First measurement of the inclusive jet cross section in p+p collisions at E CM =200 GeV
description
Transcript of First measurement of the inclusive jet cross section in p+p collisions at E CM =200 GeV
Mike Miller
First measurement of the inclusive jet cross section in p+p collisions at ECM=200 GeV
M.L. Miller (MIT)
Motivations:
1. Baseline for inclusive jet ALL
2. Constraining large-x pdf’s
Outline:
1. What do we mean by “jet”?
2. Presentation of work in progress
3. List of short- and long-term tasks towards final results
Mike Miller, for the STAR collaboration
parto
n
parti
cle
etcpe
,,,
gq,
dete
ctor
What do we mean by “jet”?
Mike Miller, for the STAR collaboration
parto
n
parti
cle
etcpe
,,,
gq,
dete
ctor
DataData
What do we mean by “jet”?
*(hep-ex/0005012)
midpoint-cone algorithm*•Search over “all” possible seeds for stable groupings
•Check midpoints between jet-jet pairs for stable groupings
•Split/merge jets based on Eoverlap
•Add track/tower 4-momenta
Mike Miller, for the STAR collaboration
parto
n
parti
cle
dete
ctor
etcpe
,,,
gq,
midpoint-cone algorithm*•Search over “all” possible seeds for stable groupings
•Check midpoints between jet-jet pairs for stable groupings
•Split/merge jets based on Eoverlap
•Add track/tower 4-momenta
DataData SimulationSimulation
GEANT
“geant” jets
“pythia” je
ts
Correction philosophy•Use Pythia+GEANT to quantify detector response
•Estimate corrections to go from “detector” to the “particle” level
pyth
ia
What do we mean by “jet”?
*(hep-ex/0005012)
Mike Miller, for the STAR collaboration
pT (track/tower) > 0.2 GeV|vertex-z|<60 cmrcone=0.4, fmerge=0.50.2< ηjet <0.8ratio<0.9(and ET-trig>3.5 GeV/c for HT)
High Tower TriggerBBC coincidence + one tower above threshold (not the best jet trigger!)
εTrig(η=0.0)=1: ET-tower= 2.5 GeVεTrig(η=0.8)=1: ET-tower= 3.3 GeV
TPC0<φ<2π1< η<1BEMC0<φ<2π0< η<1in situ MIP+e calibration
2004 p+p data set2004 p+p run
Sampled luminosity: ~0.16 pb-1
1.4 M High tower (HT) events0.8 M highly pre-scaled minimum bias (MB) events ~220k pT>5 GeV jets in HT sample before cutsCommissioned 1x1 jet patch trigger– main jet trigger in 2005+
Mike Miller, for the STAR collaboration
2004 p+p runSampled luminosity: ~0.16 pb-1
1.4 M High tower (HT) events0.8 M highly pre-scaled minimum bias (MB) events ~220k pT>5 GeV jets in HT sample before cutsCommissioned 1x1 jet patch trigger– main jet trigger in 2005+
pT (track/tower) > 0.2 GeV|vertex-z|<60 cmrcone=0.4, fmerge=0.50.2< ηjet <0.8ratio<0.9(and ET-trig>3.5 GeV/c for HT)
High Tower TriggerBBC coincidence + one tower above threshold (not the best jet trigger!)
εTrig(η=0.0)=1: ET-tower= 2.5 GeVεTrig(η=0.8)=1: ET-tower= 3.3 GeV
2004 p+p data set
TPC0<φ<2π1< η<1BEMC0<φ<2π0< η<1in situ MIP+e calibration
40 200 .GeV, rs
Mike Miller, for the STAR collaboration
Background Removal
jetT
EMCT
EE
f
:fractionenergy neutral
Datapythia+geant
Cut to remove background
Systematic uncertainty on yield from background less than ~5%
Mike Miller, for the STAR collaboration
Consider detector response (measured) to known input (true)
trueT
measT
trueT
pppresolution :
Simulation: ~25% ± 5% for 10<pT<50 GeV/c
Consistent number derived from (modest) sample of di-jet events
“geant jets” “pythia jets”
Response & bin quality factors
Simulation
Mike Miller, for the STAR collaboration
Consider detector response (measured) to known input (true)
“geant jets” “pythia jets”
generated
measured
NNpurity
accepted
:
jets all Of
Response & bin quality factors
Simulation
Simulation
Mike Miller, for the STAR collaboration
Consider detector response (measured) to known input (true)
“geant jets” “pythia jets”
Bins of width 1-sigma “purity” of ~35% over range on the diagonal.
Motivates application of bin-by-bin correction factors.
Response & bin quality factors
Simulation
Simulation
generated
measured
NNpurity
accepted
:
jets all Of
Mike Miller, for the STAR collaboration
TacceptedMB-events dp
dNN
1
:yieldevent -per raw
Reasonable agreement between data/MC10 weighted pythia samples Slope difference:
Short term: incorporated into sys. uncertaintyLong term: iterative re-weighting of MC sample
Allows data/MC comparison before luminosity
For HT, include trigger pre-scale:
trigMB
acceptedMB-events
acceptedMB-events psNN
11
Uncorrected per-event yield
4.0 , 200 rGeVs
Mike Miller, for the STAR collaboration
pythia
Tpythia
geantTgeant
T pNpM
pc
MB data statistics limited
More advanced unfolding techniques currently under study
εjet : decreases c(pT)
resolution: increases c(pT)
εTrig: ~1 e-2 at pT-jet = 5 GeV ~1 at pT-jet = 50 GeV
Bin-by-bin correction factors“measured”
“true”
Simulation
resolutionpc jettriggerT :
Mike Miller, for the STAR collaboration
Only data-statistical errors shownSystematics and ratios on next slides
Fortran code from hep-ph/0404057 (Jager et al.)rcone=0.4CTEQ 6.1μF= μR =pT
Agrees with EKS NLO to better than 1%
Towards a “corrected” cross section
TTTT dpdN
pcdtLpdpd
d
111
6.0211
±30% but invisible on this scaleMB/HT overlap for 3 bins
acceptedeventsMB
MBvertBBC
NdtL
1 ~60%
Under study
4.0 , 200 rGeVs
Mike Miller, for the STAR collaboration
* Under Study
* Under Study
* Under Study
•Dominant uncertainty: 10% change in ECal ~40% change in yield•More under long term study (see last slide)
Dominant systematic uncertainty estimatesNormalization
Pythia slope Statistics of c(pt) Background
BBC TriggerEnergy ScaleFr
actio
nal C
hang
e in
x-s
ectio
n
Mike Miller, for the STAR collaboration
“Corrected” data vs. NLO calculation
GeVs 200
40 , 200 .rGeVs
•50% systematic uncertainty from E-scale shown•Agreement (within systematics) over 7 orders of magnitude
Mike Miller, for the STAR collaboration
“Corrected” data vs. NLO calculationThings that could change this curve:1. Refined BEMC calibration (pi0
calibration)2. Improved correction for lost (no
vertex) events (MC embedded in abort-gap data)
3. Improved unfolding scheme
Issues that won’t go away for 2004:1. HT trigger: fragmentation bias,
slow turn on with pT-jet 2. Jet energy scale (need integrated
luminosity for photon-jet!)
GeVs 200
4.0 , 200 rGeVs
•50% systematic uncertainty from E-scale shown•Agreement (within systematics) over 7 orders of magnitude
Mike Miller, for the STAR collaboration
• Much work to develop methods and techniques for jet triggering, reconstruction, and analyses
• First look: – Significant pT reach (~50 GeV/c)– Agreement within large systematics with NLO calculations
• A few primary issues under study:– Refining in situ EMC calibrations (AuAu vs. CuCu vs. pp)– Overall scale corrections for no-vertex events– Improved unfolding methods– Fragmentation bias (e.g., Pythia vs. Herwig)– NLO clustering scheme
• Goal:– Bring 2004 (and 2003) analyses to efficient closure– Large 2005 data set produced, thesis analyses underway
• Long term cross section goals– Jet shapes, NLO comparisons– High stats possible improvement of large-x pdf’s
Summary and outlook
Mike Miller, for the STAR collaboration
Backup slides
Mike Miller, for the STAR collaboration
Charged Track Momentum p/p TPC 1% Charged Track Inefficiency 10% x 60% jet 6%*
Neutral Tower Energy Gain/Gain = 10% 40% in yield
Neutral Energy Inefficiency ∑ET from n+K0L/S+++ 6-10%*
MIP Subtraction 10% Correction to Jet ET 1%
Fiducial Detector Effects Edges/Dead regions 2% Jet calculation 1 tower = =0.05 3%Background Trigger Background trigger +MINBIAS
event5% in yield
Background Energy Jet + underlying Background 0Underlying Event Work in Progress NAFragmentation Comparison to HERWIG
ongoing NA
* Included in Correction Factor
Jet Energy Scale Systematics
Mike Miller, for the STAR collaboration
2004 BEMC Calibration
Typical tower response to MIP candidate in 2004 AuAu
All towers, ptrack > 1.8 GeV
Relative gain: •tower-by-tower MIP response•~250 +- 50 MeV/mip
Absolute gain:•Eta ring TPC-electrons•2<pelectron<6 GeV/c
Mike Miller, for the STAR collaboration
Theory Systematics
Changing the pdf Changing the scale
Mike Miller, for the STAR collaboration
Data vs. MC
Clear background for ratio>0.9Data/MC agreement improves at higher pt
Mike Miller, for the STAR collaboration
Data vs. MC
Mike Miller, for the STAR collaboration
Data vs. MC
Red pythia + gstarBlack dataAll error bars statistical
High tower triggerpt (jet) >10 GeV/c
Mike Miller, for the STAR collaboration
Data vs. MC
Red pythia + gstarBlack dataAll error bars statistical
High tower triggerpt (jet) >10 GeV/c