Документ взят из кэша поисковой машины. Адрес оригинального документа : http://nuclphys.sinp.msu.ru/conf/lpp14/240809/Ilina.pdf Дата изменения: Wed Sep 16 16:32:53 2009 Дата индексирования: Tue Oct 2 00:39:37 2012 Кодировка:

Study of jet transverse structure with CMS experiment at 10 TeV

Natalia Ilina (ITEP, Moscow) for the CMS collaboration CMS PAS QCD-08-002 (2009)

1

LOMONOSOV09, MSU, Moscow

24/08/2009


Outline
1. 2. 3. 4. 5. 6. 7. 8. 9. Introduction CMS jet transverse structure variables Monte Carlo predictions CMS experiment Jet transverse structure reconstructed from calorimetric data Jet transverse structure measured with charged particles Predictions of different generators: PYTHIA and HERWIG++ Estimation of the quark jet fraction Conclusions

2

LOMONOSOV09, MSU, Moscow

24/08/2009


Introduction

3

LOMONOSOV09, MSU, Moscow

24/08/2009


CMS jet transverse structure variables
We study jet transverse structure with the second moment of jet profile in transverse momentum:
< R
2 jet

> ( pT ) =

i jet



R 2 (i, jet ) pT
i jet

i



i pT

and
<
2 jet

> ( pT ) =

i jet



(

jet

- i ) 2 pT
i pT

i

i jet



<

2 jet

> ( pT ) =

i jet



(

jet

- i ) 2 pT
i pT

i

i jet



In this way:

< R2 >=<2 > + < 2 >

Summation is over: stable particles or calorimeter towers or reconstructed tracks
4 LOMONOSOV09, MSU, Moscow 24/08/2009


CMS Jet transverse structure variables
Because of the angular smearing of the calorimeter jet relative to the generator jet, < R 2 > is calculated with respect to different axes. jet To avoid the impact of jet angular resolution one can use 2nd central moments of and Instead of 2nd moments:

< c2 >=< 2 > - < > < c2 >=< 2 > - < >

2 2

And the trace of - correlation matrix:

< Rc2 >=< c2 > + < c2 >

5

LOMONOSOV09, MSU, Moscow

24/08/2009


Monte Carlo predictions
HERWIG++ HERWIG++

Both HERWIG++ and PYTHIA generators: R2 depends on pT MC predicts no dependence of R2 on both gluon and quark jets 2 <2 The same behaviour for Rc > and< R >
6 LOMONOSOV09, MSU, Moscow 24/08/2009


Monte Carlo predictions
ALL JETS GLUON jets QUARK jets

Particle level All jets

The difference in the prediction for jet shapes between PYTHIA and HERWIG++ is due to the different models for gluon jet hadronization Both HERWIG++ and PYTHIA predict nearly the same q-jets fraction for the same jet pT and .

7

LOMONOSOV09, MSU, Moscow

24/08/2009


Monte Carlo predictions
HERWIG++ Particle level

Jet transverse structure for all jets (black lines) depends on while it does not depend on for gluon and quark jets. This is due to different quark/gluon jets fraction in the rapidity regions
8 LOMONOSOV09, MSU, Moscow 24/08/2009


Jet transverse structure with CMS experiment
Weight 14000t Diameter 15m Length 21.6m Magnetic field 3.8T Tracker: Silicon Microstrips:
10 barrel layers, 3+9 fw disks, tot 200m2

Pixels:
3 barrel layers, 2 fw disks

Calorimeters: ECAL barrel&endcap: Scintillating PbWO4 crystals HCAL Plastic scintillator
9 LOMONOSOV09, MSU, Moscow 24/08/2009


Jet transverse structure reconstructed from calorimetric data
R2cal ­ jet transverse structure calculated from towers formed by ECAL+HCAL calorimeter Sources of detector related bias: · the strong magnetic field of the CMS solenoid · non-linear and non-uniform calorimeter response to hadrons · hadron shower and calorimeter tower transverse sizes

Iterative Cone algorithm (R=0.5) - for jets reconstruction JPT - JetPlusTrack algorithm for jet energy calculation (CMS PAS JME-09-002 (2009))
10 LOMONOSOV09, MSU, Moscow 24/08/2009


Jet transverse structure measured with charged particles
To avoid the magnetic field bias and bias due to calorimeter energy reconstruction we analyze the charged component shape of jets R2tr ­ jet transverse structure calculated from tracks (reconstructed in CMS tracker) R2ch ­ jet transverse structure calculated from charged particles (generated by MC generator) We restrict our study to || < 1, where the fake rate of the reconstructed tracks will be less than 2% pT > 1 GeV, where the track reconstruction efficiency does not depend on track momentum
11 LOMONOSOV09, MSU, Moscow 24/08/2009

tracks

Jet axis


Predictions of different generators: PYTHIA and HERWIG++
The comparison of generator level and reco objects for HERWIG++ and PYTHIA: tr(PT corr) and ch (PT
gen)

The average value of the biases was calculated and applied like correction to "data" points Half of the difference was incorporated in the systematics
12 LOMONOSOV09, MSU, Moscow 24/08/2009


Predictions of different generators: PYTHIA and HERWIG++
Systematics: JES: 4-6% for pT>75 GeV, varying jet pT (CMS PAS JME-09-004 (2009)) Angular resolution: much smaller than other uncertainties Tracker pT resolution: varying track pT cut by 1% (CMS PAS TRK-09-001 (2009)) Simulation bias correction: half of the difference between HERWIG++ and PYTHIA bias corrections Stat.errors: corresponds to 10 pb-1 (but data points were produced using samples with more statistics) The data sample for 10 pb-1 will allow to distinguish between predictions of different MC generators
13 LOMONOSOV09, MSU, Moscow 24/08/2009


Estimation of quark jet fraction

Stat.errors: corresponds to 10 pb-1 (but data points were produced using samples with more statistics) After achieving agreement between MC tuned predictions for vs p it can be compared with the result of simulations for quark and gluon jets
14 LOMONOSOV09, MSU, Moscow

jet T

24/08/2009


Conclusions
Jet transverse structure is sensitive to the event generator and can also be used as a quantitative analysis of the fraction of quark and gluon jets Using transverse structure measurements based on the instead of the calorimeter reduces measurement bias tracker

10 pb-1 of data collected with the CMS experiment is sufficient to distinguish between different Monte Carlo generators Pythia and Herwig++ predict no dependence of R2 on for both gluon and quark jets. The dependence of R2 for all jets is due to the difference in the quark and gluon fraction as a function of After tuning the Monte Carlo to data, the quark jet fraction can be extracted
15 LOMONOSOV09, MSU, Moscow 24/08/2009


Back-up

16

LOMONOSOV09, MSU, Moscow

24/08/2009


Tevatron results
CDF results: Phys.Rev. D71, 112002 (2005);

17

LOMONOSOV09, MSU, Moscow

24/08/2009


MC samples
Summer08 MC samples: PYTHIA6.4 tune D6T and HERWIG++2.2 10 pb
-1

CMSSW221 for analysis standard CMS single jet triggers (1E31 trigger menu) IC algo (R=0.5) for jets reconstruction JEC - L2L3 calo corrections (CMS PAS JME-07-002 (2007)) and JPT - JetPlusTrack algo for jet energy (CMS PAS JME-09-002 (2009))
18 LOMONOSOV09, MSU, Moscow 24/08/2009