benchmark qcd measurements and tools at atlas

CTEQ07 Michigan May 2007

Benchmark QCD Measurements and Tools at ATLAS

Craig Buttar

University of Glasgow


Outline

• Going from the Tevatron to the LHC– Pdfs

– Total xsect vs 2 2 xsect

– Multijets sudakov form factors

• These considerations define measurements that are needed to define the hadronic environment at the LHC– Measurements of PDFs with inclusive jet xsect

– Underlying event and minimum bias

– Multijet rates

• Implications for physics– PDFs identifying new physics

– Multijets MET?

– Underlying event Higgs; different processes


Tevatron LHC

Q2

(GeV

)

tot


Low pt physics


Minimum bias and Underlying Event: LHC predictions

Tevatron

● CDF 1.8 TeV

PYTHIA6.214 - tuned PYTHIA predictions

dN/dη (η=0)

Nch jet-pt=20GeV

1.8TeV (pp) 4.1 2.3

14TeV (pp) 7.0 7.0

increase ~x1.8 ~x3

~80%~200%

LHC prediction

Tevatron

PYTHIA6.214 - tuned

● CDF 1.8 TeV

LHC

MB onlyUE includes radiation and small impact parameter bias

dN/d in minimum bias events

Minimum bias = inelastic pp interactionUnderlying event = hadronic environment not part of the hard scatter

Pt leading jet (GeV)

Particle

density


The underlying eventTra

nsvers

e <

Nch

g >

PYTHIA6.214 - tuned

PHOJET1.12

x 3

LHC

x1.5

Extrapolation of UE to LHC is unknownDepends on• Multiple interactions• Radiation• PDFs• Striing formation

High PT scatter

Beam remnants

ISR

• Lepton isolation • Top• Jet energy• VBF


Ra

tio

<N

Tra

ckR

eco>

/<N

Tra

ckM

C>

Leading jet ET (GeV)

Reconstructing the underlying event

CDF Run 1 underlying event analysisPhys. Rev. D, 65 092002 (2002)

Njets > 1, |ηjet| < 2.5, ETjet >10 GeV,

|ηtrack | < 2.5, pTtrack > 1.0 GeV/c


Underlying event for different processes

• The underlying event for electroweak processes needs to be studied– Critical for Higgs search in

VBF

Charged Particle Density: dN/dd

0.1

1.0

10.0

0 30 60 90 120 150 180 210 240 270 300 330 360

(degrees)C

ha

rge

d P

art

icle

De

ns

ity

Leading Jet

Leading Photon

Z-boson

RDF PreliminaryPYTHIA Tune A

Charged Particles (||<1.0, PT>0.5 GeV/c)

"Transverse" Region

Jet#1PhotonZ-boson

Jet #1 Direction

“Toward”

“Transverse” “Transverse”

“Away”

Photon #1 Direction

“Toward”


“Away”

Z-boson Direction

“Toward”


“Away”

R.Field


Resolving hard and soft components

Jet #1 Direction

“Toward”

“TransMAX” “TransMIN”

“Away”

Jet #1 Direction

“Toward”

“TransMAX” “TransMIN”

Jet #2 Direction

“Away”

“Leading Jet”

“Back-to-Back”

•TransMAX and transMIN sensitive to radiation and soft UE respectively •Back-to-back sample suppresses radiation• difference between tranMAX region and transMIN in leading jet and b-2-b jet sample


Simulation of underlying event

• MC tools for simulation of underlying event– PYTHIA (UE+min bias)

– Herwig + Jimmy (UE only, pt-cut)

– PHOJET (Min bias and UE)

• All give a reasonable description of Tevatron data with tuning

• Energy extrapolation is essentially a free parameter and uncertain data required

• SHERPA also has simulation of underlying event but has been studied less


PDFs


Impact of PDF uncertainty on new physics

Similarly PDF uncertainties limits the sensitivity in inclusive xsect to BSM physics

Extra-dimensions affect the di-jet cross section through the running of s. Parameterised by number of extra dimensions D and compactification scale Mc.

PDF uncertainties (mainly due to high-x gluon) reduce sensitivity to compactification scale from ~5 TeV to 2 TeV

2XD

4XD

6XD

SM

Mc= 2 TeV

uncertainties

PDF

Mc= 2 TeV Mc= 6 TeV


Measure high-x gluon pdfs from inclusive jet cross-section

• Measure inclusive xsect to get high-x gluons

• Measure in different rapidity bins– New physics vs pdf

• Theoretical uncertainties in QCD calculation– Scale dependence

• Experimental errors– Jet energy scale


Uncertainty due to PDF

High pT PDF errors dominated by the high x-gluon. Estimates below from CTEQ6.1 error sets 29 and 30 compared to best fit (NLOJET).

At 1TeV in central region error is 10-15%


Theoretical Errors•NLO QCD cross-sections can be calculated to compare with the experimental results.

•There are errors on the theoretical prediction due to PDFs and the finite order of the calculation (renormalisation and factorisation scales).

5%-10% scale error.

From changing scale µr=µf from 0.5pT jet to 2.0pT jet


Experimental Errors

For a jet with pT=1TeV:

1% error on jet energy -> 6% on 5% error on jet energy -> 30% on

10% error on jet energy -> 70% on

Used NLOJET.

10% JES

5% JES

1% JES


Analysis – Constraining High x-Gluon

•Effect of adding simulated ATLAS collider data to gluon uncertainty in a global PDF fit (NLOGRID) Fits by Claire Gwenlan:

x x

Glu

on u

ncer

tain

ty

Reducing JES from 3% to 1%

•A very good control (1%) of the Jet Energy Scale is needed in order to constrain PDFs using collider data.


At the LHC we will have dominantly sea-sea parton interactions at low-x And at Q2~M2

W/Z the sea is driven by the gluon by the flavour blind g ->qq gluon is far less precisely determined for all x values

_

The uncertainties on the W and Z rapidity distributions are dominated by the uncertainties along gluon PDF dominated eigenvectors

Measurement of W and Z rapidity distributions can increase our knowledge of the gluon PDF key to using W,Z as luminosity monitor

Improving low-x gluon using rapidity distribution in W-decay

2.5


At y=0 the total W PDF uncertainty is ~ ±5.2% from ZEUS-S~ ±3.6% from MRST01E~ ±8.7% from CTEQ6.1MZEUS to MRST01 central value difference ~5%ZEUS to CTEQ6.1 central value difference ~3.5% (From LHAPDF eigenvectors)

W and Z Rapidity Distributions for different PDFs

CTEQ6.1M MRST02 ZEUS-S

Analytic calculations: Error bands are the full PDF Uncertainties

GOAL: syst. exp. error ~3-5%

CTEQ6.1M


PDF constraining potential of ATLASEffect of including the ATLAS W Rapidity “pseudo-data” in global PDF Fits:

how much can we reduce the PDF errors when LHC is up and running?Simulate real experimental conditions:

Generate 1M “data” sample with CTEQ6.1 PDF through ATLFAST detector simulation and then include this pseudo-data (with imposed 4% error) in the global ZEUS PDF fit (with Det.->Gen. level correction).Central value of ZEUS-PDF prediction shifts and uncertainty is reduced:

ZEUS-PDF BEFORE including W data

e+ CTEQ6.1 pseudo-data

low-x gluon shape parameter λ, xg(x) ~ x –λ :BEFORE λ = -0.199 ± 0.046

AFTER λ = -0.181 ± 0.030 35% error reduction

NB: in ZEUS-PDF fit the e± Normalisation is left free => no assumption on Luminosity measurement

ZEUS-PDF AFTER including W data

e+ CTEQ6.1 pseudo-data


Multijets


Azimuthal dijet decorrelation

Early measurement to benchmark generators particularly parton showers/higher orders

2 dijet

dijet 2

dijet=

dijet~ 2


Reconstructed di-jet azimuthal decorrelations

Selecting di-jet events:

300 < ETMAX < 600 GeV

Cone jet algorithm (R=0.7)Njets = 2, |ηjet| < 0.5, ET

jet #2 > 80 GeV,

Two analysis regions:

600 < ETMAX < 1200 GeV

J5J5

J6J6


Multijets in top events

• MC@NLO and ALPGEN agree for hardest jet

• HERWIG fails at high pt • Significant number of events

have 3 additional jets there is a discrepency between MC@NLO and HERWIG vs ALPGEN

• Key that such multijet spectra are measured

• Possible with early high energy running

Spectra include tt


Effect of multijets on inclusive SUSY studies


Summary and conclusions

• QCD benchmarks (inc low-pt processes)– Minimum bias and underlying event

• Related, need to understand for many analyses

• Baseline for HI, minijets evolution from low pt

• Do this before high lumi

– PDFs• New regime in PDFs

• Need to measure for precision SM and high-pt physics

– Multijets• Measure azimuthal decorrelations to validate simulations

• Many more jets in events tt6J+nJ

• Need to understand multiplicities


Tracking in MB events

• Acceptance limited in rapidity and pt

• Rapidity coverage– Tracking covers ||<2.5

• pT problem

– Need to extrapolate by ~x2 Need to understand low pt

charge track reconstruction

1000 events1000 events

dNdNchch/d/d

dNdNchch/dp/dpTT

Black = Generated (Pythia6.2)

Blue = TrkTrack: iPatRec

Red = TrkTrack: xKalman

Reconstruct tracks with:Reconstruct tracks with: 1) pT>500MeV1) pT>500MeV 2) |d2) |d00| < 1mm| < 1mm 3) # B-layer hits >= 13) # B-layer hits >= 1 4) # precision hits >= 84) # precision hits >= 8

pT (MeV)


Minimum bias studies: Charged particle density at = 0

LHC?

Large uncertainties in predicted particle density in minimum bias events ~x2

Measurement with central tracker at level of ~10% with ~10k events – first data

Why? soft physics, pile-up at higher luminosities, calibration of experiment


Compare abrupt and smooth pt-cut-off:Abrupt cut-off generates a Poisson distribution with too few multi-parton interactions in a single event

Compare matter distributions:uniform, gaussian, double gaussian Use double gaussian

number of interactionsn

zn

Use MB multiplicity distributions to tune fluctuations in number of events

abrupt

smooth

Uniform

gaussian

d-gaussian


Herwig+Jimmy

• Jimmy is multi-parton interaction model similar to PYTHIA

• Main parameter is pt-min• Only for hard UE cannot

model low-pt ie MB difficult to get energy dependence

• Matter distribution is determined from em form factor

• Gives a good description of CDF data with increased pt-min2.53.25GeV


Herwig+Jimmy

• Extrapolation to LHC predicts much larger UE c.f. underlying event

• No energy dependent pt-min

• Which is correct?T

ran

sver

se <

Nch

g >

Pt (leading jet in GeV)TevatronTevatron

LHCLHC

x3

x5

x4


VBF Signal (HWWll)

•forward tagging jets

•correlated isolated leptons

• low hadronic activity in central region

•central Higgs production

Tagging jet

Tagging jet

H

W

WZ/WZ/W

Important discovery channelFor Higgs in mass range120-200GeV


Model Parameter

Simple MSTP(82)=1

PARP(81)=1.9

Complex MSTP(82)=4

PARP(82)=1.9

Tuned MSTP(82)=4

PARP(82)=1.8

PARP(84)=0.5

Model CJV (eff) LEPACC (eff) All Vetoes (eff)

Simple 0.943 ± 0.003 0.610 ±0.001 0.084 ±0.001

Complex 0.885 ± 0.003 0.575 ±0.001 0.076 ±0.001

Tuned 0.915 ± 0.003 0.589 ±0.001 0.080 ±0.001

Uncertainty at the level of ~6% on CJV Pt>20GeV and ~3% on leptonGiving a total uncertaintly in the range ~8%


• Effect of UE on lepton efficiency• Vary pt-min by 3• Determine from data

• Good muons– Barrel <1.1 Pt>7GeV– Endcap 1.1<<2.4 P>9GeV

• Isolation Pt for charged tracks excluding s

with Pt>0.8GeV and R<0.3 around in - space

S.Abdullin et al (CMS) Les Houches 05

Lepton isolation in H->4

Process Event eff Default Event eff –3 Event eff +3

H4 MH=150GeV 0.775 ± 0.004 0.707 ± 0.005 0.812 ± 0.004

ZZ background 0.780 ± 0.004 0.721 ± 0.005 0.838 ± 0.004

4 random Z-inclusive 0.762 ± 0.007 0.706 ± 0.007 0.821 ±0.006


Extract effect of UE from data

• Use inclusive Z-sample, high statistics• Similar dependence to ZZ sample but small systematic shift

random

UE

benchmark qcd measurements and tools at atlas

Documents

underlying event cdf

underlying event analysisphys

gevcunderlying event

nch jetpt

jet cross section

high xgluon

highx gluonsmeasure

tevmeasure highx gluon