Post on 01-Aug-2020
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 �
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
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 3
The CMS Detector
Compact Muon Solenoid: 15m high; 22m long; 12500 t heavy
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 4
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)
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 5
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
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 6
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)
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 7
Jet Reconstruction in ATLAS and CMSJets, G. Salam (p. 27)
2. Getting the basics right
FastJetJet contours – visualised
G. Salam, 2008
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 8
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
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 9
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
[%
]
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 10
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)
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 11
Cross sections � Inclusive jet cross section
� 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)
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 12
Cross sections � Inclusive jet cross section
� 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)
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 13
Cross sections � Inclusive jet cross section
� 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
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 14
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
√s = 7 TeVInclusive Jets Anti-kT R=0.50.0 ≤ |y| < 0.5
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
√s = 7 TeVInclusive Jets Anti-kT R=0.52.0 ≤ |y| < 2.5
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
√s = 7 TeVInclusive Jets Anti-kT R=0.52.0 ≤ |y| < 2.5
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⊥
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 15
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)
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 16
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
√s = 7 TeVDijet Mass Anti-kT R=0.70.5 ≤ |y|max < 1.0
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
√s = 7 TeVDijet Mass Anti-kT R=0.71.0 ≤ |y|max < 1.5
CMS DataCT10 ⊗ NP, ∆PDF CL68MSTW2008 ⊗ NPNNPDF2.1 ⊗ NPHERAPDF1.0 ⊗ NPABKM09 ⊗ NP
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
√s = 7 TeVDijet Mass Anti-kT R=0.71.5 ≤ |y|max < 2.0
CMS DataCT10 ⊗ NP, ∆PDF CL68MSTW2008 ⊗ NPNNPDF2.1 ⊗ NPHERAPDF1.0 ⊗ NPABKM09 ⊗ NP
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
√s = 7 TeVDijet Mass Anti-kT R=0.72.0 ≤ |y|max < 2.5
CMS DataCT10 ⊗ NP, ∆PDF CL68MSTW2008 ⊗ NPNNPDF2.1 ⊗ NPHERAPDF1.0 ⊗ NPABKM09 ⊗ NP
0.7 2 4
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 17
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
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 18
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
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 19
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
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 20
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 ...
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 21
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
= 7TeVs at -134.7 pbCMS PreliminaryJet Pruning Algorithm
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
= 7TeVs at -134.7 pbCMS PreliminaryJet Pruning Algorithm
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
= 7TeVs at -134.7 pbCMS PreliminaryJet Pruning Algorithm
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 22
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)
S. Menke, MPP Munchen � Hard QCD Results with Jets @ LHC � HCP 2011, 16. Nov. 2011, Paris, France 23
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