Hard QCD Results with Jets @ LHC (ATLAS + CMS) · Total Delivered: 48.1 pb1 Total Recorded: 45.0 pb1 LHC: pp Collisions @ p s = 7 TeV since March 2010 2011 ATLAS/CMS recorded integrated

Hard QCD Results with Jets @ LHC(ATLAS + CMS)

Hadron Collider Physics Symposium 2011 Sven Menke, MPP Munchen, 16. Nov. 2011, Paris, Franceon behalf of the ATLAS and CMS collaborations

� Introduction• ATLAS+CMS

• jetreconstructionand calibration

� Crosssections• inclusive jet

cross sections

• di-jet eventstudies

� Jetproperties• sub structure

• jet mass

� ConclusionsHighest Energetic Jet in ATLAS from 2010 @

√s = 7 TeV with Ej = 3.37 TeV �

http://www.mppmu.mpg.de/~menke

http://www.mppmu.mpg.de/

Introduction � ATLAS and CMS @ the LHC

Day in 2010

24/03 19/05 14/07 08/09 03/11

]1

Tota

l In

tegra

ted L

um

inosity [pb

0

10

20

30

40

50

60

Day in 2010

24/03 19/05 14/07 08/09 03/11

]1

Tota

l In

tegra

ted L

um

inosity [pb

0

10

20

30

40

50

60 = 7 TeVs ATLAS Online Luminosity

LHC Delivered

ATLAS Recorded

1Total Delivered: 48.1 pb1

Total Recorded: 45.0 pb

� LHC: pp Collisions @√

s = 7TeV sinceMarch 2010

� 2011 ATLAS/CMS recorded integratedluminosity @

√s = 7TeV: L = 5.25 fb−1,

and L = 5.22 fb−1, respectively as of30-Oct-2011 (stop of pp-run)

� Results based on ∼ 40 pb−1 collected byeach experiment in 2010 are presentedhere

S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 2


The ATLAS Detector

A Torodial LHC AparatuS: 25m high; 44m long; 7000 t heavy



The CMS Detector

Compact Muon Solenoid: 15m high; 22m long; 12500 t heavy



Kinematics � Tevatron vs. LHC

x710

610

510 410

310 210 110 1

Q (

Ge

V)

1

10

210

310

410

y = 6 4 2 0 +2 +4 +6y = 6 4 2 0 +2 +4 +6

y = 6 4 2 0 +2 +4 +6

y = 6 4 2 0 +2 +4 +6

y)±exp(s

M = 1,2x

Q = M = 14 TeVsLHC: = 7 TeVsLHC:

= 1.96 TeVsTevatron: tM = 2m

x3

10 210 110 1

x f

(x)

210

110

1

10

Te

va

tro

n

LH

C 1

4 T

eV

LH

C 7

Te

V

)=0.118Z

(ms

α=175 GeV, µCTEQ 6.6,

g

u

d

u

dsc

b

� Q vs. x1,2 for Tevatron and LHC (left)

� red curve shows top-pair production as example

� QCD measurements constrain αs and PDF’s (here CTEQ 6.6,right)



Jet Reconstruction in ATLAS and CMS

� Traditionally all calorimeter cells are grouped in fixed size towers with∆φ×∆η = 0.1× 0.1 including all em and hadronic layers per grid point

� ATLAS and CMS use refined reconstruction methods to feed jetalgorithms

� CMS uses Particle Flow (PF)• PF reconstructs leptons, photons, charged and neutral hadrons by

linking tracks to ECAL and HCAL clusters• energy of each particle measured by all subsystems• ECAL and HCAL energy is calibrated to respective particle level

(electromagnetic or hadronic)

� ATLAS uses TopoClusters• topological 3D clustering over all 190 k calorimeter cells• individual energy blobs separated by local maxima correspond to

particle level• in 2010 em-scale clusters were used as input to jets• jet-level corrections correct for non-compensation, PileUp, dead

material



Jet Reconstruction in ATLAS and CMS

� Modern standardized jet algorithms like SISCone (JHEP 0705 (2007) 086), Kt(Nucl. Phys. B 406 (1993) 187, Phys. Rev. D 48 (1993) 3160), and AntiKt (JHEP 0804 (2008) 063)have been evaluated by ATLAS and CMS• these jets are collinear and infrared safe• are available in a standard C++ library (FastJet by Matteo Cacciari, Gavin Salam

and Gregory Soyez)• are seedless and iterative

� AntiKt in inclusive mode was found to be the most useful algorithm forATLAS and CMS• AntiKt combines like Kt pairs of objects based on a

min(

piT

2x, pj

T2x)

scaled distance metric ∆R2ij /R

2 in y − φ-space

• Kt uses x = 1 and treats objects with the smallest pT first• AntiKt uses x = −1 and treats objects with the largest pT first• the net result is that AntiKt-jets are much more regular shaped in

y − φ space and don’t suffer from the “vacuum cleaner” effect like Ktmaking them easier to calibrate

• CMS uses R = 0.5 (incl.) and R = 0.7 (di-jet), while ATLAS usesR = 0.4 and R = 0.6 (both)



Jet Reconstruction in ATLAS and CMSJets, G. Salam (p. 27)

2. Getting the basics right

FastJetJet contours – visualised

G. Salam, 2008



Jet Calibration in ATLAS

� most analyses on 2010 data used simpleEM+JES scheme• topo clusters on EM-scale as input to jet algorithms• MC based correction function f(p⊥, |η|) to restore jet p⊥ to

stable hadron level• top plot shows correction as function of p⊥ for 3 y ranges• middle plot shows systematic uncertainties on jet energy scale

(JES) for 0.3 < |y| < 0.8

� more sophisticated approaches exist• global cell weighting (GCW) applies energy-density

dependent weights to all calorimeter cells• local hadron calibration (LCW) classifies and calibrates topo

clusters as em or hadronic• global sequential calibration (GSW) modifies EM+JES by

jets-shape based correction factors• bottom plot compares light-quark – gluon-jet-response for

different calibration schemes

� LCW is used in ATLAS for missing transverseenergy and in jets for 2011

� Jet mass and substructure studies reportedhere also used this already for 2010 data

[GeV]jet

Tp

20 30 210 210×23

103

10×2

Avera

ge J

ES

corr

ection

1

1.2

1.4

1.6

1.8

2| < 0.8η0.3 < |

| < 2.8η2.1 < |

| < 4.4η3.6 < |

= 0.6, EM+JESR t

Antik

ATLAS Preliminary

[GeV]jet

Tp

30 40 210 210×23

103

10×2

Fra

ctio

na

l JE

S s

yste

ma

tic u

nce

rta

inty

0

0.02

0.04

0.06

0.08

0.1

0.12

ATLAS Preliminary

| < 0.8, Data 2010 + Monte Carlo QCD jetsη | ≤=0.6, EM+JES, 0.3 R t

Antik

ALPGEN + Herwig + Jimmy Noise Thresholds

JES calibration nonclosure PYTHIA Perugia2010

Single particle (calorimeter) Additional dead material

Total JES uncertainty

[GeV]true

Tp

30 40 100 200 300

Lig

ht

qu

ark

G

luo

n J

et

Re

sp

on

se

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1ATLAS Preliminary

SimulationPYTHIA AMBT1

EM+JES CalibrationGCW Calibration

LCW CalibrationGS Calibration

|<2.8η |≤2.1 R=0.6 Jets

tantik



Jet Calibration in CMS

� most analyses use jets with particle flow (PF)objects as input

� jet energy scale uncertainties are evaluated forthree types of jets in CMS• calorimeter only jets• jet-plus-track jets (calorimeter jets with corrections based on

associated tracks)• particle flow jets (all subsystems are used to reconstruct individual

stable particles)

� in all three cases the jet energy is corrected by aMC based correction function in jet p⊥ and η• top plot shows jet energy correction factors for central jets vs. p⊥

� scale and uncertainty are validated with di-jetevents (p⊥ balance for relative JES) and γ-jetevents (MPF and p⊥ balance for absolute JES)• middle plot shows absolute JES uncertainty for particle-flow jets

vs. p⊥• bottom plot shows total JES uncertainty for all three jet types vs. p⊥

for central jets

(GeV)T

Jet p30 100 200 1000

Je

t E

ne

rgy C

orr

ectio

n F

acto

r

0.8

1

1.2

1.4

1.6

1.8

2

2.2

Calorimeter jets

Jet-Plus-Track jets

Particle Flow jets

= 7 TeVs R = 0.5

Tanti-k

= 0.0η

CMS Preliminary

Je

t E

ne

rgy C

orr

ectio

n F

acto

r

Je

t E

ne

rgy C

orr

ectio

n F

acto

r

TT

(GeV)T

p20 30 100 200 1000 2000

Absolu

te s

cale

uncert

ain

ty [%

]

0

1

2

3

4

5

6

7

8

9

10Total uncert.Total MPF

Photon scaleExtrapolationOffset (+1PU)

Residuals

particle flow jets

0.5 & 0.7T

Anti-k

= 7 TeVs-1CMS preliminary, 2.9 pb

(GeV)T

p20 30 100 200 1000 2000

Absolu

te s

cale

uncert

ain

ty [%

]

0

1

2

3

4

5

6

7

8

9

10

(GeV)T

Jet p20 30 100 200 1000

To

tal JE

S U

nce

rta

inty

[%

]

0

5

10

Calorimeter jets

Jet-Plus-Track jets

Particle Flow jets = 7 TeVs

R = 0.5T

anti-k

| = 0.0η|

CMS Preliminary

To

tal JE

S U

nce

rta

inty

[%

]



Cross sections � Inclusive jet cross section

d2σjetdp⊥dy =

Njet∆p⊥∆y

1εL

� measured spectrum is corrected bin-by-bin for migration effects in p⊥• by Gaussian smearing with p⊥ resolution of modified power-law

spectrum obtained by fit to data (CMS)• by correction factors obtained from full detector simulations including

also detector inefficiencies (ATLAS)

� NLO pQCD predictions• NLOJet++ 4.1.2 with CTEQ 6.6 NLO PDFs as baseline (ATLAS)• NLOJet++ 2.0.1 with average of CT10, MSTW2008NLO, NNPDF2.0 NLO

PDFs as baseline (CMS)

� NLO predictions on parton-level are corrected bin-by-bin for non-perteffects due to hadronisation and underlying event• from ratio of leading log generator spectrum with and without

hadronisation and underlying event (ATLAS+CMS)




� corrections applied to data and respective uncertainties

� un-smearing factors as functionof p⊥ for various y

� mainly due to p⊥ resolution andsteeply falling p⊥ spectrum

� efficiency for reconstructing jetsas function of p⊥ for |y | < 0.3(ATLAS)




� corrections applied to NLO pQCD predictions

� non-perturbative corrections forAntiKt-jets with R = 0.5 vs p⊥for various |y | (CMS)

� similar correction factors forAntiKt-jets with R = 0.6 areobserved by ATLAS and found tobe dominated by the underlyingevent

� systematic uncertainties arederived by comparingmulti-parton-interactions andhadronisation for differentgenerators (GeV)

Tp

20 30 100 200N

on-p

ertu

rbat

ive

corr

ectio

ns0.9

1

1.1

1.2

1.3

1.4

(GeV)T

p20 30 100 200

Non

-per

turb

ativ

e co

rrec

tions

0.9

1

1.1

1.2

1.3

1.4

R=0.5 JetsTAnti-k

|y| < 0.5 |y| < 1≤0.5

|y| < 1.5≤1 |y| < 2≤1.5

|y| < 2.5≤2 |y| < 3≤2.5

� for the smaller R = 0.4 the correction is dominated by hadronisation; itsvalue is 0.8 at low p⊥ and approaches unity at large p⊥ (ATLAS)




� results for d2σjetdp⊥dy =

Njet∆p⊥∆y

1εL at

√s = 7TeV with AntiKt-jets

[GeV]T

p

2103

10

[p

b/G

eV

]y

dT

p/d

σ2

d

910

610

310

1

310

610

910

1210

1510

1810

2110

2410

uncertainties

Systematic

Nonpert. corr.

×NLO pQCD (CTEQ 6.6)

)12 10×| < 0.3 (y|

)9

10×| < 0.8 (y0.3 < |

)6

10×| < 1.2 (y0.8 < |

)3

10×| < 2.1 (y1.2 < |

)0

10×| < 2.8 (y2.1 < |

)3

10×| < 3.6 (y2.8 < |

)6

10×| < 4.4 (y3.6 < |

PreliminaryATLAS

=7 TeVs, 1

dt=37 pbL ∫ jets, R=0.4

tantik

[GeV]T

p

2103

10

[p

b/G

eV

]y

dT

p/d

σ2

d

910

610

310

1

310

610

910

1210

1510

1810

2110

2410

[GeV]T

p

2103

10

[p

b/G

eV

]y

dT

p/d

σ2

d

910

610

310

1

310

610

910

1210

1510

1810

2110

2410

uncertainties

Systematic

Nonpert. corr.

×NLO pQCD (CTEQ 6.6)

)12 10×| < 0.3 (y|

)9

10×| < 0.8 (y0.3 < |

)6

10×| < 1.2 (y0.8 < |

)3

10×| < 2.1 (y1.2 < |

)0

10×| < 2.8 (y2.1 < |

)3

10×| < 3.6 (y2.8 < |

)6

10×| < 4.4 (y3.6 < |

PreliminaryATLAS

=7 TeVs, 1

dt=37 pbL ∫ jets, R=0.6

tantik

[GeV]T

p

2103

10

[p

b/G

eV

]y

dT

p/d

σ2

d

910

610

310

1

310

610

910

1210

1510

1810

2110

2410

R = 0.4 (ATLAS) R = 0.5 (CMS) R = 0.6 (ATLAS)

� dominant exp. systematic uncertainties stem from the Jet Energy Scaleuncertainty (ATLAS+CMS) and resolution (CMS)• ∼ 10− 20% (CMS)

• ∼ 10− 30% (ATLAS)

� overall normalisation error of 3.4% (ATLAS, not included above) and 4%(CMS) from luminosity measurements



Inclusive jet cross section � comparison to different PDF sets� data and PDF over prediction (ATLAS, CTEQ 6.6: left); (CMS, CT10:

middle+right); for central (top) and forward rapidities (down)

[GeV]T

p210

310

/d|y

|T

(CT

EQ

6.6

)/dp

0σ

2/d

|y|/d

T/d

pσ

2d 0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2(CL90)PDF∆ jets. R =0.4 (0.0<|y|<0.3)

tantik

Data

CTEQ 6.6

CT10

NNPDF 2.0(100)

NNPDF 2.1(100)

MSTW2008

ATLAS Preliminary

0

0.5

1

1.5

2

102

103

pT (GeV)

Ra

tio

to

NL

O (

CT

10

) ⊗

NP

CMS L = 34 pb-1

√s = 7 TeVInclusive Jets Anti-kT R=0.50.0 ≤ |y| < 0.5

CMS DataCT10 ⊗ NP, ∆PDF CL68MSTW2008 ⊗ NPNNPDF2.1 ⊗ NP∆(NP ⊕ Scale)

20

0

0.5

1

1.5

2

102

103

pT (GeV)

Ra

tio

to

NL

O (

CT

10

) ⊗

NP

CMS L = 34 pb-1


CMS DataCT10 ⊗ NP, ∆PDF CL68HERAPDF1.0 ⊗ NPABKM09 ⊗ NP∆(NP ⊕ Scale)

20

[GeV]T

p210

/d|y

|T

(CT

EQ

6.6

)/dp

0σ

2/d

|y|/d

T/d

pσ

2d

0.5

1

1.5

2

2.5(CL90)PDF∆ jets. R =0.4 (2.1<|y|<2.8)

tantik

Data

CTEQ 6.6

CT10

NNPDF 2.0(100)

NNPDF 2.1(100)

MSTW2008

ATLAS Preliminary

0

0.5

1

1.5

2

102

103

pT (GeV)

Ra

tio

to

NL

O (

CT

10

) ⊗

NP

CMS L = 34 pb-1


CMS DataCT10 ⊗ NP, ∆PDF CL68MSTW2008 ⊗ NPNNPDF2.1 ⊗ NP∆(NP ⊕ Scale)

20

0

0.5

1

1.5

2

102

103

pT (GeV)

Rati

o t

o N

LO

(C

T10)

⊗ N

P

CMS L = 34 pb-1


CMS DataCT10 ⊗ NP, ∆PDF CL68HERAPDF1.0 ⊗ NPABKM09 ⊗ NP∆(NP ⊕ Scale)

20

� NLO predictions are systematically above the data but withinuncertainties; larger deviations at large |y | and p⊥



Cross sections � di-jet events� double differential cross-section in |y|max and m12� both ATLAS and CMS use bin-by-bin corrections in the observed spectra derived from simulations� deviations from the QCD would indicate new physics

� compositeness, excited quarks� good agreement with NLO QCD predictions is observed

[TeV]12

m

110 110×2 1 2 3 4 5

[p

b/T

eV

]m

ax

|y

d|

12

m/d

σ2

d

210

1

210

410

610

810

1010

1210

1410

1610

1810

2010

2110Systematic uncertainties

Nonpert. corr.×NLO pQCD (CTEQ 6.6)

)8

10× < 2.8 (max|y2.1 < |

)6

10× < 2.1 (max|y1.2 < |

)4 10× < 1.2 (max|y0.8 < |

)2 10× < 0.8 (max|y0.3 < |

)0

10× < 0.3 (max|y |

ATLAS Preliminary

= 0.4R jets, tantik

1 dt = 37 pbL∫ = 7 TeV, s

(TeV)JJ

M0.2 0.3 1 2 3 4

(pb/T

eV

)m

ax

d|y

|JJ

/dM

σ2

d

-310

210

710

1210

1710 < 0.5

max|y|

)1 10× < 1.0 (max

0.5 < |y|

)2 10× < 1.5 (max

1.0 < |y|

)3

10× < 2.0 (max

1.5 < |y|

)4 10× < 2.5 (max

2.0 < |y|

Non Pert. Corr.⊗pQCD at NLO

PDF4LHCave

T = p

Rµ =

Fµ

-1 = 36 pb

intL

= 7 TeVs

R = 0.7T

anti-k

CMS

R = 0.4 (ATLAS) R = 0.7 (CMS)� dominant exp. systematic uncertainties stem from the Jet Energy Scale uncertainty (ATLAS+CMS)

• ∼ 15− 30% (ATLAS)• ∼ 15% at low mass;∼ 60% at high mass (CMS)



Di-jet Cross section � comparison to different PDF sets� data and PDF over prediction (ATLAS, CTEQ

6.6: left); (CMS, CT10: middle+right); forvarious |y|max bins

[TeV]12

m

-210×7 -110 -110×2 -110×3 1 2

Ra

tio

wrt

CT

EQ

6.6

0.5

1

1.5

< 0.3max|y | ATLAS Preliminary

[TeV]12

m

-110×2 -110×3 1 2

Ra

tio

wrt

CT

EQ

6.6

0.5

1

1.5 < 0.8max|y0.3 < |

[TeV]12

m

-110×2 -110×3 1 2

Ra

tio

wrt

CT

EQ

6.6

0.5

1

1.5 < 1.2max|y0.8 < |

[TeV]12

m

-110×3 -110×4 1 2 3

Ra

tio w

rt C

TE

Q 6

.6

0.5

1

1.5 < 2.1

max|y1.2 < |

[TeV]12

m

-110×6 1 2 3

Ra

tio

wrt

CT

EQ

6.6

1

2

< 2.8max|y2.1 < |

= 0.4R jets,t

anti-k

-1 dt = 37 pbL∫ = 7 TeVs

statistical errorData with

uncertainties

Systematic

Non-pert. corr. :

×NLO pQCD

CTEQ 6.6

MSTW 2008

NNPDF 2.1

HERAPDF 1.5

JJ

0

0.5

1

1.5

2

1MJJ (TeV)

Ra

tio

to

NL

O (

CT

10

) ⊗

NP

CMS L = 36 pb-1

√s = 7 TeVDijet Mass Anti-kT R=0.70.0 ≤ |y|max < 0.5

CMS DataCT10 ⊗ NP, ∆PDF CL68MSTW2008 ⊗ NPNNPDF2.1 ⊗ NPHERAPDF1.0 ⊗ NPABKM09 ⊗ NP

0.2 0.5 3

0

0.5

1

1.5

2

1MJJ (TeV)

Ra

tio

to

NL

O (

CT

10

) ⊗

NP

CMS L = 36 pb-1



0.2 0.5 3

0

0.5

1

1.5

2

1MJJ (TeV)

Ra

tio

to

NL

O (

CT

10

) ⊗

NP

CMS L = 36 pb-1



0.4 2 3

0

0.5

1

1.5

2

1MJJ (TeV)

Rati

o t

o N

LO

(C

T10)

⊗ N

P

CMS L = 36 pb-1



0.6 2 4

� HERAPDF closest to data forboth ATLAS and CMS, butothers agree as well withinuncertainties

0

0.5

1

1.5

2

1MJJ (TeV)

Rati

o t

o N

LO

(C

T10)

⊗ N

P

CMS L = 36 pb-1



0.7 2 4



Di-jet Cross section � angular properties� Decorrelation in ∆φ between two most energetic jets is indirect way to

study radiation in di-jet events and to look for new physics

• events with only two high-p⊥ jets should have small decorrelations(∆φ ∼ π)

• events with ∆φ� π indicate more high-p⊥ jets• plot to the right shows ATLAS data compared to NLO pQCD

(NLOJet++ with MSTW2008)

� NLO pQCD calculations agree with data for ∆φ < π

� leading log simulations (Pythia, Herwig, Sherpa) describe the data tooincluding ∆φ = π

[radians]φ∆/2π /3π2 /6π5 π

]1

[ra

dia

ns

φ∆

/dσ

dσ

1/

310

210

110

1

10

210

310

410

510

610

710

810

910

1=36 pbdtL∫Data

)8

10× 800 GeV (>max

Tp

)710× 800 GeV (≤max

Tp<600

)6

10× 600 GeV (≤max

Tp<500

)5

10× 500 GeV (≤max

Tp<400

)410× 400 GeV (≤max

Tp<310

)3

10× 310 GeV (≤max

Tp<260

)210× 260 GeV (≤max

Tp<210

)110× 210 GeV (≤max

Tp<160

)0

10× 160 GeV (≤max

Tp<110

)4sαO(NLO pQCD

unc.sαPDF &

scale unc.

ATLAS =7 TeVs|<0.8y>100 GeV |

Tp jets R=0.6 tantik

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

> 2.2 TeVjjM (+0.5)

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

< 2.2 TeVjjM1.8 < (+0.4)

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

< 1.8 TeVjjM1.4 < (+0.3)

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

< 1.4 TeVjjM1.1 < (+0.25)

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

< 1.1 TeVjjM0.85 < (+0.2)

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

< 0.85 TeVjjM0.65 < (+0.15)

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

< 0.65 TeVjjM0.5 < (+0.1)

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

< 0.5 TeVjjM0.35 < (+0.05)

dijetχ

2 4 6 8 10 12 14 16

dije

tχ

/ddije

tσ

ddije

tσ

1/

0.1

0.2

0.3

0.4

0.5

0.6

0.7

< 0.35 TeVjjM0.25 <

CMS

= 7 TeVs-1L = 36 pb

DataQCD prediction

= 5 TeV+Λ

= 5 TeV-

Λ

� Angular distributions in different di-jet mass bins are sensitive tonew physics

• χ = e|y1−y2| ' 1 + |cosθ∗|1− |cosθ∗|

; correlated to inv. mass but

independent of calibration• QCD is almost flat in χ, while new physics (compositeness)

might peak at lower χ values• plot to the left shows CMS data compared to NLO pQCD

(NLOJet++ with CTEQ 6.6)• also plotted contact interactions with compositeness scale of

Λ± = 5 TeV (∼ observed limits)� NLO pQCD calculations agree with data



Di-jet events � with a veto on central jet activity

� Look at the activity inside therapidity gap between a di-jetsystem defined by• either the two leading jets in p⊥• or the two jets with the largest rapidity gap

∆y• gap fraction is the fraction of events without

additional jet activity in the rapidity intervalbetween the jets

• the veto scale is Q0 = 20 GeV or larger tostay away from ΛQCD

� plot to the right shows gapfraction for leading jets in p⊥ asfunction of ∆y for various p⊥ranges (ATLAS)

� Powheg+Pythia describes databest; HEJ deviates at high p⊥

< 270 GeV (+3)T

p ≤240

< 240 GeV (+2.5)T

p ≤210

< 210 GeV (+2)T

p ≤180

< 180 GeV (+1.5)T

p ≤150

< 150 GeV (+1)T

p ≤120

< 120 GeV (+0.5)T

p ≤90

< 90 GeV (+0)T

p ≤70

Data 2010

HEJ (parton level)

POWHEG + PYTHIA

POWHEG + HERWIG

dijet selectionT

Leading p

= 20 GeV0

Q

ATLAS

y∆

0 1 2 3 4 5 6

Ga

p f

ractio

n

0

1

2

3

4

5



Di-jet and Tri-jet events � ratio R32

(TeV)TH0.5 1 1.5 2 2.5

32

R

0

0.2

0.4

0.6

0.8

1

(TeV)TH0.5 1 1.5 2 2.5

32

R

0

0.2

0.4

0.6

0.8

1 CMS

= 7 TeVs

-1=36 pbint

L

R=0.5T

anti-K

Data

PYTHIA6 tune Z2

PYTHIA6 tune D6T

PYTHIA8 tune 2C

MADGRAPH + PYTHIA6 tune D6T

ALPGEN + PYTHIA6 tune D6T

HERWIG++ tune 2.3

Systematic Uncertainty

� Another measure ofadditional activity is theratio of the inclusive 3-jetover the 2-jet crosssections• measured as function of the total

transverse momentumH⊥ =

∑|pjet⊥|

• jets are required to havep⊥ > 50 GeV and |y| < 2.5

• uncertainties like jet energy scaleand luminosity drop in the ratio

� plot to the left shows R32as function of H⊥ (CMS)

� different simulationsagree for H⊥ > 0.5TeV



Jet properties

� Lots of different techniques to identify sub-structure inside jets mainly forboosted heavy objects

� C/A Filtering• reverse clustering of large Cambridge-Aachen (R ' 1.2) jets looking for a large drop in mass

• recluster remaining constituents with smaller R

• good for H→ bb for example

� Trimming• find small k⊥ subjets and remove those with small fraction of jet p⊥• improves di-jet mass resolution for heavy di-jets

� Pruning• run C/A or kT on jet and check min(p1

⊥, p2⊥)/pjet⊥ < zcut and ∆R1,2 > Dcut in each step

• if so discard the daughter with smaller p⊥• removes radiation from heavy particles

� N-Subjettiness• test how well a jet can be described by N 4-vectors

• add distances of all constituents to nearest sub-jet as measure

� lots of others ...



Jet properties � sub structure

� For the jet sub structure algorithms to be useful they have to be testedon QCD jets as this will be the main background

� Jet pruning is studiedin CMS for W tagging• C/A is used with R = 0.8• require at least 2 high

p⊥ > 200 GeV jets with∆φ > 2.1 and |η| < 2.5

� the plots show CMSdata compared to twoPythia tunes andHerwig++• leading jet pruned mass (top

left)• mass drop (ratio of masses

of highest p⊥ subjet to fulljet) (top right)

• subjet ∆R (bottom left)• subjet p⊥ asymmetry y =(

min(p1⊥, p2⊥)×∆R1,2mjet

)2

(bottom right)

)2 (GeV/cjet m0 20 40 60 80 100 120 140

Pro

bab

ility

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

0.22DataPythia TuneZ2Pythia TuneD6THerwig++ Tune23

= 7TeVs at -134.7 pbCMS PreliminaryJet Pruning Algorithm

jetm1m

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Pro

bab

ility

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045 DataPythia TuneZ2Pythia TuneD6THerwig++ Tune23


R∆ 0 0.2 0.4 0.6 0.8 1

Pro

bab

ility

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07DataPythia TuneZ2Pythia TuneD6THerwig++ Tune23


y0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Pro

bab

ility

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35 DataPythia TuneZ2Pythia TuneD6THerwig++ Tune23




Jet properties � jet mass

50 100 150 200 250 300

dm

[G

eV

]σ

d

σ1

0

0.002

0.004

0.006

0.008

0.01

0.012 1ATLAS 2010 Data, L = 35pb

Pythia

HerwigJimmy

Herwig++

CambridgeAachen R=1.2 jets

> 300 GeV, |y| < 2T

= 1, pPVN

PreliminaryATLAS

Jet Mass [GeV]50 100 150 200 250 300

MC

/ D

ata

0.20.40.60.8

11.21.41.61.8

Jet Mass [GeV]50 100 150 200 250 300

MC

/ D

ata

0.20.40.60.8

11.21.41.61.8 50 100 150 200 250 300

dm

[G

eV

]σ

d

σ1

0

0.002

0.004

0.006

0.008

0.01 1ATLAS 2010 Data, L = 35pb

Pythia

HerwigJimmy

Herwig++

CambridgeAachen R=1.2 jets

> 0.3qq

Split/Filtered with R

> 300 GeV, |y| < 2T

= 1, pPVN

PreliminaryATLAS

Jet Mass [GeV]50 100 150 200 250 300

MC

/ D

ata

0.20.40.60.8

11.21.41.61.8

Jet Mass [GeV]50 100 150 200 250 300

MC

/ D

ata

0.20.40.60.8

11.21.41.61.8

[GeV]jm

50 100 150 200 250 300

d M

[G

eV

]σ

d

σ1

0

0.002

0.004

0.006

0.008

0.01

0.012

0.014

0.016

0.018ATLAS Preliminary 1ATLAS 2010 data: 35 pb

Pythia MC10

Herwig/Jimmy

Herwig++

jets with R=1.0T

Anti k

> 300 GeV, | y | < 2T

= 1, pPVN

jet mass [GeV]50 100 150 200 250 300

MC

/Da

ta

0.5

1

1.5 [GeV]

12d

0 20 40 60 80 100 120 140 160 180 200

[G

eV

]1

2d

d

σ d

σ1

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04 ATLAS Preliminary 1ATLAS 2010 data: 35 pb

Pythia MC10

Herwig/Jimmy

Herwig++

jets with R=1.0T

Anti k

> 300 GeV, | y | < 2T

= 1, pPVN

[GeV]12

d0 20 40 60 80 100 120 140 160 180 200

MC

/Da

ta

0.5

1

1.5

� For CMS Herwig++ fits best, while itpredicts slightly too high masses in ATLAS

� Jet mass is studied inATLAS to identify topsand in Higgs searches• either C/A with R = 1.2,|y| < 2 and p⊥ > 300 GeV

and C/A splitting/filtering• or AntiKt with R = 1.0, |y| < 2

and p⊥ > 300 GeV and Ktreclustering

• jets are made from calibratedtopological clusters plus final jetenergy scale

� the plots show ATLASdata compared to Pythia,Herwig/Jimmy andHerwig++• invariant mass of C/A jets before

splitting and filtering (top left)• and after splitting and filtering

(top right)• invariant mass of AntiKt jets

(bottom left)•√

d12 = min(p1⊥, p2⊥)×∆R1,2

for the same jets (bottom right)



Conclusions

� rich QCD program at ATLAS and CMS� jet reconstruction and calibration

• reconstruction algorithms like Particle Flow for CMS and Topological Clustering for ATLAS show verygood performance

• challenge on 2011 data is the increased amount of PileUp

� cross section measurements• both experiments have impressive results on inclusive and di-jet differential cross sections

• most comprehensive studies of NLO pQCD predictions

• excellent agreement with NLO calculations was found

• regions of phase space not explored before already constrain PDF’s

� jet properties• very exciting field to identify heavy boosted objects

• QCD comparisons are vital to establish validity of the used methods

• first results look very promising

� we are looking forward to the increased 2011 data set and the QCDresults it will contain

S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC HCP 2011, 16. Nov. 2011, Paris, France 24


Backup slides � Topological Clusters in ATLAS

� Calorimeter showers are grouped by a topological clustering algorithminto 3D energy blobs

� Noise related threshold control seeding and growth of the clusters

� Proto-clusters are split around local energy maxima

� Example below shows a typical simulated QCD event in forward region(just the first layer) with different noise thresholds and final clusteringand splitting applied

� The algorithm suppresses calorimeter noise and leads to dynamicallysized clusters corresponding on average to 1.6 final state particles

|E| > 2σnoise |E| > 4σnoise 4/2/0 topological clusters

φ cos ×| θ|tan -0.05 0 0.05

φ s

in

×| θ|t

an

-0.05

0

0.05

210

310

410

510

FCal1C

φ cos ×| θ|tan -0.05 0 0.05

φ s

in

×| θ|t

an

-0.05

0

0.05

210

310

410

510

FCal1C

φ cos ×| θ|tan -0.05 0 0.05

φ s

in

×| θ|t

an

-0.05

0

0.05

210

310

410

510

FCal1C

S. Menke, MPP Munchen Hard QCD Results with Jets @ LHC HCP 2011, 16. Nov. 2011, Paris, France 25


Backup slides � Particle Flow in CMS

� Use all sub-systems toreconstruct stable particles• electrons from tracks linked to ECAL

clusters• charged hadrons from tracks linked to

ECAL+HCAL clusters• muons from links between tracker

and muon system tracks• photons from ECAL clusters without

links• neutral hadrons from remaining

ECAL+HCAL clusters

� only neutral hadrons needhadronic calibration ofcalorimeter energy

� residual jet energycorrections on the order of5− 10%

� top plots show track links to ECAL (left) and HCAL(right) clusters

� bottom plots show PF reconstructed energy fractionsfrom data (left) and simulation (right)

S. Menke, MPP Munchen Hard QCD Results with Jets @ LHC HCP 2011, 16. Nov. 2011, Paris, France 26


Documents

Hard QCD Results with Jets @ LHC (ATLAS + CMS) · Total Delivered: 48.1 pb1 Total Recorded: 45.0 pb1 LHC: pp Collisions @ p s = 7 TeV since March 2010 2011 ATLAS/CMS recorded integrated