%core
m, f!=38core
!m, !=0core!
Collision data alignment
=74%core
m, f!=42core
!m, !=0core!
ATLASPreliminarySCT EC
Unbiased Hit Residuals after Alignment
PixelBarrel23μm28μm
PixelEC20μm22μm
SCTBarrel36μm41μm
SCTEC38μm42μm
MCperfectalignmentCollisionalignment
TheseresidualstakeintoaccountalignmentandintrinsicresoluKon‐dominatedbymulKplesca\eringinthissample.
Bigimprovementinalignmentwithcollisiondataespeciallyforendcapswherecosmicsarenotsouseful.ApproachingtheperfectlyalignedMCresidualsalready!Finalclusteringstrategyshouldimprovewidthsfurther.
32
GooddescripKonofthedatabytheMCfortheimpactparameterdistribuKons.
[mm]0
d-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Num
ber
of T
racks
0
20
40
60
80
100
120
140
160
180
200
220
310!
Data
ND MC MinBias
[mm]0
d-4 -2 0 2 4N
um
be
r o
f T
racks
310
410
510
610 = 900 GeVs
) [mm]! sin(0z-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Num
ber
of T
racks
0
20
40
60
80
100
120
140
310!
= 900 GeVs
Data
ND MC MinBias
) [mm]! sin(0
z-10 -5 0 5 10N
um
be
r o
f T
racks
210
310
410
510
610
Impact parameters - data / MC comparisons
Tracking – strange particles
Detailedstudiesoftrackingefficiency,momentumscaleandmomentumresoluKonongoingusingstrangeparKcledecays(KS).
33
Pixel dE/dx
34
UsingtheparKcleIDfeaturesofthepixeldetector.
Q p [GeV] -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
]2
cm
-1d
E/d
x [
Me
V g
0
1
2
3
4
5
6
7
8
9
10
1
10
210
ATLAS Preliminary
900 GeV Collisions p
K
π
Pixel dE/dx
35
dE/dxforprotoncandidatesselectedusingtheΛ(Λ)invariantmass.‐
UsingtheparKcleIDfeaturesofthepixeldetector.
Pixel dE/dx
36
dE/dxforprotoncandidatesselectedusingtheΛ(Λ)invariantmass.‐
UsingtheparKcleIDfeaturesofthepixeldetector.
dE/dxdominatedbypionsforsidebands
Event with Two Primary Vertices
37
Charged particle multiplicity analysis • Analysisoverview
– Useall900GeVdatatakenwithstablebeamsandwithtrigger,trackingdetectors,andsolenoidoperaKonal• ~9μb‐1ofdatafrom13runstakenbetweenDec6‐15,2009
– ChoosetomeasureprimarychargedparKclesinthephasespace• PT>500MeV,|η|<2.5
– Foranalysispresentedhereweshowfor|η|<2.2aswearesKllworkingonthefinalhighηmaterialcorrecKons
– Requireeventshaveatleastoneofthesetracksandareconstructedprimaryvertex
– Select:330kevents/1.9Mtracks• BeambackgroundesKmatedbylookingateventstakeninunpairedbunchestobe<10‐4of
selectedevents
• WithinourphasespacewemeasurefullyinclusiveinelasKcdistribuKonstoavoidanymodeldependence– FacilitatedirectcomparisonswithMCmodels– Correctforourtrigger,vertexing&trackingefficiencytothehadronlevel– DonotcorrectfordiffracKonorextrapolateoutsideourphasespace
38
FirstpublicpresentaKonoftheseresults!
SelNoVtxN
1 2 3 4 5 6 7 8 9 10 11 12
Trigger
Effic
iency
0.98
0.985
0.99
0.995
1
= 900 GeVsData,
Analysis overview - Trigger • WeuseadedicatedMinimumBiastrigger
– ±3.56mfromIP(2.09<|η|<3.84)
– 32scinKllaKngcounters– Require>0hitsoneitherside(veryinclusive)
• Measurethetriggerefficiencyfromdata– Usinganorthogonaltrigger– BeamcrossingsignalsfrompickupsatL1
– RequirePixel/SCThitsandaloosetrackinHigherLevelTrigger(looserthananalysisrequirement)
• TriggerEfficiencywithrespecttotheanalysisselecKonisextremelyhigh
39
Nobiasesw.r.ttrackη,PT,…VerylowsystemaKcuncertainty(<0.3%).
MinBiasTriggerScinKllators
No.ofselectedtracks
d0 [mm]
-10 -8 -6 -4 -2 0 2 4 6 8 10
Tra
cks
10
210
310
410
510
SecondariesPrimariesData
d0 [mm]
-10 -8 -6 -4 -2 0 2 4 6 8 10
Tra
cks
10
210
310
410
510
SelNoVtxN
1 2 3 4 5 6 7 8 9 10
Vert
ex E
ffic
iency
0.6
0.7
0.8
0.9
1
= 900 GeVsData,
Analysis Overview – Vertexing & Secondaries • Requireareconstructedprimaryvertex
– Require>2tracksinvertex(PT>150MeV)
– Vertexingefficiencymeasuredfromdata• SystemaKcuncertainty<0.1%
– ~100%fornumberofselectedtracks>4
• Toremovesecondariesrequire:• |d0|<1.5mm• |z0sinθ|<1.5mm
• Bothd0,z0w.r.tprimaryvertex• RemainingsecondariesesKmatedusingtailsintheimpactparameterdistribuKontobe~2%ofselectedtracks
No.ofselectedtracks(PT>500MeV)
[GeV]T
p
2 4 6 8 10 12 14 16 18 20
sel
!
0.4
0.5
0.6
0.7
0.8
0.9
1
=900 GeVsND MC MinBias
Systematic uncertainty
!
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
sel
"
0.4
0.5
0.6
0.7
0.8
0.9
1
=900 GeVsND MC MinBias
Systematic uncertainty
• Tracksforanalysis
– PT>500MeV
• WeseeexcellentdescripKonofdatabyMCsowecantakethetrackingefficiencyfromMC
– NocorrecKonstoefficiency– Lotsofefforthasgoneintoassessing
systemaKcuncertaintyduetodiscrepanciesbetweendata/MC• HitmulKpliciKes(seeearlierplots)
• Individualsystemtrackletefficiencies(data/MC)
• KSstudiesindata/MC
– OverallsystemaKc~3%incentralregion
Analysis Overview – Tracking eff.
!-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
reco
"
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
nominal
10%
20%
Analysis Overview – Material studies ‐ Alotofstudiesonunderstandingthedetectormaterialfromdata
‐ Investmentforthefuture‐ MajorityoftrackinginefficiencyfrommaterialinteracKons‐ LargestsystemaKcontrackingefficiencyfromuncertaintyonmaterialinthesimulaKon‐ BenefitfromcarefulweighingofthetrackingdetectorcomponentsbeforeinstallaKon‐ Expectphotonconversionstogiveusadetailedmapofmaterial
‐ Currentlynotenoughdatatogiveusrequiredprecision
MCstudiesofthetrackingefficiencywithuniformscalingupofthematerialby10and20%(inrad.length).
AllchargedparKclesNomaterialinteracKonsinSitracker
TodemonstratesensiKvityoftheefficiencytomaterial
43
‐ KSstudiestoconstraintheamountofmaterialinthedetector‐DatafavoursthenominalMCandisinconsistentwiththeMCwith10%,20%addiKonalmaterial‐ ForsystemaKcwecurrentlytakethelargestdifferenceinefficiencybetweennominalMCandthe10%addiKonalmaterialMC
‐ Gives3%systemaKcuncertaintyontrackingefficiency
‐ Otherstudiesgiveconsistentresults
Analysis Overview – Material studies
Radius [mm]10 210
310
mean m
ass [M
eV
]0 S
K
493
494
495
496
497
498
499
500
Beam
Pip
e
Pix
el Layer-
0
Pix
el Layer-
1
Pix
el Layer-
2
SC
T L
ayers
DataMC (nominal)
MC (10%)MC (20%)
| < 1.5!|
=900 GeV)sMinimum Bias Events (
ATLAS Preliminary
)!
/ d
ch
)(dN
ev
(1/N
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5 1" ch
| < 2.5, N! > 500 MeV, | T
p
Data = 900 GeVs
PYTHIA ATLAS MC09PYTHIA ATLAS MC09c
PYTHIA DWPYTHIA Perugia0
PHOJET
ATLAS Preliminary
)!
/ d
ch
)(dN
ev
(1/N
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
!
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Ratio
0.850.9
0.951
1.051.1
1.15 Data Uncertainties
MC / Data
!
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Ratio
0.850.9
0.951
1.051.1
1.15
44
Analysis overview - Results
‐ dN/dηdistribuKon‐ ShapeagreeswellwithsomePYTHIAtunes‐ ATLASdatashowshighervaluethenallMCs
‐ MCstunedindifferentregionofphasespace
Notesuppressed0ony‐axis
[GeV]T
p1 10
]-2
) [ G
eVT
dp
! /
dchN2
)(d T p
ev N"
1/(2
-910
-810
-710
-610
-510
-410
-310
-210
-1101
10
|<2.4!CMS NSD pp |
|<2.5!ATLAS pp |
= 900 GeVsATLAS Preliminary
45
Analysis overview - Results
‐ ComparingPTdistribuKonwithCMSdata
‐ ThankstoCMSformakingthedataavailable(arXiv:1002.0621v1)
‐ ATLAS/CMStreatdiffracKveeventsindifferentways
Summary• DuringthefirstrunATLAShasbeenworkingverywellfromdetectoroperaKontodataprocessing,distribuKonandanalysis
• FirstcollisiondatahaveshownthatthedetectorperformanceisexcellentandremarkablywelldescribedbythesimulaKon(inthemostdifficultsovregime)
• PreliminaryresultsoninclusivemeasurementsofchargedparKclemulKpliciKesat900GeVhavebeenpresented.Apaperwillbesubmi\edsoon.
46
ATLASisverygratefultotheLHCteamfortheexcellentmachineperformanceandlooksforwardtohighenergydata!