![Page 1: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/1.jpg)
Luminosity measurements at ATLASand
impact of PDF uncertainties on LHC physics
Maarten Boonekamp on behalf of ATLASHera-LHC workshop, 26/5/8
May 26, 2008 Maarten Boonekamp, CEA-Saclay 1
![Page 2: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/2.jpg)
Outline
Cross-section measurements : single processLuminosityLuminosityEfficiency (scale, resolution…)Acceptance
l i l iMultiple processes : ratiosCross-normalizing experimentCross-normalizing theory
Examples:Z as case studyApplications to W, high-mass Drell-Yan, top pairs, HiggsHiggs
Discussion
May 26, 2008 Maarten Boonekamp, CEA-Saclay 2
![Page 3: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/3.jpg)
Cross-section measurements
Counting rate : BALN += εσCounting rate :
(function of)
BALN +εσ
dAddLdBdNd ⊕ εσ
(function of)fundamental parameter(s)
nuisance
Uncertainty : A
dAdLdL
BNdBdNd ⊕⊕⊕
−⊕=
εε
σσ
Assume B/N small and/or well known:Term decreases statistically
To be addressed -Auxiliary measurements
May 26, 2008 Maarten Boonekamp, CEA-Saclay 3
Term decreases statistically Auxiliary measurements
![Page 4: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/4.jpg)
Example selections : Z ee, μμ
Events (/104) in 50 pb-1
ATLAS CSC
May 26, 2008 Maarten Boonekamp, CEA-Saclay 4
![Page 5: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/5.jpg)
1 : Luminosity measurement
(material : TDR and Per Grafstrom, LHCC, May 08)
Machine estimatesMachine estimatesEarly running : 20-25%Using special calibration runs with simplified machine parameters : get to 10% or better
Dedicated experimentRelative luminosity monitors :
LUCID ZDCLUCID, ZDCLAr/Tile currents; MBTS activity…
Absolute luminosity measurement : ALFAElastic scattering at small angles : well calculable Coulomb processDedicated machine optics; low luminosity. Result scaled to normal running conditions using the monitors
Used before : UA4, but also e+e- machines (Bhabha scattering)Aim : <3%
May 26, 2008 Maarten Boonekamp, CEA-Saclay 5
![Page 6: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/6.jpg)
ATLAS forward detectors
May 26, 2008 Maarten Boonekamp, CEA-Saclay 6
![Page 7: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/7.jpg)
Absolute luminosity from low-t elastic scattering
General expression of the elastic cross-section at 0 angle:
Allows a 4-parameter fit to L and hadronic parameters σtot, ρ, b
Requires : Detecting protons at θ ~ 3.5 μrad (UA4 : 120 μrad). Special machine parameters : parallel-to-point focusing; L ~ 1027p p p p g;Edgeless detector for optimal acceptancePrecision mechanics controlling movement towards/away from beamBackgrounds low and under control
May 26, 2008 Maarten Boonekamp, CEA-Saclay 7
![Page 8: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/8.jpg)
The detector and the Roman Pot
May 26, 2008 Maarten Boonekamp, CEA-Saclay 8
![Page 9: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/9.jpg)
Performance – Test beam
May 26, 2008 Maarten Boonekamp, CEA-Saclay 9
![Page 10: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/10.jpg)
Backgrounds
May 26, 2008 Maarten Boonekamp, CEA-Saclay 10
![Page 11: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/11.jpg)
Expected performancep p~100 hours at 1027
May 26, 2008 Maarten Boonekamp, CEA-Saclay 11
![Page 12: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/12.jpg)
2 : Efficiency
Simplest example : Z production. Two isolated leptons – Tag & probe
Tag Muon: Track in Inner Detector ANDTag Muon: Track in Inner Detector AND Muon Spectrometer (+Isolation and pT-Cuts)
Probe Muon: Track in Inner Detector (+Isolation and pT-Cuts)
If this di-muon mass is near 91 GeV and Δφ>2, then the probe muon is assumed to be a real muon
muon efficiency is given by the fraction of probemuon efficiency is given by the fraction of probemuons with tracks in the Muon Spectrometer
May 26, 2008 Maarten Boonekamp, CEA-Saclay 12
![Page 13: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/13.jpg)
Efficiency results
Electron and muon channelsATLAS CSCATLAS CSC
Lepton efficiency : dεl/εl ~ 2% (50 pb-1); 0.5% (1 fb-1)Th l b k d h ff t th ffi i d t i tiThe low backgrounds have ~no effect on the efficiency determination
Cross-section : dεZ/εZ ~ 3% (50 pb-1); 0.8% (1 fb-1)
May 26, 2008 Maarten Boonekamp, CEA-Saclay 13
![Page 14: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/14.jpg)
3 : Acceptance
Total Z cross-section : which fraction of events lies within the detector acceptance?
Two factors : PDF( Z ) and PDF( e,μ | Z )
First factor : dσ/dy, dσ/dpT, related to proton PDFs and parton showersNot well knownNot well known
Second factor : angular distributions and QED/EW radiation in Z rest frame.
ll di d i f h l ( C@ O hWell predicted using state of the art tools (MC@NLO+Photos, ResBos, Horace, Winhac/Zinhac…)
May 26, 2008 Maarten Boonekamp, CEA-Saclay 14
![Page 15: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/15.jpg)
Acceptance
arXiv:0712.1199arXiv:0805.2093
Proton PDF induced uncertainty dA/A ~ 1%QCD higher orders and resummation contributes dA/A ~ 3%
Our ATLAS study; also CMS note 2006/082; Mangano, Frixione, 2004 (W production); Adam, Halyo, Yost, 2008 (Z production)
May 26, 2008 Maarten Boonekamp, CEA-Saclay 15
![Page 16: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/16.jpg)
Summary, so far
Z total cross-section:dL/L 10% <3%dL/L ~ 10% <3%dε/ε ~ 3% <1%dA/A ~ 3% irreducible at this stage
Acceptance uncertainties will play a dominant role, especially when measuring cross-section ratios where L cancels
arXiv:0802.3251
Z as luminosity monitor : account for overall normalization uncertainty~5% : this is, at best, a temporary hackhack
May 26, 2008 Maarten Boonekamp, CEA-Saclay 16
![Page 17: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/17.jpg)
Differential cross-sections
Total cross-section measurements are thus limited by the very effects we want to constrain! Differential cross sections provide more insightwant to constrain! Differential cross-sections provide more insight -acceptance uncertainties small (cf slide 14)
ATLAS CSC
~200 pb-1
May 26, 2008 Maarten Boonekamp, CEA-Saclay 17
200 pb
![Page 18: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/18.jpg)
Differential cross-sections
Total cross-section measurements are limited by the very effects we want to constrain Differential cross sections provide better constraintsto constrain. Differential cross-sections provide better constraints -acceptance uncertainties small (cf slide 14)
10 fb 1
May 26, 2008 Maarten Boonekamp, CEA-Saclay 18
~10 fb-1
![Page 19: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/19.jpg)
Ratios :cross-normalizing experimentcross-normalizing experiment
Measure R = σ / σ f : )/()/(0 REFREFREF AAdddNdNdR ⊕⊕⊕⊕= εεMeasure R σ / σref : )/()/(
0REFREFREF AANNR
⊕⊕⊕⊕=εε
Statistical terms No lumi term! Additional terms from REF
So careful : the interest of this is not always obvious!Gain : no luminosity dependence
Statistical terms No lumi term! Additional terms from REF
But additional terms from εREF and AREF
Might be good (if one expects correlated ε ~ εREF and A ~ AREF) : even more cancelation;or bad (if uncorrelated) : larger uncertainty
Conversely : when possible, define R keeping this in mind, i.e maximize correlation with REF
May 26, 2008 Maarten Boonekamp, CEA-Saclay 19
![Page 20: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/20.jpg)
Ratios (2)
Random example : σtt
ATLAS CSC
Th ti t Z d ti / k littlThe ratio to Z production, σtt/σZ, makes little senseCancels out L indeedAll other systematics are essentially independent; also add Z rate uncertaintyhence a worse result
May 26, 2008 Maarten Boonekamp, CEA-Saclay 20
![Page 21: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/21.jpg)
Ratios (3)
G ld l /
arXiv:0802.0007
Golden example : σW / σZ
Very similar experimentally isolated leptons, same pT rangeCan be selected using same trigger(difference : EtMiss)
Quark initial state; singlet final statesimilar QCD corrections
Behave similarly under PDF variationsBehave similarly under PDF variations
In σW / σZ, almost everything cancelsHence a beautiful test of QCD
May 26, 2008 Maarten Boonekamp, CEA-Saclay 21
![Page 22: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/22.jpg)
Ratios : cross-normalizing theorycross-normalizing theory
Data-driven predictions : ( )measREF
predREFpred σ
σσσ ⎟
⎠⎞
⎜⎝⎛=
Poor prediction Precise prediction Measurement
σpred can then be :compared against σmeas : e.g search for, or interpretation of new physicsUsed as input for precision measurements
May 26, 2008 Maarten Boonekamp, CEA-Saclay 22
![Page 23: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/23.jpg)
Data-driven predictions (1)
Example : W mass. Need to predict W distributions (not rates), e.g dσW/dy
Define : dyd
dyddyddyd Z
Z
WW /
/// σ
σσσ ×→
Use RMS of rapidity distribution, ryW,Z, to quantify dσ/dy and their variations
(choice not unique):
Raw prediction MeasuredPrecise prediction
arXiv:0805.2093
predict
measure
May 26, 2008 Maarten Boonekamp, CEA-Saclay 23
![Page 24: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/24.jpg)
dσW/dy
Spread of R :
41 CTEQ 6.1 PDF sets
dR
Ratio prediction ~20x more precise than raw
R = ryW / ry
Z
May 26, 2008 Maarten Boonekamp, CEA-Saclay 24
p p
![Page 25: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/25.jpg)
dσW/dy
Careful : precise but incompatible predictions!
CTEQ 6.1CTEQ 6.5
MRST 2006
CTEQ 6.6
Studied sets agree on correlations, not on central values – different starting assumptions and theoretical frameworks
R = ryW / ry
Z
May 26, 2008 Maarten Boonekamp, CEA-Saclay 25
different starting assumptions and theoretical frameworks
![Page 26: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/26.jpg)
Data-driven predictions (2)
Example : High-mass DY. Motivation:
May 26, 2008 Maarten Boonekamp, CEA-Saclay 26
![Page 27: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/27.jpg)
High-mass Drell-Yan
Current LHC uncertainty : ~6-7% for 100 GeV < M < 1 TeV and y~0
Gain a factor ~5. To do this, relate:
Specifically, write: )0()0()0( 2 ≠=×= yMymyM σσσ)0,(
)0,()0,(
)0,()0,()0,( 2 =≠×
≠=×=→=
ymyM
yMymyMyM Z
Z σσ
σσσσ
Smaller PDF dependence? MeasuredRaw prediction
chosing m, M and y such that m = MZ e-y ; M = MZ e+y
Work with Florent chevallier, in preparation
May 26, 2008 Maarten Boonekamp, CEA-Saclay 27
Work with Florent chevallier, in preparation
![Page 28: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/28.jpg)
High-mass Drell-Yan
dσ/σ (y=0)Raw (CTEQ61)
From ratio
M (GeV)
80 203040506070 10m (GeV)
0 1 2 2yZ
0 80 4 1.0 1.5
May 26, 2008 Maarten Boonekamp, CEA-Saclay 28
0.1 2.20.80.4 1.0 1.5
![Page 29: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/29.jpg)
High-mass Drell-Yan
Measured quantities:d /d (Z) l d h t h ( )dσ/dy (Z) already shown too much ( )
dσ/dm at low mass:
ATLAS CSCATLAS CSC
May 26, 2008 Maarten Boonekamp, CEA-Saclay 29
![Page 30: Luminosity measurements at ATLAS and impact of PDF](https://reader031.vdocuments.site/reader031/viewer/2022020701/61f7751eeadaa13419728a94/html5/thumbnails/30.jpg)
Summary & Conclusions
Cross-section measurementsC l t h ll i tComplete program : a challenge in every aspect
dL/L : luminosity program well underwayEfficiency, scale, resolution : many auxiliary measurementsNeed to measure distributions to minimize acceptance effects
Ratios : a possible simplification (normalization or data-driven predictions)Ratios : a possible simplification (normalization, or data driven predictions)Need to be defined carefully : eliminating L can easily introduce other, possibly larger sources of uncertaintyA good reference process should be correlated theoretically and experimentally to the target. And SM-certified
SM cross-sections : not just background control
LHC physics and PDFs : Intrinsically tiedp y y
PDF uncertainty sets : a great toolMost important application : more than error estimation, investigation of correlations among different physics processes
May 26, 2008 Maarten Boonekamp, CEA-Saclay 30
correlations among different physics processes